> Re Hea Be 628 eee oe er eee ee ee ee eee TN a ee ee rr eee, ee eee eee I Ee Pe Ce ores Si ae Lt ay Te re eae tO alg Maw Bie re eal OS 9 ames © yy eed ee PP te Piatt Ci ll em mt a tis Ha A a Pl A all ag te Se a Ml te lh lca Bee bagi tina” Mert an ea a te Aa ah ae tlt tet ot Ale to Ne Ate set OM Rech tS Ath Me heli Minny “Ree Me Me pail tm tte oe Rent YM el QB Bebe Maths viemt NEAR, oA ad eaab at othn hs, yah, onan ts tag nl Mh 9 Want hahah wilh ia Renal nthe beting ta gllnee 5 tae Mette Meg eta Reb Yate Bi me BR ny a ap = GR a Me hl ne i ng AI em te Ate Bitten Li Dh he A eG LN NY Ne lt he AR My lat Alon Map ie abl alpha, PREP Wve WPF Gen Wort rset tos Veneer We RENNER i Wier ertachs Bemreareeur err We WA. 1 viraesiryirerrcietcarerdr tir WW tr i eve oe Meet pak se Rig ee yom athe tt a Tg wy PR tm Nitin NA tm at dP te il st Al te At ih I re Mt a ah te A tt tell A te Ae oe ATRL Rt ll Ml th oh he Patient eng hes thm M0 A gent Midian Pasha Meigs Tye Sense Ge Yt Ment attend bY cb atsals «Mn low itn bgt afin A-ha! De et PAG anak mee Me Oe ee ek ae et tO OE a ell it ee ie th hn Ni ing Neder MI I Ne a ett le latte ee a ll ES as eel lente Rtn ntl Mma. jet oe = = ed = ete eet « _ ee ed tee Ane Oe Rr re yeh te Het ated, Fa aet St Ele Meme ee. ke Oat, ae De Eg ee te Be ee ~ ee OO Pw Oeste ee | fee. Ce nate + es = eT em tae ee ee aes RE MR IN NTR A AIR I ere inte ee Me, *¢ AS nates, Pat Bip a Mek Neel etn te ete te a tel ane de = ee eee Ley eee ee eee oe a re an * ate Pte ete Oe def 2 at PLA AMAA ee ee ee ~ A ee Oa EN A rR eet et, Dal ee eee te t,he Sateen, ae ee Rg det Live eee Ot ' t= WB Oe a ne he me AD Retin 9 ate Aaa ee RT ee ee a Re % No Me pte Pia ath. Sekt fe Ranta tte Rte we teats ne eae Soren ee 4 Un ipl 7M ee eae Gee ere, ey FO a oe Oe Re ee He me te Re bre Le eR ee Lg SR i Dae roth alg a ne ae ees om + _ yore we ad Mig & A Ree he tag ine Ne Me - ee Rete - ‘7. a eo a F rate eN Mu, «ss =e ot Ut LA Mt ee ee ee See teks te Le 7 we NG ate wm . hee, isha!) au Saateet? = n= * ‘> -- ae fos Me a ie ee Net Fe he fat hel eee Re ere ett A eee a m i ee bee wt eRe es, ~ — a a ee eee rel -* . aM Bt . eA het Fe The % —,. —_— A re a RR eee tet Re ee = PRN e tet ntitiadetat of. - «oe -! a _“he 2 - agit Bem S- Ngan ot tp Bhtlt, ~~ «4,7 - OT A eee TORR RAL ALL PLL Pale BM BO 5 cathe the Meme ie Re le Do os = - oT ol AE 0 at Sa See oe Fay ae Tey py Se Phe he RE MI A Rete HR te UT oe A Are 1 . en bs Mata: ¢ a: ~ : ‘ nS we te ee ~ tate wad othe teh a ee A RA Re ee A oe wee © a are = ae a es en oo ie Ove ee i ae ee ore = 5 > . <* 28 7 7 7 = . | o nr “ee Se Ne NS mM Re Ne Re te Qe wet ee ee A UN a Pn tA RR A eo re: oc - Vi ae Me RD te Tg . a _ ~ ? tn ~ " a ~~ “ - 7 . ~ 4 — Ane Phe > he ~ - a - « \ tre « y a be - ‘ . is 2 » eS et a “ re “ 2. * an ’ mee . ". “ nm 7 . » £m “ - . “ Se = we, Veet de : ‘ a a : * - —*. an = / a f ‘ = \ “- ~ tin, wr 7 ‘ . — aere - ~ ~ n . ~ ’ = ~ » * & a . iA A . ~ nN . ~ men ‘ r = i olde orca a 2st be. 3 fe 5 - ~ i= 2 . na thn aes . * atte . ms - o Ld ~ . ae w =r aw a - v - - — = Ota m « . is o ’ a r we - ~ ss ‘ 4 - = * wae = wide ro ™ . . _ * * - - - - - - - A ee +-=_- ee - -* 7 - - 7 = x *. ~ * 4.5 a Ww al at - - - La ~ aie — ~~ ‘ = a . ~ n <. ~ : - - = * nm Wrens es = : 7 _ > tee — > = o ~~ = ‘s. > Por ~ OP FON i al « “.% ta ™. Ne a a os -< - - - ad - - *+%~ ~ - r- *, “ - “ - - ™ - ™/ enw a . - - -" G - a wr me om” ‘. a \. - An i 1 * late on a ~ —* * ‘ ee a) athe 8M A. = = 7 7 Fs a oe a - x + a* 7 yi =e ¥ aad . pve) Me lias A = i ’ al - - * - -> - - ~ -- - - sa) - bes 4. ro a ee + ‘ ~* a. “whe . y= - - oe © ot tees - a 2 we ee - “~ Ror - ar 4 - Lat * - . -_—“ “ ‘ ” al -, ~ va ~ = 2 = ~ _=-* ah. ' - 7 7 - 2 be ee - ii ™. re x ee a” * Me 4 2 ey oe, = _ f - - - e = - - ~ “~ . > ea % - ‘ 4% - - wn - - of « ” 7 e ” ao aoe te ~ - a ’ . . a . ae ‘ aA a . An As . — “ =* m, « - te -- . - - - x n any 7 = = 7 - 7 . an . > Ae . ae aN m rn ote - R 7 7 7 a a = 5 = a : . . : x aMere ies 7 ee. a . 7 - - . ='s o - -_ - - - - —y a _ - + " a 7 7 “- 2 not - a ~ “! = ies ‘ - a ¥ - 2 - Fe +n 7 “ 7 oy e * = * e “ ss a4 * *- “ + 3 g _ - a os “: . 2 é a 3 = “ —* a5 n - as : . - 7 , ~ 3 Ps - - ‘ a .« $s 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 Camsruwnge, Proressors ADDISON E. VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT E. GREGORY AND HORACE S. UHLER, or New Haven, Proressor HENRY S. WILLIAMS, or IrHaca, ProFressor JOSEPH S. AMES, or Battimore, Mr. J. S. DILLER, or Wasuineron. FOURTH SERIES VOL. XLUI—[ WHOLE NUMBER, CXCIIT)}. WITH TWO PLATES. NEW HAVEN, CONNECTICUT. 1 al hf i a ral HERE THE TUTTLE, CONTENTS TO VOLUME: XTi. Number 253. Art. I.—Retarded Diffusion: and Rhythmic Precipitation ; Pee SENSOR UNE DY)... Oe cee eee Pe ek Be SUE ed IJ.—Calorimetry by Combustions with Sodium Peroxide ; WON «Ore WE OO RIT. ee) Se ee a et hi ee Sy Ill.—Hébert’s Views of 1857 regarding the Periodic Sub- mergence of Hurope; by C. ScHucHERT -_..._.---+--- 35 IV.—Lawson’s Correlation of the Pre-Cambrian Era; by pre sineers (WW ith Piste D2 0 es es a = AD V.—A Table for Linear and Certain Other Interpolations ompspectrootams; by H- HW. Merwin 422.5.) 2522,.4.--) 49 VI.—On the Preparation and Ionization of the Dialkyl- phosphoric and Benzenedisulphonic Acids; by W. A. Davee and mek MMrEy eC ee ae ee eo OT VIt.—On the Double Salts of Cxsium Chloride with Cal- cium and Strontium Chlorides; by G.S. JAMIESON .--. 67 VIII.—Crandallite, a New Mineral; by G. F. Loveutry and Pee ser Aernin. = 30 eee ke We ones eee gl 69 IX.—A Titaniferous Augite from Ice River, British Columbia; by C. H. Warren, J. A. Antawn and M. F. Conner --.- 75 SCIENTIFIC INTELLIGENCE. Page Chemistry and Physics—Determination of the Density of Solids, H. Lr CHATELIER and EF. Boaitcu, 79.—New Reagent for Free Chlorine, G. A. LeRoy: The Right Honourable Sir Henry Enfield Roscoe, Sir E. THORPE: Text-book of Inorganic Chemistry, A. F. Hotteman, 80.—Organie Chem- istry for the Laboratory, W. A. Noyes: Note on the Electrolysis of Gallium, H. 5S. UntER: The Condensation of Gas Molecules. R. W. Woop, 81.—The Cosine Law in the Kinetic Theory, M. KNupsEn, 83. Geoloyy—A preliminary paper on the origin and classification of intraforma- tional conglomerates and breccias, R. M. FreLp: Florida Geological Survey, Highth Annual Report, E. H. SeLuARDS: Study of the Morrison formation, C. C. Moor: Notes on the geology of Nelson and Hayes Rivers (Canada). J. B. TyrRELL, 8).—Sixth Annual Report of the Director of the Bureau of Mines, V. H. Mannine, 86.—An Introduction to Historical Geology, W. J. MILLER, 87. Miscellaneous Scientific Intelligence—Report of the Secretary of the Smith- sonian Institution, C. D. Watcort, 87.—Field Museum of Natural History, F. J. V. Sxirr, 88. Obituary—H. H. W. Pearson: H. Mosy, 88. ty CONTENTS. Number 254. : Page Arr. X.—The Water Content of Coal, with Some Ideas on the Genesis and Nature of Coal; by KH. Mack and G., A. Hypnerr: 220. fore 22 ee, ee XI.—An Apparatus for ae mining Freezing Point Lower- ing; by R. G. Van Name and "W..G. Brown 2. loa 110 XIJ.—The Sodium-Potassium Nephelites; by N. L. Bowen 115 XIII.— Pottsville Formations and Faunas of Arkansas and Oklahoma; by K> ft: Matarr!)... 2-257. 2 133 XIV.—A Study of Two So-called Halloysites from Georgia and Alabama; by P. A. vanDER MEULEN ._._.__.__-- 140 XV.—Methods in Reversed and Non-reversed Spectrum Interferometry (continued); by C. Barus .-.....-...-. 145 XVJ.—On the Identity of Hamlinite with Goyazite; by W. Ws S0HALLER 15. -eolicgal he: ee eee 163 SCIENTIFIC INTELLIGENCE. Chemistry and Physics.—Occurrence of Free Monoxide in the ‘‘ Floaters” Kelp, 8. C. Lanepon, 165.—Atomic Weight of Lead of Radioactive Origin, T. W. RicHarps and C. WapswortTs, 3d: Engineering Chemistry, T. B. STILLMAN, 166.—Qualitative Analysis, E. H. S. Barney and H. P. Capy: X-Ray Waye-Lengths, M. Sreepaun, 167.—General Physics, W. S. FRANKLIN and B. MacNoutt, 168.—Cosmical Evolution, Critical and Constructive, Second Edition, E. McLennan, 169.—An Introduction to Astronomy, New and Revised Edition, F. R. Mouuton, 170. Geology and Mineralogy.—Some Geophysical Observations at Burrinjuck, L. A. Cotton, 170.—Geology of Cincinnati and Vicinity, N. M. Frn- NEMAN: Field Geology, F. H. Lawes, 172.—The Fundamental Prin- ciples of Petrology, (KE. Weinschenk), A. JoHannsEN: L’Oural du Nord; Le Bassin des Rivieres Wagran et Kakwa, L. Duparc et M. TIKANOWITCH: Etude Comparée des Gites Platiniféres de la Sierra de Ronda et de l’Oural, L. Duparc and A. Grosser, 173.—Etude Cristallo- graphigue et Optique d ’un certain nombre de Minéraux des Pegmatites de Madagascar et de Minéraux de 1]’Oural, R.-C. Sasot: The Geological His- tory of Australian Flowering Plants, Errata, E. C. ANDREWS, 174. Obituary.—W. Exuis: C. Rem: D. Oxtiver: A. M. Worrtsineton: J. - WriGuHtTson, 174. CONTENTS. Vv Number 255. Page Arr. XVII—A Method for the Determination of Dissocia- tion Pressures of Sulphides, and its Application to Covellite (CuS) and Pyrite (FeS,); by E. T. ALLEN and TE 1S [eel CG's WT ee ea Rm gS > a Pa eee LY 4) ~~ XVITI.—-Beecher’s Classification of Trilobites, after Twenty Seas: by. HAMMOND] ?oste ee ee Che OG XIX.—Origin of Formkohle; by J. J. STEVENSON ---- ---- 211 XX.—On the Etching Figures of Beryl; by A. P. Honzss.. 223 XXI.— Plotting Crystal Zones on the Sphere; by J. M. Brake 237 XXII.—Note on the Age of the Scranton Coal, Denver asim Colorado; by.G. B. RIcHARDSON: 2... 2l22.. 22. 2. 243 SCIENTIFIC INTELLIGENCE. Chemistry and Physics— Reminiscences, C. F. CHANDLER, 245.—Removal of Barium from Brines used in the Manufacture of Salt, W. W. SKINNER and W. F. BavGcHMan: Ammonium Chloride as a Food for Yeast, C. H. HorrmMan, 246.—General Chemistry, H. P. Capy: Generalized Relativity and Gravitation Theory, 247.—Ball Lightning, M. E. Marutas, 248.— Laboratory Course of Practical Electricity, M. J. ARcHBOLD: National Physical Laboratory, Report fur the Year 1915-16, W. F. Parrott, 249. Geology—Investigations of Gravity and Isostasy, W. Bow1z, 249.— Prodro- mus of Nicolaus Steno’s Dissertation Concerning a Solid Body Enclosed by Process of Nature Within a Solid, J. G. Winter, 200.—Atlantic Slope Areas, P: G. SHELDON: Iowa Geological Survey, Annual Report, 1914, with accompanying papers, G. F. Kay: New York State Museum, Twelfth Report of the Director, J. M. CLARKE, 201.—Review of the Geology of Texas, J. A: Upprn, C. L. Baker, and E. Bos, 252: Annual Progress Report of the Geological Survey of Western Australia for the Year 1915, A. G. Martnanp: Economic Geology, H. Riss, 252. Miscellaneous Scientific Intelligence—Annual Report of the Superintendent. United States Coast and Geodetic Survey, E. L. Jonss, to the Secretary of Commerce, W. C. REDFIELD, for the fiscal year ending June 30, 1916. 2038.—Fundamentals of Psychology, W. B. PirtsBury: Mechanisms of 22 ane Formation: an Introduction to Psychoanalysis, W. A. WHITE, 204, v1 CONTENTS. Number 256. Page Art. X XIII.—Lava Flow from Mauna Loa, 1916 ; by T. A. JAGGAR, SPS oS. LS ee ae ore XXIV.—The Formation of Salt Crystals from a Hot Satu- rated Solution; by HE: Tarun Lone. -os 2.0... ee 289 XX V.—An Upper Cretaceous Fulgur; by B. Wapz_---.- _ 293 XXVI.—A Middle Eocene Member of the “Sea Drift ” ; by HW. Bae yee ee ee 298 XXVII.—Correlation of the Mississippian of Ohio and Penn- sylvania; by “W.A: Verwikser. 2). ll.) Se XXVUL—A New Labyrinthodont from the Triassic of. Pennsylvania; by. W. J. SINCLAIR... -. 2 41)-222)ee XXIX.—On the Calcium Phosphate in Meteoric Stones; by GPS MERRFEL:. 3. dee ae ee et a 322 XXX.—Crystals of Pyromorphite; by E. V. Saannon .__._ 325 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—The Penfield Test for Carbon, W. G. MixtTER and F, LL. HatcH. 3827.—New Volumetric Method for the Determination of Cobalt. W. D. Encue and R. G. Gustavson, 328.—Determination of Molybdenum by Potassium Iodate, G. S. JAMieson: Fixation of Nitrogen, J. EK. Bucuer: Volatilization of Potash from Cement Materials, E, ANDERSON and R.J. NESTELL. 529.—Dispersion and the Size of Molecules of Hydro- gen, Oxygen, and Nitrogen, L. SILBERSTEIN, 330.—Lubrication of Resist- ance-Box Plugs, J. J. MANLEY, 331.—Unipolar Induction, E. H. KEnnarp: Model Drawing, C. O. Wricut and W. A. Rupp, 382.—Teaching of Arith- metic, P. KLAPPER, 3933. Geology and Mineralogy—American Fossil Cycads, Volume II, Taxonomy, G. R. Wre_anp. 335.—New Zealand Geological Survey, P. G. MoreGan, 335.—The Gold Belt South of Southern Cross, T. BLATCHFORD, etc., 336.— Relationship of the Tetracoralla to the Hexacoralla, W. I. RoBinson: A new genus and species of the Thecidiinae, ete., J. A. Thomson: Tertiary formations of Western Washington, C. E. Weaver: Paleontologic contri- butions from the New York State Museum, R RUEDEMANN, 337.—Great Eruptions of Sakura-jimi in 1914, B. Koro: Synantectic Minerals and Related Phenomena, J. J. SEDERHOLM, 338.—Bibliography of Australian Mineralogy. C. ANDERSON: Rings, G. F. Kunz: Economie Geology, H. Rres: The Geological History of Australian Flowering Plants, E. C. ANDREWS, 339. Miscellaneous Scientific Intelligence—Indians of Cuzco and the Apurimac, H. B. Ferris, 339.—Carnegie Institution of Washington, R. S. WooDWARD, 340.—Observatory Publications: Publications of the Museum of the Brooklyn Institute of Arts and Sciences: Tables and Other Data for Engi- neers and Business Men, C. EK. FerRIs, 342. Obituary —W. Brerepe: T. Purpis: N. H. J. MInumre, 342. CONTENTS. Vil Number 257. Page Arr. XXXI.—The Geology of the Lau Islands; by W. G. ieee oe Oe fs ees Soe ey aa XX XII.— Origin and Age of the Ontario Shore-line,—Birth of the Modern Saint Lawrence River; by J. W. SPENCER 351 XX XIII.—Granite Bowlders in (?) the Pennsylvanian Strata ea amsags by Wied Lay MNHOPE Ee 2-82 oe ay S68 XXXIV.—An Oligecene Camel, Poébrotherium anderson Mesp.. Oy Bet RON hn Se ele t kbc a alae sao! 381 XXXV.—Sand Fusions from Gun Cotton ; by C. EK. Munror 389 XXXVI.—Electrolytic Analysis with Small Platinum Elec- trodes; by F. A. Goocn and M. Kopayasui-_--_-_.------ 391 XXXVII.—Crystal Drawing and Modeling; by J. M. Buaxer 397 XXXVIII.—Normal Anomalies of the Mean Annual Tem- perature Variation; by H. ArcrowskI .-..---------. . 402 SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Attempt to Separate the Isotopic Forms of Lead by Fractional Recrystallization, T. W. Ricnarps and N. F. Hatt, 409.— Manganese in Soils, M. O. Jonnson, 410.—Preparation of Sulphurous Acid, E, Hart: Gas Chemists’ Handbook: Union of Glass in Optical Contact by Heat Treatment, R. G. Parker and A. J. Datuanpay, 411.—Flame Spec- trum of Iron, G. A. HemMsaLecu, 413.—Nature of Matter and Electricity, D. F. Comstock and L. T. TRovanp, 414.—Electric and Magnetic Measure- ments, C. M. SmirH: Recreations in Mathematics, H. E. Licks, 415. Geology and Mineralogy—The Andes of Southern Peru, I. Bowman, 416.— Mount Rainier, a Record of Exploration, 417.—Publications of the United States Geological Survey, G. O. Smita, 418.—West Virginia Geological Survey, I. C. Wurte: Inorganic Constituents of Marine Invertebrates, F. W. Cuarks and W. C. WHEELER: Synopsis of American Early Tertiary Cheilo- stome Bryozoa, F. Canu and R.'. BassuEer, 419.—Note on Goyazite, O. C. FARRINGTON: Elements of Mineralogy, Crystallography and Blowpipe Analysis, A. J. Moses and C. L. Parsons: Optical Character of Sulphatic Cancrinite, E. S. Larsen, 420. Miscellaneous Scientific Intelligence—Morphology of Invertebrate Types, A. PETRUNKEVITCH: Microbiology—Text-book of Microorganisms General and Applied: Growth in Length: Embryological Essays, R. ASSHETON, 421.—Respiratory Exchange of Animals and Man, A. KrocH: Memoirs of the Queensland Museum, Vol. V, R. Hamytyn—Harris, 422. vill : CONTENTS. Number 258. Page Art. XX X1X.—The Geology of Pigeon Point, Minnesota; Ri: A, DALY... 2 oe ee 423 XL.—On the Temperature Coefficient of a Heterogeneous Reaction; by R.-G.-Van Nan S222 22) eee 449 XLI.—A Sail Fish from the Virginia Miocene; by E. W. BERRY 222 2) P20 g2 Lo oF 461 XLII.—Eakleite, a New Mineral from California; by E. S. LARSEN. 220. so. Ss ose PSUS ie ho) a XLITI.—Rensselerina, a New Genus of Lower Devonian Brachiopods; by C. O. Dunpar. (With Plate Il)_.... 466 XLIV.—Local Versus Regional Distribution of Isostatic Compensation; by W.2 Bowim 2. _..=._2 2.) 2a XLV.—The Constitution of Melilite and Gehlenite ; Py BW. OW. CraARK EB) Deb ee ha ae Se SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Extraction of Potash from Silicate Rocks, W. H. Ross: Preparation of Uranium Dioxide, C. L. Parsons, 485.—Theoretical Chemistry from the Standpoint of Avogadro’s Rule and Thermodynamics, W. Nernst: A Text Book of Thermochemistry and Thermodynamics (O. Sackur), G. E. Grpson, 486.—X-Ray Band Spectra, DE BRoGLin: Finite Collineation Groups, H. F. BLicHFELDT, 487. Geology and Mineralogy—Thirty-seventh Annual Report of the Director of the United States Geological Survey, G. O. Smiru, 488.—Publications of the United States Geological Survey, G. O. Smiru, 489.—Report of the State Geologist on the Mineral Industries and Geology of Vermont, 1915-- 1916, G. H. Prerxins: Illinois Geological Survey, F. W. DEWotrF, 490.— Geology and Economic Deposits of a Portion of Eastern Montana, J. P. Rowe and R. A. Witson: A Preliminary Report on the Alkali Resources of Nebraska, E. H. Barsour: Recent and Fossil Ripple Marks, E. M. KInDLE, 491.—The State of the Ice in the Arctic Seas, C. I. H. SPEER- SCHNEIDER: Annual Report of the Government Geologist of South Aus- tralia for 1915, L. K. Warp: The Efficient Purchase and Utilization of Mine Supplies, H. N. Stronck and J. R. Brntyarp, 492.—The Ordovician and Silurian Brachiopoda of the Girvan District, F. R. C. Reep: Notes on Cincinnatian Fossils, A. F. ForERSTE: New Mineral Names, W. E. Forp, 493. Miscellaneous Scientific Intelligence—National Academy of Sciences, 495.— State Sanitation, a Review of the Work of the Massachusetts State Board of Health, G. C. WurpPpLe: The American Year Book, a Record of Events and Progress, 1916, F. G. Wickware, 496.—Explorations and Field Work of the Smithsonian Institution in 1916, 497. — Obituary—A. HaGcue: G. Masses: J. G. Darsoux: R. H. Tippeman, 497. THE 1 ##$.AMERICAN JOURNAL OF SCIENCE, Epiror: EDWARD S. DANA. ASSOCIATE EDITORS Prorgssors GEORGE L. GOODALE, JOHN TROWBRIDGE, | W. G. FARLOW anp WM. M. DAVIS, or Camsrwcz, PROFESSORS ADDISON E. VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT E. GREGORY AnD HORACE S. UHLER, or New Haven, Proresson HENRY S. WILLIAMS, or Irnaoa, ~ Proressor JOSEPH S. AMES, or Battimorz,/ \\ Mr. J. S. DILLER, or Wasuineton. / o> FOURTH SERIES Sto not hese VOL. XLIII—[WHOLE NUMBER, CXCIII]. | , No. 253—JANUARY, 1917. aah WITH PLATE I. NEW HAVEN, CONNECTICUT. SES a oes — | THE TUITLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. ita Published scent Six dollars per year, in advance. $6.40 to countries in the - Postal Union ; $6.25 to Canada. Single numbers 50 cents. Entered as ‘second- class matter at the Post Office at New Haven, Conns, Ate the Act of sae as d, 1879. 2 ey. - : Si o «> hs nt LIST OF CHOICE SPECIMENS. = Garnets, all known localities. 50e. to $5.00. Bets Topaz, Mexico; Utah; Colorado; Siberia; Japan. 50¢.to ~~ $10.00. : sys Opal, Australia; Mexico; Nevada; Honduras; Hungary. $1.00 to $15.00. ee Emerald, Ural Mts.; Bogota; Tyrol; North Carolina. $3.00 to $30.00. Beryls, all colors, North Carolina; Maine; New Hampshire; Ural Mountains; Brazil. $1.00 to $10. 00. Pink beryl, Caran: $2.00 to $10.00. . Alexandrite, Ural Mountains; loose xls. and matrix speci- mens. $4.00 to $22.50. : Quartz, New York; Arkansas; Colorado; Montana; North - Carolina; Switzerland; Japan. 10c. to $5.00. : ape Quartz, with enclosures of rutile, actinolite, chlorite, bysso- lite, etc., etc. Japan; Madagascar; Ural Mountains; Brazil. ~ 75e. to $5.00. 3 Pink quartz, Maine; Connecticut; South Dakota; Madagas- car. 50c. to $3.00. Amethyst, North Carolina; Michigan; Montana; Colorado; Mexico; Brazil; Ural Mountains; Hungary. ‘5c. to $5.00. Azurite, Arizona; New South Wales; Chessy, France. 50c. to $5.00. Anglesite, iodyrite, stolzite, pyromorphite, cerussite, smith- sonite, embolite, Broken Hill, N.S. W. $2.50 to $10.00. Pyrite, Colorado; Utah; French Creek, Pa.; Franklin Fur- nace; Elba. ‘dc. to $5.00. ee Amazonstone, Colorado; loose xls. and groups. ‘dc. to $5.00. Zaratite, Lancaster Co., Texas, Pa. 50c. to $1.00. Natrolite, Aussig, Bohemia. $1.00 to $2.50. Chrysoprase, Tulare Co., Porterville, Lower California; pol- ished specimens. %2.50 to $7.50. : Calcite, honey colored, Joplin, Missouri. $2.00 to $4.00. Calcite, Schneeberg, Saxony. $1.50 to $2.50. Chiastolite, Lancaster, Massachusetts and Madera Co., Cali- fornia, loose crystals and crystals in matrix. 25c. to $2.50. ALBERT H. PETEREIT 81-83 Fulton St, New York sae THE ores fe AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES.] aoe Arr. 1.—Retarded Diffusion and Rhythmic Precipitation ; by J. Sransrisip, Geological Dept., McGill University, Montreal. Historical. Owr1ne to the fact that the greater part of the literature of this subject is in the German language, it has been considered advisable to give the following somewhat detailed account of it. The rhythmic precipitates formed by diffusion of reagents in gelatine or other media, now generally known as Liesegang rings, were first described in a publication which is not generally available* (1). The precipitates appear to have been shown to W. Ostwald by Liesegang, as a result of which Ostwald pub- lished a short note regarding them (2). The ring-formed pre- cipitates were formed by diffusion of silver nitrate from a drop placed upon a gelatine containing potassium chromate. Ost- wald considered that the diffusion outward of the silver nitrate and the inward diffusion of the chromate resulted in the formation of silver chromate, which was present, at first, as a super-saturated solution in the meta-stable state. Consequent upon the setting up of the labile state a precipitate formed removing the silver chromate in excess of saturation and the diffusion continued with similar results. As the removal of - silver chromate continued the solutions became continuously more dilute, so that the precipitates were formed at continu- ously greater distances apart. In 1908 Morse and Pierce (8) regarded the precipitation as certainly due to a super-saturation effect. They produced by diffusion ring precipitates of mercurous chromate, lead chrom- ate, and Berlin blue, which had been obtained previously by Liesegang, and also ring precipitates of lead sulphate, silver * See literature references at the end of this paper. AM. JOUR. ee var SERiES, Vou. XLIII, No. 258.—January, 1917. 2 J. Stansfield—fetarded Diffusion and earbonate, silver pyro-phosphate, silver thiocyanate, silver bromide, cobalt hydroxide, barium chromate, mercurous brom- ide, and carbon di-oxide, the last-named as rings of bubbles. They performed experiments of a quantitative character, using capillary tubes containing a weak potassium chromate solution in gelatine, which were dipped into silver nitrate solu- tions of different strengths. The precipitates appeared sud- denly as bands across the capillary tubes, and were observed and measured by means of a cathetometer, the times being observed at the formation of the precipitates. Assuming that the chromate solution is completely dissociated, they proved i Jistance , mathematically that the sat for all the layers in all tubes with the same initial concentrations is constant. They gave twenty-three tables of observed figures, which show this con- stancy, within the limits of experimental error. At a lower temperature, with the same initial concentrations, the distances appear to be greater, i. e. the diffusion is more rapid at higher temperatures. In a saturated solution, for equilibrium Ag,’ x CrO,’ = &. In an unsaturated solution, for equilibrinm Ag,’ X CrO, = hp AeOrOe is theresa boundar y which cannot be overstepped in a super-saturated solution without precipitation following ? e. QI Ase >< CrO) = Ia Morse and Pierce found that for a given concentration of one ion a certain definite concentration of the other causes pre- cipitation. They calculated that when precipitation takes place the concentrations of the solutions are mee for the chrom- ( N : ate and ae for the silver. Both solutions must be fully disso- ciated at these dilutions. They caleulated, further, that H is a definite super-saturation limit, which is 1-4 x 107° gram molecules per liter, at 16°C. They found that the diffusion is quicker in gelatine than in water, the diffusion constant cal- culated for silver nitrate being 1:24 ems. per day, that observed being 1:54 (They also suggested that the precipitate is col- loidal, but this appears not to be the case.) Small finer white lines are mentioned as forming between the main precipitates, which become red and then are oblit- erated when the main precipitate forms over them. These are said to be due to impurities, and are not formed in pure gela- tine. They found that gelatine is not essential to the formation of the banded precipitates, but that capillary tubes containing aqucous solutions give similar results, and that the precipitates Rhythmic Precipitation. 3 break up, say after about four in series have been formed. When the experiment is performed in gelatine they are held in place and do not break up. In 1904 Hausmann (4) carried out experiments with capillary tubes containing gelatine solutions of silver nitrate dipping into solutions of sodium chloride, ete., and found that the heights to which the precipitates extend in any given time depend on the concentration of the silver nitrate and of the sodium chloride, ete.,also. Liesegang had claimed that equiva- lent weights of sodium chloride, strontium chloride, potassium bromide, and potassium iodide diffuse to equal heights in equal times. Hausmann found that the distances varied with the reagents, with their concentrations, and with the concentra- tions of the gelatine. In some of the reactions banded pre- cipitates were formed. Mention is made of little brown bands between the main ones and also to secondary rings in the upper part of the tube, which are broader than those in the lower part and increase in breadth upwards. With increase of gelatine concentration the bands were found to be brought closer together. (Curves given pp. 117, 118, loc. cit.) The precipitation of metallic sulphides is discussed. Certain compounds are formed in the colloidal state, e. ¢., silver iodide, mercuric oxide, ferric hydroxide, copper hydroxide, and copper ferro-cyanide. The following substances were formed in non- banded erystalline precipitates:—barium sulphate, strontium sulphate, calcium sulphate, barium oxalate, silver oxalate, thal- lium chloride, bromide and iodide. In certain reactions com- pounds separate out which are not stable at atmosplieric temperatures under normal conditions, e. g. the yellow form of mercuric iodide, which is stable above 126° C., and certain thallium salts. Attention is called to the results obtained by Larsen on cooling a salt solution whose concentration decreases at higher temperatures. Regular layers are formed with increasing dis- tances between them, and having decreasing concentrations, upward (5). With regard to the banded precipitates Hausmann suggests that the compounds are present in the colloidal state before precipitation which only ensues upon super-saturation. By diffusing different chlorides against a constant strength of silver nitrate he showed that the distances diffused were independent of the metal combined with the chlorine, depend- ing only on the silver and chlor-ions, their speeds of diffusion and concentrations. As the concentration of the chloride decreases it is able to penetrate less and less into the (slightly J less than =a) silver nitrate. Weaker solutions of copper and + J. Stansfield—etarded Diffusion and iron chlorides penetrate farther, perhaps because hydroxy] ions are present. Cadmium chloride diffuses more slowly in a stronger solution, possibly due to association. By diffusing different salts of silver into gelatines contain- ing the same reacting salts, in the same concentrations, the reactions were found to proceed at the same speed regardless of the combined ion, showing that the silver ion is the import- ant one for the reaction. Solutions of potassium chloride, bromide and iodide were diffused against silver nitrate, the distances being measured, with the following results: Ratio of distances diffused <2 1:034 Vel. of diffusion 1.0366 Cl AgBr Br —2—— 1:017 ie 5 ea I 1°0185 AglI I NEC Ce Gla Other cases gave equally satisfactory results, while some gave discordant results. Hausmann thus proved that the reaction depends only on the ions involved in the precipitate, and not on those combined with them, and that the speed of the reaction depends on the ion which diffuses into the jelly and not on the one which is present in the jelly. He also confirmed the result of Morse ; distance . : and Pierce (see page 1) that “Se constant. He pointed out that this constitutes the first non-electric method of deter- mining diffusion velocities. In 1905 Bechhold (6) called attention to the importance of rhythmic precipitation in the development of layers of silica, horn, and calcium carbonate in sponges, the development of layered calcareous shells in the perforated Foraminifera, in gasteropod and lamellibranch shells, the development of layers in the bones of some vertebrates, and its importance in intus- susception in plants. Bechhold agrees with Ostwald concerning the super-satura- tion idea. He used ammonium chromate or bichromate in making ring precipitates and found that owing to the fact that silver chromate is slightly soluble in ammonium salts that the bands are more widely spaced. He mentioned the strong con- tractive forces which are set up in gelatine when silver nitrate is added to it, and suggested that these forces may have some- thing to do with the formation of the rings. But this appears not to be the case because under crossed nicols there is not a maximum brightness between the bands. Rhythmic Precipitation. po He found that on diffusing ammonium chromate into a silver solution in gelatine no precipitate forms. In the diffusion of silver nitrate by increasing the ammonium nitrate thicker rings were obtained. In 1907 Liesegang (7) showed that a later set of rings may be formed cutting across an earlier set, apparently without being disturbed or influenced by the rings already present. Close examination shows that the earlier formed rings are extended in comb-like forms to a very small distance by the second deposition. But the “sowimg” action of the first formed precipitates is almost negligible. In 1912 Hatschek (8) produced by diffusion in gelatine, agar- agar or in silicic acid precipitates of calcium sulphate, and car- bonate, barium carbonate, sulphate and chromate, lead chloride, iodide, bromide, chromate, ferro-cyanide, ferri-cyanide, stron- tium carbonate, sulphate, phosphate, oxalate and silicofluoride, copper phosphate, cadininm sulphide, manganese ferro-cyanide, silver bichromate, and silicofluorides of sodium and potassium. Some of these were always formed in layers and some always as macroscopic erystals. According to Hatschek the results are always the same, no matter. which ion diffuses into the jelly containing the other ion. With a view to testing the super-saturation theory of Ost- wald he impregnated a jelly with lead iodide and potassium iodide, diffusing a solution of lead into it. The rings of lead iodide were formed just as though the medium were not sown with lead iodide. He holds that super-saturation cannot be used as a general explanation of the formation of precipitates in layers. The work of Liesegang quoted above would appear to invalidate this conclusion. In 1914 Liesegang (9) claimed that the presence of a small amount of acid and of gelatose is necessary for the formation of the rings in gelatine, none being formed in pure gelatine. If the amount of acid is increased the rings do not appear, so that a definite amount is essential. By increasing the amount of acid present the silver chromate, which is soluble in acid, is enabled to diffuse to a greater distance before precipitation takes place. The result is the bands are more widely spaced, and are thicker. Finally, by greater increase of the acid the silver chromate forms as a continuous mass. Spiral bands were produced, and an examination of some of the older published figures reveals the fact that they are spira! and not actually rings as they were thought to be. Liesegang mentions the presence of spirals on the retina in some birds. When the drops of silver nitrate from which the diffusion proceeds are not circular, cracks appear, cutting the bands, and are free from precipitate. This shows how short spaces may 6 J. Stansfield— Retarded Diffusion and separate the ends of bands without the formation of precipi- tate by “sowing.” The formation of the rings is due to the production of super-saturation at innumerable individual points. Thus the spiral form does not afford any difficulty to the super-saturation hypothesis. Bradford (10) has recently investigated the effect of the by- products of the reactions resulting in the formation of the ring precipitates, and by varying the concentration of the by-pro- duct was unable to affect their formation in any way. Hat- schek had shown that the particles formed in the ring precipitates are larger than those obtainable in aqueous solu- tions. Bradford snggests that this phenomenon is due to adsorption by the precipitate of substance dissolved in the gel, which merely serves to retain the precipitate in place. Bradford finds that the distances between the layers formed are roughly inversely proportional to the molar strength of the reagent. in the gel, and not to that of the diffusing reagent. ‘Tn a more recent paper (11) the same author has developed the idea of adsorption in connection with the formation of banded precipitates. It might be suggested that the term pseudo-stratification would be preferable to stratification as used by him. Retarded Diffusion and Rhythmic Precipitation. By reason of its application to geology the subject of rhythmic precipitation has assumed an importance to geologists which cannot be neglected. In connection with an investiga- tion of Eozoon the writer had occasion to consider the question of the causes governing the formation of rhythinie precipitates. A series of experiments has been made in order -to afford a basis for a comparison with banded structures in rocks. The experiments were carried out using diffusion from a drop placed upon a gelatine layer on a glass plate, as described by Liesegang. The gelatine was prepared according to his direc tions by preparatory washing in several portions of distilled water, to remove some but not all of the impurities, then dis- solving i in distilled water, a convenient concentration for use at a room temperature of 12° C. being 1 gr. of gelatine in 35 ce. water. The solution was used after standing one or two days, reheating the jellied mass to bring it to the ‘liquid state again, care being taken both in the solution and re-solution of the gelatine to avoid over-heating. Care was taken to keep the relative gelatine concentrations the same in any set of experi- ments. This was done by adding 1 ¢.c. of a solution ten times as strong as the one to be investigated to 9c.¢. of the gelatine soln- tion. Thus.a gelatine solution of known str ength was obtained, containing any “desired concentration of the re eagent to be examn- Rhythmic Precipitation. ii ined. This was poured out on to several glass plates and allowed to set, after which drops of the second reagent, of different strengths, were placed upon them. ‘These solutions gradually diffused through the gelatine and precipitates were formed, sometimes continuously, sometimes in separated bands. (See figs. 1, 2, and 3.) It is found that with different concentrations of the diffus- ing reagent (the one in the drop) against a constant concentra- tion of the reagent in the gelatine the diffusion goes at different rates, the stronger solutions diffusing more quickly and to greater distances than the weaker. Again, by keeping the diffusing reagent constant and varying the concentration of the opposing reagent the diffusion is retarded by increasing the opposing concentration, and also the total distance of dif- fusion is decreased. When the molecular concentrations of the solutions approach closely there is a definite limit to the distance to which the stronger solution is able to diffuse. In many of the experiments the gelatine dried before the limit was reached, if any exists, with the solutions of more unequal concentrations, but with the solutions mentioned above the limit was clear, and was reached, in many cases, within a few hours. The following tables for different experiments, showing the distances to which diffusion has taken place with different strengths of the reacting solutions, illustrates the poimt. The figures enclosed in a double line are for experiments using one batch of gelatine and are comparable with each other, but not strictly comparable with those within another double line, by reason of possible differences in the preparation of the gela- tine. The experiments which yielded clearly separated bands of precipitate are indicated by a small ring within the squares belonging to them. It is seen that the production of the bands in gelatine takes place best when a strong silver solution dif- fuses against a weak chromate solution. In those cases which did not show clearly separated bands, the precipitate appeared to be continuous, but examination with the hand lens or some- times with the microscope showed that the bands were present, but that they were more closely spaced, so as to appear con- tinuous. Some cases are clearly banded, to the naked eye, but with the stronger solutions this fine banding of the apparently uniform silver chromate can only be seen under the micro- scope. When clearly separated bands are formed, our experiments have sometimes shown these passing into fragmentary banded layers toward the outside. In some cases this may be quite extensive, corresponding to ten or more bands. In other cases, 8 J. Stansfield—Retarded Diffusion and AgNO; 8N ex | 4n | oN N | N/2 | N/4 | N/6 | N/10 | | | | | 235 | 145 | -09 | 08) 4 nena | N/Q | stipes ee a terry hes ee | | N/4 66 59 ‘425 | -28 2 | | N/6 775 | -705 | 5 | 495 | 258 | ae 111 | -865 | (-875)| 725 | -55 2 1 a N/20 Teayveslaleh Peles Poceaee ory IP sada Ge) en — || N/40 Ape (86s Jolene t-te cao alleeso 18) il |) ena 1A}, Psy | TG aE, eg || Ga | Sr ea | ee 1/80 (1°35). 1°32 0 (1-21), 1-21, 65 | 54 | 81 25 N/100 | 154. | 1-470 ree 128, | 1-270 -84.| 6. | -585| -41 || N/200| e eg! aac 085 1-075.) Ms || ‘N/2 bes | Say ee ari | N/4 | 375 | +365 325 of ll | ‘N/e Az AO ty 4: | 3 19 ae Ree N/10 | -625 | -5 | (-825)) -45 | -ga5 || 18 | 1 ae _— | N20 | 795 | +625 -6 5 2) 295 | :16 | -1 — || N/40 | $ 199 | “Wo | Who | B85) “Blo || “B45 |0 175 | 1 || gee By 755 | “TB | 675] (-76)o || (-445)| -805 | 225 | 2 “y/o | (79)o | -%6o | (710 | -715. |~-ag5 | 385 | 225) 19 | N/100 “85 Be 8, | To] “740 || “S40| 4250 -85 | -28 |i W200] | Ke CrO, | —hythmic Precipitation. 9 this is followed, on the outside, by an area in which the pre- cipitate occurs as a continuous non-banded area containing the precipitate as small granules. In other cases the fragmentation is not so prominent and the non-banded area of “ disseminated ”’ precipitate is followed outward by “secondary” bands of precipitate, which differ from the “ primary” ones within, m that they are evenly spaced and not farther apart, with greater distance from the center, and in being made up of large, irregularly shaped orains of silver chromate, sometimes separate from each other, whereas the “ primary” bands are continuous and consist of exceedingly minute granules of silver chromate. (See fig. 1.) There is always a clear “halo” outside the circle of diffusion which appears to be free from precipitate, but under the micro- scope is seen to contain minute grains of silver chromate. In some cases “subsidiary” bands have been noticed in this “halo.” They correspond to the very fine lines noticed by the earlier workers and ascribed, by them, to the presence of impurities in the gelatine. A set of experiments was made with silver nitrate in the gelatine and potassium chromate as the “dominant” reagent in the central drop, having the higher molecular concentration. The results in this case were remarkably interesting. The fol- lowing table gives the scheme of concentrations used. Numbers 26 to 80 showed only isolated spots or granules formed here and there in the gelatine as the chromate diffused outward. Numbers 21 to 25 showed the beginning of a very indefinite arrangement of the spots in bands. Numbers 16 to 20 showed an inner set of apparently continuous bands consisting of large granules of silver chromate followed by a space with the pre- cipitate as isolated granules, and outside this again the arrange- ment in bands as in 21 to 25, but more clearly shown. (See fig. 4.) The inner bands are less well shown the stronger the chro- mate solution. The outer bands often show discontinuity and Table (p. 8) showing distances diffused in a radial direction, measured in centimeters. The head of the column shows the strength of the silver nitrate solution diffusing and the strength of the potassium chromate solution in the gelatine is indicated at the extreme right. Upper left part of table— distances diffused after 6 hrs. 15 mins. Upper right, after 6 hrs. Lower left, after 1 hr. 50 mins. Lower right, after 2 hrs. The results which show discrepancies are indicated by parentheses. These are due to inaccuracies in measurement. Those which recur in the same positions in both parts of the table are due to incorrect measurement of the size of the original drop, the others to inaccuracies in those individual measurements. The figures on the right side of the table are comparable with each other, and those on the left side with each other, but as they were obtained from experiments using two separately prepared batches of gela- tine, the figures on the left are not comparable with those on the right, except in a ‘general way. 10 J. Stansfield—fRetarded Diffusion and relative displacement of the bands, which are doubtless due to inequalities of diffusion. Both the inner and outer bands are more widely spaced toward the outside. Numbers 12 to 15 show similar bands to those in 16 to 20. Numbers 1 to 11 show three similar areas, but in 1 to 7 there appears to be a continuous precipitate round the drop. This K.CrO, N/10 | N/20 [oS) Co NAG Se rad ~ N/20 4 Di 6 N/40 8 el) Peel Saleh AgNO; |N/80 | 12 | 18 | 14 | 45 N/i00| 16 | 17 | 18 | i9 | 20 | N/200'| <2) 12a) Pastel yet sos 30 ri) lor) vw ~ Ci) (o.6) (OS) [deo 6/400. \ The strength of the potassium chromate in the drop in any experiment is given by the heading of the column in which the number falls, the strength of the silver nitrate in the gelatine being given at the extreme left, thus in the case of number 15 normal potassium chromate diffused into a gelatine containing an eightieth normal solution of silver nitrate. is seen to have bands on the surface as with the stronger silver solutions cited above. N show O The experiments with the weakest silver solutions( that bands are not produced unless a certain concentration is present. From this it may be argued that the areas between the two sets of bands, in numbers 1 to 20, had an insufficient concentration of silver ions to allow of the formation of bands, as aresult of the removal of silver ions by precipitation, the pre- cipitation being followed by decrease of concentration outside the bands as a result of diffusion of silver ions inward to take the | Rhythmic Precipitation. 11 place of those precipitated. The result is somewhat analogous to the thinning of a sheet of rubber by stretching. With the diffusion of the chromat-ions farther outward, however, a sufti- cient concentration of silver ions to allow of the formation of bands will again be met with and the * secondary ” bands will result. A similar explanation would apply to the unbanded layer in results cited above with diffusion of silver against a weaker solution of chromate. (Sce figs. 1 and 2.) It is possible, how- Hines al: Fig. 1. x3. The result of diffusion of 6N silver nitrate into a gelatine containing N/200 potassium chromate. The drop is surrounded by an area of apparently continuous precipitate, which actually shows fine bands on its surface. This passes outwards into the zone in which the bands are clearly separated, and become more widely spaced. Outside this zone is a non- banded zone of granular precipitate, beyond this a zone of granular precipi- tate arranged in evenly spaced bands. This is followed by a clear *‘ halo.” It is probable that the potassium chromate may not have been thoroughly mixed with the gelatine so as to give uniform conditions. At one point the non-banded zone is crossed by a narrow strip of bands. This is doubtless due to the presence of a streak of gelatine richer in potassium chromate than the rest. The unequal extensions in a radial direction of the inner banded zone may be explained in a similar manner. 12 J. Stansfield— Retarded Diffusion and ever, that this explanation is too simple, and that the hydrions present may have to be taken into account. Another question should be discussed at this point. In all the bands and in the non-banded layer in the experiments with dif- fusion outward of chromate, and in the outer bands and the granular non-banded area in the case of outward diffusion of iG: Fic. 2. x5. A portion of the same (fig. 1) enlarged to show the character of the precipitate in the ** granular” zones. silver, the precipitate consists of large isolated granules some- times taking on irregular shapes, in the more pronounced and well-marked bands, probably as a result of accretion of several granules. (See figs. 1 and 2.) But in the “ primary ” bands of the outward silver diffusion the precipitate is made up of exceedingly numerons and very small granules. In watching Rhythmic Precipitation. 13 the development of one of these bands under the microscope it was seen that the very numerous small granules were devel- oped apparently independently of each other. The band grew quite slowly, being extended laterally. As it developed the central part extended itself in advance of the two sides, and MiGs or Fie. 8. x0. Theresult of the diffusion of 8N silver nitrate into a gelatine containing N/60 potassium chromate. It shows the close bands (apparently a continuous mass of precipitate) passing outward into separated bands, which are broken toward the outside, and surrounded by a zone of ‘‘ granu- lar” precipitate and then by a clear ‘‘ halo.” The bands are broken by sev- eral clear channels. The bands on opposite sides of these are not directly opposite to each other, and they bend inwards toward the drop from which the diffusion took place. These may be explained by uneven distributions . of concentrations of chromate or silver ions, or of both. these followed, sometimes more slowly, sometimes more rapidly on the central part. Again, a little distance away from the 14 J. Stansfield—Letarded Diffusion and small cloud of granules, but along the line of the band, another cloud may appear, more thickly crowded with oranules i in the central line of the band. This would develop as before, in both directions, finally uniting with the other advancing part. The precipitate on each side of the central part gradually thickened. Thus, while the result is a series of bands of pre- (OME ae Fie. 4. x21/4. The result of the diffusion of N/10 potassium chromate into a gelatine containing N N/100 silver nitrate. (No. 19.) There isan inner set of close bands round the drop, which appear to be continuous precipi- tate, in the photograph ; outside this zone is one in which the precipitate occurs as isolated granules. and this is fellowed by bands in which the pre- cipitate consists-of separate large granules. Examination with a lens will show this. The outer bands are broken by clear channels, the bands being offset with regard to each other on opposite sides of the channels, in some cases. Both the inne? and outer sets of bands become more widely spaced outwards. cipitate, of apparently striking regularity, yet each separate small portion is deve eloped independently of the rest at such point as the super-saturation boundary is over-stepped. We are able to understand more clearly, now, how the development of spiral forms may ensue, and liow breaks may occur in the bandas, with accompanying displacement of the bands (see’ fig. 4), and also the presence of cracks passing through the r=) . bands, which are free from precipitate, and along both sides of Rhythmic Precipitation. 15 which there may be a bending of the bands. (See fig. 3.) These are doubtless due to irregularities of the diffusion, and perhaps, in part, to irregular distribution of certain impurities or disturbing factors. The quite leisurely deposition of the very fine silver chromate granules suggests that each one is formed almost independently and kept independent, at least for some time, by the gelatine. The experiments of Hatschek and Liesegang investigating the possible sowing action of pre- cipitate already in the gelatine have shown how little this action needs to be taken into account in the formation of the Liese- gang rings. The question as to whether this isolation of pre- cipitated granules is entirely due to the action of the gelatine K, Cr O4. i, next arises. An experiment was made with diffusion of silver into a chromate-bearing gelatine to which a small amount of citric acid had been added. No banded precipitate was formed but a continuous precipitate in the colloidal state. That is to say, by increasing the concentration of hydrions the number of points at which precipitation takes place had been increased and the size of the particles resulting from precipitation had been decreased. (The definition of a colloid suspension being that the particles range in size somewhere between the limits of ‘Lu and ly in diameter.) In the case where the chromate solution is diffusing outwards this action of the hydrions appears to be prevented or very much curtailed, possibly owing to the potassions present. Also in the “secondary” bands of the silver diffusion this action appears to have been prevented. This is also probably due to the “dominance” of potassions relatively to the hydrions in 16 J. Stansfield—fetarded Diffusion and that part of the gelatine. Liesegang (9) has shown how gradual increase of hydrions present in a silver diffusion against chro- mate, in gelatine, results in the formation of broader bands, 1. e., not only increases the number of centers of formation of the precipitate, but also increases the area over which these may form. In a case where diffusion of silver from a drop and of dies. mate from a line drawn around the drop took place through a space free from both reagents, the first precipitate did not form at the point where the solutions first met. The silver diffusion Fic. 6. 0 4D D a is quite visible as there is a clearly marked line in the gelatine at the outward front of the advancing solution. The first formed precipitate was situated about one millimeter within this line, i. e. nearer the center of the silver drop. This would suggest that the halo which surrounds the silver diffusion is an area in which the concentration is below the super-saturation boundary. Rate of Diffusion a Controlling Factor. From the foregoing it is seen that the explanation offered by Ostwald (see page 1) has been confirmed by later work. By the diffusion outwards of the silver and inwards of the chro- mate a depletion of the reservoirs of those ions takes place, i. e., the lonie concentration is gradually reduced, with the result that the silver ions are enabled to diffuse to a greater distance before the precipitation concentration is attained. Lehythmic Precipitation. LT Thus ionic concentration is an important governing factor in determining the distances between the bands of precipitate. Rate of diffusion is also an important factor, which has not been given adequate consideration in the past. Considering the case of silver chromate deposition, the rate of diffusion of the silver solution is greater than that of the chromate solution. So that, after the formation of one band the silver ions pass through it and outward beyond it, or continue onward from DISTANCE Fie. 7. Diffusion of silver nitrate against potassium chromate. dL SN AgNO; / a KE GrO., = 60 II, 4N AgNO; oF do Hie BING OO! 4/ do VA Fe Por Ns do We NEG edo.) ef ~ K2CrQ,. their position in front of it, before meeting the chromat-ions which have not yet crossed the free space. At such point as the concentration is high enough to give labile conditions pre- cipitation ensues. The distance of this point from the last formed band must therefore depend upon the relative rates of diffusion of the two solutions. Thus, if other disturbing con- Am. Jour. Sct.—FourtH SeRiES, Vou. XLITI, No: 253.—January, 1917. 18 J. Stansfield— Retarded Diffusion and ditions, such as gelatine concentration and amount of acid or other solute present, remain constant the distance between the rings depends upon the ionic concentrations and also upon the relative rates of diffusion. It thus appears that if in any reac- tion the increase in distance between the bands due to the effect of the difference of rates of diffusion could be counter-balanced by an equal effect in the opposite direction due to coneentra- tions, the spaces between the bands would remain constant. Fic. 8. DISTANCE. Fig. 8. Diffusion of normal lead nitrate against potassium chromate of different strengths. Agate structure has been ascribed to the formation of Liese- gang rings by diffusion of solutions through gelatinous silica. The discovery of gelatinous silica in a cavity in the Simplon tunnel is of interest in this connection. On examining certain agates the writer was struck by the fact that many of them show bands which are equally spaced and not at successively increasing distances, as in the ordinary Liesegang rings. The writer does not consider that agates in general can be held to be produced by the Liesegang reaction ; out of a large number examined but very few seem to be explicable only along these lines, by far the greater number allowing of other explanations. Each specimen of agate needs to be examined by itself, the history of each one being, of necessity, a separate entity, and only comparable with others by an accidental repro- duction of a similar series of events. Rhythimie Precipitation. 19 Considering the formation of bands of silver chromate in outward diffusion of silver, if the concentrations for precipita- tion are A, for the silver ions and OC, for the chromat-ions and X, for the silver chromate, then for equilibrium A,C, = kX, (1) k will vary with temperature. Consider a space into which diffusion is taking place, just in front of a band of precipitate which is newly formed. The Hireeege DISTANCE. Palettes Fic. 9. Simple diffusion (no reaction). silver ions are “ dominant” and the chromat-ions in excess of saturation of silver chromate have been removed from a cer- tain space on each side of the band. The silver ions being “dominant” the value A, is large compared with C,, so that the silver ions may be regarded as starting their diffusion from the precipitate band (or perhaps some plane outside it), while the chromat-ions begin their diffusion from some plane at a dis- tance D outside the band. (See fig. 5.) It should be noted that in keeping with the generally accepted conception regard- ing aqueous solutions the reagents involved will diffuse partly as molecules and partly as ions. When silver ions are referred to it is understood that each of these is associated with an oppositely charged nitrat-ion the two charges neutralizing each other. For the purposes of discussion of the reaction with the 20 - J. Stansfield—fetarded Diffusion and potassium chromate it is convenient to speak of the silver ions and chromat-ions, but the fact should not be lost sight of that these ions do not diffuse as isolated entities, but each charged ion must be associated with a charge of equal amount and opposite sign borne by some other ion or ions. Suppose that the concentrations of the silver and chromate ions at the planes from which they begin to diffuse are a and ¢ respectively, that the rates of diffusion are V, and V,, and that DIsTANCE Fic. 10. Simple diffusion. a precipitate is formed at some plane P ata distance d from the original silver diffusing front. Then the concentrations of silver and chromate ions at P are T Fas = and noes respectively. But the produet of these two is the precipitation value, so that ee has a value which is susceptible of measurement, because A,.C, = k.X,. (1) Assuming that the distance D remains about the same for two or three consecutive bands, the condition for the forma- tion of a precipitate is from (1) VaViae Sua VT eg oh a (2) a(D — d) or, since V, and V, are constants for given concentrations and for a given temperature, the relation between the initial con- Rhythmic Precipitation. 21 centrations a, and c, and the distances between successive bands is of the form d(D — d) = K.ac_ where K is a constant. (3) It seems legitimate to assume that D varies very slowly over the region oecupied by a few consecutive bands. In these cir- cumstances the variation of @ with the product ac is exhibited graphically in fig. 6, the curve being a parabola having a 1/4D Ks It will be seen from this curve that as the concentration pro- duct a.c decreases, the distance between successive bands will diminish or increase according as d 1/2D so that the usual result is that the bands are formed at successively increasing distances apart. An illustration of this is seen in the inner part of fig. 1. The outer part of the same figure illustrates equal spacing of the bands, which is only rarely obtained. An example of the third case, where the bands become successively closer, is discussed below. (See page 24.) In some of the experiments described above measurements of distances of diffusion were made over an extended period, the results being plotted in the form of curves, with times as abscissae and distances as ordinates. These curves bring out clearly the way in which diffusion is prevented by approach of the molecular concentrations of the two reacting solutions toward the same point. They also show that in those cases where the diffusion proceeds rapidly at first there is a remark- ably sudden drop in the rate of diffusion, and that this drop coincides with an almost uniform distance of diffusion. Some of these curves are reproduced in figs. 7 and 12. A compari- son with figure 12 appears to indicate that the flat portions of the upper curves in figure 7 are due to a change in the vis- cosity of the gelatine. A comparison of the curves for silver and lead solutions against the same concentrations of chromate solution shows that the lead diffuses the more slowly. This is in agreement with the higher rate of diffusion of the silver solution in pure gelatine. From this a consideration of the possibility that the speed of 22 J. Stansfield—Letarded Diffusion and = + Rs ~ g R e, <= = S us uy = = = ft Se = DISTANCE Fic. 11. Simple diffusion of silver nitrate ; N of different strengths. Fie. 12. Reaction 4N AgNO; / 200 K.CrO, in gelantines of different strengths. I. 10 grs. gelatine in 53°5 cc. of gelatine solution. iD do 171 ce. GA do 3d71 cc. do do a i gg. Rhythmie Precipitation. 23 the lead nitrate solution in gelatine may be less than that of the potassium chromate solution led to a measurement of the rates of diffusion of several solutions, in pure gelatine. Well- washed gelatine was placed on glass plates, allowed to set, and drops of different reagents of different strengths were placed uponthem. The reagents diffused outwards. In some cases, e.g. silver nitrate, lead nitrate, ferric chloride, and sodium hydroxide, the diffusing solution had a clearly visible outline in the gela- tine, which is susceptible of measurement. The measurements were made as before, by means of a millimeter scale, estimat- ing to tenths of a millimeter, taking care to eliminate parallax. The diameter of the drop was taken and the diameter of the circle of diffusion (in the case of the precipitates above, to the outer limits of the precipitates). Subtracting the diameter of the drop from the total diameters of diffusion, and dividing by two the radius of diffusion is obtained. It is the rate of elongation of this radius which has been measured and plotted, and which is called the radius of diffusion here. The chromate solutions did not give a diffusing front which was clearly visi- ble and susceptible of easy measurement. The yellow color due to the chromat-ions became gradually fainter away from the drop, coming to an indistinet and diffuse margin. ‘There- fore, in the curves given in fig. 9 those for the chromate solu- tions are dotted, since they are not as reliable as the other curves. The curves (see figs. 9, 10, 11) show that a strong solution of any reagent diffuses at a greater rate than a weaker solution of the same. They show that silver nitrate diffuses more rapidly than potassium chromate solutions of the same molec- ular concentrations. (The crossing of the silver and chromate curves cannot be regarded as established, for the reason stated above. Also the apparent straight lines of the chromate curves cannot be considered final determinations. The sudden stop- page appears to be correct, though this may be due to invisi- bility of the chromat-ion below ‘a certain dilution.) It is certain that the lead nitrate solutions diffuse more slowly than the potassium chromate solutions of equal molecular strengths. It was considered that this pair of reagents, the lead in out- ward, the chromate in inward diffusion, offered a good chance of testing the theory that with such a couple bands of precipi- tate might be formed which would become closer together out- wards, instead of farther apart. Several different strengths of these reagents were tried against each other, the precipitates formed being apparently continuous to the naked eye. In the case of normal lead nitrate against N /200 potassium chromate an apparently continuous precipitate was formed. The hand lens showed no banding, but under the microscope the desired 24. J. Stansfield—Retarded Diffusion and result was observed. Toward the outer part of the precipitate a very tine banding was seen and the bands were closer together outwards. An attempt to reproduce this was not successful, owing to different hygroscopic character of the atmosphere, but a recent paper by Bradford (10) describes an experiment in which the diffusion of N/5 lead nitrate against N /10 potassium chromate, in a agar gel. gave rise to bands of precipitate which became more closely spaced in the later deposited layers. The first eight bands were 1:°15"™ apart, the ninth one being -9™™ from the eighth. No chromate was left between the bands. The same strengths of lead nitrate and potassium chromate used by Bradford were tried in gelatine. Three gelatine layers were placed in a test-tube, the lower one being a N/10 solution of potassium chromate, the middle layer being clear gelatine, and the upper one a N/5 solution of lead nitrate. Another test-tube was prepared with the same solutions reversed, the lower one being lead nitrate-and the upper one potassium chromate. In both cases the chromate solution diffused the more rapidly so that the first formed precipitate was nearer to the starting point of the lead solution than to that of the chrom- ate solution, while in both cases the later formed precipitate was on the side remote from the lead solution. With the chromate solution above the precipitate was a continuous mass, but when the lead solution was above the precipitate was formed in bands, which became closer and closer together, pass- ing into a continuous mass. These experiments were repeated with the same results. This illustrates a controlling effect of gravity, which does not come into play when the diffusion takes place horizontally, as in the glass plate experiments. Miscellaneous Diffusions. In comparing the rates of diffusion of ferric chloride and sodium hydroxide the observation was made that ferric chloride possesses in a marked degree the property of inducing coagula- tion in gelatine, a property possessed by silver and lead nitrate solutions in a much smaller degree. This coagulant property of ferric chloride is well known. In comparing the rates of diffusion of the two reagents it was found, contrary to expecta- tion, that the sodium hydroxide diffuses much more slowly than ferric chloride. (See fig. 10.) But the diffusion of the sodium hydroxide is not a simple case of diffusion. A reaction takes place with the gelatine, and this is at once apparent in the unexpected nature of the curve. Diffusion of normal ferric chloride against N/100 and N/200 sodium hydroxide—/stronger solutions prevent the setting of the gelatine)— gave a colloidal precipitate of ferric hydroxide. Lehythmie Precipitation. 25 This is surrounded by a broad clear zone marked by a clearly visible circumferential line. The nature of this zone is un- known, at present. A set of nine diffusions was carried out with 5N, 2N and N potassium iodide against N/20, N/40 and N/80 lead nitrate, and another set of nine with 2N, N, and N/2 lead nitrate against N/20, N/40 and N/80 potassinm iodide. Probably owing to the state of the gelatine (not thoroughly free from acid) the precipitates were colloidal and not visibly banded except in the cases of 2N and N potassium iodide against N /20 lead nitrate, which showed fine rings to microscopical examina- tion. They all showed that the diffusion proceeds more rapidly with greater difference of concentrations of the two reagents and all showed a broad indefinite band around the drop in which the precipitate was thinner, and a broad band on the outside of it in which the precipitate was thicker. This was especially marked im the cases where potassium iodide was the diffusing reagent (the one in the drop). It was found that several of these plates developed a banded structure i1mme- diately npon being immersed in a solution of sodium hydroxide, which was used to clean the plates after they had dried. Thus, although the precipitate is present in colloidal form it is arranged in layers or bands, though these are invisible until some such reaction as that described shows their presence. Effect of Different Gelatine Concentration. The reaction 4N silver nitrate against N/200 potassium chromate was carried out in gelatine of different strengths, obtained by dilution of the one stock. Rings were formed, and the rates of diffusion were measured. The plot of these results confirms the result of Hausmann (loe. cit.) that increase of the gelatine concentration retards the diffusion. The curves are almost parallel and the rates differ so slightly that the curves only begin to diverge very slightly toward the right. The measurements recorded in these curves are only rough, but they suggest that more careful measurements may be able to detect a numerical relation between the concentra- tion of the gelatine and the amount of retardation. Summary and Conclusions. After giving a historical account of the work of earlier writ- ers, the results of certain experiments are given, the main points of which may be summarized as follows: 1. The rate of diffusion of a reagent of given strength is retarded by increasing that of the reagent in the gelatine. Also, the total distance to which diffusion takes place is de- 26 J. Stansfield— Retarded Diffusion. creased by a similar increase of strength of the reagent in the gelatine. 2. With too close approach of the concentrations of the two reagents, silver nitrate and potassium chromate, a continuous precipitate is formed, bunt under the microscope, the surface of this is seen to be finely banded. These fine bands follow the same rules as regards spacing as are found with the separated bands. 3. Separated bands are best produced by the diffusion of a strong silver solution against a weak chromate solution. (For other reagents, these conditions may not be universal.) 4. Similar results are obtained by diffusion of a strong chromate solution against a weak silver solution. But the particles formed are larger in this case. 5. Increase in distance between bands has been held to be due to progressive dilution of the reagents. Rate of diffusion is shown to be an important controlling factor, and that under certain conditions the bands may be equally spaced, or may be spaced at decreasing distances. These conditions are discussed. 6. The rates of diffusion of different reagents in pure gela- tine are found to differ; thus potassium chromate diffuses more rapidly than a lead nitrate solution of the same molecular con- centration, and a silver nitrate solution more rapidly than a potassium chromate solution. Also, a strong solution of any given reagent diffuses more rapidly than a weak solution. Bibliography. Liesegang, R. E., Chemical Reactions in Gelatine. Dusseldorf, 1898. . Ostwald, W., Zeitschr. phys. Chem., vol. xxiii, 1897, p. 365. Morse, H. W., and Pierce, G. W.. ibid., vol. xlv, 1908, p. 589. Hausmann, J., Zeitschr. anorg. Chem., vol. xl. 1904, p. 110. Larsen, Ann. d. Phys., vol. iv. folge 9, p. 1186. Bechhold, H., Zeitschr. phys. Chem., vol. lii, 1905, p. 185. Liesegang, R. E., ibid., vol. lix, 1907, p. 444. Hatschek, E., Zeitschr. Chem. Ind. Koll., vol. x, 1912, p. 124. Liesegang, R. E., Zeitschr. phys. Chem., vol. Ixxxviii, 1914, p. 1. . Bradford, S. C., Sci. Prog., vol. x, No. 89, 1916, p. 369. . Bradford, 8. C., Biochem. Jour., vol. x, 1916, p. 169. FG Sas EN ae de mes en le ld Ee W. G. Mixter—Calorimetry by Combustions. 27 Arr. Il.— Calorimetry by Combustions with Sodium Peroxide; by W. G. Mixer [Contributions from the Sheffield Chemical Laboratory of Yale University. | Fusion with sodium peroxide is the only way known for finding the heat of oxidation of elements which do not burn in oxygen and which form oxides insoluble in acids. The method is adapted to the determination of the heat of forma- tion of the oxides of a metal and also the heat of combination of metallic oxides with sodium oxide. The writer has used the method the past ten years and can now describe it more com- pletely than was done in any of his papers. The method is indirect and the heat effect sought is not the observed effect ; hence burning in condensed oxygen is prefer- able where possible. For example, when carbon is burned with sodium peroxide the observed heat (z) is the result of the following reaction 2Na,O, + C= Na,CO, + Na,O =a and # equals the heat of formation of carbon dioxide plus the heat of combination of carbon dioxide with sodium oxide and less the heat required to separate two atoms of oxygen from two molecules of sodium peroxide, thus «= C + 20 + (Na,O + CO,) — (2Na,O0 + 20) and C +20 =x — (Na,O + CO,) + (2Na,O + 20) Moreover, many substances do not give with sodium peroxide sufficient heat to fuse the mixture and hence some readily com- bustible substance, such as sulphur or carbon, must be added, which gives in many cases the larger part of the total heat effect. As yet we have only a few results obtained by fusion with sodium peroxide to compare with those by other methods. They are Sodium peroxide method Other methods C + 20 = CO, + 96°4* 94°7* Ti + 20 = TiO, (amor.) 215°6¢ TiO,(crys.) 218°4¢ 38Ke + 40 267.5} 265°2] 2Na +8 + 40 326°78 328°6| 3Na + P + 40 451°4 452°4 * This Journal, xxix, 130; ibid., xix, 484. t+Ibid., xxvii, 343. ¢ Ibid., xxxvi, 55. § Ibid., xxvi, 125. . || Thomsen. 28 OW. @. Miater—Calorimetry by Combustions The values for G + 20 = CO, + 94:7, Ti + 20 = TiO, erys. + 2184 and 3Fe +40 = Fe, 0. crys.) + 265-2 were obtained by combustion in oxygen. The other values are derived and have the mean error of several experiments. Both values for C + 20 are for acetylene carbon. One reason for the higher value found in the sodium peroxide method is this : The car- bon and peroxide were mixed in a mortar, thus allowing the peroxide to absorb a little moisture which added to the heat of the fusion. The amorphous TiO, used in the experiments which gave 215°6 was heated to redness for an hour. Appar- ently it has nearly the same heat of formation as the crystalline form. The value 267-5 for 3Fe + 40 is derived from the results of fusions of iron, ferrous oxide, ferric oxide and the mineral magnetite respectively with sodium peroxide, and 265:2 was the result of burning iron in oxygen. The result for 2Na + 8 +40 is derived from the heat of the reaction of sul- phur with sodium peroxide and the heats of formation of SO, and Na,O. The value 451:4 for 3Na + P + 40 was derived In a similar way. 452°4, given in the Physikalisch-Chem- ische Tabellen is derived from Berthelot’s data. Sodium peroxide absorbs water rapidly from the air and hence it should be exposed as little as possible as the hydrated peroxide will give more heat with a combustible than the anhydrous. One of two samples which gives off the less oxy- gen when fused is the better one. The error from water con- tent is small in good peroxide especially when carbon, for example, is added to make a-mixture fuse because the heat effect of the carbon has been found for the carbon and perex- ide used. The writer has obtained peroxide in pound packages, containing according to the seller 92 to 95 per cent of Na,O,,. To ensure uniformity in composition four or five pounds are mixed thoroughly in a large stoppered jar. Then it is rapidly placed in about half pound lots in flasks with necks which are narrowed to half an inch, and the necks are drawn off and hermetically sealed. The flasks containing peroxide should be stored in a metal box as a precaution against fire in case of breakage. For convenience in using, the peroxide is put into an eight ounce bottle having a smooth, unground neck with a smooth red rubber stopper. Various substances may be added to a peroxide mixture to increase the temperature of the fusion. The writer has used successively acetylene carbon, sulphur and lampblack. Pure rhombohedral sulphur in fine powder would appear to be the best of the three, but it becomes electrified when shaken in the bomb with the other ingredients and sometimes sticks to the bomb and is not completely oxidized. Sulphide is formed and occasionally free sulphur is left. When the bomb is much Ee se EEE with Sodium Peroxide. 29 blackened by a fusion with sulphur the heat result is low. Acetylene carbon is the ideal substance to use, but difficult to obtain. It is constant in composition after heating to expel hygroscopic moisture, and any unburned is not dissolved when the fusion is treated with water or acid and may be collected on a Gooch filter and weighed. Commercial lampblack nearly ash free is prepared as follows: It is heated for two or three hours to 1000° or higher, sifted when cool and shaken in a large bottle to ensure uniformity in composition. A portion for a calorimetric experiment is heated in a crucible, best in an electric furnace, until the top of it is a faint red to expel moisture, then allowed to cool in a desiccator. Finally it is quickly weighed and placed in a bomb. As the lampblack changes in composition with the intensity and time of heating, eare should be taken to heat the different portions used uni- formly. Lampblack gives a quicker combustion, often of explosive violence, than acetylene carbon or sulphur. Thus far the writer has found no lampblack left in fusion with sodium peroxide. One part of carbon requires 13 parts of pure sodium perox- ide for combustion and it is best to take about 20 parts in determining the heat effect of the carbon or lampblack. For the combustion of sulphur doubie the calculated amount of peroxide should be used. Oxygen is often evolved in a com- bustion from the action of an acidic oxide on the sodium per- oxide and the heat required to set it free from the peroxide is added to the observed heat. This correction, 1:73 g—cal. for 1° of oxygen at 0° and 7607", is derived from Beketofi’s Na, + O = 100-26 Cal. and de Forcrand’s Na, + 20 =119°8 Cal. The writer has tested and found no carbon dioxide in the oxygen given off in considerable quantities from fusions of mixtures of sodium peroxide, an acidic oxide and lampblack. It is best to use large quantities of substances in calorimetric determinations, not only because the errors are less, but because a large fusion remains liquid longer than a small one and hence the combustion is more likely to be complete. A mix- ture of 20 to 50 grams giving a heat effect of 10 to 20 Cal. answers well. The bomb, fig. as is sterling aieee It is 3/32 in. in thick- ness. The inside diameter at the top is 15/8 in., at the bottom 14/8 in., and the length not including the top is 3 le Amat Tt iS slightly conical for convenience in fitting the expanded top of the cup a, to make a dust-tight joint. A fusion in the cup cools more slowly than when in contact with the cold bomb and hence the reaction is more complete. The cup is tine sil- ver and weighs 15 to 30 grams. The top and fittings are brass. The top is 23/4 in. in diameter and 5/8 in. ~ thick 30 W. G. Miaxter— Calorimetry by Combustions Figs. 1 and 2. 9 with Sodiwm Peroxide. Si except the rim which is 1/4 in. The gasket slot in the top is 5/32 in. wide and 1/8 in. deep. It should fit the top of the bomb so that the lead gasket will not flow under pressure. The gasket is easily made by placing a disk of lead 1/25 in. in thickness on the bomb and then pressing the top into place by means of the screws. There are eight screws 5/16 in. in diam- eter. Four are sufficient except for high pressures. The screws should turn easily with the fingers and are best greased with tallow. A 4 in. solid wrench is a convenient one for tightening them. The tube 0 hasa length of 7 1/2 in., includ- ing the screw ends. The narrow part of it is 3/16 in. external and 1/20 in. internal diameter. The lower screw joint of 0 is “made tight with soft solder, and for the upper one joining the valve shown in the fig. 3 sealing wax answers. The tube c¢ is soldered in the tapering hole in the brass top. It is shown full size in fig. 2. The insulated rod in it has a small short tube on the lower end for the plug which fastens the iron wire d@ of fig. 1. The upper end of the tube has a glass tube 1/2 in. in length and is packed with dental phosphate cement. The mid- dle of the tube is filled with a flexible cement of caoutchoue and beeswax, and the rest with phosphate dental cement. To prevent the fusion, in case it is thrown against the top of the bomb, from closing the hole in the tube 0, it is covered by a thick disk, e, of pure silver, which is held in place by three fric- tion lugs. The ignition wire d should weigh at least 20 mlg. If less is taken it should be weighted with a bit of silver, other- wise the oxide formed when the iron burns will not drop off into the peroxide mixture. The sterling silver bomb weighed when made 472 grams and after eight years’ use 465 grams. The loss is due to corrosion, especially by sulphur, and to polishing. The total weight of the brass work excluding the thick top of the tube c is 397 grams. The lead gasket weighs 10 grams. The writer has two nickel-plated German silver cans for holding the water of the calorimeter. The smaller can measures 5 1/2 in. diameter, has a depth 9 1/4 in., and weighs 50°6 grams. The dimen- sions of the larger one are 61/2 in. and 9 1/4 in., and it weighs 51°2 grams. The water equivalent of a calorimeter and can may be calcu- lated from the specific heats of the metals in it, or may be deter- mined by the method of specitic heat. By the latter way 285 and 281-2 grams were obtained for a steel calorimeter and can; calculated 284-7 grams. ‘The specific heat of the metals are quite accurately known, hence a calculated hydro-thermal equivalent of a calorimeter is likely to be more accurate than an experimental one. * This Journal, xix, 425. 32 W. G. Mixter—Calorimetry by Combustions The apparatus shown in fig. 3 is designed for use in a room of varying temperature. A is made of very thin tinned iron (sheet tin) or tin foil, and B is a copper tank, holding in the annular space about 20 liters of water. It is tinned on the sur- face opposite A, and has one hole in the top for the stirrer, one for a Fi aométer and another for adding water. The EiG. 3: AN (Zaz ‘ 1 SSS NY | il HOP Lee Pe Hy ‘Il TION a rs calorimeter can E is supported by the wooden ring F. The cover © is made of two semi-circular pieces of wood. The wooden parts are varnished with shellac. The rod D rests in a cup and is inclined about an inch so that the propeller will give a rotary motion to the water. The bulb G is for collect- ing gas that may be given off in an experiment. It has a capacity of 500 to 700° and is connected with the bomb by,a small, thick-walled rubber tube and with H by a large rubber tube. Both rubber tubes should be securely fastened by wir- ing. The lower stop-cock is large so as to allow the water to flow rapidly between the bulbs. If the room and apparatus are colder than desired the latter may be warmed by a lamp flame with Sodium Peroxide. 33 against B while the water in it is stirred. When the jacket water is rapidly raised 4° or 5° it will be some time before the empty can E is warmed approximately as much. To save time E may be heated by putting into it a closed flask of hot water. If the can E is colder than the water in B when the calorimeter water is cooler than the jacket water, the tempera- ture of the calorimeter may not rise at first and will not rise regularly for some time. And it is better not to let the water in the calorimeter remain long before a combustion on account of loss by evaporation. Manipulation—The sodium peroxide (weighed in a glass- stopped weighing bottle)-and the substance to be burned are placed in the open bomb, which is at once covered with a plate glass cover to keep out moisture from the air. Then the cover is clamped to the bomb by two screws and a wooden piece with a bit of rnbber under the middle of it. The ingredients are thoroughly mixed by shaking and then the cover is replaced by the top of the bomb. Before tightening the screws the air is displaced by passing about 200° of dry oxygen* through the tube 0, fig. 1. Next the screws are carefully tightened so as not to strain them and the valve is closed. If, however, oxygen from the fusion is to be collected it is left shghtly open so that gas may pass slowly. Then the bomb is adjusted as shown in fig. 8, and the required amount of water is poured into the water can. The stirrer is started and the temper- ature is noted each minute. When it is rising regularly the mixture is ignited by a current passing four 32 candle- power lamps. The temperature usually falls regularly after thirteen minutes and is observed six minutes longer in order to get the rate of fall. Finally the bulb containing the oxygen set free by the fusion is disconnected and the bomb opened and placed in a beaker of water. Rapid evolution of oxygen shows that the mixture contained an excess of peroxide. After the fusion has disintegrated the bomb is removed from the beaker. If an hydroxide insoluble in water is formed it is dissolved by nitric, acetic, or hydrochloric acid as may be best. If some unburned substance remains it is collected on a Gooch filter, washed with water and then ammonia to remove any silver chloride present and its weight is found. If any gas is col- lected it is brought to atmospheric pressure and known temper- ature, then the stop-cock of the bulb G, fig. 3, is closed and the rubber tubing removed. The weight of the bulb full of water, less the weight when partly filled with the oxygen collected, equals the number of cubic centimeters of the gas. The weight of the oxygen is found in the usual way. The silver cup a, fig. 1, is usually easily removed from the bomb after an experiment. Sometimes it is necessary to heat * Tron wire burned in air does not always ignite the peroxide mixture, Am. Jour. Sci.—FourtH SERIES, Vou. XLII, No. 253.—January, 1917. 3 34 W. G. Mixter— Calorimetry by Combustions. the bomb in order to loosen it. The bomb and cup are cleaned with strong hydrochloric acid, washed and polished. The top is cleaned in the same way but does not require polishing. The silver cup is often partly melted by a sodium-peroxide combus- tion and a number should be provided. The top of the cup is easily expanded, if desired, by pressing it against the hemis- pherical bottom of the bomb. The following experiments* illustrate the use of the sodium- peroxide mixture: iron! pooner eee 1°754 1754 3°000 grms. Suliphune: pete eae 1°500 1°500 1°500 1°500 Sodium peroxide..-- 20° 19° 20° 22° Water equivalent of SYSUCIM tac eee oe 3080 3078 3188 4108 Temperature interval 4:016 3°573 3°423 3°180° Heateect) 2a) 2-4 12369 10998 10912 13063° Heat effect of sulphur — 7905 — 7905 — 7905 — 7860 3 Ogishke conglomerate « quartzite, Hemlock, volcanics quartzite 3 48 = oh ee Be | She great Serves == yl a rt Unconformity {| Unconformity Unconformity Unconformity Unconformity Uneonformity 3 a mee | ny i ee a SS | = qe Steeprock Series O " ' 7 —— a 23 a . 5 lle uartzite ewe slate, ruce Horne Dae) a o | Bruce ¢ Voleanics, limestone (fossils), Eo rey (a Yale 2 ) ime stone Bad River limestone, | Randville dolomite, Randville dolomite, Kona dolomite, Quartzite, limestone, 1@) 4 (Epoch, Sories) quartzite, conglomerate Sunday quartzite Sturgeon quartzite Sturgeon quartzito Mesnard quartzite graywacke, limestono, i conglomerate, qu 5 =a 8 EPILAURENTIAN INTERVAL Major unconformity | Major unconformity Major unconformity Major unconformity | Major unconformity Major unconformity Major unconformity Major unconformity Major unconformity Major unconformity Major unconformity Major unconformity Major unconformity 4 LAURENTIAN, REVOLUTION Granite gnolss Granite gneiss Granite gneiss Granite gneiss Granite gneiss Granite Granite gneiss Granite gneiss Granite gneiss Granite gneiss Granito gneiss Granite gneiss > | Grenvitte iI ia | Grenville Cy mest Granville § (Epoch, Series) a |iE =e a el een sto a en oO 5a KKEEWATIN Keewatin Keewatin Keowatin Keewatin Keewatin Keewatin Keewatin eee we Keewatin Keewatin Keewatin aed bo (Epoch, Series) Sche $ 3 ZS Ile chis (Cs. . teal. eat [ cal ES t—Cormmrations— —great)chemicat-conce N 5 Courentenine Contchichin | 1 Yio Oreq wnat red fo a ’ ror + & written in place of rX in the table. The transformed table can then be used as in eases Land II. For a certain flint glass spectrograph k = -09 p, and for a small quartz spectrograph & = — ‘02 u. If great dispersion is obtained by means of a train of prisms the table may not be carried to a suflicient number of figures. A suitable table can, however, be written easily in the same way this table was written. The departures of the true values of X from those obtained by means of the linear equation and the table are less than the departures from Cornu’s (or Hartmann’s) formula. Suppose there are so few comparison lines (less than 4 or 5) on a spectrogram that a sufliciently accurate departure curve cannot be made. By means of the table a much more accurate interpolation can be made by using the table and the following formula* than can be made by means of Cornu’s formula. Suppose we have Q,, A,, A,, d,, d,, d, for 8 lines, we read from the table »’,, »’,, 2’, corresponding to 2, etc., and write ae pnt Oy ae ee uw, — 1, rv — n', and solve for C. 7 Then put d, (known) and »’, (unknown) in place of d, and n, and we have a on =n : : mene GET cl ar Finally read off X, corresponding to 7’,. *See J. Wash. Acad. Sci., iv, 467, 1914. oe — Je) SOMVIHWNP WWE STS OSDVIHOMRwWMwDH ae o|~ =) w > CORIAackWwWworH HT. F. Merwin—Interpolations on Spectroyrams. TABLE I, Oe lage Quin 8G 4 2 bonne hens By, 9 oo = J ! “Wie? D1 59, H. £. Merwin—Interpolations on Spectrograms. A 0 1 2 3 4 5 6 ri 8 9 Diff. 236 °21 828 805 783 760 7388 715 693 671 648 626 aH 603 581 559 5387 515 492 470 448 426 404 22 38 aoe. S00) 358 ol6 29) 27a. mone ee) 208 186 . 39 “21 165: 143° 122 100! 079: 0b 036" 015 994 - oe 240 °20 951 929 908 887 866 845 824 8038 782 ‘761 21 41 740 719 699 678 657 686 616 595 574 354 42 588 512 492 471 451 480 410 390 369 349 43 829 3808 288 268 248 228 208 188 168 148 20 44 20 128 108 O88 068 048 028 009 989 969 950 45 19 980 911 891 871 852 832 818 798.774 754 46 735 716 696 677 658 639 620 600 581 562 Aq 548 524 505 486 467 449 480 411 392 373 19 48 B04 800 “oli. 290 (210. 26 242 224 20ae ter, 49 19 168 150-132 "4138 095. 077> 058 040. “0200s *250 18 985 967 949 931 918 895 877 859 841 823 18 51 805. .%87 --769- Vol F3838> 715-- 698 680). 662. G45 52 627 609 591 574 556 539 521 504 486 469 58 451-434-417. -399- -382°- 36b. 347-7330) 313,296 54 219) 2612. 244 OP S10. 19S al 7 159 142 125 b kr 5d 18 109 092 075 058 041 024 O08 991 974 957 56 -17 941. 924 907 891 874 858 841 .825. 808: 792 57 V6. 759° 743. “727-710 -. 694 677 662 645 ees 58 =~ 613 5986 ~580.~ 564 ~548) 9582)" 516 500 48245 468 16 59 452 436 420 404-388 (372 301 S41) lo2er 30S 260 293 a 278. eeG2tse40 2 col 21d) 200 5 184.268 158 61 197.137. 129°. 106.091 -075 060: ..045. 029) 014." 554 62 16 984 968 958 9388 922 907 892 877 862 847 63 S82 Mel 7. BOL tA TOr ea pie I TAO eee aaa a 15 64 682 668 6538 6388 623 608 594 579 564 549 65 5384 520 505 490 476 461 447 432 418 403 66 388 374 359 345 3381 316 3802 288 2738 259 67 245 231 216 202 188 174 159 145 1381 117 68 "16.108 089 075 061 047 033 019 005 991 977 14 69 15 968 949 985 921 907 894 880 866 852 888 270 S207 ell F907. A840 ad ia 1A fon omens val 689 675 662 648 635 622 608 595 582 568 ie 555 541 528 515 502 489 475 462 449 4386 73 492,409. 3896) 2aGae o10n a0) 644. oo lanolomoon 74. 292). 279: 266%2.200) 4240 nected ele. 2201, ASSenre 13 "5 162 150 137 124 111 099 086 073 060 048 76 15 035 022 009 997 984 972 959 946 934 922 hh 14 909 897 884 872 859 847 834 822 810 797 % 780° 778. 761 748 786" 724" 711 699 Gers ovo 79 663 650 688 626 614 602 590 578 566 554 -280 542 580 518 506 494 482 470 458 446 484 12 81 422 410 399 387 3875 363 351 339 327 316 82 304 292 281 269 257 246 234 222 210 199 83 187 176 164 155 141 130 118 107 095 084 84. ‘14 072 061 049 O88 027 016 004 993 981 970 85 ‘138 959 947 9386 925 914 9038 891 880 869 858 86 847 835 824 8138 802 791 780 769 758 747 87 7386 725 714 '708 692 681 670 659 648 637 11 88 626 615 604 594 5838 572 561 850 539 7528 89 518 6507 496 -485 475 464 454 448 432 421 *290 411 400 390 3879 369 3858 348 337 3827 316 91 306. 290... 28) ..2744.264 2253) 6243 eae eects 92 201.191.180.170 260.150 139" 1292 19-108 HI, EL. Merwin—Interpolations on Spectrograms. 53 vy 0 i a 3 4. 5 6 7 8 9 Diff. °293 13 098 O88 O77 O67 O57 047 0387 027 O17 006 94. 12 996 986 976 966 956 946 986 926 916 906 10 95 896 886 876 866 856 846 8386 826 816 806 96 TIO MISCO 6G Ot MAD ATS Ee 12 bo Lie 708 7 698 688 678 669 659 649 640 630 620 610 98 601 591 581 [2 O02. Oda. DLoMnotn oes ole. 99 5U5 495 486 476 467 457 448 488 429 419 “300 AAO A400" 39h 382" 372° "368 800" 13044 aan oeD O01 816 307 297 288 279 270 260 251. 242 238 02 224 214 205 196 187 178 168 159 150 141. 03 132 123 114 105 096 086 O77 068 059 050 04. "12 041 032 023 014 005 996 987 978 969 960 9 05 °11 952 948 984 925 916 907 898 889 8s0 872 06 863 854 845 837 828 819 810 801 793 784 O07 ETO GOO FOL 749) 74034.782 0 728 6 7140706 * 69% 08 689 680 671 6638 654 646 637 629 620 612 09 603° )95 7 580-078 569 O61. O52 44 535527 -310 518 510 502 498 485 476 468 460 451 448 11 435 427 419 410 402 398 385 377 368 360 12 Boe petds Bon Ofte elo oll 302.294) 286 278 13 DO Ol 263.245: 237) § 2998 2202219) 2044.196 14 LSS LOO 721164 156 A468 140 1388" 124-916 8 15 108 100 092 084 076 069 061 053 045 037 16 “11029 021. 018; 005: 997, 990. 982.974 966) 958 LT "10 950 948 935 927 919 912 904 896 888 880 18 873 865 857 850 842 8384 827 819 811 803 19 TOC S60 we 1 OU) eel (ond GO OS miO0 Cel 43 oO l8 *320 720 713 705 697 690 683 675 668 660 653 a1 645 638 630 622 615 607 600 5938 585 578 22 570 565 556 548 541 534 526 519 511 504 23 497 489 482 475 467 460 458 446 488 481 24 424 417 409 402 395 388 3881 38738 366 359 25 aoe 044 837 3830 3823 316 30609 302 295 288 26 ZO he cori coG ) 2902524245 1 238.1. 281) 22204 217 On 210 203 196 189 182 175 168 161 154 147 7 28 140 1383 126 119 112 105 098 092 O85 O78 29 071 064 057 050 043 036 0380 023 016 009 °330 10 002 Ba 10° 002 934 867 801 735 670 606 542 479° 416 68-62 *34 09 354 292 231 172 112 053 995 937 880 823 62-57 “By 08 766 710 655 601 546 498 440 387 335 283 56-52 36 08 282 181 131 081 081 982 933 885 837 790 51-47 “BT 07 743° 697 651 605 559° 514 470 426 382 339 46-43 °38 07 296 253 211 169 127 086 045 005 965 925 43-40 °39 06 885 846 807 768 730 692 654 617 580 543 39-37 “40 5OT ATi: 435, 1399 © 364.338 294 209 225 191 36-34 “41 06 157 124 091 058 025 992 960 928 896 864 34—32 *42 05 8838 802 771 741 710 680 650 620 591 562 31-29 “43 0383 504 475 447 418 390 362 335 307 280 29-27 "44 05 2538 226 199 173 147 120 094 069 043 017 27-25 “45 04 992 967 942 917 893 868 844 820 796 772 25-24 “46 749 726 702 679 656 633 610 588 565 543 24-22 Ai Deir 499 (477 24002 434 4191391) - 370! - 349 328 22-21 “48 307 286 266 245 225 205 185 165 145 125 21-20 “49 04 106 O86 067 048 029 010 991 972 954 935 20-18 50 03 917 898 880 862 844 826 808 791 7738 755 18-17 Soil! 738 721 704 687 670 658 636 619 608 586 17-16 540 OH. EB. Merwin—JInterpolations on Spectrograms. - a 0 10 DOR BCR Bea hae Apr ee eae Diff. "02 ‘03 570 553 5387 521 505 489 473 457 442 426 16-15 08 410 3894 3879 364 3849 334 319 304 289 274 15-14 "O4 209 244 230 216 201 187 172 158 144 180 18-14 uay9) ‘03 116 102 O88 O75 061 047 038 020 O07 993 14-13 “06 "02 980 967 954 941 928 915 902 889 876 863 13 5 851 888 826 813 801 789 776 764 752 740 58 727 715 708 691 680 668 G56 644 6338 621 12 59 610 599 587 575 564 553 542 531 519 508 60 497 486 475 465 454 448 432 421 411 400 11 61 390 380 369 359 349 388 328 318 307 297 62 987 277 967 B57 947) O87 987 918 208198 68 188 179 169 160 150 140 131 121 112 102 64 093 084 075 065 056 047 088 029 020 O11 65 02 002 993 984 976 967 958 949 941 932 993 9 66 ‘01 914 906 897 889 880 872 863 855 846 838 67 830 822 813 805 797 789 781 772 764 756 68 748 740 732 724 716 708 701 69% 685 677 8 69 669 661 653 646 638 631 623 616 608 601 70 593 586 578 571 563 556 549 541 534 527 71 520 513 506 498 491 484 477 470 463 456 "2 449 442 485 428 421 414 407 400 393 386 7 7 380 373 366 359 352 346 339 332 326 319 74 313 306 299 293 286 280 273 267 280 254 5 247 241 285 228 222 216 209 203 196 190 oF 184 178 172 165 159 158 147 141 135 129 ay 123 117 111 105 099 093 087 081 075 069 6 ‘78 063 057 O51 045 039 033 027 022 016 010 +h0 “01 004.:2998 4992) 987s. 98 e016 229710 962 950 esas ‘80 "00 947 942 9386 931 925 920 914 909 903 898 DEVELOPMENT OF THE FORMULA. Consider a narrow beam of light bent twice in the same direction and dispersed by a very small prism and formed into a spectrum which is photographed withont intervening lenses on a flat plate. Then if 7 is the angle of incidence upon the prism, 7 the angle of refraction from the prism, A the angle of the prism, 2 the refractive index of the prism, 2’ and 7” the angles of incidence and refraction within the prism, 8 the angle the emergent ray inakes with the photographic plate, d the distance on the plate from the normal to the back face of the prism to the image of the emergent ray, A etc., constants, Then siny = 2sin?@’, and 2)? =A —7’ Then sin 7 = 2 sin (A — 7’) Or sinz = n(sin A cos?’ — cos A sin 7”) sin Z ve sin?’ But sin, ame cosy aye n n Then sin r = sin A(4/x* — sin?i’ — sin Zz cot A) (1) ; d sin Also sin7 = : (2) A H. FE. Merwin—Interpolations on Spectrograms. 55 Therefore d= A ia al /n? — sin*t — sin ¢ cot A) (3) sin B A! , => which may be written d= “ey (fn? — B—C) (4) There are three conditions which obtain in the practical application of equation 4 to spectrograms: first, 6 is not far from a right angle (80° to 100°) ; second, n* is between about 24 and 3:0; third, the ideal relations assumed in obtaining equation 4 are not significantly violated by the optical system of a spectrograph. Then we may write as a first approxima- tion* sinB =1,and Wn? — B= Fr? —G which gives d = A'(Fn? —G —C) or d= Hv — K (5) The relation between refractive index and wave-length for wave-lengths that are freely transmitted is given by the equa- tion 2 b 2 Wi ee Bite arrange ee (6) Then fromiand6 d@=—Ah+ K(s— — en’) (7) The quantity in parenthesis, call it »’, can be evaluated and tabulated for various values of A by taking the constants already known. Then we may write d= Kn’— WR (8) That is, d,= Kn', —f (8a) However, as shown later, not the exact values of 0, ¢ and e, but such approximate values as are best adapted to the easy cal- culation of values of »’ need be used. For quartz each of these constants is nearly ‘01. This value was used in com- puting the accompanying table by means of Barlow’s tables of squares and reciprocals. These values of n’ are so near those for flint glass that the tables may be used in interpolating spectrograms from glass prisms. In most cases interpolations made by using formula 8 in con- nection with the tabulated values of A and 7’ will be as exact as desired, especially if one or more comparison lines besides _the two required for the determination of A and F# are used to determine a deviation curve. : * n? — B is very nearly equal to Ln — M, so that an equation, 6a, may be written d= Nn—P. The error introduced by assuming sin § = 1 is much larger than the error from either of the other assumptions. 56 A. Eh. Merwin—ITnterpolations on Spectrograms. But in some eases, e. g., Lil. ante, a second approximation is required. The resulting formula also is a straight line and is used with the same table. The second approximation is based upon the considerations which follow. The graph of n’ against X has gradually changing curvature. We may make use of this fact in applying formula 8 if we simply increase or de- crease by the same amount all the values of ) along the margin ‘O1 of the table:> Vhussin the table 7. ——— rX — ‘OL —O1X’. But we may write n’ = I —h(r + ky’ and reproduce the values of n’ provided & is small compared with A. Trials have shown that a small value of / is all that is required to make the deviations from formula 8 very small or negligible; 7, ¢ and A do not require evaluation. Then Aje= In (80) Formulas 8a and 8) apply to a spectrum from a single prism. If a train of prisms is used in producing the spectrogram, equation 3 is still a true equation if 2 is the angle of incidence upon the second and each succeeding prism. But 2 is not con- stant, as in the case of a single prism, but it decreases with 2X. This has the effect of slightly changing the relative values of d and A. but the changes are taken care of (if necessary) in equation 80. 7 Geophysical Laboratory, Carnegie Institution of Washington, Washington, D. C., October 9, 1916. Drushel and Felty—Preparation of Acids. 57 Art. VI.—On the Preparation and Lonization of the Dialkylphosphoric and Benzenedisulphonic Acids; by W. A. Drusnet and A. Rh. Ferry. [Contributions from the Kent Chemical Laboratory of Yale Univ.—cclxxxyv. | Tur acids studied in this investigation are the lower dialkylphosphorie acids of the general type R,HPO,, simple monobasie acids, and the three unsubstituted, dibasic, isomeric benzenedisulphonie acids of the formula C,H,(SO,H),. Loniza- tion measurements of only one of the dialkylphosphoric acids have been reported in the literature.* Preparation of matertals.—(a) The dialkylphosphorie acids and their alkali salts. For making conductivity measurements the free dialkylphosphorie acids and their sodium salts were used. The trialkyl phosphatest were prepared by the action of the proper sodium alcoholate upon phosphorus oxychloride in the presence of ether according to the equation: 3RONa + POC], = R,PO, + 3NaCl. The precipitated sodium chloride was filtered off and washed with ether. The ether was dis- tilled off from the filtrate and the trialkyl phosphate was recovered from the residue and purified by fractional distilla- tion, using diminished pressure for the tripropyl phosphate. The trialkyl phosphates were decomposed by concentrated aqueous barium hydroxide and the barium tetra-alkyl phos- phates were purified by crystallization from water. The anhydrous barium salts dried to constant weight were decom- posed by the theoretical amount of sulphuric acid, taking care to avoid an excess, and the aqueous acid solutions were exactly neutralized with pure sodium hydroxide. The aqueous solutions of the three dialkyl sodium phosphates were evap- orated in platinum. The sodium salt of dimethylphosphoric acid was recrystallized and dried to constant weight at 110-120°. The sodium salts of the diethyl- and dipropylphosphorie acids could not be recrystallized but were prepared by evaporating their aqueous solutions in platinum and drying to constant weight at 140-150°. These two salts are hygroscopic and must be weighed from weighing bottles. The three sodium dialky] phosphates were made up in N/8 solutions for conductivity work. The dimethyl-, diethyl- and dipropyiphosphorie acids were prepared for conductivity work by treating weighed amounts of the pure barium salts with the theoretical quantities of sulphuric acid and making up the solutions to the proper normality, verifying the normality of each solution by titration * Van Hove, Bull. Acad. Roy. Belg., 1909, 282-294. + Limpricht, Ann. d. Chem., cxxxiv, 347, 1865. 58 Drushel and Felty—Preparation, etc., of with standard sodium hydroxide. The dipropylphosphorie acid was also prepared from the sodium salt by adding the theoretical amount of sulphuric acid, evaporating the solution on the steam bath and extracting the free acid from the residue with absolute alcohol, the alcoholic solution was diluted with water, concentrated in platinum on the steam bath and finally made up to the proper normality with distilled water. The solution was found to be free from sulphate and its normality was checked up by making a titration with standard sodium hydroxide solution. (6) The isomeric benzenedisulphonic acids and their alkali salts. For the preparation of these acids and their potassium salts methods were chosen which give each acid and its salt entirely free from the acid or salt of either of its two isomers. This is important in the present investigation, for in order ‘to compare the strengths of the three isomeric benzenedisulphonic acids it is necessary that each acid should be prepared entirely free from its two isomers. The m-benzenedisulphonie acid is the only one of the acids prepared by the direct sulphonation of benzene. When benzene is sulphonated benzenemono- sulphonic, m-benzenedisulphonic, s-benzenetrisulphonic, and possibly under special conditions, a little p-benzenedisulphonie acid may be formed; but by properly regulating the tempera- ture and concentration of the sulphuric acid the only products are the monosulphonie acid, the mm-disulphonic acid, a little excess of sulphuric acid and a trace of benzene. The benzene was removed by water, the excess of sulphuric acid as barium sulphate, and the strongly acid aqueous filtrate was then con- verted to the potassium salt by exact neutralization with’ potassium hydroxide or carbonate, and evaporation to erystal- lization. By a number of recrystallizations the pure potassium m-benzenedisulphonate was obtained in good yield, the mono- sulphonate tending always to remain in the mother liquor, Potassium m-benzenedisulphonate was readily converted into the pure acid by the following procedure. The potassium salt was dried at, 200°—220° to remove the molecule of water of crystallization and the anhydrous salt was then mixed with the theoretical amount of powdered phosphorus pentachloride and warmed. After the reaction was over the mixture was treated with cold water to remove the sodium chloride formed in the reaction, the insoluble diacid chloride was filtered off, dried and purified by reerystallization from ether. The diacid chloride was obtained in quantitative yield and was converted to the acid by heating for an hour with distilled water. This acid solution was then treated with a little more than the theoretical amount of pure silver oxide necessary to remove the hydrochloric acid formed in the hydrolysis. Care was exercised Dialkylphosphorie and Benzenedisulphonic Acids. 59 in preparing the silver oxide to wash it thoroughly and repeatedly with boiling distilled water in order to remove every trace of alkali and sodium nitrate resulting from the precipitation of the silver oxide by the action of sodinm hydroxide upon a dilnte hot solution of silver nitrate. The free disulphonic acid containing a little silver disulphonate was separated by filtration from the silver chloride and treated with hydrogen sulphide in excess. The precipitated silver sulphide was fiitered off and the filtrate heated to boiling to remove the excess of hydrogen sulphide. The solution was tested for silver and for halogen and was found to be free from both. It was an aqueous solution of pure m-benzenedisulphonic acid of about N/5 concentration. The solution was standardized against decinormal sodium hydroxide and then diluted with freshly distilled water to N/8 concentration for beginning the conductivity measurements. For the introduction of the “AS acid group (SO,H) into the benzene ring in positions other than meta to acid groups already in the ring several methods have been pro- posed. P. Klason* acted upon the diazo salt of sulphanilie acid with alcoholic potassium sulphide to introduce sulphur in the para position to the sulphonic acid group. He obtained the potassium salt of p-thiophenolsulphonic acid (KSC,H,SO,K, 1, 4) which on oxidation with potassium per manganate gave potassium p-benzenedisulphonate. O. J. Blanksmat and J. J. Polak{ prepared the ortho- and _paranitrobenzenedisulphides by the action of sodium disulphide upon the corresponding chlornitrobenzenes. These disulphides (NO,C,H,S.SC,H,NO,) on oxidation gave the corresponding nitrobenzenesul phonic acids (O,H,NO,.SO,H). In 1890 the xanthate method of hk. Leuckart§ was published. This was shown by Erhardt,| Ruhnau{ and others to be a general method of preparing various sulphonic acids free from their isomers. Polak,** Waltert+ and Armstrong and Worleytt used this method for the preparation of o- and p-benzenedisulphonic acids. None of these investigators, however, described the xanthate method with sufficient particularity to make a fair yield of the disul- phonie acids possible without further experimentation. Since the disulphonie acids can be obtained entirely free from their isomers it was deemed desirable to determine more closely the * Klason, Ber. d. d. Chem. Gesell, xx, 350, 1887. + Blanksma, Rec. trav. chim. Pays: Bas, xix, 111, 1900. { Polak, Rec. trav. chim. Pays-Bas, xxix, 419, 1910. § Leuckart, J. pr. Chem., xli, 179, 1890. | Ehrhardt, J. pr. Chem., xli, 184, 1890. *] Ruhnau, J. pr. Chem., xli, 184, 1890. ** Polak, L..¢, ++ L. E, Walter, Proc. Chem. Soc., xi, 141, 1895. {t{ Armstrong and Worley, Proc. Roy. Swc., xc, 86-7, 1914. 60 Drushel and Felty— Preparation, etc., of conditions which give the best yield of these acids. The xanthate method uses as starting out material in each case the corresponding amidobenzenesulphonic acid, and involves the following steps to obtain the potassium salt of the benzene- disulphonic acid: (1) diazotization of the amidosulphonie acid, (2) action of the diazo compound with potassium xanthate to form the xanthate ester, (3) hydrolysis of the xanthate ester to form the potassium salt of thiophenolsulphonic acid and its oxidation to the disulphide of potassium benzenesulphonate, and finally (4) oxidation of the disulphide to potassium ben- zenedisulphonate by means of potassium permanganate. The steps and conditions for the formation of the potassium o- and p-benzenedisulphonates are essentially the same, starting with o- and p-amidobenzenesulphonie acids. 7 Several points are of importance in carrying out the pro- cesses previously mentioned. Diazotization of the amidoben- zenesulphonic acids by passing N,O, gas through a suspension of the amidosulphonie acids in cold water is unsatisfactory since the amidobenzenesulphonie¢ acids as well as their diazo compounds are only slightly soluble in cold water, and an attempt to diazotize in this way results merely in the super- ficial diazotization of the solid particles, and rarely results in the diazotization of more than 20 per cent of the amidoben- zenesulphonic acid, even after passing the N,O, for several honrs. The most satisfactory method proved to be a slight modification of Fischer’s process. The amidosulphonie acid was treated with about an equal weight of water and to this was added a cold saturated solution of the theoretical amount of sodium nitrite, resulting in immediate diazotization with evolution of heat, pr obably according to the equation: C-HNH,.S0,8 + NaNO,= €C,B NO SON - Ox ne mixture was at once treated at reom temperature with the eal- culated amount of concentrated hydrochloric acid with stirring. Reaction took place at once with the separation of the diazo compound: O,H,N,OH.SO,Na + HCl = O,H,N,SO, + NaCl + heOs do make the separation of the diazo compound as complete as possible the mixture was chilled in an ice bath, then quickly filtered on a porcelain funnel with suction and washed with a little ice-water. The diazo compound was made into a thin mush with water and added in small portions to the caleulated amount of potassium xanthate dissolved in about ten times its weight of water and heated on the steam bath to 65° or 70°. Vigorous reaction took place at once with copious evolution of nitrogen after the addition of éach portion of diazo compound. At the temperature used the xanthate ester formed was immediately hydrolyzed chiefly to the corre- sponding thiophenolsulphonate. These reactions may be Dialkylphosphoric and Benzenedisulphonic Acids. 61 expressed by the following equations: O,H,N,SO, + KS.CS.- OkKt = C,H,SO,K.SCSOEt + N,, and C,H,S0,K.SCSOEt + H,O = C,H,SH.SO,K + COS + EtOH. Hydrolysis was com- pleted by evaporating the reaction mixture to dryness on the steam bath, during which process the thiophenolsulphonate was oxidized by the air chiefly to the disulphide, KSO,0,H,S.- SOC,H,SO,K. The disulphide prepared as described always contained an orange-red dye resulting from the diazo reaction. This dye is soluble in alcohol while the disulphide is not. The disulphide was therefore purified from the dye by twice precipitating the disulphide from its saturated aqueous solution by the addition of alcohol. The trace of color still remaining could be removed later by means of animal charcoal. On oxidation the disul- phide yields the corresponding benzenedisulphonate. The oxi- dation was best effected by the use of hot saturated aqueous potassium permanganate, which was added in about theoretical amount in small portions in the course of 20 to 30 minutes to the cold dilute aqueous solution of the disulphide. The reac- tion evolved considerable heat but the temperature was not at any time allowed to rise above 60°. Toward the end of the oxidation the solution retained the purple color of the perman- ganate for 10 to 15 minutes on the steam bath. This slight excess of permanganate was finally destroyed by the addition of a few drops of alcohol. The pean mixture was some- what diluted by the addition of water and allowed to stand on the steam bath until the hydrated oxide of manganese had well settled out. The oxide of manga- nese was then easily removed by filtration with suc- tion and was well washed with hot water on the porcelain funnel. The combined filtrates were concentrated on the steam bath to crystallization and allowed to stand over night for the potassium benzenedisulphonate to crystallize out. If any of the reddish dye persisted at this stage it was removed by animal charcoal. The py-benzenedisulphonate was further purified by recrystallization from water. A 60 per cent yield of potassium p-benzenedisulphonate may be obtained from sulphanilic acid by the xanthate process under favorable condi- tions. A convenient method for the purification of the o-ben- zenedisulphonie acid is to convert the potassium salt into the barium salt by the action of saturated aqueous barium chloride, and recrystallization of the barium salt from hot water. This method was described by J. J. Polak.* The o-benzenedisulphonie acid was prepared from o-amido- benzenesulphonie acid by the xanthate method. The amido- sulphonic not being obtainable in the market, it became * Polak, Rec. trav. chim. Pays-Bas, xxix, 418, 1910. 62 Drushel and Felty—Preparation, etc., of necessary to prepare it in the laboratory. In its preparation from aniline a few important points must be observed in order to obtain the final product in pure condition and in good yield. Several modifications were introduced into the processes described in the literature, increasing the yield and improving the purity of the products. Parabromacetanilide was _pre- pared from aniline by a modification of Hiibner’s* method. One mol of aniline, one mol of acetic anhydride and two mols of glacial acetic acid were mixed and allowed to stand until cold. To this mixture one mol of bromine was added in small portions with cooling. When the last portion of bromine was added crystals began to separate out and the mixture was immediately poured into a liter of cold water with stirring. On recrystallizing the precipitate from aleohol snow white p-bromacetanilide was obtained in almost quantitative yield. The p-bromacetanilide was converted to p-bromaniline-o-sul- phonie acid by heating at 170°-180° with the theoretical amount of sulphuric acid of specific gravity of 2:0 in an oil bath until the acetic acid was completely eliminated, instead of heating the mixture directly over a free flame as sug gested by Kreis.+ In this way charring was entirely avoided and the yield was practically quantitative. The method of removing the bromine from p-bromaniline-v-sulphonic acid suggested by Kreist was also modified in an important point, shortening the time ‘required and improving the quality of the product. Under the conditions described by Kreis boiling with zine dust in alkaline solution for nine hours was found insufficient to remove completely the bromine from 60 grm. of p-bromaniline- o-sulphonic acid. A 60 grim. portion of this sulphonic acid was boiled for three hours with 600° of sodium hydroxide solution containing 15 grm. excess of sodium hydroxide and 25 grm. of zine dust. At the end of three hours’a titration of 5° of the solution showed that over 70 per cent. of the bromine had been removed from the benzene ring. Fifteen grams more of sodium hydroxide and 10 grm. of zine dust were now added and the boiling continued for two and a half hours, when a titration showed that the bromine was removed quantitatively from the benzene ring. In other particulars the method was carried out as described by Kreis, obtaining a good yield of pure o-amidobenzenesulphonic acid. The p-benzenedisulphonie acid was prepared froin its purified potassiuin salt by the procedure described for the m-benzene- disulphonic acid. The o-benzenedisulphonic acid was prepared directly from its purified barium salt by the action of the theo- * Hiibner, Ann. d, Chem., ccix, 305, 1881. + Kreis, Ann. d. Chem., celxxxvi, 377, 1895. + Ibid., cclxxxvi, 379, 1895. Dialkylphosphoric and Benzenedisulphonic Acids. 63 retical amount of sulphuric acid. The purity of the three disulphonic acids was determined by converting portions of their purified potassium salts into the acid chlorides and taking the melting points of the acid chlorides* recrystallized from ether, and by converting the acid chlorides to the acid amides and taking their melting points. For beginning conductivity measurements the three isomeric disulphonic acids were made up in N/8 solutions with freshly distilled water, and for the determination of the strengths of the acids as catalytic agents in the hydrolysis of ethyl acetate they were made up in N/10 solutions. The normality of the acids was checked up in every case by making titrations with standard sodium hydroxide. Conductivity and hydrolysis measurements. The apparatus used in this work was the usual conductivity apparatus, a Wheatstone bridge previously calibrated, an ordi- nary Ostwald cell whose constant was determined with N/50 potassium chloride, an alternating induction coil with high fre- quency vibrator and telephone receiver, a standard plug resist- ance box, a pair of carefully calibrated 10°™* pipettes and a thermostat kept at 25°. (a) Lonization of dialkylphosphoric acids.—The molecular conductivities of the sodium salts of the dialkylphosphoric acids were first determined at different dilutions starting with N/8 solutions. From these measurements the values of pw, were determined by the method of graphic extrapolation described in Ostwald—Luther.t To obtain the values of zp, for the acids the corrected values of the sodium salts were diminished by the ionic conductivity of the Na ion (51) and increased by that of the H ion (847). The molecular con- ductivity of the dimenthylphosphoric and dipropylphosphoric acids were determined in this way. In attempting to dehy- drate the sodium diethylphosphate slight decomposition occurred giving a value for w, abnormally high. This value was therefore rejected and the value of uw, for the correspond- ing acid determined by Van Hovet was used. The molecular conductivities of the three homologous dialkylphosphorie acids were measured for dilutions ranging from N/8 to N/1024. The value of « for dipropylphosphoric acid obtained by the graphic method was verified by direct measurement at high dilution. The mean of the two values (389°5) was used in * Ortho-, 143°. Meta-, 638°. Para-, 139°. + Ostwald-Luther, Physiko-Chemische Messungen, (2d ed.), 414. { Van Hove, Bull. Acad. Roy. Belg., 1909, 282-294. Van Hove determined by the conductivity method the dissociation of diethylphosphoric acid, but no previous work is reported on the dissociation of its homologues. His average value for k was 10:03 x 10-°. 64 Drushel and Felty— Preparation, etc., of the calculations. From the values obtained by direct measure- ment and calculation the degrees of ionization of the acids expressed in per cent were found, also the affinity constants of the acids were calculated from the formula of Ostwald, k= «a'/(l—a)v. The values found for the dialkylphosphorie acids are recorded in Table I. TABLE I. A. Degrees of ionization of dialkylphosphoric acids. 25°. Dilu- MesHPC, EKt,.HPO, Pr.HPO, tion per cent per cent per cent v Lv dissociated pe dissociated be dissociated 8 262°6 69°3 236°8 62°2 2217-2 58°25 16 292°3 feral 274°4 T21 266°7 68°4 32 31770 83°7 304°6 80°0 296°9 76°1 64 335°0 88°4 328°d 86°26 325°9 83°56 128 348°'0 9135 344°7 90°51 342°0 87°68 256 256°'0 93°95 354°9 93°20 354°2 88°70 012 362°7 95°7 362°5 95°20 36671 93°86 1024 367°8 97°05 365°7 96°03 379°9 95°20 Hao oom. Ha 379°0 379°0 389°5 (Graphic) (Van Hove) (388 — graphic) (391 — direct) B. Affinity constants of the dialkylphosporic acids. 25°. Dilution Me.HPO, Et.HPO, Pr,HPO, k k k 8 19:5) 52510 OTE ats Oe 10°d) SCI 0e 16 LO. Solr ASG Schon 979. Sls 32 gS a Wes es WO ee 10S = SCO S Sie (b) Lonization of the isomeric benzenedisulphonic acids.— The benzenedisulphonic acids represented by the general formula C,H,(SO,H), are dibasic and are dissociated similarly + to sulphuric acid at different dilutions into the ions H, and C,H,(SO,H)SO,, and finally into H, H and C,H,(SO,),. That is, at moderate concentration one H ion splits off and on dilu- tion the second H ion is formed. Considerable work has been done on the sulphonic acids of benzene derivatives, but conductivity measurements of the simple unsubstituted benzenedisulphonie acids have apparently not been previously reported. Ebersbach and Ostwald* deter- mined the conductivities of ortho- and meta-toluidinsulphonie * Ebersbach and Ostwald, Zeitschr. phys. Chem., xi, 617-8, 1893. Dialkylphosphoric and Benzenedisulphonic Acids. 65 acids (C,H,N H,CH,(SO,H),) in order to ascertain the effect. of the different positions of the methyl group on the dissociation of the acids. Armstrong and Worley* studied a number of sulphonic derivatives but their work on the simple benzenedi- sulphonic acids was confined to the use of the three isomeric acids as catalytic agents in the hydrolysis of cane sugar in order to determine their relative strengths. They found the meta- acid slightly stronger than the para-acid and both very much stronger than the ortho-acid. Our results obtained both by the conductivity method and the hydrolysis method are only partially in agreement with the results of Armstrong and Worley. Both methods of measurement gave concordant results and show the para-acid to be a little stronger than the meta-acid and both very much stronger than the ortho-acid. Our results are really in better agreement with Armstrong and Worley’s structure theory for the explanation of the catalytic action of the three isomeric benzenedisulphonic acids than their results, although these acids, as shown by our conductiv- ity measurements, require no special structure theory to explain their catalytic activity. The general ionization theory amply covers the case of the three isomeric benzenedisul phonic acids. The results of our conductivity measurements are given in Table If. TABLE II Conductivities and degrees of dissociation of the isomeric benzenedisul- phonic acids. 25°, Dilu- Ortho-acid Meta-acid Para-acid tion per cent per cent per cent v Lv dissociated bv dissociated ay dissociated 8 317°2 83°09 339°2 88°64 343°3 89°47 16 329°4 86°28 351°5 91°63 353°8 92°21 32 341°3 89°39 359°3 93°66 361°6 94°24 64 351°9 92°17 367°3 95°75 370°5 96°56 128 361°1 94°59 374°2 97°52 375°2 97°78 256 367°5 96379 380°1 99°09 380°1 99°06 512 374°7 98°06 383°'6 100° 383°1 99°84 1024 OLED DO CO OM ar te caret ees cee 383°7 100° 2048 381°7 100° 2048 381°6 100° By single dilution. (c) Catalytic action of the isomeric benzenedisulphonic acids wn the hydrolysis of ethyl acetate——The ionization of these acids as determined by the conductivity method was con- firmed by using them in decinormal concentration in the hydrolysis of ethyl acetate at 25°. In order to determine pos- itively the relative strength of the meta- and para-acids dupli- * Armstrong and Worley, l. c. Am. Jour. Sc1.—Fourts Series, Vou, XLIII, No. 253.—January, 1917. 5 66 Drushel and Felty—Preparation of Acids. cate measurements were made for these two acids. For the purpose of comparison hydrochloric acid in decinormal strength was also used for the hydrolysis of ethyl acetate. Tuitrations were made in the usual way and the constants were calculated from the usual titration formula for reactions of the first order. The results obtained with these acids are given in Table III. TABLE ITI Isomeric benzenedisulphonic acids as catalytic agents in the hydrolysis of ethyl acetate. Acids used in decinormal solutions at 25°. HCI Ortho-acid Meta-acid Para-acid 10°K 10°K 10°K 10°K 64°4 67°2 (76-0) 73°8 65°0 65°2 F(a 74°8 63°3 pre 1252 Tou 65°4 65°1 12:0 75°6 64°0 64:0 73°9 far 63°6 Rass’ 72°4 fou 62°2 64°6 HELE (7773) 64:0 65°2 73°0 74°35 72°9 duplicate 74°50 duplicate Summary. 1. The lower homologous dialkylphosphorie acids may be prepared by the action of barium hydroxide upon the trialkyl esters, and the decomposition of the barium salts with sul- phurie acid. 2. The degree of dissociation of the dialkylphosphoric acids decreases regularly with the merease in the molecular weight of the alkyl groups. 3. The isomeric benzenedisulphonic acids may be prepared, each entirely free from its isomeric forms, by a proper choice of methods. In this connection the xanthaté method of con- verting amidobenzenesulphonic acids into the corresponding benzenedisulphonic acids gives good results under favorable conditions. 4. The catalytic activity of the isomeric benzenedisulphonic acids is in the order ortho-, meta-, and para-. The relative catalytic activity of the three acids requires no special expla- nation based upon a structure theory, but is in agreement with the general ionization theory. G. S. Jamieson—Double Salis of Cesium Chloride. 67 Arr. VII.—On the Double Salts of Cesium Chloride with Calcium and Strontuum Chlorides ; by GrorcEe 8. JAMIE- SON. [Contribution from the Sheffield Chemical Laboratory of Yale University. | THE object of this work was to investigate the double salts of cesium chloride in combination with the alkali earths metals. It should be observed that no barium double salt was obtained, but this was not surprising since barium generally shows little tendency to form double salts. Many attempts were made to obtain double cesium alkali earth bromides, but without suc- CeSs. The double salts to be described are as follows: 2CsCl.CaCl,.2H,O 5CsCl.28rCl,.8H,0. Cestum-Calewum Chloride.—This salt was obtained readily from concentrations of 1-4 molecules of cesium chloride to 1-4 molecules of calcium chloride. 80 grams of cesium chloride and 40 grams of anhydrous calcium chloride were dissolved in 100° of hot water. Dur- ing the night a crop of colorless slender prisms separated. The crystals were filtered by suction, pressed thoroughly on filter paper, and allowed to stand in a desiccator over calcium chloride for about 4 hours before making the analysis. This double salt was found to be very soluble in water. After the removal of the first crop of crystals, 40 grams more of anhy- drous calcium chloride, previously dissolved in 50° of water, were added to the solution. A second crop of the double salt was obtained. Some of these crystals were 2°5°™ long. An- other 40 grams of calcium chloride in 50° of water were added to the solution and a third crop of the double salt was obtained. The three preparations gave the following results by analysis: Calculated for Found 2CsCl.CaCl..2H.O 1 2 3 Jl, Gene eee rae 744% oll A Age ee eae 29°34 GIGS OTe nal nS ie a 8°26 8°43 8°46 8°60 Ops es SS 1 54 OG Cesium-Strontium Chloride.— A. solution containing 2 mole- cules of cesium chloride to 1 of strontium chloride yielded the double salt. When the concentration was one molecule of ce- sium chloride to one of strontium chloride, only strontium chloride crystallized, and when the concentration of the solution 68 G@. 8 Jamieson—Double Salts of Cesium Chloride. was in the proportion of 8 molecules of czsium chloride to one of strontium chloride, czesium chloride separated from the solu- tion. . 100 grams of cesium chloride and 79 grams of crystallized strontium chloride were dissolved in about 150° of hot water and allowed to cool slowly. . After the solution had stood for some hours, a crop of very thin leaf-like crystals with a pearly luster was obtained. Two other crops of this double salt were obtained by concentrating the solution slightly each time after the previous crop of crystals had been removed. The following results were obtained by analysis : Calculated for Found 5CsC1.2SrCl..8H.O i 2 3 A B H,O BND Ea gg Aa pes, ad 11:03% 10°66 10°88% Oe hs et are 24°50 24°48 24°33 24°13 RSP cetc ce tym aia ee 13°43 13°36 13°36 13°24 13°20 Bee ge ee) va perpen ya a) 51°28 The analyses agree well with the calculated formula, and as the crops were carefully examined under the microscope and. appeared to be homogeneous, there seems to be no doubt as to this formula, which is unusual for double halogen salts, and which does not agree with the type shown by the cesium eal- cium salt. Loughlin and Schaltler— Crandalliite, a New Mineral. 69 Art. VIII. — Crandallite, a New Mineral; by G. F. LovugHuin and W. T. ScHALLER.* Introduction.—To the long list of unusual minerals in the Tintic mining district, Utah, which has recently been augmented by A. H. Meanst, another is here added—crandal- lite, named after Mr. M. L. Crandall, until recently engineer for the Knight Syndicate of Provo, Utah, who did much to aid in the recent study of the district by the U. 8. Geological Survey. Occurrence.—Crandallite is a hydrated phosphate of alumi- num and calcium, and has apparently resulted from the altera- tion of a pre-existing non-fibrous mineral similar to goyazitet (hamlinite). The new mineral was found by Mr. Loughlin in vein material on the dump of the Brooklyn mine, in the mon- zonite area of the district 14 miles east of Silver City. The mine workings were inaccessible and only a small amount of ore was available for study on the dump. In this ore the new mineral was very scarce. Crandallite occurs in compact to cleavable masses without distinct crystal outline, and partly fills irregular-shaped cavi- ties in a quartz-barite ore aggregate, resting indifferently on any of these minerals and to a minor extent replacing them. The cavities, some of which are almost completely filled by erandallite, measure from a few millimeters to six centimeters in diameter. The ore minerals of the vein include ‘principally pyrite with considerable enargite and small amounts of galena and zine blende. The crandallite is covered by a crust of tenorite half a millimeter thick (partial analysis of impure sample gave: CuO, 80°12; H,O, 4:02; P,O,, 1:44; insol., 2°94), which in turn is coated by a film of greenish copper minerals. According to these relations crandallite is later than the undoubted primary (hypogene) minerals of the vein, and earlier than the common secondary (supergene) minerals. So tar as its composition is concerned it may be either the latest of the primary or the earliest of the secondary minerals. The hand specimen suggests that crandallite has a platy structure, yielding smooth cleavage surfaces, but when any of * Published by permission of the Director U. S. Geol. Survey. + Means, A. H., Some new mineral occurrences of the Tintic mining dis- trict, Utah, this Journal (4), xli, 125-180, 1916. ¢{ The suggestion was made by W.T. Schaller (ibid., (8), xxxii, 359, 1911; U. S. Geol. Survey, Bull. 509, p. 70, 1912) that hamlinite was probably identical with goyazite. The name goyazite (1884) has priority over the perhaps better known name hamlinite (1890); Farrington (ibid., xli, 358, 1916) has recently questioned this proposed identity. See paper in the fol- lowing number. 70 Loughlin and Schaller—Crandallite,a New Mineral. these plates are crushed and examined microscopically it is seen that they are composed of fibrous aggregates, the fibers, extremely fine, being generally in radiated groups. The con- clusion reached is that crandallite is a fibrous mineral, result- ing from the alteration of a pre-existing mineral of similar chemical composition to which belong the platy structure and smooth cleavage surface. Description.—The exterior surface of the crandallite lining, which is about a millimeter thick, is uneven to imperfect botryoidal, but the cross sections of the lining show a distinct straight lamellar structure, the individual plates being about a fourth of a millimeter across and considerably less in thickness. The part of the lining next to the rock is compact and waxy looking whereas the last formed part of the lining seems to be more distinctly and coarsely crystallized and somewhat whiter in color. When crushed and examined under the microscope, however, the two parts appear the same and the apparent dif- ferences seen in the hand specimen can no longer be noted. The crushed fragments, observed under the microscope, have no regular boundaries and are irregular in shape with only a faint suggestion of the lamellar structure seen on the hand specimen. The new mineral on the hand specimen closely resembles a crust of very fine-grained dolomite or calcite ; in thin section or in crushed fragments, with nicols crossed, a striking resemblance to some chalcedony is noted. The color is white to light gray with shadings into yellow and brown. The streak is white. The luster is dull, some- what greasy in the more compact, and somewhat pearly in the coarser lamellar variety. Almost opaque on the hand speci- men, the crushed fragments are transparent to opaque under the microscope, the opacity being caused by minute indeter- minate inclusions, many of them apparently unreplaced ore minerals. The mineral, free from any inclusions, is colorless and transparent in thin pieces. The cleavage on the hand specimen is distinct and the cleavage faces have a decidedly pearly luster. The direction of the cleavage is basal, such fragments, which are unaltered, being isotropic and yielding a uniaxial positive interference figure. This basal cleavage is believed to be the cleavage of the original mineral (goyazite 4), whose alteration has yielded crandallite. The cleavage of crandallite could not be determined. The density was not determined, on account of the scarcity of the material and because of the many impurities. The brittle mineral has a hardness of about 4. Optical properties.—The white or gray crandallite is color- less and non-pleochroic under the microscope. The material analyzed contained a considerable amount of impurities, chiefly quartz, with a little barite and traces of sulphides. The Loughlin and Schaller —Crandallite, a New Mineral. 71 crandallite was in irregular granular or cryptocrystalline masses, many of which showed distinctly on high magnitica- tion a fibrous structure in large part radiating. The individual fibers are very minute and it was not possible to isolate on the glass slide a fragment or fiber which was composed of only one unit. The refractive indices of the mineral showed a minimum of 1:°585 and a maximum of 1°595, the birefringence of the material ranging from zero to a maximum of about 0-01. Some of the material appears amorphous, but it may be very Bie.) 1 Fig. 2. Fie. 1. A portion of a hexagonal crystal, basal section, showing concen- tric, radiating fibrous structure. A pseudomorph of crandallite (fibrous) after goyazite (2). Fie. 2. A crystal plate showing fibrous structure, and with a birefracting center with isotropic rim. A pseudomorph of crandallite (fibrous) after goyazite (2). finely eryptocrystalline. The fibers apparently have parallel extinction and the elongation is negative. Some of the crandallite on the hand specimen which appeared better crystallized than the material analyzed, was crushed and examined optically. In this crushed material were found several irregularly shaped to poorly defined hexagonal plates which were isotropic, uniaxial, positive, and which on edge gave parallel extinction and a moderate birefringence (estimated as about 0:01-0:02). The refractive index varied from about 1°605+:005 to 1°62. Some of these plates were uniform in structure and isotropic; others showed a concen- tric, radiating, fibrous structure (fig. 1), and were either iso- tropic (index about 1:605+-:005) or else feebly birefracting. A single crystal plate showed an inside area which was feebly birefracting, the fibers giving parallel extinction while the out- side rim was isotropic, although both parts showed a distinct fibrous structure as shown in fig. 2. | 72 Loughlin and Schaller— Crandallite, a New Mineral. Pyrognostics.—Heated in a blowpipe flame, crandallite decrepitates somewhat, then exfoliates slightly and fuses to-an opaque white enamel, coloring the flame intermittently a pale green (phosphorous) with occasional flashes of red (calcium, strontium). In a closed tube decrepitation occurs with the liberation of water. Soluble in acids. Chemical composition.—The results of the analysis are as follows, the material being dissolved in acids and the insoluble residue filtered off and weighed. The insoluble residue con- sisted of 97 per cent SiO,, the remainder being barite with traces of sulphides. Analysis of Crandallite. [W. T. Schaller, analyst. | Same with insoluble Analysis deducted Ratios Insol. 35°18 . AVVO. 2516 38°71 0°379 0°379 1°96 or 20°98 CaO 4°88 7°50 "134 SrO 1°44 Fon 021 | "178 92 or 10:92 MgO 0°61 0°94 "023 P.O 17°61 27°09 eo SO, 0°47 3°80 O47 215 1°11) or toa H,O— 0°84 1:29 072 H,O + 12°26 18°86 1°048 1:048 5°43 or 5X 1°09 100°40 100°40 An inconclusive test for rare earths (probably cerium) seemed to show their presence to a slight extent but the iden- tification was not verified. Barium could not be detected in the acid-soluble part, neither was any barium sulphate precip- itated on solution of the mineral. The barium sulphate in the insoluble residue was in relatively large cleavage plates, plainly derived from the mineral barite. Genetic relations.—The ratios of the analysis yield the formula CaO.2Al,0,.P,0,.5H,O with slight replacement of CaO by SrO and MeO, and of P.O, by SO, [or more exactly af (POY by (SOn ie This type of tae is slightly different from that of goya- zite (hamlinite), Crandallite 2Ca0.4Al1,0,.2P,0,.10H,O Goyazite 25r0.3A1,0,.2P,0,.7H,O but is of the same type as that of gorceixite. Crandallite Ca0.2Al1,0,.P,0,.5H,O Gorceixite BaO. 2 Al. 10 P, O. 6H, O Loughlin and Schaller—Crandallite, a New Mimeral. 78 In an earlier paper it was suggested* that gorceixite was the barium equivalent of goyazite (hamlinite) and that its formula should be written 2Ba0.3A1,0,.2P,0,.7H,O. The original analysis gives ratios which do not support this suggestion but lead to the formula BaO.2Al1,O,.P,O,.5H,O. Thin sections of gorceixite are stated to have shown an aggregate of minute, colorless, irregular grains, therefore, apparently not fibrous. The question thus arises: What is the systematic position of goreeixite? Does it belong in the alunite-beudantite group with goyazite (hamlinite), plumbogummite, and florencite, and would another analysis on pure material agree better with the type formula, 2RKO.3A1,0,.2P,0,.7H,O ; or is gorceix- ite, with the formula BaO.2A]1,O,.P,O,.5H,O, a fibrous altera- tion product of a pre-existing unknown mineral of the goyazite type of formula; or is gorceixite a non-fibrous mineral, with the formula BaO.2A]1,O,.P,O,.5H,O, as independent genetically as goyazite and not a fibrous alteration product of some pre- existing mineral of a similar chemical composition @ The formulas of goyazite and gorceixite have been discussed by Farringtont in a recent paper. The three analyses (by H. W. Nichols) of Brazilian favas given by Farrington represent well the composition of the favas but are of no value for the elucidation of the composition of any mineral. His material was not examined optically to see if it was homogeneous or free from impurities. The analysis of Fava No. 2. baftled interpretation so that Farrington concluded that “this fava was evidently a mixture.’ Of Fava No. 3 he finally states: ‘“ Of its optical properties nothing can be stated since, unfortu- nately, all the substance of the fava was used for the chemical analvsis.”’ The homogeneity and character of transparent minerals can be so easily and quickly determined by the present day facil- ities for microscopic examination (imbedding in oils of known index) that it is a matter of regret that mineralogical papers are still published with detailed and careful chemical analyses of material which has not been first submitted to microscopic examination. The fibrous plumbogummite, described by Lacroixt (hitch- cockite ? ), may represent a distinct species related to plumbo- summite as crandallite is to goyazite. A sulphate mineral possibly related to crandallite was described * Schaller, W. T., The alunite-beudantite group, this Journal, (3), xxxii, 309, 1911. In Mineralogical Notes, Series 2, U. S. Geol. Survey, Bull. 509, ps 10; 1912. + Farrington, O. C., Studies of Brazilian favas, this Journal, (4), xli, 355, 1916. ¢ Lacroix, A., Minéralogie de la France, vol. IV, p. 586, 1910. 74 Loughlin and Schaller— Crandallite, a New Mineral. by Lindgren* as a basic hydrous strontium sulphate which oceurs as an impalpable white powder covering granular celestite, in a vein on level 7 of the Ironclad Mine, Cripple Creek, Colorado. Under the microscope the mineral was seen to consist of short and extremely delicate white fibers of very feeble double refraction and an extinction which is probably parallel. An analysis by W. T. Schaller of a very small quantity gave: SrO, 2571; CaQ; 15-1; ALO,/:13°9 y MeO) <2-5 >) SOn ieee H,O (107°), 0; H,O (ignition), 24:1; total, 94:6. No water was lost at 100° and at 260° only 149 per cent H,O was driven off. No test for phosphoric acid was made and the material was too limited in quantity and of too undetermined purity for these figures to have much more than a qualitative value. The relations of these minerals, especially of the fibrous ones to those not tibrous (those of the alunite-beudantite group), are not at all clear, but the composition of crandallite seems to indicate that there is a detinite group of fibrous minerals which are closely related in the type of formula to those of the alunite-beudantite group and yet seem to be very distinct there- from and to probably result from their alteration. * Lindgren, W., and Ransome, F. L., Geology and gold deposits of the Cripple Creek district, Colorado. U. 8. Geol. Survey, Prof. Paper 44, p. 125, 286, 1906. Warren, Allan and Conner—A Titaniferous Augite. 5 Art. [IX.—A Titaniferous Augite from Ice River, British Columbia ; by CHartes H. Warren and Joun A. ALLAN, with a Chemical Analysis by M. F. Conner. In the course of an extended study of the nephelite-syenites and their associated rocks, occurring in the Ice River District of British Columbia,* one of the authors (Allan) has described several melanocratic types of rock, occurring as basic differ- entiates from the nephelite-syenite magma, particularly near the contacts. In several of these (jacupirangites, etc.), a rather strongly pleochroic pyroxene, containing abundant, rod-like, black inclusions, was noted. On study, it was found to be an augite of somewhat unusual characteristics, and it has accord- ingly been thought well to publish a brief description of it. In several specimens, representing the melanocratic types, this augite was noted, frequently associated with a pale green, augitic pyroxene.t It was also commonly associated with a brown, barkevikitic hornblende, biotite, nephelite, apatite, sphene and ilmenite. Fortunately, one specimen was found which consisted very largely of the augite—an augitite, in fact—and from this specimen material, suitable for micro- scopie study and chemical analysis, was obtained. _ In the specimen alluded to, the augite was associated with a small amount of apatite, occurring as rounded or elongated grains between the augite crystals, with ilmenite, as irregular grains, sometimes in the augite, sometimes about its margins, with a little biotite, and with a little original, brown horn- blende. In places a little alteration of the augite to hornblende was noted, and occasionally small amounts of other alteration products were seen, among them calcite. On the whole the material was quite fresh. In the hand specimen the augite is black in color and of rather irregular habit, although there is a well-marked tendency toward the prismatic development. The grain is rather fine, the length of the prisms rarely exceeding 2™™. The cleavage, while distinct, is rather poorly developed. There is perhaps a suggestion of a schistose texture present. In thin-section, the grains have, in general, irregular forms, although the tendency to elongation in the direction of the vertical axis is obvious. The cleavage is poorly developed. The majority of the grains show minute, black, rod-like inelu- sions arranged in two distinct series. In many of the grains the inclusions are very abundant, and are a striking character- istic ; in other grains they are less numerous, or almost absent. * Geology of the Ice River District, B. C., Thesis for the Degree of Doctor of Philosophy, Massachusetts Institute of Technology, 1912. + Attempts were made to isolate this green pyroxene for chemical analysis, but without entirely satisfactory results with the material at the time avail- able. 76 Warren, Allan and Conner—A Titaniferous Augite. One of the series of rods lies parallel to the vertical crystal- lographic axis; the other follows a direction which makes an angle of about 74° with the first series (on sections parallel to (010)) and appears, therefore, to follow the direction of the edge, 001-010. In the absence of well-defined cleavage lines, the first set of inelusions can be used very weil to, measure the extinction directions on. In view of the black color and opac- ity of the inclusions, and the very high titanium content of the augite, it seems pretty certain that the inclusions are ilmenite. It may be noted, that when some of the inclusions are viewed on end, they appear to be minute tubes, filled with some trans- parent material. The augite grains also contain some minute black inclusions which le at random, and also a few larger, irregular or rounded masses of what seems to be also ilmenite. The pleochrism is rather strong, and is as follows :— a = reddish or pinkish-brown (with a violet shade in thick fragments). 8 = reddish or pinkish-brown (with a violet shade in thick fragments). y = pale, bright yellow (a little brownish shade in thick frag- ments). The absorption is:— a < B >. The extinction was measured against the series of black inclusions lying parallel to the vertical axis in several thin- sections of grains showing 4 maximum double refraction, and was found to range for y ~ oc’ from 41°9° to 42°6°, or an aver- age of 42°3°. This value was obtained by measuring from the direction of the lines of inclusions to the position of maximum darkness in white light. Like many titaniferous augites this pyroxene has a very strong dispersion of the bisectrices, and anomalous interference colors appear on either side of the posi- tion of maximum darkness. In the present case, the change was quite sharp, and experiment showed that the results obtained in this way were more consistent than were those obtained by using monochromatie lights. The interference figures, obtained in convergent light, show a very strong, inclined dispersion ;—red greater than blue. One of the hyperbolas appears black, while the other consists of a strong, red, inner band (convex side) and an equally strong, blue band on the concave side. The width of the two colored curves covers together about five divisions on the micrometer scale used, which is equivalent to about 9° angular. Two sections were found which were nearly norma] to the acute, positive bisectrix, and with these sections the optic angle in the air was measured under the microscope using the Wright micrometer scale. On account of the breadth and lack of sharpness of the hyperbole, both in white light and Warren, Allan and Conner—A Titaniferous Augite. T7 also in monochromatic light, no great accuracy can be claimed for this measurement, and it was deemed sufficient to measure it in white light alone, taking the readings from the middle of the black brush passing through the one optic axis to the division between the strong red and blue parts of the other brush, this division being fairly well marked—the sharpest thing about the interference figure in fact. One section gave QE— 58°5°, the other, 59°5°, or an average of 59°. In sections near the optic axis A, the interference color as seen in parallel light between crossed nicols is a peculiar dull, deep blue or bluish gray ; in those almost perpendicular to the axis, the color is still more dull and has a gray or brownish tint. The sections nearly normal to the acute bisectrix have also a pecu- liar dull brownish- or purplish-gray color. These interfer- ence colors are, as a matter of fact, quite characteristic of the thin-slices of the rocks carrying this augite. Two of the indices of refraction were determined by the immersion method, and were found to be, for sodium light, a = 1°725, y = 1:746. This makes the double- refraction equal to 0°021. This is agreement with the value estimated from the thin-slice, which was about 0°020. A very careful separation of the augite from the minerals accompanying it was carried out by means of the barium- mercuric-iodide solution. ‘The final product, examined under the microscope, showed a very satisfactory degree of purity, there being very little attached or included hornblende or biotite present, and only traces of other incidental impurities. The mineral thus prepared was analyzed by Mr. M. F. Conner, of Ottawa, with the following results :-— Augite from a Augite from Monchiguite dike, Ice River Syenite, B. C. Rio de Janeiro. Conner, Analyst. Hunter, Analyst. % Ratio. % Ratio. BIO ula kee 2 41°80 0°697 44°55 0°742 BAO ater es) O30 ‘091 7°86 077 RO iso? ia 4 5°44 "034 3°81 "024 Oi ie ei ay 3°30 "046 4°53 063 MoO) 2 is)... WOS82 "238 12°71 °318 NO ei ep FEA 22°89 "409 20°84 2372 [Cael Baa a 0-16 PRO spe ht 2282 1°10 060 Oe os 4°84 "060 2°85 "035 AVEO) esas ye 0°10 002 38 "006 99°75 Nag 1:29 "021 K,O "49 "005 Sp. Gr.=3°39 99°31 Sp. Gr.=3°284 78 Warren, Allan and Conner—A Titaniferous Augite. A search of the literature relative to the chemical composi- tion of the augites has so far failed to find any which are very close to the one under consideration. The nearest approach is found in an augite from near Rio de Janeiro, analyzed by Hunter and described by Rosenbusch.* The latter gives its properties as follows :—Extinction y on c’, 40°; dispersion strong; pleochrism, 6 = y = reddish- violet, a = yellow- rose. While the Ice River augite is much higher in titanium and contains no alkalies, there is otherwise considerable similarity in the composition of the two. In both, the silica is rather low, the alumina, ferric iron and lime are high, as is also the titanium. The latter is very high in the Ice River variety, although it is true, that some part of the titanium is present in the form of the rod-like inclusions, and not in the augite itself. It is impossible to say how much TiO, should be deducted from the angite, but if it amounts to as much as one or even two per cent, which is not improbable, it would still leave the titanium high. A combination of the various oxides into groups yields the following :— Ice River Augite. Rio de Janeiro Augite. SiO, + TiO, = 0°757 0:77 RO+H,O = 0-755 0-785 (+R,0) R,O, = Oe) 25 0°101 SiO, + TiO,:R,O,;RO+H,0 = SiO, +TiO,:R,0,:RO+R,0 = 6 SOOO: Meek: fe one (Okeroa OO) The titanium has all been included, which in the ease of the Ice River augite unquestionably makes it somewhat too high. The water and alkalies are here included with the RO. It is noticeable that, as might be expected, the relative proportions of the RO, and RO are about equal. The ratio of the RO, in the one augite is 6:1, and in the other, 7:1, or nearly so, thus bringing out a rather sharp distinction between them. It has occurred to the authors, that at the time of crystalh- zation of the augite, the ilmenite molecule, RTi0O,, analogous to the metasilicate molecule, RSiO,, may have crystallized isomorphously with it, and subsequently becoming unstable in that state, have separated ont—unmixed. In any case the tita- nium would be combined with the RO (FeO), and hence would not affect the relative values of the ratio, RO,: RO, as caleu- lated. It would still leave the excess of R,O, in the Ice River mineral, as compared with the other, as marked as before. In conclusion it may be said that the augite described above differs in its microscopic and chemical characteristics from any augite known to the writers and is believed to be of interest in extending somewhat our knowledge of the titaniferous augites. Geological Department, Massachusetts Institute of Technology, Cambridge, Mass., August, 1916. *Tsch. Min. Petr. Mitt., xi, p. 460. This augite occurred in a monchi- quite dike in gneiss at Rio de Duro, Sera de Tingua, near Rio de Janeiro. Chemistry and Physics. 79 SCIENTIFIC INTELLIGENCE. I. CHEMISTRY AND PHYSICS. 1. The Determination of the Density of Solids.—HEnry Lz CuaTELieR and F. Boaitcu observe that the determination of the specific gravity of solids is one of the simplest of physical meas- urements, but that this determination is usually made in an inex- act way, so that it is unusual that the density of a solid is known with an accuracy of one part in 100. The treatises on chemistry hardly ever give exact numbers for densities. For example, in Moissan’s large treatise we find for common substances occurring frequently in a state of great purity the following statements in regard to densities : Quartz, between 2°55 and 2°74 Natural galena, varying from 7°26 to 7°70 Artificial galena, varying from 6°9 to 7°5 Natural blende, varying from 3°5 to 4:2. Now, in regard to quartz, deLépinay and Buisson have shown that its specific gravity is absolutely fixed, and is 2°6507 with an uncertainty of only one unit in the fourth decimal. The authors discuss the probable cause of our shocking igno- rance in connection with these constants, and they conclude that while impure substances and insufficient weights of samples may sometimes cause errors, the principal cause of inaccuracy is the adhesion of a thin layer of air to the surface of the solid. They observe that this effect is well known to be enormous, as in the “< flotation ” processes of separating sulphides from other miner- als. Upon experimenting with various liquids, they have found that carbon tetrachloride, benzol and petroleum ether do not give this difficulty, and they recommend that water should never be employed for the purpose. They have devised a very simple and convenient apparatus for determining specific gravities. It consists of a glass tube of about 5™™ interior diameter, graduated in cubic centimeters and their tenths, with a bulb at its lower extremity and placed upright. The tube is charged with the liquid, the level is read, either by the eye alone or by means of a cathetometer, and the weighed solid, thoroughly freed from dust by sifting, is introduced by means of along funnel. The increase in volume, which should amount to at least about 3°°, is deter- mined by reading the new level of the liquid. Using carbon tetrachloride, with coarse and fine material in each case, galena gave the results 7°584 and 7590, while zinc blende gave 4:079 and 4°079. When water was employed, even in cases where it was attempted to remove the adhering air by exhausting the air above the liquid, the results were very unsatisfactory.— Comptes Rendus, clxv, 459. H. L. W. 80 ' Seventifie Intelligence. 2, A New Reayent for Free Chlorine.-—The detection and determination of active chlorine in public water supplies has become important on account of the frequent use of hypochlor- ites in purifying such waters. The classical reagent potassium iodide with starch paste is generally used for this examination, but G. A. LeRoy proposes the employment of the chlorhydrate of hexamethyl-triparaminotriphenylmethane for the purpose. When this reagent is added to the extent of a few thousandths to a potable water containing traces of free chlorine, a violet coloration is formed immediately and this varies in proportion to the amount present. The reagent will show about 3 hundred- millionths of chlorine, while potassium iodide and starch show only about one ten-millionth.— Comptes Rendus, clxv, 226. 1 Fi 3. The Right Honourable Sir Henry Enfield Roscoe, A Bio- graphical Sketch, by Sir Epwarp TuHorrPe. 8vo, pp. 208. Lon- don, 1916 (Longmans, Green & Co. Price $2.50 net).—The distinguished author of this biography, who calls himself “a grateful pupil, an attached co-worker, and a lifelong friend ” of the eminent subject of his essay, has produced a book of unusual interest. Roscoe’s name is well known to chemists on account of his text-books, and his chemical researches, particularly his inves- tigations dealing with the compounds of vanadium, and his explanation of constant-boiling acid solutions. But this story of his life shows that his activities were important in many direc- tions. He did effective service in the development of Owens College, Manchester, he was active in scientific societies and in public service, and gave much time to the advancement of pop- ular and technical education. After teaching for nearly 30 years at Owens College he was elected a Member of Parliament in 1885 and held his seat for 10 years. He died in 1915 at the age of 83 years. His biographer characterizes him as a strenuous high- minded man, of large aims and generous impulses, who spent his abilities and energies unstintingly in promoting the welfare of science and the good of his kind. The book contains a fine por- trait of Roscoe as an elderly man. H. L. W. 4. A Text-book of Inorganic Chemistry ; by A. F. HoLtemMay, Issued in English in Codperation with H. C. Cooper. 8vo, pp. 521. New York, 1916. (John Wiley & Sons, Inc.)—This is the fifth edition in English, the first of which appeared 15 years ago, of an excellent and widely-used text-book. ‘The present issue has been thoroughly revised and many parts have been rewritten by the American collaborator, so that the recent achievements in chemical science receive consideration. For instance, the results of T. W. Richards on the atomic weight of radioactive lead iso- topes are mentioned. There is no doubt that the book presents the whole subject in a very satisfactory way. However, it is decidedly advanced in its treatment of physico-chemical topics from a mathematical standpoint, but it is stated that notwith- standing the appearance of differential formule in the book, it is Chemistry and Physics. | 81 believed that a student who is unfamiliar with the calculus should have little difficulty in understanding the meaning and use of such formule, provided he is willing to take the author’s word for the solution of the equations. He LE. We 5. Organic Chemistry for the Laboratory, by W. A. Novyess. 8vo, pp. 292. Easton, Pa., 1916 (The Chemical Publishing Co.). —This guide for organic preparations now appears in its third edition, revised and slightly enlarged. As is well-known, the book presents a liberal number of classified and well-selected preparations, with full and clear directions and abundant refer- ences to the literature. There are 132 numbered exercises, many of which involve several operations, from which more or less extensive courses of laboratory work may be selected. 4. L. w. 6. Note on the Electrolysis of Gallium; by H. S. Unter (communicated).—After the article by Philip E. Browning and myself had appeared (vol. xlii, p. 389, November, 1916), I met with a difficulty in the electrolysis of gallium which caused me to lose much time, and hence it seems desirable to present this brief note of warning for the benefit of other investigators who may be working in the same field. In order to increase the purity of some gallium which had been obtained by the electrolytic method outlined in the above mentioned paper, the metal was dissolved in a hot. solution made from equal volumes of con- centrated nitric acid and water. The resulting solution was then made strongly alkaline by the addition of caustic soda so that the gallium hydroxide was thoroughly dissolved. An attempt to electrolyze this solution was unsuccessful, from the practical point of view, for only an insignificant mass of gailium deposited on the cathode in the course of several days. The addi- tion to the electrolyte of several times its volume of a very con- centrated aqueous solution of sodium hydroxide increased the rate of deposition, to some extent but not enough to be compara- ble with the rate obtained with alkaline solutions containing no nitrates. ‘To prove that no blunder had been made, the solution was neutralized with hydrochloric acid and the precipitated gal- lium hydroxide was thoroughly washed and thus freed from sol- uble compounds. The precipitate was then dissolved in an excess of caustic soda, after which treatment the solution electrolyzed with great facility. The entire experiment was repeated, com- mencing with 1:05 grams of metallic gallium. Only 4:5™2 were ‘deposited on the cathode in 112°75 hours. After neutralization, and so forth, the metal was recovered with promptness. It is certain, therefore, that the presence of sodium nitrate greatly retards the electrolytic deposition of gallium. Although no quantitative comparisons were made, the rates of deposition with and without the presence of chlorides seemed to be about the same. The warning is,—avoid nitrates. iss 5 MO 7. The Condensation of Gas Molecules.--By using cadmium instead of mercury vapor R. W. Woop has performed some very interesting and instructive preliminary experiments on the con- Am. Jour. Sci.—Fourts Series, Vou, XLIII, No, 253.—January, 1917, 6 82 Scientific Intelligence. densation and reflection of gas molecules. The advantage of cadmium over mercury consists in the fact that films of the former metal remain solid and in situ at room temperature whereas films of the latter liquefy and form globules at tempera- tures convenient for making observations. “We shall omit the experiments relating to the cosime law of reflection because they are not as delicate as those of Knudsen (vide in/ra). A long glass tube was sealed off at one end and blown out into a relatively large bulb at the opposite end. Near the bulb the tube was constricted for a length sufficient to ensure the trans- mission of an approximately one-dimensional flow of vapor mole- cules into the bulb. The cadmium was placed in the small end of the glass system and was heated by a gas-flame 3 or 4™™s high, burning at the orifice of a glass tube drawn down to a fine capil- lary The tube and associated bulb were kept in communication with a Gaede pump during the experiments, since a good vacuum is necessary. When the bulb is kept at room temperature no trace of a deposit appears, but when the wall of the bulb opposite the inlet is cooled with a pad of cotton wet with liquid air a small deposit of metal begins to form at once. If the degree of exhaustion is sufficiently high, prolonging the expériment ‘for 15 or 20 minutes causes no sensible increase in the diameter of the circular zone of condensed metal, whereas if the vacuum is inferior the diameter of the spot increases due to the deflection of the cadmium vapor by the residual gas molecules. It is evident, therefore, that with a high vacuum the molecules of the vapor shoot across the bulb without spreading laterally. If, at the beginning of the experi- ment the cold pad is placed against the bulb at any point 90° from the inlet a large deposit of irregular shape forms immedi- ately, showing that the bulb is filled with cadmium vapor having three-dimensional motion. Even if the cold cotton wad is held against the bulb for only a second or two the deposit starts to form and continues to build up indefinitely after the pad has been removed. Hence, the chance of reflection at room temperature of cadmium vapor from a cadmium surface is zero, that is, con- densation occurs at the first collision. If the surface of the bulb opposite the inlet is subsequently cooled with liquid air the small circular spot mentioned above forms at once and increases in thickness after removal of the cotton wad. No further deposi- tion of cadmium molecules occurs on the metal at the side of the bulb, for the polar deposit, when once formed, serves as a trap for the incident parallel stream of molecules. The fact that molecules of cadmium vapor can experience a large number of reflections from glass without condensation was neatly shown in the following manner. A glass tube was bent at right angles in a number of places and blown out as a thin-walled bulb at the end remote from the source of vapor. Condensation took place in the bulb when it was cooled with liquid air. Tubes with more than a dozen bends were used, proving that a large Chemistry and Physics. 83 number of reflections are possible, since the number of bends gives the minimum number of reflections suffered by a molecule before reaching the condensation bulb. The greater the number of bends the longer was the time required to produce in the bulb a film of definite thickness. Consequently a certain fraction of the total number of molecules in the column of vapor must have fallen by the wayside. In a particular tube deposits of equal density were obtained in the bulb, at a chosen bend, and ata bend still nearer the source, in 15 mins., 2 mins., and 10 secs., respectively. Although the density of the cadmium vapor is con- siderably decreased by its passage along the bent tube, no visible deposit can be detected on the uncooled walls. Microscoyic exami- nation of a bulb which had been kept at room temperature and in communication with the source of vapor for 40 mins. showed dis- crete aggregates of metal which seemed to be clusters of very minute crystals. ‘The temperature necessary for the formation on glass of a homogeneous film of cadmium was found experimentally to be in the neighborhood of —100°C. The corresponding “critical ” temperatures for iodine and mercury are roughly —60° and —140°, respectively.— Phil. Mag., xxxil, p. 364, October, 1916. Ri iS 10 &. The Cosine Law in the Kinetic Theory.—F¥rom his earlier experiments on the flow of gases through capillary tubes Knud- sen concluded that all of the 2 molecules which strike an element of surface from a solid angle do leave this surface in such a man- ner as to have their velocities uniformly distributed over every azimuth. More precisely dz = 1~'n cosxdw, where x denotes the angle between the normal to the surface and the direction of the axis of dw. This is equivalent to putting a certain fraction J, which occurs in Maxwell’s original theory, equal to unity. Later simultaneous and independent experiments by Knudsen and R. W. Wood have apparently confirmed Knudsen’s cosine law. Since, however, these investigations involved the comparison of different surface densities, which cannot be determined with ease and accuracy, a more recent and conclusive method of attacking the problem, devised and employed by Martin KNupsEn, deserves notice. The innovation consists in assuming the validity of the cosine . law and then finding the form of surface over which the distribution of condensed vapor would be uniform. This method is very sensitive because any appreciable departure from constant surface . density can be readily detected. It can be shown by very simple analysis that if the molecules obey the cosine law and are reflected from one portion of the inner surface of a spherical cavity they will be uniformly distributed over the rest of the same spherical surface. Accordingly a glass bulb was blown in as nearly a spherical shape as possible and was drawn out at one point in the form of a tube which communicated with the source of mercury vapor, with the air pump, etc. The lateral tube and the auxiliary apparatus were designed in such a way as to cause a narrow stream of mercury vapor to enter the bulb diagonally and 84 Scientific Intelligence. then strike the spherical surface obliquely at the other end of the chord. ‘To prevent condensation the region of reflection was kept warm by means of an electrically heated metal trough the lower end of which fitted the outside of the glass bulb very closely. The temperature of the water in the trough was main- tained at 30°C. so that the temperature of the glass reflecting surface was probably between 0° and 20°. Below the zone of contact of the heating trough the rest of the bulb and a portion of the auxiliary apparatus were kept immersed in liquid oxygen. This liquid was contained in a glass vessel with unsilvered walls to enable the making of visual observations without changing the temperature of the bulb and its accessories. When the mercury in the auxiliary system had been warmed for a very short time the layer of condensed mercury on the walls of the inlet tube became completely opaque and, after a few min- utes, the condensed vapor in the bulb became visible. The layer in the bulb appeared to be uniformly distributed from the instant when it first became discernible until it had attained almost com- plete opacity. ‘It was not possible to detect anywhere a strue- ture or a difference in transmissive or reflective power, save at a zone about 1™™ wide which surrounded the heating vessel and which remained absolutely clear and transparent.” After the experiment had continued for “a good 20 minutes” the mercury deposit was so dense that an incandescent lamp could barely be seen through it. The heating currents were then broken, the liguid oxvgen and frost removed, air was admitted to the bulb, and the apparatus was allowed to come to room temperature. The film of mercury, now liquid, had a very characteristic and uniform opalescence. At the expiration of 24 hours the diffrac- tion rings formed by the mercury droplets, when the bulb was illuminated by a distant electric light, were examined and found to have the same appearance for all parts of the bulb which had been previously frozen. In order to obtain quantitative measurements the experiment was repeated and continued until a thicker deposit of mercury had formed. By skilfully manipulating the bulb the mercury layer was collected from definite spherical zones by causing an auxiliary globule to run around over the region in question. By weighing the mercury and measuring the corresponding areas the surface density of the selected regions was obtained directly. The zones were chosen at such azimuths, with respect to the direction of the incident stream of mercury vapor and the warmed spot, as © might be expected to produce the greatest differences in surface density in case the cosine law did not hold true. ‘The numbers found, however, were constant within the limits of experimental error. Consequently the cosine law is valid even when there is no temperature equilibrium, for the temperature of the reflecting surface was below 20° C. while that of the incident mercury vapor was above 80°.—Ann. d. Physik, vol. xlviii, p. 1113, Feb., 1916. BH. So te Geology. 85 II. Grounoey. 1. A preliminary paper on the origin and classification of intraformational conglomerates and breccias; by Ricnwarp M. Fieitp. Ottawa Nat., Vol. XXX, 1916, 23 pages.—An interest- ing paper explaining the origin of several kinds of intraforma- tional conglomerates. The author also defines conglomerates, glomerates, tectibreccias, and bioglomerates. Phenoclasts are defined as the fragments and rocks of which the foregoing are composed. “ Intraformational conglomerates and breccias seen at Chambersburg, Bellefonte and Tyrone, Pennsylvania, are of extremely shallow water origin; in fact, their formation postu- lates an emergence from the sea such as is common under tidal action. . . Mud-cracked beds and intraformational breccias are in certain cases one and the same thing.” OFS 2. Florida Geological Survey, Highth Annual Report ;, KK. H. SELLARDS, State Geologist. Pp. 168, pls. 31, text figs. 14, 1916.— Besides the administrative report and an account of the mineral industries of Florida during 1915, this volume has three memoirs treating of Cenozoic vertebrates. O. P. Hay describes twenty- two species, of which nineteen are turtles ; E. H. Sellards describes anew Miocene fauna of five species, and four vertebrates from the Pliocene, and discusses the Pleistocene vertebrates from the state. He also presents a bibliography of the literature treating of Floridian fossil vertebrates. ‘The last paper, also by the state geologist, describes in greater detail the human remains and asso- ciated fossils from the Pleistocene of Florida first announced in this Journal last July. C. 8. 3. A study of the Morrison formation ; by CHARLES CRAIG Moox. Ann. N. Y. Acad. Sci., Vol. XX VII, 1916, pp. 39-191, pl. 6—The Morrison formation originally “had an extremely wide distribution, which may have amounted to four or five hundred thousand square miles,” and is celebrated for its many and striking dinosaurs, among them the largest of all land ani- mals. ‘The author here brings together our knowledge of the Morrison deposits and their distribution east of the Rocky Mountains, and seeks to determine whether they are of Jurassic or Comanchian age. His conclusions are as follows : “ It appears, then, that the Morrison commenced as a continental deposit in the western areas of its occurrence in early Comanchean time (or possibly latest Jurassic), and that it spread outward as it was built up, the uppermost and easternmost beds being laid down in [later] Comanchean time . . . If the above interpretation of the Morrison be anything like the truth, it seems probable that the Morrison merged into the marine [Comanchian] deposits in the southeastern areas, such as Texas, and that the Morrison in its southeastern and eastern areas consisted of true delta deposits ” (172). C. S. 4, Notes on the geology of Nelson and Hayes Rivers ; by J. B. Tyrrevy. Trans. Roy. Soc. Canada, Ser. III, Vol. X, 1916, pp. 1-27, pls. 1-5, 2 text figs.—An interesting address, particu- 86 - Serentifie Intelligence. larly the description of the glacial deposits on the west side of Hudson Bay. The oldest continental glacier was the Patrician, followed by the Keewatin, and finally the Labradorean. Thir- teen post-glacial Hudson Bay shore-lines are described, ranging from 190 to 600 feet above the present sea-level. .. 5. Sixth Annual Report of the Director of the Bureau of Mines, Van. H. Mannine, for the year ending June 30, 1916.— As will be remembered, Dr. Joseph A. Holmes, under whose direction the United States Bureau of Mines took shape and was developed to a high degree of efficiency, died on July 13, 1915, and his place was taken by the present Director, Dr. Van. H. Manning, who now presents the annual report for the year end- ing June 30, 1916. The work of the Bureau, as heretofore, is concerned perhaps first of all with efforts to insure the health and safety of the miners. It is stated that during the year, 8,400 miners were trained in mine rescue and first aid ; numerous acci- dents were investigated and a large number of men rescued. Further, the health conditions in mining towns have been inves- tigated, as regards, for example, the presence of the hookworm in California mines, and the prevalence of tuberculosis in Montana, and pathological conditions elsewhere. The Bureau is also actively engaged in endeavoring to accom- plish a greater degree of economy in the various mining and metallurgical processes. Perhaps the most important of these are concerned with the fuel supply of the country. It is shown that the petroleum industry, which in 1915 produced over 281,000,000 barrels, valued at about $180,000,000, or over 65 per cent of the total for the world, needs a most thorough investiga- tion as to the economy of production and elimination of waste, since it is estimated that the known supply is likely to be exhausted in less than thirty years, at the present rate of produc- tion. Economy in coal mining is not less important, as well as the attainment of the most efficient results in the use of the fuels. Much is being done now in these directions, but a more liberal supply of funds is called for, in order to accomplish the best results. As another example of the activity of the Bureau to be mentioned is the extraction of several grams of radium from ¢ar- notite ore at the Denver plant, this being accomplished at an esti- mated cost of $40,600 per gram. The publications of the year include a large number of bulle- tins, technical papers, and miners circulars. The following bul- letins have been received since the last list published (see vol. xli, pp. 838, 84): No. 105. Black damp in mines; by G. A. Burrert; I. W. Rosertson, and G. G. OBERFETLL. Pp. 88. No. 106. The technology of marble quarrying, by OLIvER Bow tes, 1916. Pp. 174, 12 pls., 33 figs. No. 108. Melting aluminum chips, by H. W. Grtuerr and G. M. James, 1916. Pp. 88. No. 116. Methods of sampling delivered coal, and specifica- tions for the purchase of coal for the Government, by G. 8S. Pore. Pp. 64; 5 pls., 2 figs. Miscellaneous Intelligence. 87 No. 118. Abstracts of current decisions on mines and mining, reported from October to December, 1915, by J. W. THompson. Pp. 74. No. 126; the same from January to April, 1916, by J. W. THompson. Pp. 90. No. 134. The use of mud-laden fluid in oil and gas wells, by J. O. Lewis and W. F. McMurray. Pp. 86; 3 pls., 18 figs. 6. An Introduction to Historical Geology ; by Wiu1am J. Mituer. Pp. xvi, 399, with 238 figures.. New York, 1916 (D. Van Nostrand Company).—In 1915 appeared the MHistorical Geology by Schuchert, earlier in 1916, Cleland’s Geology, Physi- caland Historical, and now a third book on the earth’s history by Miller of Smith College. While the subject is handled differ- ently by each of these authors, there is a similarity of treatment that is most marked in the first two books. Miller’s book is unlike the other two in that it is not accompanied by a text treat- ing of dynamic and structural geology. He aims at using a small number of technical terms, especially the names of fossils. As a rule, only the ordinal and class terms of organisms are given in the text, but the legends of the illustrations of fossils all have their specific names “in the interest of scientific accuracy [the names are not always correct] with no thought that these are to be remembered by the student.” The paleogeographic maps are those of Willis and De Lapparent. The book looks well, and the illustrations as a rule are good, adequate, and up to date. In general, it may be said that a great deal of information is here compacted into small space, and that the book is one of facts, unadorned to stimulate interest on the part of the undergraduate student. Ill. Misce,rLAngeous Screntiric INTELLIGENCE. 1. feport of the Secretary of the Smithsonian Institution, CuarLes D. Watocort, for the year ending June 30, 1916.—The most interesting announcement contained in Dr. Walcott’s report is the record of the gift, by Mr. Charles L. Freer, of a sum of $1,000,000 in cash, for the immediate erection of a building for the permanent preservation of the collection of art objects pre- sented by him to the Institution in 1906, and since increased by further gifts. The building will be of granite and located at the southwestern corner of the Smithsonian reservation ; it is expected that the work of construction will soon begin. The original col- lection consisted of about 2,300 paintings and other objects of art, and has since been increased to 5,346 items, including Amer- ican paintings and sculptures, the Whistler collection, and Orien- tal paintings, pottery, bronzes, and jades from China, Korea, Japan, and other Asiatic countries. During the past year the Institution has carried on the usual series of explorations and researches in the different lines of sci- ence. ‘These include the work of the Secretary in the Yellowstone Park and from there north in the Belt Mountains east of Helena. 88 . Scientific Intelligence. The work in the former area had to do with the investigation of the depositions from the geysers and hot springs as influenced by algze and possible bacteria; also the -collection of specimens of these deposits and of silicified wood for the National Museum. The investigations also include the study of the Paleozoic deposits of the Mississippi valley by Dr. Ulrich, work in the Ohio valley by Mr. Springer, and in Pennsylvania and Virginia by Dr. Wherry. Zoological and botanical explorations have also been carried on in St. Thomas, South America, and at several points in the Far East. Of the orants from the Hodgkins fund, the most important is that to Professor Angstrom for the study of noc- turnal radiation. Earlier results were published in 1915, and later ones have been extended to the Far North during the arctic night. The work of the Institution in its regular departments has gone forward as usual, except so far as the war has interfered with the Bureau of International Exchanges ; this work is summarized by the Secretary and the same subjects are discussed at length by the various gentlemen in charge in Appendixes I-VIII. As to the work of the Astrophysical Observatory, it may be noted that Messrs: Abbot and Aldrich have designed an instrument called the pyranometer, so constructed as to measure accurately the intensity of skyhght by day and the radiation from the whule sky at night. The tests which have been made with it prove its accuracy and general use, and further indicate that it may be suitable even for measurement of radiation in deep shade, as in forests and greenhouses. 2, Held Museum of Natural History ; FRreprrick J. V. SxiFF, Director. Annual Report of the Director to the Board of Trustees for the Year 1915. Pp. 74; 14 pls. Botanical Series. Vol. Il, No. 11. I. Contributions to North American Kuphorbiacee—VI; Il. Vegetation of Alacran Reef; by Cuartes F. Minitspaves. Pp. 401-431; maps and illustra- tions. Geological Series. Vol. Il], No. 10. Catalogue of the Collec- tion of Meteorites; by O. C. FARRINGTON. Since the earlier catalogue of 1903 (see vol. xvii, p. 329) the Field Museum has acquired the Ward-Coonley collection of 620 falls and an aggregate weight of 2,495 kilograms, bringing the total of the Chicago collection up to the impressive total of 657 falls and 7,560 kilograms. OBITUARY. Proressor Henry H. W. Pearson, the able botanist of Cape Town, South Africa, died on November 3 at the age of forty-six ears. ; Prorressor Henrik Moun, for many years director of the Norwegian Meteorological Service and an active contributor to the subject of meteorology, died in Christiania on September 12 at the age of eighty-one years. ——_ Wii's N, ATURAL SCE 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. J-148. Price List of Rocks. Mineralogy: J-109. Blowpipe Collections. J-74. Meteor- ites, J-150. Collections. J-160. Fine specimens. Paleontology: J-134. Complete Trilobites. J-115. Collec- tions.. J-140. Restorations of Extinct Arthropods. Entomology: J-80. 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-185. Bacteria Slides. Taxidermy: J-138. Bird Skins. J-189. Mammal Skins. Human Anatomy: J-16. Skeletons and Models. General: J-155. List of Catalogues and Circulars. Ward’s Natural Science Establishment 84-102 Colle#@ Ave., Rochester, N. Y., U.S. A. £ The American Journal of Science ESTABLISHED BY BENJAMIN SILLIMAN IN 1818. ConTRIBUTORS should send their articles two months before the time of issuing the number: for which they are intended. The title of communications and the names of authors must be fully given. Notice is always to be given when com- munications offered have been, or are to be, published also in other Journals. Thirty: separate copies of each article will be furnished to the author free of gost and without previous notice from him. They will be provided with a plain cover (but with reference to volume and year). If the author orders separate copies, they will be understood to be in addition to the thirty mentioned above, and he will receive a bill for the extra expense involved, as also for that of a printed cover (with title, etc.), when this is specially ordered. These charges will conform to the following schedule; but will be’ increaged if there are plates. No. Copies. 50 100 200 300 | 500 8 pages.--..-----. $2.25 $2.75 $3.50 $4.25 $5.50 |. 0" aha ios le ea em 3.50 4.25 5.25 6.25 8.00 Be her eee! ne et 4.75 5.75 7.00 8.25 10.50 Cone, 2p $1.00 $1.25 $1.75 $2.25 $3.00 (@s" During the Paper Famine the above prices may be somewhat increased (e. g., from 5 to 10 p. c.). C O-N-TAN-ES: Art. I1.—Retarded Diffusion and Rhythmic Precipitation ; by J. STANSFIELD a ca le -Ii.—Calorimetry by Combustions with Sodium Peroxide ; | by W.'G. MIxtar 2 S20 oe ee ee ee III.—Hébert’s Views of 1857 regarding the Periodic Sub- . mergence of Europé:-by €. Scnuchert 2 ssc 35 _ IV.—Lawson’s Correlation of the Pre-Cambrian Era; by A.C. Lanne (With Plate 1) 22322 ae 420 V.—A Table -for Linear and Certain Other Interpolations on Speetrograms ; by H. E. Merwin_..- --.-.---.-:- 49 VI.—On the Preparation and Ionization of the Dialkyl- phosphoric and Benzenedisulphonic Acids; by W. A. Drusgen and-A,; BR. RELTY (020230 = eS eee 57 VII.—On the Double Salts of Cesium @bloride with Cal- cium and Strontium Chlorides; by G.S. Jamreson __-. 67 VIII.—Crandallite, a New Mineral; by G. F. Loventin and W.T: SGaatieh ee 69 IX.—A Titaniferous Augite from a River, British Columbia; by C. H. WARREN, | J. A. Atuan and M. F. Conner. 75 SCIENTIFIC INTELLIGENCE. % Chemistry and Physics—Determination of the Density of Solids, H. La CHATELIER and F, Boaitcu, 79.—New Reagent for Free Chlorine, G. A LeRoy: The Right Honourable Sir Henry Enfield Roscoe, Str E. THORPE: - Text-book of Inorganic Chemistry, A. F. Hotteman, 80.—Organic Chem- istry for the Laboratory, W. A. Noyes: Note on the Electrolysis of . Gallium, H. 8. Unter: The Condensation of Gas Molecules, R. W. Woop, 81.—The Cosine Law in the Kinetic Theory, M. Knupsen, 83. Geology—A preliminary paper on the origin and classification of intraforma- tional conglomerates and breccias, R. M. Freip: Florida Geological Survey, Eighth Annual Report, E. H. SELLARDS: Study of the Morrison formation, C. C. Moox: Notes on the geology of Nelson and Hayes Rivers (Canada), J.B. TyRRELL, 85.—Sixth Annual Report of the Director of the Bureau of Mines, V. H. MANNING, 86.—An Introduction to Historical Geology, W. J. MILLER, 87. Miscellaneous Scientific Intelligence—Report of the Secretary of the Smith- sonian Institution, C.D. Watcortt, 87.—Field Museum of Natural History, Fi do. SRIFE; 88. Obituary—H. H. W. Pearson: H. Mouy, 88. | bLIWTALY, WV. YW. INAL. LViUSCUln. RMAUaae iirc Established by BENJAMIN SILLIMAN in 1818. THE AMERICAN JOURNAL OF SCIENCE Epirors: EDWARD 8. DANA. ~ ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or Camsrivaz, Proressors ADDISON E. VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT E. GREGORY AND HORACE S. UHLER, or New Haven, Proresson HENRY S. WILLIAMS, or Irnaca, Proressor JOSEPH S. AMES, or Batrimore, Mr. J. S. DILLER, or Wasuinerton. ~ FOURTH SERIES VOL. XLII—[WHOLE NUMBER, OXCIII}. No. 254—FEBRUARY, 1917. NEW HAVEN, CONNECTICU ee ke _ THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. stal Union ; $6.25 to Canada. Single numbers 90 cents. eee 3 1879. ‘VOL. XLT. : FEBRUARY, 1917. = Published monthly. Six dollars per year, in advance. $6.40 to countries in the ntered as second: class matter at the Post Officé at New Haven, Conn., under the Act LIST OF CHOICE SPECIMENS Apatite, Renfrew, Ontario, Canada. 50c. to $2.00. Corundum, var. ruby and sapphire, Franklin Furnace, N. J.; Orange Co., N. Y.; Montana; Litchfield, Conn.; Canada. 50c. to $10.00. Ruby Corundum, Macomb Co., N. C.; loose crystals. 25¢. to $1.00. Sapphire, loose crystals, Montana; Ceylon; Australia. 50c. to $5.00. Ruby, loose crystals, Hast India, $3.00 to $5.00 ; matrix specimens, $10.00. to $20.00. Stibnite, Japan; Felsobanya, Hungary. $1.00 to $10.00. Tetrahedrite, Kapnik, Hungary; Cornwall, England. $1.00 to $5.00. Scheelite, Zinnwald, Bohemia; Cumberland, England. 50ce. to $3.00. Benitoite and Neptunite, iit Bernnidine Co., Cal. $1.50 to $8.00. Vanadinite, Kelly, N. M.; Yuma Oo., Arizona. 50e. to $4.00. Cassiterite, Sciniaokemaale Bohomias Khrenfriedersdorf, Saxony. Toe. to $4.00. Rutile, Graves Mt., Georgia; Chester Co., Pa. 50c. to $5.00. Cooks crystals and crystals in matrix from Georgia.) | Diamond, loose crystals, Africa; German Africa; Brazil. $3.00 to $10.00. Diamond, in matrix, Kimberley, S. Africa. $15.00 to $35.00. Zircon, India, all colors, water-worn pebbles. 25c. to $2.00. Epidote, Untersalzbachthal, Tyrol; Riverside Co., Cal. ; Warren Cox, N. Y.; Arendal, NO, 7oe to $8.00. Zircon, cnet NC: ; Renfrew, Canada; Niedermendig, Tye Ural. 25c. to $3.00. Peridote, Indian Reservation, pebbles. 25c to $1.00. Olivine, Island of St. John, Red Sea, Egypt; crystals from $1.00 to $3.00. Sphene, Tilly Foster, N. Y.; Habachthal, Salzburg; Renfrew, Canada. $2.00 to $5.00. Pyromorphite, Ems, Germany; Phoenixville, Pa. $1.00 to $5.00. Dioptase, Kirghese Steppes, Siberia. $3.00 to $20.00. Dioptase with Plancheite, Guelab, German S. W. Africa. $6.50 to $10.00. Chrondrodite, Amity, Ellenville, and Tilly Foster, N. Y. 75c. to $5.00. Pyroxene, Phillipstown and Lewis Co., N. Y.; Franklin Furnace, N. J. ; Burgess and Eganville, Canada. $1.50 to $3.00. Wernerite, Burgess, Canada; Arendal, Norway. $1.00 to $2.50. Tourmalines, Norwich, Vermont; Maine; LEdenville, Gouverneur, Pierrepont and New York City, New York. 75dc. to $3.00. Orthoclase, Essex Co., N. ¥.; Colorado; Carlsbad, Bohemia. $1.00 to $2.00: a Carnotite, Telluride Co., Colorado, 25c. to 75c. Calciovorborthite, Telluride Co., Colorado. 50c. to $2.00. Beaverite, Beaver Co., Utah. 25c. to $1.00. ALBERT H. PETEREIT 81-83 Fulton St., New York City By THE ee AMERICAN JOURNAL OF SCIENCE ee O UL ES Mkt eS: | Arr. X.—The Water Content of Coal, with Some Idzas on the Genesis and Nature of Coal; by Epwarp Mack and GA, Horm. Partl. Water Content of Coal in relation to its Origin and Constitution. : Ar the present time our understanding of the nature of coal is largely limited to a knowledge of the chemical elements present in the coal substance, while there is little known about the chemical compounds present, or about the physical prop- erties of coal. Since the economic utilization of coal depends mainly on destructive distillation, it is of first importance to have definite information about the constitution of the various kinds of coal, and much attention has been given to this sub- ject, especially during the past few years, from the standpoint of distillation at low temperatures and pressures and by the use of extractive solvents. Commercially it has been found desirable to know the mois- ture eontent of coals and chemists adopted the usual analytical method for moisture, namely, drying the sample in an oven at 105°C. But it soon developed that the results were not con- cordant or reproducible, and for that reason several committees have been appointed to look into the question. Among these committees was one which reported to the Eighth Interna- tional Congress of Applied Chemistry in 1912.* At that time it was recommended that in making moisture analyses a special oven be used which would provide proper temperature control, since there is sometimes a variation of 15° C., or more, in dif- * Holloway and Coste, Report of Sub-Committee 10 of the International Commission on Analysis. Also, Hillebrand and Badger, Proc. 8th Intern, Cong. Appl. Chem., vol. x, 187 Am. Jour. Sc1.—Fourts Srerizs, Vou. XLII, No. 254.—Frspruary, 1917. 7 90 Mack and Hulett— Water Content of Coal. ferent parts of an ordinary air oven. Furthermore the mois- ture of the air passing through the oven was found to be a factor as well as the depth of the coal in the crucible, the dry- ing agent used in the desiccator in which the coal and crucible were cooled, etc. In short it was found necessary to specify numerous empirical details of manipulation, and even when these were strictly adhered to, it seemed impossible to obtain very satisfactory checks. Indeed, in this connection Hillebrand and Badger have emphasized the fact that loss of moisture is not the only change which occurs on heating a sample of coal at 105° C. in a drying oven. Other factors which may operate to cause a change of weight are: the sensitiveness of the powdered coal to the atmospheric conditions; the loss of volatile matter other than water; the taking up of oxygen which may be either oxygen added dir ectly to the coal substance, or oxygen combined with carbon or with hydrogen, and split off to form respectively carbon dioxide and water. Some coals are very easily ox1- dized, e.g. Archibald and Lawrence* have showed that the error introduced into the value of the moisture determination by the effect of oxidation may in extreme cases be very large. Moreover it can readily be seen that with different coals the moditying factors may assume the most varying values with respect to each other, and it is therefore desirable that as many as possible of them be eliminated. Yet, in spite of all its defects the present method will likely be retained for trade purposes in the commercial examination of coals. The fact remains that there is still a very decided need for a method which would give the real moisture content of coal and which would serve as a standard of reference. This would be of importance in research, especially in con- nection with the study of the constitution of coal. On account of the scientific importance of a definite know!l- edge of the moisture content of coal, many methods have been proposed, in which any effect due to oxygen is excluded, e. g.: drying in a vacuum over sulphuric acid or phosphorus pen- toxide for several days, heating the coal sample in a stream of dry nitrogen, hy drogen, or carbon dioxide; treatment of coal with calcium carbide and measurement of volume of acetylene gas generated; heating in boiling xylol, toluol, and similar liquids with measurement of the volume of water which distils over with the oil; collection of water given off by absorbing it in anhydrous calcium chloride or hot hme; using the reaction of water with the Grignard reagents; and several others. It is not necessary to discuss the advantages and disadvantages of * Determination of Water in Coals, by E. H. Archibald and J. H. Law- rence in J. Ind, Eng. Chem., iv, 258-62. Mack and Hulett — Water Content of Coat. 91 these various methods here, as they have been treated at length in the literature.* We will, however, consider at this place the method by which coal is dried in a vacuum over such dehydrating agents as concentrated sulphuric acid or phosphorus pentoxide. This method has been regarded as most reliable, since there is no oxygen in contact with the coal and also since it gives higher results than most other analytical methods. But it does not follow that it really gives the true water content of coal, for it is based on the assumption that by dehydrating coal in a vacuum at room temperatnre all of the water may be removed by these reagents in experimental time. This, however, involves a fundamental point in this problem; as a matter of fact, we have found that dehydration at room temperature still leaves with the coal a considerable portion of its water content. The following experiments, in which we used widely different kinds of coal, were made to test this point. The samples were powdered (20-30 mesh), weighed out in crucibles and permitted to stand in large vacuum desiccators (2-3™™ pressure) over fresh phosphorus pentoxide for 7-14 days or until repeated weighings from day to day showed con- stant weight to +0°2 milligram. The desiccator was provided with a small electric heater into which the coal crucible just fitted so that the coal could be heated inside the desiccator when desired. A very short thermometer (from a set) was adjusted so that the bulb was imbedded in the coal. After the coal had been desiccated until it showed no further loss, it was warmed up gradually and maintained finally at a definite tem- perature for a few minutes. In every case there was a marked loss of water from the coal which had apparently attained equi- librium with the first hydrate of phosphorus pentoxide at ordinary temperature. The accompanying table gives the results. None of the above losses resulting from the application of heat was due to the evolution of gases (other than water) since the amount of these was determined and corrections made. This was done as follows: Previous to heating, each of the samples had been run in duplicate, one crucible with its contents being placed in the electric heater just described, the other being heated to the same temperature (or 10°-30° higher) in an apparatus where the gas was collected and its volume and weight measured. The weight of the gas evolved in each case was subtracted from the total loss on heating and the amount of water lost by each of the six samples was thus obtained. The corrections were small, lying between the * For example, see Huntly and Coste, The Determination of Moisture, Jour, Soc. Chem. Ind., xxxii, p. 62. 92 Mack and Hulett— Water Content of Coal. TABLE I, Powdered coal dried in vacuum desiccator over phosphorus pentoxide with percentage loss of water : New Pitts- Thli- Wyo- River burgh nois ming Lignite Peat Weight of 2 each 1°6623 1°6131 1°3704 1°5853 1°7465 1°2875 sample Time : 1 hr. 45°98% 3 hrs. 175% 6°40% 5 hrs. 126% | 40-09% and day 152° 129° 7-66 «97174. 40-76 eee Sry ie 1°63 | |148 7°86. - 19°68) AIO 4th « 1°69.> 1:61) 4098 1285 bth 7-97 12°92. 41-20 iets 6th. 172 4 70 | (thi 1°74 1°76 8°08 12°93 56°16 8th “ 174 1°78 8°04 9th “ 8°02 TO Gh. = 12°92 41°32 56°18 AS ss 41°38 ithe = 41°39 On heating 3minup 13min 4min 12min 26min 10 min gradually to 125° upto upto upto up to upto for 230° 1a 5” 10° 190° 115. 213% 210% 858% 13:°20% 42°91% 56°63% 15 more 11 more 11 more min up min up min up to 205° to 200° to 170° 10:29% 13:36% 56914 Note—Weighings on successive days were made only roughly at 24 hour intervals. limits of 0°09 per cent. for New River and 0°61 per cent. for lignite. It has been shown* that there is not a measurable decompo- sition of these kinds of coals when heated in a vacuum to 250° for several hours—a point we will consider more in detail below. Since none of the additional water expelled by heat- ing above room temperature could come from a synthesis of water from oxygen and hydrogen, and since it did not behave like combined water or water of crystallization but showed a typical solution or adsorption curve,t the water must, there- * H.C. Porter and G. B. Taylor, The mode of Decomposition of Coal by Heat, Proc. Amer. Gas. Inst., vol. ix, 234, 1914. + Porter and Ralston. Mack and Hulett— Water Content of Coal. 93 fore, have been present as such. Later work indicates that, even by heating in the manner just described, we did not get all the water out. In our definition of the water content of coal this water must be included, consequently the method of drying over phosphorus pentoxide does not give the true water content, nor do any of the other methods which have been used, so far as we know. There does not appear to have been any attempt by workers in this field to state the problem clearly, that is to consider what is the precise significance of the term “ the water content of coal.” In order to get a working definition of the water content of coals let us consider the problem from the stand- point of the origin of coal. It has been fairly well established that the starting point of coal in the natural process of its making was peat. There is considerable evidence which indicates that this peat was formed in a manner somewhat analogous to the peat of the present age, though on a much vaster scale.* From year to year the decaying vegetable matter accumulated at the bottom of the swamp, and as it gradually sank below the surface of the water, it was protected from the oxygen of the air, and from certain ferment organisms which, together with the oxygen, were the active agents in the chemical changes that trans- formed the vegetable matter to peat. The rate of accumula- tion depended on the rate of the vegetable growth, and the efficiency of its preservation. As the deposit grew, the deeper portions became more dense, and according to a rough esti- mate, it. required something like a century to form a thick- ness of one foot of such dense peat. With regard to the changes taking place in the transforma- tion of peat to coal our knowledge is still largely speculative. One of the most favored ideas at present is that pressure plays the most important role in these changes—a pressure due to the overlying weight of debris and rock which covered up the swamp; or due to a metamorphic change such as a compres- sion in the earth’s crust. According to this view, the peat subjected to compression gradually became in successive stages lignite, sub-bituminous, bituminous and finally anthra- cite coal and graphite. Starting with peat and going down the series, chemical analyses show that extensive chemical changes have taken place,t the net result of which was a pro- gressive elimination of volatile matter such as carbon dioxide, carbon monoxide, methane and water, and in such relative pro- * White and Thiessen, The Origin of Coal (page 53), Bulletin 38, Bureau of Mines, 19138. +O. C. Raiston, Graphical Studies of the Ultimate Analysis of Coal. Paper presented at Cincinnati meeting of Amer. Chem. Soc., and pub- lished as a Bulletin of the Bureau of Mines. 94 Mack and Hulett— Water Content of Coal. portions that there was a continuous diminution in the per- centage of oxygen present in the substances which make up the coal. It must, however, be noted that water cannot be removed from peat or coal to any great extent by pressure, but as a result of the chemical reaction involved in the transformation of peat to coal the physical properties of the mass changed so that it was possible for the water partially to escape. We may question whether too much importance has not been assigned to the influence of pressure as a factor in the change from peat to coal; for pressure plays little, if any, role in the chemical reactions which take place in a condensed system, elther as regards the kind of the reaction or the rate. How- ever if the pressure causes a rise in temperature then the rates of the chemical reactions taking place are markedly accel- erated. Pressure has undoubtedly aided somewhat in the mechanical elimination of the more volatile and liquid pro- ducts of the chemical reactions, and has been a factor in giv- ing coal its firm, compact qualities. The possibility that pressure may have caused a rise in the temperature of the peat-like substance has been suggested. It: is, however, to be remembered that the mere subjection of a’ body to a pressure strain will not generate heat unless there is an actual decrease in the volume of the substance compressed, in accordance with ‘the energy equation : ? Peay == 2 calories: Heat is produced by the action of pressure through a distance, represented here by a volume change. And in the case of peat, coal and the intermediate substances, the deposit (from the old swamp) either yielded suddenly to an enormous pres- sure, breaking or crumpling, in which case the heat produced, would be dissipated or conducted off fairly rapidly by sur- rounding rock ; or the deposit and the rock above and below it. were pressed in such a way that they were not forced to yield suddenly, but remained under tke strain for a long time.* For this latter case we may make an approximate calculation. Assume that we have a deposit of dense peat 5 meters thick and that it is compressed at the rate of about one-half a meter a century, 1. e. one-tenth of its volume the first hundred years. Suppose the peat layer to be buried one kilometer beneath the surface, under a mass of debris, which had an average specific gravity of 3:0. Then the pressure on the peat per square centimeter would have been, 1000 (m) X 100(cm) X 3 grams 1 1000 (grams) * White and Thiessen, The Origin of Coal, pages 106 and 128 of the chap- ter ‘‘ Regional Metamorphism of Coal,” = 300 kilograms Mack and [Tulett— Water Content of Coat. 95 or about two tons per square inch. The amount of heat developed in a hundred years in each cubic centimeter of the peat substance would have been : 1000 (m) X 100 (cm) X 3 gm. X 0°5 meter 7 = o l ] S. 4-5 (m) X 100 (cm) X 424 (mech. equiv. of heat) Deh oat e If the specific heat of the deposit was -4, the heat produced during a hundred years would have been sufficient to raise the temperature of the deposit only 2°. Even assuming that the peat was subjected to a continuous prodigious pressure a thousand times as great as that here calculated, still the increase in temperature would have been only about 2° a month, even with perfect insulation: but peat and the adjacent material saturated with water are quite good heat conductors. There- fore, the heat must have been conducted off almost as fast as it was developed and it is difficult to see how the prospective coal bed could have been kept above the temperature of the surrounding rock strata for any considerable length of time. If, however, as a result either of folding or slipping of strata (friction), a higher temperature was maintained locally, there might have been a sufficient increase in the rate of the chem- ical reactions taking place to have caused a comparatively rapid alteration of peat- to coal. But this would be local in its effects and could be detected. Igneous intrusions do not appear to have played any great role. (See The Origin of Coal, p. 101.) The foregoing considerations are supported by the experi- ments of Dr. I’. Bergius.* who actually succeeded in the labora- tory in transforming cellulose or peat into a product very much resembling, and having the composition of, coal. Hitherto all attempts in this direction had failed because, when the peat or other raw material was heated in a closed vessel with the object of making it into coal, it was found impossible to control the temperature of the reacting particles, that is, to avoid a superheating of the materials due to the exothermic reactions, and for this reason the final products resembled charcoal or coke rather than coal, all the volatile and resinous substances having been driven out or decomposed by exposure to too high a temperature. In Bergius’ experiments this difficulty was overcome by heating peat together with a large amount of water, where the heat capacity of the liquid water was such that it served to absorb the excess heat from the reacting peat particles which otherwise would have been super heated. Thus the water acted as an excellent thermostat, making it possible to control the temperature at which the chemical re- * Jour. Soc. Chem. Ind., xxxii, p. 462, 1918. 96 Mack and Hulett-— Water Content of Coal. actions in the peat were taking place. In working above 100° C. it was of course necessary to maintain the pressure at or above that of the vapor pressure of the water in order to keep the water in the liquid state, but we fail to see that the pressure played any role in the chemical reactions. The interesting point of Bergius’ work is that he showed that the same product was obtained by heating for 8 hours at 340°, as for heating 64 hours at 310°. In other words the rate of reaction was donbled for every rise of 10°, which is the characteristic increase of rate for chemical reactions. On this basis, assuming that the chemical changes involved in the transformation of peat to coal in the natural process took place at the average temperature of 10° C., Bergius calculated that to form bituminous coal from peat, 2°° x 8 hours was the time required, or about 8 million years, an estimate in good-enough agreement with the calculations of the geologists. Bergius gives this figure with all “due reserve,” but it will readily be seen that this line of evidence, while perhaps not absolutely conclusive, gives very strong support to the idea that the transformation of peat to coal was the result of a series of slow chemical changes proceeding at about the same temperature as the earth’s surface, and quite independent of the presence or absence of pressure. Bergius showed that vegetable matter or peat could be changed to a residue having the same composition as bituminous coal, by reactions which require a definite “‘ temperature-time ” interval, but the product was an amorphous powder when separated from the water. Bergius further showed that this residue could be transformed to a substance having not only the chemical composition but also the texture and properties of anthracite, by subjecting it to a temperature of 340° and a pressure of 5000 atmospheres for some time. At this tempera- ture the chemical changes were quite rapid and the residues were compressed to a substance with the hardness of anthracite. It is quite possible that both a great pressure and a high temperature, considerably above ordinary rock temperature, have played a réle in the formation of our anthracite coals. The comparative suddenness in the change, from a chemical standpoint, from bituminous to anthracite coal, is indicated in the graphical studies of the Ultimate Analyses of Coal,* and is in line with these results of Bergius. It is not necessary to assume either pressure or any higher temperature than ordinary rock temperature for the trans- formation of peat through the various forms of coal to bitu- minous coal. The cellulose and protein substances of the original vegetable matter are gradually changed by the action * Ralston, Tech. Paper, 93, U. S. Bureau of Mines, 1915. Mack and Hulett— Water Content of Coat. 97 of fungus bacterial ferments to so-called “ humus” substances, while at the same time, but more slowly, the wax and resin- like matter decomposes, giving rise to hydrocarbons. The microscopic examination of peat and coals by Thiessen (I. ¢.) has revealed a fine spongy structure made up of a mat of resin granules, spores, and pollen exines, humic and woody remains. The original plant cells having been collapsed, there remains a mass which has passed through various colloidal stages. In the first stages peat may be largely a hydrosol. Accom- panying the slow chemical changes, reactions which take place of themselves with a decrease in free energy, there is an elimi- nation of water, carbon dioxide, methane, etc., and the mass becomes more largely a hydrogel which by the further loss of water and volatile products can harden to bituminous coal. The chemical reactions involved are generally of the nature of the formation of small molecules from larger ones, a change almost invariably accompanied by an increase in volume, and the effect of pressure in such cases is rather to oppose than aid the chemical reactions. The foregoing considerations indicate the importance of the “water content of a coal” in questions concerning its present status and constitution. In addition to the water originally present, the slow chemical reactions continually form water as one of the decomposition products, and consequently there must be an elimination of water from the coal beds during the slow transformations from peat to anthraeite. Nevertheless coals are normally dry. We may dismiss any idea that hydro- carbons, resinous bodies, waxes, carbon residues and humus substances combine with water in any such way as water of crys- tallization, but if these substances form a colloidal solution with water as one of the phases the matter is entirely comprehensible. The organic substances from peat to anthracite are in an ex- ceedingly finely divided state and present an enormous surface for the water phase. It is generally recognized that at the surface of heterogenous phases there is a concentration layer, and in the case of a colloid the total surface is so great that the actual amount of water concentrated in the film may be a con- siderable percent of the total mass. Some interesting estimates have been made of the thickness and nae of such films. For the thickness Lewis* assigns 5 x 107° ¢ This seems rather large, for taking the alheper of the water molecule as 0°2 — 1:0 x 107° em.,¢ we have a film 500-2500 times as thick as a water molecule. A more reasonable figure is that of Taylor,t 0°6 x 10-* cm. Sutherland gives *W. C. MeC. Lewis, On the Nature of the Transition Layer between Two Adjacent Phases, Phil. Mag., xx, p. 509, 1910. +W. Sutherland, The Molecular Constitution of Water, Phil. Mag., Nov. 1900, p. 473; and W. Ostwald, Grundriss der Kolloidchemie, 1912, p. 32. tW. W. Taylor, The Chemistry of Colloids, 1915, p. 225. 98 Mack and Hulett— Water Content of Coal. 1-26 10-° cm. If these latter values are to be regarded as probably nearer the truth, then the film cannot be more than a few molecules thick. The density of the film has been put (for 25° C.) at about twice that of water, namely at 2. ° It is to be noted, however, that the first few layers of molecules next to the colloidal surface represent a greater density than the layers more removed, which in fact probably grade off inaperceptibly into the bulk water, making it hard to say just exactly where the “ film”’ really begins. If we consider one cubie centimeter of the or ganic sub- stances which make tp coal, free from water and ash, it would have a density of not much over 1. If the eube were divided into little cubes 10-°cm. on each edge—a particle readily detected by the ultra-microscope—the area of the faces of these little cubes would be 600 square meters. If we suppose that they are covered with a layer of water molecules say 1x10~* em. thick with a density of 1, although it is probably greater than that, there would be -06° of water on these faces in a con- densed condition such that it would show little or no vapor pressure. If the layer of water were 1/10 of the thickness of the cube or 10 molecules thick with the increased size of the cubes, there would be about °7° or 60 per cent water intimately associated with this original cubie centimeter of coal substance and the vapor pressure of this water would undoubtedly be much below that of a plane surface of water at the same tem- perature. The substance would no doubt take up water from ordinary air and “ feel dry.” Some of the particles of coal are pr moe as small or smaller than we have assumed—others larger, the humus material in general being in the colloidal condition outlined. In the progress of the transformation of peat to bituminous coal, as the chemical reactions slowly result in the disappearance of the humus, the waxes, resins, etc. forming carbon residue (compo- sition of anthracite), carbon dioxide, volatile hydrocarbons, water, and the various other products, there results a er adual but marked decrease in the water content propor tional to the decreased colloidal surface. | The true water content of any coal might thus be taken as a rough measure of its humus content, and vice versa.” The condition of water on “adsorption” surfaces was first studied by Van Bemmelent who experimented in particular with silicic acid and ferric oxide gels. The method used was to place weighed amounts of the gels in desiccators where a definite vapor pressure was maintained by sulphuric acid of a * John Harger, Jour. Soc. Chem. Ind., xxxii, p. 460, 1913, suggested that the humus material contained most of the water found in coal. +J. M. Van Bemmelen, Die Adsorption, 1910. Mack and Hulett— Water Content of Coal. oo suitable density. There was no evidence of a combination between the water and gels. The results showed that the water was continuously removed as the surrounding water ‘vapor pressure was lowered, and in general when the vapor pressure was increased the gel would take on water; but this “hydration” curve fell above the dehydration curve, that is, the vapor pressure of a gel with a given amount of water had different values depending on its previous history and on whether it was being dehydrated or hydrated. Porter and Ralston* have found this same phenomena with different vari- eties of coal, so that in this respect also coal behaves like a col- loid. Whether or not there is here a coagulation of colloidal particles during dehydration of such a nature that there is less surface available when hydrating, the fact is that the process is not reversible, and if we allow a coal sample to lose or gain water we are not ina position to return it to the original con- dition and water content, with a definite vapor pressure. These facts and considerations make it necessary to fix some base line or starting point in determining the moisture content of a coal. There seems to be no evidence that extraneous water pene- trates the coal beds or coal. It would be only locally that spring water, or an unusually warm zone, might alter the normal content of a coal bed. In the main the water content would depend on the per cent and texture of the humus and colloidal matter in the coal, both of which are determined by its previous history. It would seem therefore entirely satis- factory to define the water content of coal as the amount of ' water present in that coal as it is found im the bed. Care must be taken that the sample is from a fresh break, and that it does not lose or gain water during the snbsequent handling and preparation of the analytical sample. On exposure to air coal begins to lose or gain moisture depending on the kind of coal, the temperature, and the mois- ture content of the surrounding air. In view of the forego- ing considerations we could not expect attainment of a condi- tion of equilibrium even with fixed air conditions or even attainment of a “steady” state. If the air is saturated with moisture any coal takes on moisture, the amount and rate depending on the nature and thickness of the water layer on the humus and colloidal particles and on the per cent of humus present. If this absorbed layer is very thin it has a very low vapor pressure, increasing with the thickness of this layer, and approaching but never reaching the vapor pressure of a plane water surface. Under ordinary conditions air is generally not saturated with water vapor and for the most part is far from it, so that generally a coal loses water on exposure to air. As * Tech. Paper 113, U. S. Bureau of Mines, 1915. 100 Mack and Hulett— Water Content of Ooal. the rate of loss of water is dependent on the thickness of the film and surface available we find the loss most pronounced in the lignites,* where ordinary moisture determinations are quite useless unless the sample has been properly collected and° handled. As we approach the anthracite end of the coal series the adsorption surface and moisture content are small and vapor pressure such that there is relatively little loss or gain under ordinary conditions. — In removing water from coal by the use of such drying agents as sulphuric acid or phosphorus pentoxide, time is gained by reducing the coal to a fine state of division, and by using a vacuum. ‘The rate of loss is a maximum at first, fall- ing off rapidly as the adsorbed water film becomes thinner and finally ceasing when the adsorbed film has been reduced to a vapor pressure about equal to that of the dehydrating agent. With our best dehydrating agents such as H,SO, and P,O, it would seem to be impossible to remove the films of water which are only a few molecules in thickness. Lewest con- siders that a certain sub-bituminous coal which loses 15 per cent of moisture at 105° is composed of humus 35 per cent, hydrocarbons and resinous matter 30 per cent, carbon residue 20 per cent,—or over 50 per cent of this coal might be con- sidered as colloidal and according to the approximate calcula- tion previously made the adsorption surface is suflicient to retain some 3 per cent of water in a layer one or two mole- eules thick, so that it would not be removed by vacuum desic- cation with our best dehydrating agents. Many assumptions were necessary in making the calculations but they should give an idea of the order of magnitude and in fact they agree satis- factorily with our experimental results in Table I. It was to be expected from the first that a rise of tempera- ture would increase the vapor pressure of the adsorbed film, thus yielding more water. For it is well known that the con- densation of water on fine materials is accompanied by an evolution of heat, i. e., the process by which water in mass passes into a film having a greater density, is exothermic. Consequently, according to the law of Le Chatelier, the appli- cation of heat to such a film as that on the colloidal humus must result in diminishing the density and the surface tension of the film, with a corresponding increase in the vapor pres- sure. In fact it has been shown that the density of film water approaches that of water in “bulk” as the temperature of both is raised the same amount, and of course at the critical tem- perature 365° the density of both become equal to that of water vapor. *G,_H. Frankforter, Jour. Amer. Chem. Soc., xxix, p. 1488. + Lewes, The Carbonisation of Coal, p. 48. Mack and Hulett— Water Content of Coal. 101 These considerations gave us a fairly definite conception of .the water content of coal and also indicated the line of attack which promised a successful solution of the problem of deter- mining the water. We cannot determine the water content of a sample of coal by vacuum desiccation at ordinary temper- atures on account of the difficulty of maintaining a sufficiently low vapor pressure in the desiccator and especially on account of the slowness of the escape of the last but considerable part of the water from the coal. If, however, we keep the vapor pres- sure low but heat the coal, we increase the vapor pressure of the film of adsorbed water, and so may make the time required for its removal as short as we wish. Oxygen must of course be ex- cluded and in a vacuum the rate of loss of water is somewhat greater. We have devised several satisfactory methods for accomplishing the desired results, but have studied particularly some questions raised by such procedures. Our preliminary experiments, ‘able I, showed that vacunm desiccation, with reagents which maintained the vapor pressure of water at about -01™", still left very considerable portions of the water in the coal, which however came off rapidly when the coal was raised some 190° in temperature, and while it was con- cluded that this probably removed most of the water from the coal we recognized the possibility that some water could be retained until a temperature approaching the critical tempera- ture of water was employed—but it was necessary to consider another result of heating coal. Coal is a conglomerate of con- stituents, some of which at least are decomposing as outlined above, and forming among other products water. The rate of these reactions at ordinary temperatures is immeasurably slow, but Bergius’ work shows that the rate doubles for each increase of 10°. Therefore we must avoid the conditions of tempera- ture and time at which the reactions are capable of yielding measurable amounts of water. The problem was then to find out whether we could heat the coal for a ‘“‘ temperature-time” interval such that the water film could be decreased to negligi- ble amount, without accelerating the decomposition reactions of coal substances to an extent such that a measurable amount of water would be formed. For the question of the preliminary decomposition of coal sub- stances we have some data from the work of H. C. Porter and G. B. Taylor.* These workers determined the amount and nature of the volatile products obtained from coal at 250°, 350° and 450° when maintained at these temperatures from 7 to 13 hours and at very low pressure. The following table shows some of the results which are of interest here. Four types of coal were used. * Proc. Amer. Gas. Assn., vol. ix, p. 234, 1914. 102 Mack and Hulett— Water Content of Coal. New River Pittsburgh Illinois Wyoming Per cent gas liberated NG Oy ee eee eT i "10 °35 1°04 Per cent CO,+CO in Sue AS cer ae “49 “70 “71 "85 Per cent gas liberated ab ODO ple eee °33 910 1°2 8 ib Per cent CO,+CO in the gag Vase See “49 "21 "45 64 At higher temperatures the amounts and rate of formation of gas increase rapidly. In addition to the above, “ tar ” and water were volatilized, but they may have been present as such in part or largely, and in view of the large percentage of CO, and CO in these first gaseous products it may well be that they also were largely present as such in the surface film, held much as is the water. This would be particularly true of the CO,. The moisture contents of these four kinds of coal increase, from the New River to Wyoming, in the same ratio as do the gas losses at these temperatures. So we may con- sider that the adsorption surfaces vary in about the same ratio, and it is entirely probable that these first gases were largely adsorbed along with the water film. There is no doubt that there is some decomposition of coal constituents at 250° but it is very slow and small in amount for a period of 6-7 hours, while for an hour at this temperature there would be formed only a trace of water, but at 350° for a comparatively short time we would have measurable amounts of water formed by decomposition of coal substances. Our work confirms these Mack and Hulett— Water Content of Coal. 103 observations and shows that a temperature even less than 250° for a reasonable time is sufficient for our purpose. Our first experiments were carried out with a very simple apparatus. A glass “alembic” of the form shown in fig. 1 was weighed, and after the lower bulb (5° capacity) was filled with coal, it was reweighed to get the weight of coal. The tube was now evacuated to about 2™™ pressure and sealed off. 10 15 20 25 30 ey) 40 Fie, 8. New River Coal a Re ras) = o @ An Temperature Aime eke 405%, 13077) 200° 108 Mack and Hulett— Water Content of Coal. of the large bulb and the burette and a vacuum ereated. The mercury, after this, was adjusted to the level L, and the coal which had been cooled and evacuated by connection with the pump through the three-way cock, was then joined to the burette by turning the cock. The gas from the heated coal collected in the large bulb and could be measured at any time in the burette by closing the cock and raising the leveling bulb. The coal bulb was heated in the vapor bath up to the point C, and the water from the coal condensed at the opening of the capillary tube and was rapidly forced by the slight pressure iu the coal bulb to traverse the capillary. The water collected on the top of the mercury, but a measurement of the volume of the water was not very satisfactory. We obtained the weight of the gas by transferring it from the burette by means of the mercury into a partially evacuate tube which had been weighed. The weight of the gas was subtracted from the total loss from the coal bulb to give the weight of water. The determina- tions made on the three coals with the alembic method were repeated, using the second method, the two sets of results agreeing very well, as fig. 7, with the results at 184°, will show. During the first few minutes water was removed more rapidly from the coal bulb of the second apparatus than from the first, but the flat portions of the curves are almost identical. If the final values represented by the flat portions of the time curves in figs. 2, 3, and 4, are plotted as ordinates against the temperatures to which the coal was heated, as abscisse, the curves shown in figs. 8,9 and 10 are obtained. The marked flattening of the curve at the higher temperatures indicates that very little more water could be removed from the coal by any further heating, though some water probably still remains on the adsorption surfaces even at the highest temperature, especially since pressure of the water in the alembic tube is appreciable, even when the alembie bulb is cooled. Below 250° very little gas indeed camie from the coals used and, as we have pointed out, it is probable that a considerable portion of this gas was merely adsorbed on the large surfaces of the colloidal coal. We were not particularly concerned with the gases evolved during the heating of the sample coal, except in so faras it enabled us to know whether or not decom- position of the coal was taking place. We did however make determinations of the amounts of gas liberated in the various experiments and the work of Porter and Taylor on these coals give us sufficient information on their composition. Summary. We have attempted to define the true moisture content of coals and have made some determinations on three typical coals, ee ee ee oe Mack and Hulett— Water Content of Coal. Fig. 9. Q, ae ‘E U 3) U - 100° 140; 180° 220° Ilinois Coal Percent H,0 Tem berature GO” yagO>) 10G) 120") 140° 160° 180° 200° 220° 240° 109 110 Mack and Hulett— Water Content of Coal. by a method which should give a close approximation to the true moisture content of these coals. This work has suggested several ideas on the nature and genesis of coal. It is evident that there is practically no definite information about the relation of moisture to such sub- stances as coal and im fact to all organic and colloidal substances. Laboratory of Physical Chemistry. Princeton University, June, 1916. Art. XJ.—An Apparatus for Determining Freezing Point Lowering ; by BR. G. Van Name and W. G. Brown. [Contributions from the Kent Chemical Laboratory of Yale Univ.—cclxxxvi. | Tue apparatus here described and figured operates, like the familiar Beckmann method, on the undercooling principle. It differs from others of its type both in the manner of effect- ing the undercooling, and in the fact that the whole process, undercooling, inoculating, freezing, and remelting in prepara- tion for a new determination, is carried out in a vacuum walled container (Dewar tube). With any type of freezing point apparatus the need of effec- tive heat insulation of the mixture of solid solvent and solution during the period of temperature observation is self-evident, and the use of a Dewar tube as container is a recognized feature of at least one well-known method, that of Richards. The convenience and accuracy of the Richards method are unquestioned, but it has the disadvantage shared by others of its type, that the concentration of the solution whose freezing point is measured cannot be definitely fixed beforehand. For this and other reasons methods of the undercooling type are in certain cases to be preferred, both for very exact measure- ments and for those of lower accuracy. The form of apparatus which we have used is represented in fig. 1, which shows all the important parts, but omits the mechanism for operating the stirrer, and a few other minor details. The Dewar tube, of about one-half liter capacity, is surrounded by a projecting jacket of tin, AA, and is closed by a cork stopper, B, perforated in four places to admit (1) the Beckmann thermometer, (2) the stem of the glass stirrer C, (3) a short tube, not shown in the figure, ordinarily kept stoppered, through which inoculation is effected, and (4) the cooling device D. This cooling device, consisting of two concentric Tak Van Name and Brown—Freezing Point Lowering. Hie 2: Fia. 1. 112 Van Name and Brown—Freezing Point Lowering. glass tubes through which cold brine® is circulated, is the essen- tial feature of the apparatus. | As the brine enters by the inner tube its temperature is shown by a small thermometer. At the beginning of a deter- mination the inner tube is drawn up slightly so that there is a free passage between the two tubes at the bottom. The brine passes down the inner tube, rises between the inner and outer tubes, thus exerting its cooling effect on the liquid outside, and finally escapes through the side tube. When the desired degree of undercooling has been obtained, the inner tube is lowered until in close contact with the outer tube at the bottom, where it has been ground to a fit, thus stopping the circulation through the lower part of the cooling system. This is the position shown in fig. 1. The lowering of the inner tube uncovers a hole in that tube, previously covered by the rubber stopper E, and the brine now circulates only through the upper portion of the cooling tube which projects from the apparatus. So long as the inner tube is in this lower position the upper part of the cooling device is thus kept at a tempera- ture below that of the solution, and by an amount which is easily regulated. This is an important point since it insures that such transfer of heat as occurs through the cooling tube shall be outward rather than inward, so that instead of pro- moting leakage of heat into the apparatus, as it might otherwise do, the cooling device actually tends to compensate for such leakage as occurs elsewhere. Fig. 2 shows the details of the cooling device, though with- out its thermometer. The inner tube is here shown raised, in the position for undercooling. All the upper part of the cool- ing tube, down to a point below the level of the liquid into which it dips, is surrounded by a narrow vacuum jacket, so that the cooling effect is exerted only below the surface of the liquid. This prevents accumulation of frost on the cooling tube above the liquid, which would be apt to cause premature freezing, and therefore permits the use of a much colder brine, making the whole determination a great deal more rapid. An unjacketed cooling tube can be used with equally accurate results, but only at a considerable sacrifice in convenience. The brine is drawn through a strainer of fine wire gauze from a large ice and salt mixture placed slightly above the level of the apparatus, and circulates through the system by gravity, the rate of flow being regulated by screw pinch-cocks on the rubber connecting tubes. It is collected in another * It is assumed here and in what follows that we are dealing with the meas- urement of freezing point of water solutions, since this is by far the common- est and most important case, and is the one to which our practical experience with the apparatus has been confined. The adaptability of the method for other solvents, of widely differing melting points, is, however, quite obvious. Van Name and Brown—Freezing Point Lowering. 113 receiver and from time to time returned by hand to the freez- ing mixture. A side branch in the tube which conveys the brine to the apparatus is connected with the water main so that when desired the brine temperature may be raised by admitting water in regulated amounts. ‘To insure thorough mixing of the two liquids they pass, just beyond their point of meeting through a roll of fine wire gauze, so arranged as to be easily accessible for cleaning (fig. 2, G). After each determination, the ice which has separated is ‘melted by shutting off the brine altogether and permitting water alone to flow through the cool- ing tube for a short time. The stirrer is driven by a motor, and except for a brief interval at the time of inoculation, is kept in motion through- out the experiment. The successive steps in determining the freezing point of a solution are as follows: The solution (of known concentration) is placed in the apparatus, the stirrer set in motion and the flow of brine started. When the desired degree of undercooling has been reached the inner tube of the cooling device is lowered to the position shown in fig. 1. After waiting a short time for temperature equilibrium the thermoin- eter is read (this reading is used only in calculating the degree of undercooling) and the inoculation effected in the usual way, whereupon the temperature rises and becomes constant at the freezing point. Owing to the good heat insulation the con- stancy of the freezing point is almost as perfect when working with a solution as it is with pure water. This eliminates the necessity for considering the so-called “‘ convergence tempera- ture,’ and makes the calculation of the correction for under- cooling simple and free from uncertainty.* The effective heat insulation is due not only to the use of the Dewar vessel but also to the fact, mentioned above, that the flow of brine through the upper part of the cooling tube tends to compensate for the heat leakage and the heat gener- ated by stirring. With regard to the possibility of using this form of apparatus for very accurate measurements we would call attention to the fact that by suitable provision for a closer regulation of the temperature, rate of flow, and path of the brine, this compen- sation could be made approximately complete, thus permitting the temperature readings to be made under practically ideal conditions. The accuracy would then be limited only by that of the temperature measurements. * This correction is based on the degree of undercooling, the heat of fusion of the solvent, and the heat capacities ( (a) of the liquid, “and (b) of the con- tainer, stirrer, thermometer, etc. In practice (a) is so large compared with (d) that a rough estimate of the latter is generally amply “sufficient for the purpose. 114 Van Name and Brown—Freezing Point Lowering. For most purposes, however, a comparatively rough compen- sation is all that is needed. With our apparatus as at first used there was a tendency, under ordinary working conditions, for the compensation to be incomplete, an unfrozen liquid showing a slow rise in temperature. This was corrected in a sufficiently effective manner by filing a small notch in the lower edge of the inner tube of the cooling device, thus permitting a very slight flow of brine at this point even when the ground sur- faces were in contact. We have put the method to a practical test in an extended series of determinations of the freezing points of water solu- tions. A Beckmann thermometer with a 5° scale was used throughout so that it was impossible to observe temperature differences of less than 0:001°, but within the limits of accuracy so imposed the results were very concordant and satisfactory. A few results obtained with two different solutions prepared from a carefully purified sample of cane sugar are given below by way of illustration : Gms. sugar Observed Corrected per ff Under- £.\p: Molecular 100gms.H,.O° lowering cooling lowering Avy. lowering 4°274 0°239° 1815, Opes BS 0°236 0°94 0°233 O88) 1°86, sf 0°235 0°67 0°233 8°521 0°487 LO 0°481 oe 0°489 1°48 0°480 ° 09 ec 0°485 0°88 0°480 ipa ees 3 0°488 1°27 0°480 The thermometer used had been carefully compared with a similar one certified by the U. 8. Bureau of Standards, and the observed freezing point lowerings as given in the second col- umn of the table have been corrected for the errors of the thermometer. Those in the fourth column have been further corrected for the undercooling. Considering the lack of sensi- tiveness of the thermometer employed, and the fact that it was not kept in ice-water when not in use so as to avoid the effect of slow volume changes in the glass, the values found for the molecular lowering agree about as well as could be expected with the best available values for such solutions, as determined by Raoult and others. - N. L. Bowen—Sodium-Potassium Nephelites. 115 Arr. XII.-—Zhe Sodium-Potassium Nephetites ; by N. L. Bowen. INTRODUCTION. Amone rock-forming minerals, especially those of the inter- esting alkaline rocks, nephelite takes a prominent place and for this reason has attracted considerable study on the part of mineralogists and petrologists. This study has revealed in nephelite considerable chemical complexity the explanation of which has been and, to some extent, still is a matter of con- troversy. The simple compound NaAISiO, is now rather generally regarded as the fundamental molecule of nephelite, but the composition of the natural mineral always departs from this markedly, showing a large content of potash, often con- siderable lime, and a variable excess of silica above the ortho- silicate ratio. A study of the fundamental compound and of the lime content of nephelite has already been made at this laboratory.* The present paper describes the continuance of this work in the study of the potash-bearing nephelites or the binary system, NaAl]SiO,— AlsiO,, The End Members. NaAlSi0,.—The compound NaAlSiO, can be prepared by sintering together Na,OO,, AI,O, and SiO, in the proper pro- portions at a low temperature (about 800°) to prevent loss of soda. Sintering must be repeated several times with inter- mediate grindings and then the whole may be raised above the melting temperature. The product is a clear homogeneous glass of the appropriate composition.t This glass crystallizes at low temperatures to the hexagonal form, nephelite, and at high temperatures to the triclinic (?) form, carnegieite. The inversion temperature of 1248° + 5 obtained in the earlier work was confirmed during the present investigation, as was also the melting temperature of carnegiete, 1526°. Pure sodium nephelite is hexagonal and negative. The re- tractive indices are wo = 1°537 + 002; «€ = 1°533 + 002. The density is 2°619 at 21°. Carnegieite always shows a complicated polysynthetic twin- ning. There are often two intersecting sets of lamella sug- gesting the microcline twinning and again three sets giving an hexagonal aspect. It has been considered triclinic, though it could be, perhaps, orthorhombic or monoclinic. The refractive *N. L. Bowen, this Journal, xxxiii, 551, 1912. +The reasons for this procedure. are given in the former paper, this Journal, xxxiii, 552, 1912. 116 NV. L. Bowen—Sodium-Potassium Nephelites. indices are y = 1°514 4. 002, a = 1°509 +. -002. It is biaxial with negative optical character and 2V —12°-15°. The density is 2°513 at 21°. Glass of composition NaAl$iO, has a refractive index 17510 + -002. A-ALSiV,.—The preparation of the pure potash compound is not so readily accomplished. The same precauticns of sin- tering at a low temperature before raising above the melting point are not adequate in tls case to prevent volatilization of the alkali, partly because the potash is much more volatile than soda and partly because of the high melting-temperature of the potash compound. In raising the charge to this temperature which is somewhat above the melting-point of platinum, con- siderable potash is lost. The pure compound cannot, then, be prepared in the dry way simply by mixing the components and melting them, though its preparation by this method has been claimed. Especially should one not attempt to prepare it by fusion in a platinum crucible, for the crucible melts and runs away when the sintered charge is still intact. In a poorly mixed charge melting can be obtained at comparatively low temperatures, silica and potash fluxing together and leaving much alumina undissolved, but this is not melting of kalio- philite. Nevertheless one can obtain a product with only a moderate deficiency in potash by sintering at a low tempera- ture as described and then rapidly raising to the melting temperature in an iridium furnace. The glass so obtained is apparently not very different in viscosity from the nephelite glass and crystallizes about as readily when not too rapidly cooled. In this way one obtains crystals which correspond in properties with the natural mineral kaliophilite, always, how- ever, with a little foreign material since the composition does not lie exactly at the compound K AISi0O,,. ) The artificial kaliophilite so obtained is hexagonal and nega- tive, w = 1°532 + 002, e= 1527+ 002. The prism and base are predominant with only an occasional truncation of the edge by a pyramidal face. One sees, then, the hexagonal basal sections and quadratic prismatic sections precisely as in nephelite. With the aid of potassium tungstate as a flux good erystals of kaliophilite were prepared at about 1300°. The largest of these were somewhat contaminated with the flux and showed higher indices of refraction than those given above, but the smallest crystals were found to agree exactly in ‘all their properties. Small crystals of kaliophilite were prepared hydrothermally by Morey in this laboratory. Kahlbaum’s potassium aluminate and potassium silicate were used. These were placed with NV. L. Bowen—Sodium-Potassium Nephelites. iT water in a gold crucible in a steel bomb after the usual method.* The temperature was kept at 600° and the pressure of water vapor at about 1000 atmospheres for 24 hours. An analysis of the kaliophilite crystals obtained is stated below. I II STO tut SIE 380. 738-07 Mle@eal Hii aka, 32:0 1118920 Oe ON Digit 29° 29-79 AONE Mesre ieh 953 “Tet hayas. 99°9 100-00 I. Artificial kaliophilite. Analyst N. L. Bowen. IJ. Theoretical composition, K A1Si0,,. The optical properties were exactly as given above. Formation of leucrte.—In attempting to prepare kaliophilite _ by heating glass of composition K AlSi0, at about 800° with potassium | tungstate occasional crystals of kahophilite were obtained but the great bulk of the product was a material of weak double refraction and conspicuous polysynthetic twinning. The refractive index was 1°509. This value, together with the polysynthetic twinning, gave the impression, at first, that a form of the potash compound corresponding with car negieite had been obtained. These properties agree, however, with those of leucite and an occasional crysta! suggested the typical icositetrahedron of that mineral though the majority of them were in branching forms. The amount of material was insufii- cient for analysis so a small portion was converted to glass in the oxy-hydrogen flame and the refractive index of the glass determined. The index was found to be about 1490 and to correspond with that of glass made by melting leucite from Vesuvius whereas the index of glass of composition KA1Si0, is 1508. The crystals were thus proved to be leucite and not a new form of KAISiO,. At 800° with potassium tungstate as a flux, then, KAISiO, glass is largely converted to leucite crys- tals whereas at 1300° it erystallizes as kaliophilite. Orthorhombic form of KALSiO,—In the charges heated in the iridium furnace occasional cr ystals were seen with polysyn- thetic twinning. These were thought to be kaliophilite twinned on a pyramidal face because the indices were sensibly those of kaliophilite. In the later work on intermediate mixtures it was found that those rich in kaliophilite always showed this twinned form when quenched from high temperatures. These better crystals were found to be biaxial. The crystals are usually interpenetration twins giving an hexagonal section *G. W. Morey, Jour. Am. Chem. Soc., xxxvi, 217, 1914. 118 =. L. Bowen—Sodium-Potassiceum Nephelites. divided into six sextants that extinguish together in opposite pairs. Occasionally polysynthetic twinning similar to that in albite is shown, but this appears to follow the same law as the previous case with the same prism face acting repeatedly as composition face instead of adjacent faces as in the case of penetration twins. Crystals of this orthorhombic form were made by fusing together silica, alumina and potassium fluoride over a Bunsen burner after the method of Duboine.* The twinning was not so frequently developed as in the crystals. obtained at high temperatures but occasionally appeared in typical form. The refractive indices are y = 1°556, a = 1°528, 2V = 39 + 3° negative. The product obtained by this method is not strictly pure but contains an amount of foreign material estimated under the microscope to be about 5 per cent. This consists of isotropic octahedra whose refractive index is about 1540. The corre- spondence with the compound K,A1,SiO, as described by Wey- bergt is complete. | Analysis of this mixture resulted as follows: fo) 6 © ERM ek Aa PT SS 36°7 38°07 NIK Oia er oath pe ptt he Spas 32°21 Ke: tree ere nes 29°72 99°8 100°00 I Product obtained from fusion with KF. Analyst N. L. Bowen. II KAISi0,, theoretical composition. It will be noted that the analyzed material is low in silica and high in alumina and potash and calculation shows that it corresponds with a mixture of 93 per cent KAISiO, with 7 per cent K,A],SiO,. The analysis and the microscopic examination together may therefore be regarded as definitely proving that the foreign material is K,AJ,SiO, present to the extent of about 5-7 per cent and that the twinned material corresponds with kahophilite in composition. Glass of composition K AlSiO, (approximately) has a refrac- tive index for sodium light of 1508 + -002. | A number of supposed other forms of KAISiO, have been encountered by various investigators, but the material was always poorly crystallized and the measured properties were hardly sufficient to individualize them as species. No crystal-- line forms of KAISiO, other than the two described were encountered during the course of the present work. The * Bull. Soc. Min. Pr., xv, 191, 1892; ‘‘In Canada-balsam sehr schwer sichtbar,” Weyberg, Centralblatt Min. 1908, p. 329. NV. L. Bowen—Sodium-Potassium Nephelites. 119 alteration product obtained by Stephenson is very similar to if not identical with the twinned form of KAI]Si0,.* Intermediate Mixtures. By mixing the alkaline carbonates, silica and alumina, in the proper proportion, sintering and then fusing, the intermediate mixtures containing up to 40 per cent of the potash compound were made satisfactorily. In mixtures richer in potash, however, the temperature which must be reached to obtain a homogeneous melt is higher so that a little alkali is lost. The loss increases as the potash end is approached. An attempt was made to avoid this difficulty by using kaliophilite made hydrothermally by Morey, and nephelite made in the ordinary way as the ingredi- ents of the mixtures. Even with this method the same loss of potash was encountered in those mixtures whose melting required a high temperature. The results from the potash-rich mixtures are, therefore, to some extent unsatisfactory. This loss of potash is not, however, so serious as to obscure the rela- tionships involved, though it does interfere with the accurate fixing of temperatures. A study of equilibrium in these mixtures was made in the ordinary way by the method of quenching. JBesides the loss of potash noted above certain difficulties were encountered in the optical determinations which are, perhaps, worthy of note. The hexagonal forms of the two components nephelite and kaliophilite are so nearly identical in properties that when obtained as small erystals embedded in glass, as they are in quenchings, it is impossible to determine whether the crystals are kaliophilite or nephelite or of an intermediate composition. As arule in such a ease one can obtain a clue to the composi- tion of the erystals by determining the composition of the glass, but in the present instance the extreme glasses and all intermediate glasses have nearly the same refractive indices and the composition cannot, therefore, be determined by such a method. To determine optically the question of solid solu- tion between the components, and its limits, one crystallizes a glass of intermediate composition and examines the product » to see whether it is homogeneous or not. In the present sys- tem, however, one cannot determine whether the fine-grained product so obtained consists of homogeneous mix-crystals or whether kaliophilite and nephelite are present as distinct phases. Studies of equilibrium are often made in systems of various kinds with little observation of the phases themselves. The work is almost entirely a measurement of the temperatures at which changes of phase take place in various compositions, the change being recorded, say, thermally or dilatometrieally. When these temperatures are plotted against composition it is * Jour. Geol, xxiv, 197, 1916. 120 WW. L. Bowen—Sodium-Potassium NM ephelites. usually possible to deduce a complete equilibrium diagram. Dependence had to be placed on this method to a considerable extent in the present case. The temperatures at which begin- ning of melting and inversion and completion of melting and inversion take place were measured by the quenching method. These, when plotted against composition, give a complete equilibrium diagram from which the composition of the erys- talline phases can be determined even though this cannot be done by optical means. Even in these determinations difficul- ties were encountered as a result of the near agreement of the refractive index of all the glasses with that of carnegieite. Thus when one takes erystalline nephelite-like material and holds it at successively higher temperatures, a temperature is finally reached at which a small trace of interstitial material of low refraction is formed but one is at a loss to decide whether it is carnegieite or glass. by raising the temperature the size of the interstitial patches can be increased until their nature is determinable. Unfortunately, however, one cannot be sure that small patches which were carnegieite have not, with the rise of temperature, been changed to patches of glass, indeed there are some mixtures in which this does take place within a narrow range of temperature. Then again there is the difficulty of determining the temper- ature at which carnegieite mix-crystals begin to melt, for the first trace of glass cannot be found under the microscope on account of the correspondence of refraction in glass and carnegieite. The microscope will detect the glass only when enough has been formed to give definite isotropic patches. It was found, however, that the first appearance of glass could best be deter- mined macroscopically in this case. Before the formation of olass the material is a white, opaque, sintered cake but on the formation of glass the interstices become filled with it, and since the index matches that of the crystals the charge becomes transparent and to all appearances entirely glassy even when under the microscope no unmistakable glass can be detected. Some difficulty is experienced, too, in determining the tempera- ature at which final solution of carnegieite takes place. Rare crystals of carnegieite in glass are easily overlooked, for they appear merely as indefinite birefringent patches without distin- guishable outline against the glass. These difficulties were finally solved, especially as a result of the assembling and correlation of all the results, but during the course of the investigation they were often very discouraging. Results of Quenching Hauperiments. The results of the quenching experiments may now be given in tabular form. (Table I.) NV. L. Bowen—Sodium-Potassieum Nephelites. . 121 TABLE I. Beginning of inversion of nephelite to carnegieite. Composition Initial Temper- Ne Kp* condition ature Tine Result Op. 1d)! CF ystalline 1276 1hr. hexagonal form persists unchanged he 1282 “« trace of carnegieite formed 90 10 . 1302 ‘“¢ hexagonal form unchanged Saea 2 1306 ‘“< trace of carnegieite 85 15 he 1322 ‘¢ hexagonal form unchanged canes i 1327 “< trace of carnegieite 80 20 of 1345 “¢ hexagonal form unchanged Cea ore 1350 ‘““ trace of carnegieite 715 25 ‘i 1366 ‘< hexagonal form unchanged Ca ce 1370 “< trace of carnegieite 70 30 in 1384 ‘“‘ hexagonal form unchanged ay : 1388 ** trace of carnegieite Completion of inversion of nephelite to carnegieite. 95 5 eels 1335 2 hrs. a little nephelite loge 1340 1/2 hr. all carnegieite 90 10 cb 1370 2 hrs. a little nephelite persists Gli Gr a 1375 1/2 hr. all carnegieite 85 15 op 1892 2hrs. a little nephelite persists es a 1397 1/2 hr. all carnegieite Beginning of melting of carnegieite. 95 5 ella 1448 J1hr. all carnegieite ah ae 1452 “ = crystals and glass stringers Siar t 1550 “ marked increase of glass Completion of melting of mix-crystals. 70 30 crystalline 1420 1/2 hr. glass and crystals Pema 1424 “« glass only 66 40 e 1463 “* glass and crystals aes mS 1467 “ — glass only 50 50 co 1520 “ glass and erystals pareere 1524 “ glass only 40 60 er 1570 “« glass and crystals Bera pine % 1574 «< glass only 30 70 pe 1629 “¢ glass and crystals as enen a 1633 ‘“¢ glass only Discussion of Results. The mixtures containing 50 per cent or more of KAISiO, always show a little glass even when quenched from tempera- tures below the eutectic. This behavior is the result of the deviation from true composition to which reference has already been made. Concerning the method of determining the tem- perature of beginning of melting in such material a word of explanation is, perhaps, necessary. ‘The glass formed at mod- erate temperatures occurs throughout the crystalline grains as shreds which are revealed only after close scrutiny under the microscope. These shreds do not increase very materially in amount as the temperature is raised and appear, moreover, to be of a highly viscous, presumably siliceous, glass which ocea- NV. L. Bowen—Sodiwm-Potassium Nephelites. 128 sions uo sintering of the powdered charge. On raising the temperature further, however, a point is finally reached at which rapid increase of glass takes place for only a few degrees rise of temperature. This glass has, moreover, quite different properties. It evidently flows comparatively freely and fills all the interstices of the powdered charge which becomes, even to the naked eye, obviously semi-vitreous. This temperature Fie. 1. WT. PER CENT KALSIO Nl Alsio ("26 20 a0 50" 60, 76 80 90 . A at which rapid increase of glass is first observed is regarded as the true temperature of beginning of melting. The method involves the personal equation to some extent. The temperature is probably somewhat lower than that which the true composition would show. Nevertheless there can be no question that the true curve would have a steep slope and approximate in position the determined curve shown in fig. 1. The temperature of completion of melting of these somewhat impure mixtures is easily determined in the ordinary manner. On account of deviation of the mixtures from the true com- position the various points are somewhat too low, no doubt. The effect of a small amount of impurity in lowering the tem- 124 WV. L. Bowen—Sodium-Potassiwm Nephelites. perature of beginning of melting may be very great if the impurity has a low melting point, whence the small amounts of glass found in these mixtures at moderate temperatures. The effect on the temperature of completion of melting, the so-called melting point, is, however, a more or less direct fune- tion of the amount of impurity. Since the deviation from true composition is small the position of the curve as far as it has been determined is believed to be substantially correct. The equilibrium diagram is of a peculiar type, though among those deduced by Roozeboom as possible in a system involving solid solution and inversion.* Between the potash compound and carnegieite there is a cicediie but with nephe- lite a complete series of solid solutions is formed. As was noted formerly, solid sclution cannot be definitely proved by optical means but such is the only possible interpretation of the thermal results. There is a continuous rise in the inversion temperature of nephelite as far as 35 per cent K A]SiO, amount- ing in all to more than 150°. Such achange of inversion temper- ature can be accomplished only through solid solution. At 35 per cent KAISi0, the inversion curve gives place to a melting curve which shows a continuous rise in temperature of begin- ning of melting. If the possibility is entertained that an hiatus may exist in the series in those mixtures close to kalio- philite whose beginning of melting cannot be accurately fixed, then it may be pointed out that there should be a correspond- ing break in the liquidus. There is, however, no break in the liquidus at least up to about 1580°. In the neighborhood of this latter temperature there probably is a break, but this is occasioned not by the appearance of kaliophilite as a separate phase but by the appearance of the orthorhombic form of KAISiO,. Solid solution as far as about 70 per cent KAISi0, is established, therefore, beyond all question, while none of the evidence furnishes any reason for doubting complete solid solution. It may seem at first thought that the evidence of the natural minerals is against complete solid solution. We _ have, for example, nephelite with upwards of 32 per cent KAISiO, in solid solution but none with a greater amount. Then at the other end we have a maximum of 10 per cent of NaAISiO, in kaliophilite. The interval is unbridged among known natural minerals, but this does not mean that no such minerals are possible. Nephelites which contain, say 20 to 30 per cent KAISiO,, are never found in contact with kaliophilite in such a manner as to indicate approximately simultaneous formation. Only such an association would prove that the nephelite was incapable of taking up more KAISiO,. For these reasons we *Zs. phys. Chem., xxx, p. 426, 1899. ee NV. L. Bowen—Sodium-Potassium Nephelites. 125 must conclude that the moderate amount of KAISiO, usually found entering into nephelite is the result of a deficiency of supply rather than of any inability on the part of nephelite to take up greater quantities in solid solution. Crystallization of a Typical Mixture. The phenomena observed during the cooling of a mixture in which perfect equilibrium is attained are, perhaps, of sufficient interest to be described in detail. A mixture of 80 per cent NaAISiO, and 20 per cent KAISiO, begins to crystallize at 1438° with separation of crystals of carnegieite containing 6 per cent KAISiO, (see fig. 1). As the temperature falls the liquid changes in composition towards the eutectic and the carnegieite crystals increase in amount and become richer in potash content. At the eutectic temperature, 1404°, they con- tain about 16 per cent KAISiO, and are then joined by hex- agonal crystals (nephelite) containing 35 per cent KAISiO,. The temperature remains constant until all the liquid has dis- appeared. Then as the temperature falls carnegieite immedi- ately begins to invert to nephelite and both change in composi- tion, becoming richer in soda, until at 1348° all the carnegieite has changed to nephelite which now has the composition - NaAlSiO, 80 per cent, K AISiO, 20 per cent. Relation between the hexagonal aud orthorhombic forms of K AlSiO,.— Both forms of KAISiO, can be obtained at low temperatures. Kaliophilite has been obtained hydrothermally by Morey and several others at temperatures in the neighbor- hood of 500°. The orthorhombic form was prepared in a simi- lar manner by Lemberg.* Both forms can likewise be obtained with the aid of various fluxes at moderate temperatures. Nevertheless a number of facts point to the stability of kalio- philite only at low temperatures and of the orthorhombic form only at high temperatures with the inversion temperature somewhere in the neighborhood of 1540°. In the present mixtures when working without fluxes the orthorhombic form was obtained only in mixtures containing 70 per cent or more of K AIlSiO, and in these only at high temperatures. More- over, ordinary uniaxial kaliophilite formed by crystallizing glass made in the iridium furnace, is converted at 1550° into excellent twinned crystals of the orthorhombic form. The change is here facilitated by the formation of a little liquid as a result of the deviation of the material from the true composi- tion. Pure kaliophilite made by Morey suffers a change at about the same temperature ora little lower, 1540°. In this ease no liquid is formed and for that reason the erystals are not as typically developed, but the quadratic section of kalio- * See Z. Weyberg, Centralblatt Min., p. 401, 1908. 126 = WV. L. Bowen—Sodium-Potassium Nephelites. philite with its parallel extinction is broken up into two or more areas which are in twinned relation with each other and whose extinction has no relation to the elongation. At about 1540°, therefore, or possibly somewhat lower, the hexagonal kahophilite is converted into the orthorhombie form. The formation of the orthorhombic form at low temperatures with fluxes is not at all incompatible with its stability at high tem- Wires et 1400 1200 20 30 40 5O 60 7O 80 10 90 NAALSIO, wr bene KALSIO,g peratures only, for many such forms can be so obtained, notable examples being the forms of silica.* The writer has not, how- ever, been able to convert the orthorhombie form into the hexagonal form, but this does not necessarily preclude an enan- tiotropic relation between them, for it is not uneommon to experience great difficulty in converting the high temperature form into its low temperature equivalent. *C, N. Fenner, this Journal (4), xxxvi, 359, 1013: Se ee ee ee N. L. Bowen—Sodium-Potassium Nephelites. 127 The orthorhombic variety is encountered in the mixture with 20 per cent NaAISiO, only at a considerably higher tem- perature, 1580°. The inversion point is, therefore, raised, which means that the low temperature, hexagonal form takes more NaAISi0, into solid solution than does the high tem pera- ture form. This is precisely as one would expect it to be, since the hexagonal form is so like the corresponding form of NaAISiO,, and, indeed, forms with it an unbroken series of mix-erystals. On the basis of these results fig. 2 is presented as represent- ing a partly hypothetical but very probable equilibrium dia- goram of the complete system. Some of the curves are shown in broken lines because they are less wel] supported by ascer- tained facts than the others, though they are believed to offer the only possible interpretation of the results obtained. The Composition of Natural Nephelite. It is now established beyond question that natural nephelites are of variable composition. That the molecules NaAISiO, and KAISiO, are fundamental constituents of nephelite may be regarded as definitely decided by the present investigation. Concerning the excess silica in nephelite above the ortho- silicate ratio there is still some controversy. Most investiga- tors believe that NaAISiO, and KAISiO, are fundamental and that the variable excess of silica is to be ascribed to the pres- ence in variable amount of a more silicious molecule. The suggestion that this molecule is NaAIlSi,O, was first made, apparently, by Clarke* and later stated in terms of the more modern solid solution theory by Schallert and the writer. Thugutt,§ however, would still assign a definite formula to nephelite and assumes that it is sNa, Al,Si,O,,.4Na,Al,0,.3K,Al,Si,O,, and that any deviation from this formula is the result of an admix- ture of products of its own decomposition, He therein ignores the fact that the later students of the composition of nephelite have taken the greatest precautions in selecting only absolutely fresh material for analysis.| He likewise fails to consider the fact that one can prepare nephelite showing the same variability by dry fusion, under conditions absolutely precluding the possibil- ity of aqueous decomposition.{ No consideration is given, *F. W. Clarke, U. S. Geol. Survey Bull. 125, 18, 1895. + Jour. Wash. Acad. Sci., i, 109, 1911. {This Journal (4), xxxiii, 49, 1912. $C. R. Soc. Sci. Varsovie, VI Année. Fasc. 9, 862, 1913. || Morozewicz, Bull. Acad. Sci. Cracovie, 958, 1907. “| Wallace, Zs. anorg. Chem., Ixiii, 1, 1909 ; and Bowen, this Journal (4), Xxxili, 00, 1912. 1928 WN. L. Bowen—Sodiwm-Potassium N. ephelites. moreover, to the fact that NaAISiO, and mixtures of it with KAISiO, can be prepared in a form identical with nephelite and, indeed, occur so in nature in the so-called pseudonephelite, whereas the molecules chosen by Thungnutt do not occur in forms at all suggesting nephelite. Several objections have been raised to the suggestion that the albite molecule, NaA1Si,O,, is the siliceous molecule pres- ent in nephelite. Thugutt points out that albite has never been isolated from nephelite, which is perfectly true, but neither has albite been isolated from labradorite. The sugges- tion of Foote and Bradley that the question of the condition of the excess silica should be left open* is safe, to be sure, but there are nevertheless good reasons for preferring the albite molecule. ; It was formerly considered, if a triclinic mineral was taken into solution by an hexagonal mineral, that the triclinic mineral must be dimorphous, must have an hexagonal modification. But this idea is now known to be contrary to well-ascertained facts. We may take the case of anorthite in solid solution in nephelite, which extends as far as 35 per cent anorthite, yet anorthite has no hexagonal modification. It seems to the writer, in the light of recent studies of the actual atomic structure of crystals, to be more reasonable to assume that the anorthite atomic pattern, though of triclinic symmetry, is nevertheless sufficiently close to the hexagonal symmetry of nephelite or sufficiently amenable to modification that, under the influence of the fields of force existing within a nephelite crystal, it is capable of being so modified as to conform with the hexagonal pattern, though not in unlimited amounts. Is it surprising, then, that the albite atomic pattern should be capable of precisely the same modification by a nephelite crystal, up to a certain limit, when we consider the fact that albite and anorthite are so nearly alike in pattern that they form an unbroken series of mix-erystals ? Or if we look at the question from the point of view of phase equilibria and remember that albite and anorthite in any system form not two phases but the single phase, plagioclase, then it becomes quite inappropriate to speak of the solubility in nephelite of anorthite and of albite except as limiting values of the solubilities of the phase, plagioclase. Any nephelite forming in a magma together with plagioclase must, of necessity, dissolve the amount of both anorthite and albite (i. e. of the phase plagioclase) that is required by the condi- tions prevailing, sufficient opportunity for equilibrium to become established being assumed. It may be said, then, that theoretical considerations afford * This Journal (4). xxxiii, 439, 1912. N. L. Bowen—Sodium-Potassteum Nephelites. 129 the strongest additional reasons for stating the composition of nephelite in terms of the molecules NaAISiO,, KAISi0,, CaAl,Si,O, and NaAISi,O,. These appear, moreover, to be entirely adequate for the purpose, apart from the small iron content. In the following are tabulated the percentages by weight of the above four constituents in some typical nephe- lites. For various reasons the sums of these constituents are not 100, an important one being that H,O seems to replace the alkaline oxides to a moderate extent in some nephelites. I II iil IV Vv VI MELO? VOTE IX Nees 65°0 63°0 56.0 ‘70°0 68°0 64°5 73°5 70°0 9°5 Gp opOrte ac O oO meno Ota Ore eso iba On s2:O =’ SEO 73 pag ea oO ai 220 0°0 2°d 4°0 2°0 4°0 1°5 JS Ch eke lame aes) 2°0 OOF 1670 9°5 5°0 Opn 0 0°0 I. Nephelite from Wausau, Wis. Weidman, Geology of North Central Wisconsin, Wisconsin Survey Bull., xvi, 1907, p. 317. Analysis I. II. Same. Analysis IT. III. ‘‘Pseudonephelite” Zambonini, Zs. Kryst., li, 606, 19138. IV. Nephelite from LEikaholmen, Norway. Foote and Bradley, this Journal, xxxi, 27, 1911. V. Nephelite from Mariupol. Morozewicz, Bull.. Acad. Sciences Cracovie, 958, 1907. VI. Nephelite from Coimbatore, India. Mem. Geol. Surv. ind,)xxx pt. 3, 187, 1901. VII. Nephelite from Kuusamo. Hackman after Ramsay, Bull. de la Commis. Geol. de Finlande No. 11, 9, 1911. VIII. Nephelité from Monte Ferru. Washington and Mer- win, Jour. Wash. Acad. Sci., v, 391, 1915. IX. Kaliophilite from Vesuvius. Zambonini, Mineralogia Vesuviana Appendix, p. 23. Dependable analyses of nephelite are apparently not very numerous and no very definite general conclusions can be drawn from them. Nephelite III shows no albite, i. e. has no excess silica above the orthosilicate ratio, and II has very little,* showing that excess silica is not essential. When the albite is high the anorthite tends to be low (I, IV, V and VIII) and when the anorthite is high the albite tends to be low (II and III). The suggestion is that the members of the former group were formed from solutions containing sodic *Some of the albite of II occurs as minute inclusions. Weidman, Wis. Survey Bull. XVI, p. 318. It might be assumed, therefore, that the nephe- lite should be saturated with albite and that the analysis is faulty. It is possible, however, to imagine conditions under which saturation would not occur even with albite inclusions. 130 WW. L. Bowen —Sodium- Potassium Nephelites. plagioclase and those of the latter group from solutions of calcic plagioclase, but it is impossible to be sure that this is true from the data at hand. The nephelites average upwards of 12 per cent plagioclase in solid solution. The peculiar twinned nephelite described by Esch, from Etinde voleano, West Africa, seems to correspond definitely with the orthorhombic form of the potassium-rich nephelites.* Esch observed slightly inclined extinction and therefore con- siders his mineral triclinic, but the correspondence is too great in other respects to render it likely that his form is distinet from the artificial form. It is to be noted that the nepheiin- ites described by him, though dominantly soda-rich, are some- times rather rich in potash as well, running to leucitites. Pos- sibly, then, some of the nephelite is potash-rich and closely related to the artificial varieties showing the same twinning. It was suggested in a former paper that Esch’s mineral might be carnegieite with its properties modified by solid solution, but this suggestion now seems quite unjustitied.t General Considerations. The mineral nephelite proves to be of rather complex consti- tution, a common feature of a number of rock-forming min- erals. The micas, pyroxenes and amphiboles exhibit this characteristic in even more marked form; indeed their consti- tution has not yet been satisfactorily explained. The difficulty is the result of the binding up, within one crystalline phase or mineral, of several of the components} of the magma. While the complexity of the individual crystalline phases is greatly increased their number is correspondingly decreased and, in some respects, a simplification of the process of crystallization results. This fact is the key to the problem of the “ gesteins- serie.” It is the crystallization from the magma otf these min- erals of continuously varying composition which results in the formation from a single magma of a series of rock-types show- ing that consanguinity which is found to characterize the petro- oraphic province. Nephelite syenite is undoubtedly the most important of the alkaline rocks. Alkalic feldspar is a prominent constituent and through it nephelite syenite is related to sub-alkaline types. The suggestion has been offered in another paper that the nephelite syenites are intimately related to the mica-bearing sub-alkaline rocks, biotite granites. They are considered to * Sitzb. Berl. Akad., xviii, 400, 1901. +N. L. Bowen, this Journal (4), xxxiii, 572, 1912. {Components in the specialized phase rule sense, not synonymous with constituents, § The Later Stages of the Evolution of the Igneous Rocks, Jour. Geol. Supplement vol. xxiii, 55, 1915. NV. L. Bowen—Sodium-Potassium Nephelites. 181 be probably a residuum from the granite magma, especially rich in volatile components. From the present work we obtain detinite proof of a fact long suspected, viz. the exist- ence in nephelite of the molecule KAISiO,. This commonly occurs in amounts of 15 per cent or more and is the same mole- cule that plays a fundamental role in the formation of the micas. At the same time it is not to be expected that any definite solution of the relation between two such rocks will be accomplished experimentally except in aqueous systems. The methods of attacking such systems, combining high tem pera- tures and high pressures, have been worked out by Morey* and his investigations will be extended to more complex sys- tems with a view to the solution of this and related problems. The existence in nephelite of an average of more than 10 per cent plagioclase is of importance in connection with such rocks as iolite and nephelinite. Though free from plagioclase as a separate mineral, it is plain that one cannot consider the crystallization of the magma except as a part of a plagioclase- bearing system. This system would include likewise the plagioclase rocks with which nephelinite is normally associated and consideration of the crystallization of such a system serves to emphasize the relationship of the types. The inversion temperature of nephelite does not prove to be very useful for the determination of the temperatures of for- mation of minerals. The pure compound NaAl]Si0O, inverts at 1248° but all the materials that it takes into solid solution occa- sion a sharp rise in the temperature of inversion. In order to appear in the carnegieite form, separation would have to take place at least as high as 1825°-1350°. The non-occurrence of carnegieite may, therefore, be considered as proof, if any be needed, that the separation of nephelite from magmas has always taken place at temperatures below those named. The occurrence of the pseudo-hexagonal, twinned form of nephelite in the nephelinite described by Esch is not to be con- sidered as evidence that separation took place at high temper- atures. It is true that this formis believed to be a high temper- ature form of potassium-rich nephelites or kaliophilites but, unlike carnegieite, it can form at low temperatures also. More- over it will persist at these temperatures just as the forms of silica, tridymite and cristobalite do. Summary. The present paper gives the results of an experimental investigation of the binary system NaAISiO,KAISiO,. The soda compound occurs in two enantiotropic “forms, nephelite * Jour. Am. Chem. Soc., xxxvi, 215, 1914. 132 WV. L. Bowen—Sodium-Potassium Ne ephelites. and carnegieite, with an inversion point at 1248°. The high temperature form, carnegieite, melts at 1526°. The potash compound shows two forms, kaliophilite, isomorphous with nephelite, and an orthorhombic form with twinning analogous to that in aragonite. The orthorhombic form is apparently stable at temperatures above 1540° and melts 1 in the neighbor- hood of 1800°. The potash compound has a entectic with carnegieite at 1404°. With nephelite it forms an unbroken series of solid solutions. It is concluded, therefore. that NaAIlSiO, and KAISi0O, are the fundamental molecules of natural nephelites. But, in addition to these, nephelites contain variable amounts of plagioclase in solid solution, the plagioclase varying from albite to anorthite, the latter accounting for the lime content and the former for the excess silica of the natural mineral. The composition of nephelite should, therefore, be expressed in terms of the four molecules NaAIlSiO,, KAISi0O,, NaA1Si,O, and CaAl,Si,O,. Reference is made to the petrogenetic importance of the occurrence of the last two molecules, viz. plagioclase, in nephelites. Geophysical Laboratory, Carnegie Institution of Washington, Washington, D. C., December 4, 1916. K. F. Mather— Pottsville Formations and Faunas. 133 Arr. XIII.— Pottsville Formations and Faunas of Arkansas and Oklahoma ; by. Kirtiey I. Marner. Pottsville formations of the Boston Mountains. Tuer Boston Mountains form an escarpment which overlooks the Ozark Plateau and extends from the vicinity of Muskogee, Oklahoma, eastward past Fayetteville, Arkansas, to Batesville, Arkansas. Their structure is monoclinal, the beds dipping gently toward the south. The strata exposed in the outlying erosion remnants and the terraced front of the mountains are in the main shales and sandstones with occasional limestones. Of these, the Morrow group embraces two units, the Hale formation and the Bloyd shale. The latter contains two lime- stone members, the Brentwood and the Kessler. The Morrow group rests unconformably upon the Pitkin limestone, or in places on the Fayetteville shale, and is overlain unconformably by the Winslow formation. Both unconformities are more pronounced toward the north and practically disappear toward the south where the group is lost to sight beneath the Wins- low. From east to west there is a change in lithology; in Oklahoma the Morrow is largely limestone while toward Bates- ville it is dominantly clastic. Limestone lenses in the Hale formation and the Brentwood and Kessler limestone members of the Bloyd carry an abundant marine fauna which has been described elsewhere.* A thin seam of coal in the shales be- tween the two upper limestone members contains a fossil flora of Sewanee affinities. The early Pottsville age of the group is further demonstrated by the marine fauna, in which a large number of residual Mississippian invertebrates are associated with forms of distinctly Pennsylvanian aspect. Potisville formations south of Arkansas River. Arkansas River and its tributaries in eastern Oklahoma and western Arkansas occupy a geosynclinal basin composed of Pennsylvanian strata. The lowest beds of the series are exposed along the southern margin of the basin, between the Arbuckle Mountains in south central Oklahoma and the Oua- chita Mountains farther east. Throughout this region the Atoka formation is especially conspicuous, both because of its great thickness, 3000 to 6000 feet, and because of the recur- rence of massive sandstone beds in the midst of the shales of which it is dominantly composed. Toward the north it is * Mather, K. F., The fauna of the Morrow group of Arkansas and Okla- homa, Denison Univ., Sei. Lab. Bull., vol. xviii, pp. 59-284, 1915, 134. K. F. Mather— Pottsville Formations and Faunas. overlain by the successive formations of the “ Arkansas Coal Measures.””* Along the eastern flank of the Arbuckle uplift, the Atoka is underlain conformably by the Wapanucka limestone+ which approximates 150 feet in thickness and rests with probable dis- conformity upon the Caney shale. Farther west, the Franks conglomerate is apparently the near-shore equivalent of the Wapanucka, Atoka, and possibly even higher strata as well. It overlaps the Caney and is found resting upon the early Paleo- zoic strata to the southwest. Traced eastward, the Wapanucka limestone at first thickens slightly and then decreases in thickness and comes to an end north of the Ouachita Mountains near the Arkansas-Oklahoma line. In the Arbuckle region, the Caney is over a thousand feet in thickness and overlies the Woodford chert, of probably Devonian age. Over a part of the area, however, the Syea- more limestone intervenes between the Woodford and the Caney and is correlated with an undifferentiated lower division of the Caney in other localities. In the vicinity of the Ouachita Mountains, the succession is somewhat different. The Wapanucka limestone disappears in southern Oklahoma and the Atoka formation rests directly upon the Caney shale, there about 800 feet thick. The Caney overlies with apparent, and probably real, conformity the Jack- fork sandstone and Stanley shale, each about 5000 feet in thiek- ness. Farther east, in ‘Arkansas, the Caney also disappears, and the Atoka rests upon the Jackfork.| The Fauna of the Wapanucka limestone. Through the courtesy of Mr. C. R. Thomas, of the Oklahoma Geological Survey, I have been enabled to study a representa- tive collection of fossils from the Wapanucka limestone. The fossils were obtained at several localities in and near the Atoka and Coalgate quadrangles midway between the Arbuckle and Ouachita Mountains. * Collier, A. J., The Arkansas coal field, U.S. Geol. Survey, Bull. 326, Dal2, 1907. Snider, L. C., Geology of East Central Oklahoma, Okla. Geol. Survey, Bull, ls, i. 8, 1914. + Wallis, | By LS Geology and economic value of the ° Wapanucka limestone of Oklahoma, Okla. Geol. Survey, Bull. 28, 1915. 1 Matt. oid. MG, Geology of the Arbuckle and Wichita Mountains, U. 8. Geol. Survey, Prof, Paper 31, p. 34, 1904. § Taff, J. A., Grahamite deposits of southeastern Oklahoma, U. 8S. Geol. Survey, Bull. 380, p. 289, 1910. Leg A. H., The slates of Arkansas, Geol. Survey, Arkansas, p. 48, 1909. K. F. Mather— Pottsville Formations and Faunas. 135 The complete faunal list is as follows :— Pentremites angustus Hambach Cromyocrinus n. sp. Archeocidaris sp. Rhombopora lepidodendroides Meek Cystodictya sp. Rhipidomella pecosi (Marcou) Productus nanus Meek and Worthen ? Productus morrowensis Mather ? Spirifer opimus Hall Spirifer goreii Mather Squamularia perplexa (McChesney) Spiriferina transversa (McChesney) Composita ozarkana Mather Composita gibbosa Mather Nucula parva McChesney Parallelodon sp. Myalina orthonota Mather Aviculopecten arkarsanus Mather ? Aviculopecten sp. Kuphemus carbonarius (Cox) Griffithides morrowensis Mather? All the genera listed above are present in the fauna of the Morrow group, and 17 of the 21 species have been found in the limestones of that group. This includes all the forms spe- cifically identifiable with the exception of the new species of Cromyocrinus. It is significant that none of these is known from the Kessler horizon without also being present in the Hale or Brentwood, while several of the Wapanucka forms are not known to occur in the Boston Mountains region above the Brentwood member. Moreover, none is recorded from the Hale limestone lenses, which is not also known from the Brentwood or higher strata. (Nucula parva is frequently found in the ‘‘ Coal Measures ” of the Mississippi Valley.) That the Wapanucka is the homo- taxial equivalent of the Brentwood is further indicated by the association of Pentremites angustus and Spiriferina trans- versa, residual Mississippian types, with Lhzpedomella pecosi, Productus nanus | ?, Spirifer opimus, Squamularia perplexa, and Huphemus car ‘bonaré ‘gus, proemial Pennsylvanian forms, and Composita gibbosa, a characteristic Brentwood species. It necessarily follows that the Wapanucka is of early Pottsville age. The fauna of the Atoka formation. The Atoka formation is nearly everywhere barren of fossils ; indeed it is highly probable that the formation is in part of non-marine origin. There are, however, local beds of calcareous sandstone near the base of the formation, which are crowded 136 K. F. Mather Pottsville Formations and Faunas. with the remains of marine organisms. These are known to outcrop near Clarita, Oklahoma, in the northwest corner of the Atoka quadrangle. The fossiliferous strata were discovered by Dr. L. C. Snider, formerly Assistant Director of the Okla- homa Geological Survey, and to him [ am indebted for the opportunity of studying their fauna. The fossils are preserved entirely as casts and moulds in the weathered portion of the outcropping beds from which all eal- careous material has been leached. The individual specimens are indiscriminately intermingled and many were broken before fossilization took place. It is quite clear that most of the shells were transported by waves and currents until they collected in quiet water, where they were buried beneath sandy sediments. No less than forty-two identitiable forms have come to hand from the one locality. These are listed below. Campophyllum torquium (Owen) Rbombopora sp. Prismopora concava Mather Chonetes choteauensis Mather Chonetes levis Keyes Productus fayettevillensis Mather Productus gallatinensis Girty ? Spirifer rockymontanus Marcou Squamularia perplexa (McChesney) Composita wasatchensis (White) Nucula parva McChesney Yoldia glabra Beede and Rogers ? Parallelodon pergibbosus Mather Caneyella? n. sp. Pseudomonotis precursor Mather Myalina cuneiformis Gurley ? Myalina orthonota Mather Schizodus affinis Herrick Schizodus wheeleri (Swallow) ? Schizodus telliniformis Girty Deltopecten occidentalis (Shumard) Aviculopecten arkansanus Mather Plagiostoma? acosta Cox Pleurophorus subcostatus Meek and Worthen ? Pleurophorus n. sp. Pleurophorus sp. Astartella n. sp. EKuphemus carbonarius (Cox) Worthenia tabulata (Conrad) Orestes sp. Euconospira turbiniformis (Meek and Worthen) Strophostylus remex (White) Platyceras parvum (Swallow) K. F. Mather— Pottsville Formations and Faunas. 1387 Aclisina ? sp. Orthoceras sp. Metacoceras sp. Pronorites lect var. arkansasensis Smith Gastrioceras listeri (Martin) Gastrioceras hyattianum Girty Gastrioceras angulatum Girty Gastrioceras carbonarium von Buch Gastrioceras kesslerense Mather The Atoka fauna as thus made known includes six forms not specifically identifiable and three new species, so that there remain thirty-three species known to occur elsewhere. Twenty of these have been described among the Morrow fauna of the Boston Mountains. One additional form, Pronorites cyclolo- bus var. arkansasensis, is known elsewhere only from the upper strata of the Morrow group in Carroll County, Arkansas. Another species, Strophostylus remex, occurs in the Lower Aubrey group of Utah, a horizon which may be closely corre- lated with the Morrow. The remaining forms are present in somewhat higher strata in this and adjoining regions. Four of the Atoka species, Prismopora concava, Myalina cunerformis ?, Worthenia tabulata, and Gastrioceras kessler- ensé, are especially characteristic ‘of the Kessler member of the Morrow as distinct from the lower fossiliferous beds of that group. The last named of these four is the most com- mon species in the Atoka fauna. On the other hand, two species, Pseudomonotis precursor and Parallelodon pergib- bosus, are noteworthy as representative Brentwood forms. The evidence is quite conclusive that this fossiliferous por- tion of the Atoka formation cannot be much, if any, younger than the upper beds of the Morrow group. The deposition of the Atoka sediments must, therefore, have begun in early Pottsville time. The age of the Caney, Jackfork, and Stanley formations. The proper correlation of the Caney shale, the Jackfork sandstone, and the Stanley shale present unusual difficulties because of the somewhat conflicting nature of the data avail- able. G. H. Girty in his description of the Caney fauna* has presented all phases of the problems involved. There have been found in the Stanley shale obscure plant remains which David White states are of Carboniferous age ; he concludes that “it is probable that they belong either in the upper part of the Mississippian or in the lower E Pottsville, but this point *Girty, G. H., The fauna of the Caney shale of Oklahoma, U. S. Geol. Survey, Bull. 3777, 1909. Am. Jour, Scl.—Fourts Serizs, Vou. XLII, No. 254.—Fresrouary, 1917. £38 - AO Mather Pottsville Formations and Faunas. requires additional paleontologic data for its determination.”* Six or seven thousand feet above the Stanley plant horizon there is found the marine invertebrate fauna of the Caney shale. This fauna is undoubtedly of Upper Mississippian age and the Caney shale may be correlated directly with the Moorfield shale, Batesville sandstone, and Fayetteville shale of the Boston Mountains.+ It is further demonstrable that the Caney of the Arbuckle region and the Caney of the Ouachita Mountains are one and the same formation. So carefully did Girty entertain all possible solutions of the really baffling problem, which would be presented if the Stanley flora should prove when better known to be of Pottsville age, that his conclusions have been misinterpreted. J. B. Wood. worth states, ‘‘ The fauna of the Caney shale is marine. Girty very ouardedly referred the beds to the Pottsville, which reference Ulrich (in a note to the author) later proved to be correct on stratigraphic and faunal evidence.”t Ulrich considers the Caney shale to be of Pennsylvanian age,$ but the evidence upon which his conclusions are based has not been published. Upon his authority, also, Purdue has referred the Jackfork and Stanley to the Pennsylvanian.| In this regard he follows the lead of Branner, who long ago included all the strata above the Arkansas novaculite in the ‘Lower Coal Measures,’ and reported the total pee of Pennsylvanian sediments in Arkansas as 23,780 feet.4] With the facts now available concerning the footie’ of the Wapanucka and Atoka formations, the chain of evidence would appear complete. The correlation of the Caney fauna with late Mississippian faunas in northern Arkansas is further strengthened by the determination of the early Pottsville age of the immediately overlying strata. Pennsylvanian sedimen- tation in the Arbuckle and Onachita regions evidently began with the deposition of the Wapanucka and Atoka formations. The Jackfork and Stanley must be of middle or late Missis- sippian age, and their enormous thickness is the result of unus- ual conditions of sedimentation maintaining in the Ouachita region at that time. The accurate placing in the time scale of the initiation of Stanley sedimentation must await further paleontologie discov- * Quoted by Girty, loc. cit., p. 8. + Girty, G. H., The fauna of the Moorfield shale of Arkansas, U. S. Geol. Survey, Bull. 439, p: 20, 1911. t Woodworth, J. B., Boulder beds of the Caney shaies at Talihina, Okla- homa, Bull. Geol. Soc. America, vol. xxiii, p. 457, 1912. § Ulrich, E. O., Revision of the Paleozoic systems, Bull. Geol. Soc. Amer- ica, vol. xxii, p. 352, foot-note; also pl. 29, 1911. | Purdue, A. H., The slates of Arkansas, Arkansas Geol. Survey, p. 48, 909. “| Branner, J. C., Thickness of Paleozoic sediments in Arkansas, this Journal (4), vol. ii, pp. 229-236, 1896. K. F. Mather— Pottsville Formations and Faunas. 139 eries, but it is now evident that that sedimentation could not have been contemporaneous with the uplifting of the Arbuckle region.* The orogenic disturbance which resulted in the rapid accumulation of the ten or twelve thousand feet of clastic beds composing the Stanley and Jackfork cannot be included with the “‘Culmides” of Chamberlin. It probably was a feature of the events which mark the transition between the Waverlyan and Tennessean systems of Ulrich. Instead of one great orogenic revolution closing the Paleozoic era there would appear to be at least four such disturbances: Waverlide, Cul- mide, Hereynian, and Appalachian. It is perhaps significant that the relation between the epoch of refrigeration implied by the ice-borne bowlders in the lower portion of the Caney shalet and the Waverlide disturbance is analogous to that betweer the much more extensive ‘ Permo- Carboniferous” glaciation and the more violent Hercynian revo- lution. Finally, it should be stated that the third leg of the paleon- tological tripod, the evidence accorded by vertebrate fossils, supports the conclusions which have been based upon the invertebrate remains. According to EKastman¢ the character of the Caney fish remains indicates their Upper Mississippian age so far as the evidence goes, though it is admittedly shght. Queen’s University, Kingston, Canada. October 20, 1916. * Chamberlin, R. T., Periodicity of Paleozoic orogenic movements, Jour. Geol., vol. xxii, p. 333, 1914. + Taff, J. A., Ice2-borne boulder deposits in mid-Carboniferous marine shales, Bull. Geol. Soc. America, vol. xx, pp. 201-202, 1907. ¢ Hastmam, C. R., Brain structures of fossil fishes from the Caney shales, Bull. Geol. Soc. America, vol. xxiv, pp. 119-120, 1913. 140 Vander Meulen—Two So-called Halloysites. Art. XIV.—A Study of Two So-called Halloysites from Georgia und Alabama; by P. A. vANDER MEvLEN. Tue term halloysite is usually applied to a massive, clay-like or earthy mineral, with a conchoidal fracture, waxy luster, and made up essentially of silica, alumina, and water. Accordin to Le Chatelier the composition probably is 2H,O, A1,O,, 2810, + Aq., or silica, 43°5 ; alumina, 36°9; water, 19-6. Analyses of this type of clay have been published from time to time, many of which show similarity to the theoretic com- position, but no other tests have usually been recorded on the samples thus analyzed. , Some time ago, Prof. Ries placed at my disposal, for chem- ical investigation, two samples of sedimentary clay, which have gone under the name of halloysite. Both possessed the usually accepted properties of halloysite.* They were almost pure white, had a more or less conchoidal fracture, and became somewhat translucent when placed in water. The first of these came from Chattooga County, Georgia. On close examination, it was observed to have on the surface, and lining cracks,a number of very fine needle-like crystals, and in some of these cracks also a small amount of a black powdery substance which proved to be an oxide of manganese. The clay itself was entirely homogeneous under the microscope, appear- ing either amorphous or exceedingly fine-grained, and free from the small crystals mentioned above. A number of the largest of these crystals were selected and tested by blowpipe methods. They were infusible, and became opaque on heating, had a hardness of about three, gave much water in the closed tube, contained no silica, but much alumina. These tests indicated that the crystals were hydrargillite. ‘Several of the best of these, about 1 mm. in length and 0:1 mm. in diameter, were selected, and the angles between faces in the prism zone measured. The angles corresponded with those of hydrargillite, described by Broggert on crystals of Norwegian material. The end faces were so small as to render measure- ment impossible. The indices of refraction determined by the minimum deviation method were found to be « = B=1°554 and y == 1:576. These values are somewhat higher than those obtained by broégger. A sample of the clay, free from crystals, and also from the black oxide of manganese, was finely ground and analyzed. The specific gravity of the fine powder was determined by the pyenometer method, the air being removed by means of a * Dana, p. 688, 6th ed. + Zeitschr. Kryst., xvi, 49. Vander Meulen—Two So-called Halloysites. 141 vacuum pump in the usual manner. The results on the air dried material are given below (No. I of table). The second clay (No. II of table), which comes from the Fort Paine Chert formation in Northern Alabama, contained no good crystals, but showed a very few small nodules of a substance resembling bauxite in appearance. The results of the analysis and the ‘specific gravity are given herewith. Analyses of Halloysite. I II III VE ee Fs) 85 BOM) 4. A8308 | 41-69%) _438°18¢ Al,O, ~ hood Stee ieee 44°38 39°94 30°88 39°21 Me Ao is ‘the tr 37 "15 ew dE none none 06 none G26 ee 8 tas ae tr. 2 pha Be cee INO Th 1:18 25 a We 08 [CC oe Paeeees eee none none > ioe Be none H, O below 108° C.-_- 1°61 1°28 7:97 3°39 H, @ above 108°... 16°68 15°04 F450 GS: Rofale.st wrt). 3" 99°62 100°02 99°98 100°24 Sp. gr. 200. 2-497 2°44] as 2-460 The results of these analyses show the clays to have a water content below that of halloysite, while the specific gravity is higher, for that of halloysite is usually given as 2°0-2°2. An analysis” of halloysite (No. III) from Horse Cave, Ky. by E. C. McNeil* is given for purposes of comparison. Analysis No. IV, made by the author, of a sample of clay which goes under the name of halloysite, from Grubb Mines, near Roan- oke, Virginia, is included because it has a somewhat higher alumina-silica ratio than does halloysite of the usually accepted composition, but agrees with it in the properties ascribed to this substance by Dana. The clays above described also contain a higher percentage of alumina than halloysite, and agree closely in chemical com- position with the high alumina flint clays described by Greaves- Walker.t These clays are, therefore, not halloysite. There is a good deal of doubt concerning the compounds present in clay mixtures such as those under discussion, espe- cially when the material is so very fine-grained that a micro- scopic examination does not even give a clue as to any of the constituents. It is exceedingly unsafe to calculate the mineral composition of a clay from its chemical analysis.t The com- pound assumed by many to be present in practically all clays is kaolinite, to which the formula H,A1,Si,Q, is usually assigned, *U.S.G.S. Bull. 591, p. 345. + Tr. Am. iia Soe., viii, 297. {H. Ries, Econ. Geol., ix, 402, 1914. 142 Vander Meulen—T'wo So-called Hatloysites. and many white clays, after washing, closely approximate this formula, In the case of the Chattooga clay the high alumina content together with the presence of hydrargillite crystals on the surface points strongly toward the probability that it eon- tains this latter compound, The Na,O might perhaps be con- sidered as some undecomposed mineral sueh as albite, or it may be there in some other form, In any ease it is better to leave Pra, 1, PLRCENT WATER LosTr it out of consideration, and to assume that the clays are mix- tures chiefly of kaolinite and hydrargillite, with lesser amounts of other substances. If the relative amounts of these com- pounds be caleulated, the results given below are obtained, Chattooga Alabama Olay Clay Kaolinite ...... 77'08% 93°12% Hydrargillite .. 21°32 4°83 WV BLOT wae take 7 aU 16] By employing dilute sulphurie acid at a fairly low tempera- ture it was thought possible to dissolve from such mixtures Vander Meulen—Two So-called Halloysites. 143 the hydrargillite, leaving the kaolinite unattacked. For this purpose, sulphuric acid of four different concentrations was employed. These were 10, 15, 20, and 25 per cent. In each case 0°2 gram of the finely powdered clay was digested with 50 grams of the acid. The temperature was held at 55°-60° C., and the digestion continued for three hours with occasional stirring. The solution so obtained was filtered. Alumina and silica were precipitated with ammonium hydroxide from the hot solution, collected on a filter, dried, ignited in a platinum crucible, and weighed. The contents of the erncible were treated with a few drops of sulphuric acid, and fumed down with a little hydrofluoric acid, ignited, and again weighed. The final residne was considered to be alumina and the loss in weight due to the hydrofluoric acid treatment, silica. The results of these experiments are given below. Chattooga clay Alabama clay Conc. of acid. 441,03 4ZS3i02 %Al,03 ZSi02 10% 8°50 2°75 2°65 1°€5 15 9°40 2°27 3°10 1°45 20 10°55 3°20 4°50 2°20 25 10°20 2°00 4°55 1°45 Even at the relatively low temperature employed, some decomposition of the hydrous aluminium silicate must have taken place. The results do not show any regularity, and no specific conclusions can be drawn from them, except perhaps, that the use of dilute sulphuric acid for the rational analysis of clays is of doubtful value. Brown and Montgomery,* and others, have shown that kao- jin when dehydrated loses bnt little water at 300° C., and loses most of its chemically combined water between 450° and 500° C. It appeared interesting to study the dehydration of these clays in an analogous manner. Approximately 0°5 gram portions of the finely powdered clays were weighed out in platinum crucibles and heated side by side. In each case the heating was continued until the loss in weight on reheating for fifteen minutes became less than 0-2 mg. This was found to be the nearest approach to equilibrium that could be reached in a reasonable time. Up to 200°C. the erncibles were heated in an air oven, but for higher temperatures a small gas-mufile furnace was employed, the temperature being measured with a thermo-electric pyr- ometer. The probable error in temperature readings was less than 5°C. Weighings were made at intervals of about 50° C. until all the water had been removed. Curves were plotted (fig. 1) using per cent of water lost as abscisse, and tempera- ture as ordinates. *Tr. Am. Cer. Soc., xiv, 709. 144 Vander Meulen—Two So-called Halloysites. A comparison of these curves is interesting. The Chattooga clay, which contains approximately 21 per cent of hydrargillite, probably in a colloidal form, lost about six per cent of water below 300°. The bulk of the water was ‘removed between 450° and 500°. This amount of water would be lost if the hydrargillite gave up most of the water combined with it, below 300°. The water remaining combined with the alumina appears to be removed at the temperature at which most of the water is removed from the kaolinite, between 450° and 500°. The Alabama clay, which contains only about five per cent of hydrargillite, also loses a small amount of water below 300°, after which the bulk is again given off between 450° and 500° . In both cases the last traces are removed only above 900°. Conclusions. The chemical analysis of the two clays shows that they con- tain less water and more alumina, and to have a higher specific gravity than halloysite. They may be regarded as mixtures of kaolinite and hydrargillite, with a amounts of other sub- stances. It also points to the fact steady emphasized by others, that some of the high alumina clays, instead of being composed chiefly of some hydrous aluminium silicate other than kaolinite, are probably in many cases mixtures of kaolinite and some aluminium hydrate, like hydrar gillite. Under the conditions of the experiments, digestion with dilute sulphuric acid for the purpose of determining the con- stituents of a clay mixture is of doubtful value. Dehydration of a clay mixture containing hydrarg villite tends to remove most of the water of the hydrargillite below 300°; the remainder is driven off along with the chemically combined water of the kaolinite between 450° and 500.° The last traces of water are removed only above 900.° A erystalline substance in the Chattooga clay was shown to be exceptionally well crystallized hydrargillite, and its indices of refraction were found to be slightly higher than those usually given for this substance. Mineralogical Laboratory, Corneii University. C. Barus—Non-reversed Spectrum Interferometry. 145 Art. XV.—WMethods in Reversed and Non-reversed Spectrum Interferometry (continued); by Cart Barus.* 11. Prism methods without grating.—A more interesting method, in some respects, in which the grating is entirely dis- pensed with, is shown in fig. 15. Z is the beam of white light from a collimator, P a refracting prism (here with a 60° prism angle), JZ and WV the opaque mirrors with either or both on a micrometer, P’ a silvered reflecting prism (here right angled). The telescope is at Z and should have high magnification. The rays Z are refracted into abc and a’b’c’ and the two spectra observed by the telescope at Z. Each of the prisms should be.on three adjustment screws, as well as the mirrors. P must be revolvable slightly around a vertical axis and capa- ble of fore and aft motion. /P’ is preferably a large prism placed on a tablet. The ravs 6 and 0’ are made collinear before /’ is inserted and both the rays e¢ and c’ must come from near its edge. The fringes are strong and large and lie within a relatively remarkably wide transverse strip. This may be ten or twenty times as wide as the )),D, doublets, which in view of the small dispersion are hardly separated. or the same reason, moreover, the range of displacement of J/ within which fringes are visible, rarely attains half a millimeter. Within this the fringes grow from the fine hair lines, usually oblique, to their maximum coarseness. Apart from the small range of displacement, these fringes are available for measurement. If both mirrors Jf and JY are on micrometers, they may be brought forward or the reverse, alternately, and the range in- creased 5 or 10 times. To change the form of the fringes, the first prism, P, may be tilted slightly on an axis parallel to Z7) fig. 15. The fringes then pass through a maximum in the vertical direction (linear phenomenon). Fore and aft motion of P rotates the fringes partially toward the horizontal; but, as a rule, the component beams } and 0’ pass beyond the edge of /’ and the fringes vanish. Just before this (the spectra separating), the strip within which the fringes lie, widens enormonsly. In other words, the breadth of the phenomenon depends on dit- fraction, not on dispersion, so that even though the prism P scarcely separates the LD lines, the striated strip has about the same width as when it is produced by highly resolving grat- ings. | * Abridged from a Report to the Carnegie Institution of Washington, D.C. See this Journal, pp. 402-420, November, 1916. 146 C. Barus— Methods in Reversed and It is preferable to use sunlight directly (without along focus condensing lens), as there is a superabundancee of light. The best results are attained with a long collimator. A spectacle lens with a focal distance of one meter is excellent. The range of displacement of J/ is not increased, but the spectra and fringes become very sharp. If with the large collimator the spectra are just separated in the field of the telescope, by fore and aft motion of /, a magnificent display appears resembling Figs. 14-19. a thick twisted golden cord. With further separation confocal elliptic fringes often cross the gap, as in fig. 16. Here a and # are graphs suggesting the wave-lengths of the two spectra, g being the gap or deficient overlapping. The appearance in the tele- scope is shown at y, Sand S’ being the spectra. When the fringes are erect, huge vertical furrows may lie in the gap. When the gap is closed, the linear phenomenon reappears. These enlarged fringes vanish, however, within one-fourth mil- limeter of displacement at JZ. In further experiments, screens s, s’ (fig. 15) were placed in the paths of the pencils 5, 6’, so that they were compelled to pass through vertical slits, 1 to 3™™ wide, in the screens. In this way the interfering rays were identified. The first vertical ee ee ee ee | ee Non-reversed Spectrum Interferometry. 147 hair line fringes came from rays about 5"™ behind the edge of the prism P’. Hence the pencils were here about 1:2 apart when they entered the telescope. The largest and last of the fringes come from close to the edge of P’. The experiment was varied as follows: supposing both screens s and s’ placed as far to the rear as the visibility of fringes permits, let the former, s, be slowly pushed forward. The fringes then con- tract from the very broad set, fig. 17, case 1, to the strong and narrow set, 2 (which is a mere line for a full wave-front), and then expand again to 3. If now s is left in placeand s’ moved forward, slowly in the same way, the identical contraction and expansion, 1, 2, 3, is reproduced. The screen s’ may then be left in place and s in turn slowly moved forward with the same results, etc. (there may be six alternations), until finally the effective parts of the pencils } and 0’ are beyond the edge of the prism PP’. In case 2, the two slits s and s’ are obviously _ symmetrical to the interfering rays, whereas in cases | and 3 the diagonally opposite edges of the slits, s and s’, limit the efficient pencils rigorously to a sheet. A similar result (passage of case 2 into 38, fig. 17) may be produced by moving / forward, the case 3 appearing just before the pencils 6 6’ leave the edge of P’. Again, when J/ is moved rearward, when both 6 and 0’ are near the edge of P"', the cases 2, 3 are obtained. In general the width of the diffraction pattern increases without changing the size of fringes, as the width of the available wave-front decreases. 12. Displacement parallel to rays—It now becomes of im- portance to test the range of displacement as modified by the angle of reflection, increasing from 6=0. It is therefore desirable to make a few direct measurements. The angle @ at P, fig. 15, was found to be about 49° 45’, so that the total angle at MJ is 6 = 40° 15’. Mand JV are both on micrometers, with the screws normal to their faces. P’ is on a micrometer with its serew parallel to 0b’, so that this prism is shifted right and left. The range of displacement was found at ee ae ot O42 har 04K 939; = 076°, i about ==) 07s) Wy = "140, where x = 2e cos (90 — @)/2 and 2y are the corresponding path differences between the inception and evanescence of fringes. With a very fine slit, 2¥ was possibly smaller (see fig. 18). The question at issue is thus in the first place, how the value of 2y compares with «; for in the former case the angle 6 is effectively zero. In other words, when J is displaced from M to M’ over a distance e, the pencil 4, fig. 18, changes to 6, and is soon lost at the edge of P’; whereas, when / Is dis- placed in the direction 6b’, over a distance y, the rays } and 0’ 148 C. Barus—WMethods in Reversed and do not change their point of impact at the prismatic mirror P’. Jf pp represents the principal plane of the objective of the telescope and /’ its principal focus, there should be no acces- sory effect for the case y as compared with the case #. TABLE I. REVERSED SPECTRA. Refracting prism, 6 = 49° 45’; B=4-6x10-" (assumed); z=H(u—1)+2BE/?? Remarks jes e is 2y Zz LL | cm. | Sem. || Sem. a, 1em: em. Mirror right, Plate* right | °736 | -203 | °381 | -405 | -3901 | 1°526 cra = Lett, lett | “f | 208 | -39L | °408 |+°0195 Tight, riots of e203 sso e020 | ee SG: lets <<. Jett eee 20 | 389 AOL co Tieht, -. °F riohtne 136,) 206 oop +06 mated Ere Fo lett [ee | 207 | “seo ae 0G « yight, “> right | 434 | *194 | 9345/0491 /"\-a3 ia Geen er latte left’ | se) “7964 286)| 250m) = Enon ss Left, te Gprehite eS * Te WG T° 237k Seer ee | * Plate slightly wedge-shaped. + Different part of micrometer screw. Results bearing on this subject are given in Table I, in which the displacement, e, observed at df and at J, as well as the displacement y at /’, are recorded, when a plate of glass of thickness / is inserted normally to the rays 6,0’. The cor- responding air path difference computed from £, p, B, dr, should be z, nearly. This is about the value (2y) observed, remembering that to set the micrometer, fringes of a particular pattern must be selected. The rotation of fringes being but 90°, or less, there are no fiducial horizontal lines. The values of « computed from @ and e, however, certainly fall below z, being about 6 per cent and 3 per cent short of it in the two cases, respectively: or again w is -019™ and -014 (about 5 per cent) smaller than the mean values observed for 2y. This extra 5 per cent of path difference can not be an error of observation, or of adjustment, but must be interpreted as the path difference added, when the pencil shifts towards the edge of the prism (w) instead of being stationary as in y. In cases of inverted spectra moreover (next paper) « is usually in excess of 2, and the shift is the other way. The deficiency in #, though not equally marked, is present in observations both on the right and left side of the prism P’. 13. Breadth of efficient wavesronts and apparent uni- Sormity of wavetrains. Leotalion of fringes.—It follows from Non-reversed Spectrum Interferometry. 149 fig. 18, that if JV is displaced to M/’, over a distance ¢, the pencil 6 is displaced parallel to itself over S = 2e sin 6/2 where 6 = 90°—@. The pencil c is then displaced parallel to itself over a distance St hae i fDi Since 0 = 49° 45’, 6/2 = 20° 7’ and therefore s = 2e X °344 = ‘T e,nearly. If the rotation of fringes is but 90°, either s (or s/2) is also the breadth of the strips, or patches of like origin which, when sliding over each other more or less, produce the fringes. This may be treated from a graphic point of view as follows, a theory not being aimed at. In fig. 18a, let a@ and 6 be two patches of light of like color and origin at the objective pp, fig. 18, producing inter- ferences at the focus /, fig. 18. Hence the fringes will be arranged in the direction /, fig. 18a, at right angles to the line joining a and 6. Since a and 6 here correspond to ¢ and c’ in fig. 18, let @ be continually displaced to the right, as indi- cated by the arrows. In proportion as the positions ab, a’b’, a”b”, are taken, the fringes must pass by rotation from 7, into meme 7. ete. ; 1. e.. over about 90°. In the present experi- ment, ¢, fig. 18, can never pass across c’, for they are essen- tially separated by the edge of the right angled prism /”. Hence the rotation can not exceed 90°, for the vertical through @ cannot cross the vertical through 6. This is not the case when a grating replaces ?’, as in fig. 14; nor is it the case when, as in an earlier paper, inverted spectra are treated, and the patches a and 6 slide along the edge of the prism. Insuch cases fig. 18@ may be continued symmetrically, toward the right (mirror images) and the limit of rotation is therefore 180°. All these suggestions are borne out by experiment. Moreover if the first prism P, fig. 15, is tilted slightly on an axis parallel to ZZ, a (fig. 18a) will be lowered and b raised. If @ and 6 are on the same level, the fringes are always vertical and pass through a vertical maximum, when ab isa minimum. On the other hand, if @ and 6 are not in the same level, as in the figure, fore and aft motion brings the rays ¢ and ¢’ (fig. 18) to or from the edge of the prism P’. Hence the case ab passes into a”b”, or the reverse; in other words the fringes pass through a horizontal maximum when abisaminimum; ete. This is also shown by experiment. The experiment made by moving screens with slits, forward or rearward, successively, by which the appearance and evanes- cence of fringes may be repeated through several cycles, is next to be explained. Here it is merely necessary to remember that 150 C. Barus—Methods in Reversed and the spectra ¢ and c’ are reversed, or that the colors of like origin and wave-length are successively farther apart. When the screens are alternately moved therefore, the same phenome- non is in turn produced in slightly different colors. But as ab continually increases whereas the efficient breadth of the strips does not, the fringes soon pass beyond appreciable smallness. When as in the earlier methods but a simple grating is used with two successive diffractions through it, the patches a and } are obviously in the same level when the longitudinal axes of spectra coincide. Hence the fringes are essentially vertical. In the experiment with screens, s, s’. fig. 15, it is obvious that path difference remains constant. The distance from the same wave-front in the pencils 6 and 0’, fig. 18, to the prin- cipal plane pp, is always the same; but pencils different in lateral position are successively selected. On the other hand, when the prism P’ is moved in the direction y, parallel to 60’, path difference only is introduced, while the pencils selected remain the same. Supposing the ordinary conditions of visi- bility (magnification, ete.) to remain unaltered, the wave- fronts are, as it were, explored in depth as to their uniformity ; i. e. the distance is apparently recorded, throughout which a wave train consists of identical wave elements. Effectively, however, the rapidity with which fringes decrease in size beyond visibility is directly in question. Finally when the opaque mirror JZ (or VV) is moved from J/ to MW’, both effects occur together. Path difference vw = 2e cos 6/2 is introduced and the pencil is displaced from 6 to 0’. 14. film Grating.—-The method of two gratings was now again resorted to, except that the first at G, fig. 3, was a- film grating. This attempt failed in my earlier work, when but a single film grating was used for the two diffractions, because of insufficient light. In the present case, two gratings (G’ being reflecting) are employed, and the method succeeded at once. The first grating constant was D = 10~°X167™; observations were therefore necessarily made in the second order of G’, so that the spectra are not as intense as with prisms. But the fringes are perfect and may be made as large as desirable, with but two in the breadth of the spectrum, for instance. The range of displacement was found to be about 6 millimeters under the best conditions (arrows). If both J/ and JV are successively displaced in the same direction, the total displacement available between the hairlike fringes at the extremes is about 1°5°" for each mirror. At these extremes the two patches of light on the grating G’ may have been separated by several millimeters. The nature of the transformation from arrows to the oblique striations would be well reproduced, if equidistant vertical wedges were moved from right to left, or the reverse, behind a vertical slit. Non-reversed Spectrum Interferometry. 191 The surprising success obtained with the film grating at short distances induced me to test similar methods at long distances. Figure 23 is an apparatus of this kind, in which Z is the white beam incident from a collimator, G and G’ are the transmit- ting gratings, 17, WV, m,n, pairs of opaque mirrors, 7’ the tele- scope. ‘The undeviated ray, d, is screened off. The component paths a+b+e, a'+b’+c' were each about 4 meters long. The method of adjustment again consists in bringing the shadow of the thin wire across the slit, into the same position of the spectra seen in the telescope when the spectra coincide. For this pur- pose the adjustment screws for horizontal and vertical axes on M, N, m, n, must be actuated together. To facilitate this tiresome work, with the observer at Z, long levers brought from m and n, with their ends near his hands, as well asa lever from G’ (fore and aft motion) were useful. Since the adjustment screws at Jf and J are already within reach, it is thus easy to bring any Fraunhofer line to the middle of the field and to make these fields overlap, with the guide wire cen- tral in both. The fringes were found after some searching and seemed to be of D, DY, breadth, a strip of oblique lines of the usual char- acter. They were not brilliant and hard to recover when lost. The Fraunhofer lines were still disagreeably blurred. On exchanging the gratings (weaker ruled glass grating at G and film at G@’), though the dispersion was smaller, the bril- lianey of spectra was greatly improved. On cutting down the incident beam at the collimator and near G, to a breadth of not more than -5°", the fringes were acceptable and capable of high magnification. They remained visible for a displacement of 5 millimeters at the micrometer at JZ. With fore and aft motion of G’, the fringes rotated as usual from fine vertical hair lines, through the horizontal (probably arrow-shaped forms of maximum size,) back again to hair lines. Here the excursion of G’ was about 1°5°". On tilting the grating G’ in its own plane and readjusting J/, the rotation is through the vertical maximum (the linear phenomenon). The film grating may be used by reflection, on adapting the method fig. 14, for this purpose, with a ruled grating or prism at P and the film grating (with its ruled toward P) at G. Ifaruled grating is put at P, the spectra and fringes are good ; but naturally there is deficient illumination. Neverthe- less a strong telescope may be used and a range of displace- ment of 4™™, at J, is available. This may be increased indefinitely by using a micrometer at JZ and JV alternately. The chief difficulty was the (incidentally) unequal brightness of spectra. Again, the method of fig. 15, apart from the drawbacks to 152 C. Barus— Methods ay Reversed and which that method is incident, succeeds almost perfectly, both in the first and second order spectra. The fringes are strong and clear. An Ives grating of high dispersion () = 167 x 107° em.) was tested. A prismatic method with auxiliary mirrors to accommodate the dispersion of the grating was also successfully tried. A concave reflecting grating may be replaced by a film grating used as a reflecting grating, with entire success. The ruled side of the film should be free (without cover glass), but the reversed side cemented on plate glass, as usual, and the latter placed towards the telescope. The prism /, in other words, admits an abundance of light, so that even the loss in reflection from the film is not serious. Sunlight should be used without a condensing lens; or if the latter is added, the light leaving the telescope is to be narrowed laterally. 15. Non-reversed spectra.—The prismatic method of cleav- ing the incident beam of white light is available for the super- position of non-reversed spectra, under conditions where the paths of the component rays may have any length whatever. It is thus an essential extension of the method, fig. 19, given in a preceding paper (P/’ prisms, JZ, mirrors, Gp, Ives prism grating, Z’ telescope), where the path differences were essentially small and the spectra reversed. In fig. 20, P is the first prism cleaving the white beam, Z, diffracted by the slit of the collimator. d/ and J are the opaque mirrors, the former on a micrometer. [for greater ease in adjustment, the second prism /” is here right angled, though this is otherwise inconvenient, since the angle 5 = 90° — @, is too large. The rays reflected from /’ impinge normally on the reflecting grating G(D = 200 x 107°) and are observed by a telescope at Z. P,P’, JZ and W are all provided with the usual three adjustment screws. P’ must be capable of being raised and lowered and moved fore and aft. The field is brilliantly illuminated. When the path difference is sufficiently small, the fringes appear and cover the whole length of superposed spectra, strongly. They are displaced, with rotation, if JZ is moved normally to itself. As first obtained the fringes were too close packed for accurate measurement. But experiments on the displacement of the mirror J, for successions of 40 fringes replacing each other at the sodium lines, showed a mean displacement of 39 x 10-° em. per fringe. The computed value would be » = 2 cos 6 /2 assuming 6=90°—@¢. The difference is due both to the small fringes which are dificult to count and to the rough == 4 10m Cm: Non-reversed Spectrum Interferometry. 153 value of 6. The range of measurement is small (if MZ only moves) not exceeding 1°5 millimeters for a moderately strong telescope. Usually but one-half of this displacement is avail- able, as the fringes increase in size (with rotation) from fine vertical hair lines to a nearly horizontal maximum, and then abruptly vanish. ‘This is only one-half of the complete cycle. If we regard the component beams, @ 6 ¢ and a’ b’c’, as being of the width of the pencil diffracted by the slit of the collima- Figs. 20-28. tor, it is clear that the maximum size of fringes will occur, when ¢ and ¢ are as near together as possible; furthermore, that as J/ moves toward P’, ¢ continually approaches ¢’, until 6 drops off (as it were) from the right angled edge of the prism P’. To get the best conditions, i. e., the largest fringes, e must therefore also be moved up to the edge of P and very sharp angled prisms be used at both P and P’. The largest fringes (lines about 10 times the D), D, distance) obtained with the right angled prism were often not very strong, though otherwise satisfactory. Much of the light of both spectra does not therefore interfere, being different in origin. Results very similar to the present were described long ago* and found with two identical half gratings, coplanar and * Phil. Mag., xxii, pp. 118-129, 1911 ; Carnegie Publ. No. 149, chap. vi. Am. Jour. Sct.—Fourts Series, Vou, XLIII, No. 254.—Frsruary, 1917. ot 154 C. Barus—Methods in Reversed and parallel as to rulings, ete. when one grating was displaced normally to its plane relative to the other. The edges of the two gratings must be close together; but even then the fringes remain small and the available paths also. Strong large fringes, but with small paths, were obtained by the later method* of two identical transmitting gratings, superposed. If the prism /P’ is right angled (a special case of fig. 19), it may be rotated as in fig. 21, so that the rays c and ¢’ pass off towards the observer. ‘They are then to be regarded through an Ives prism grating G and a telescope at 7. This method admits of much easier adjustment. With the component beams a 6, a’ 6’, coplaner, horizontal and of about equal length in the absence of the prism /’, the latter is now inserted with its edge vertical (rotation) and the white slit images in 7 (without G) superposed, horizontally and vertically. G is then added and the micrometer at J/ or VV manipulated till the fringes appear. As above, they are largest when ¢ and c’ are as nearly as possible coincident and vanish as horizontal fringes at the maximum. The case of fig. 20 was subsequently again tried on the large interferometer, the distance P to /—J being about 2 meters. G, in these experiments, was a concave grating and 7’a strong lens near the principal focus of G. The adjustment for long distances is not easy. The equilateral triangle of rays, a, a’, 6’, 6, should be first carefully levelled, the edges of P and P’ being on the median line. With @ placed at the proper dis- tance, the two spectra seen at 7’ will usually be quite distinct in the field. They should show the shadow of the black line across the slit, at the same level in the spectra. The longi- tudinal axis of the spectra may then be made collinear by slightly tilting the edge of /’ to the vertical, on a horizontal axis, with the adjusting screws. Jf and J are then rotated on a vertical axis till the D lines coincide. Small changes may be completed at JJ and VV. The fringes when found are usually strongly, but very fine, less than the D,D, distance in width. I have been able to increase them toa width of 2D, D,, but they are then faint. The two illuminated strips on the grating may even be an inch apart; but the fringes are as usual larger, when this distance is the smallest attamable (virtual coincidence). The grating may be moved fore and aft without effect. As JV is displaced on its micrometer, the interferences are first seen as vertical hairlike striations, which gradually en- large, rotate and vanisii just before reaching the horizontal and at maximum size. The range of displacement did not exceed 15 for this rotation of 90°. Sinee WV and 7 are close together, the manipulation is convenient here, but with another * Physical Review, vii, 1916, p. 587, 1916; Science, xlii, p. 841, 1915. Non-reversed Spectrum Interferometry. 155 lens at Z’ the phenomenon could be traced further on the MM side. To secure a smaller angle of incidence and reflection, 6/2, at D/, fig. 22, the combination of a silvered 20° prism, P, and a 30° prism, /’, was tested. J and J are the opaque mirrors, G the concave grating with its focus at Z’ for inspection by a strong lens. Z is the incident beam of white sunlight from the collimator, which is split into the component pencils abcd and a’b’c'd’ and interfere at 7. The results, however, were about the same as above, the range of displacement at / for 90° of rotation of fringes being about 15°. As a and b make angles, @ and ¢’, with the line of symmetry ZL’, was about 10°, 6= d'— d. At a subseqnent opportunity I made further trials with the paired prisms of 20° and 30°, but failed to increase the fringes above about D,D,/2 width. Two micrometers, one at J/ and the other at /V, were installed, and moved forward in alternate steps, within a range of over 2°, naturally without modify- ing the fringes. These are now observed on both sides (WV and M), each with the micrometer which is manipulated. One may note in passing that the two screws are being incidentally compared. It is noteworthy that the 30° prism at P’ is no marked improvement as to range of displacement over the 90° prism at P’, previously used. In other words, the effect of decreas- ing the dangle of reflection, 5, at J/ is, unexpectedly, of small importance in relation to the range of displacement at J/. This result has already been accentuated in other ways, above, $13. 16. Won-reversed. spectra. Restricted coincidence.—In fig. 24, the white ray Z from the collimator is diffracted by the grating G and the two spectra @ and a’, thereafter reflected by the parallel opaque mirrors J/ and J, to be again diffracted by the grating G’. The rays are observed by a telescope at 7’. If the gratings G, G’ have nearly the same constant, it is obvi- ous that the field of the telescope will show a sharp white image of the slit, for each mirror. If JZ V G G’ are adjusted for symmetry by aid of the adjustment screws on each and the rul- ings are parallel, the two white slit images will coincide hori- zontally and vertically. If now a direct vision spectroscopic prism, or a direct vision prism-grating (”, is placed in front of the telescope, the superposed white slit images will be drawn out into overlapping non-reversed spectra, which will usually show a broad strip of interference fringes. The equation is NN=2e cos 6/2—Ye sin @. 156 C. Barus— Methods in Reversed and This equation is not obvious, as for constant A, the distance being G and G’ measured along a given ray (prolonged) for any position of J/ or JV is also constant. The equation ma be corroborated by drawing the diffracted wave-front at G’ for M and MV’, which cuts off a length 2e sin 6 from a’. Since sin 6=A/D, if D is the grating space, the last equation becomes n=2e/D or per fringe de=D/2 a remarkable result, showing that the displacement of the mir- ror J per fringe is independent of wave length and equal to Fig. 24. half the grating space. An interferometer independent of 2X and available throughout relatively enormous ranges of dis- placement is thus at hand. It may be shown that it is also independent of the angle of incidence at G. To change the size of fringes it is necessary to rotate the grating G (relatively to @) on a horizontal axis normal to itself. They then both rotate and grow larger, attaining the maximum of size when the fringes are vertical. Fringes quite large and black may be obtained in this way. To show the close relation of the present experiments with one reflection, to the earlier work with crossed rays and two reflections, experiments may be made with homogeneous light. Accordingly the sodium are with a wide slit was installed. Strands of fringes with nodules were obtained as before. These rotated in marked degree (180°) from vertical hair lines, through coarse vertical strands with horizontal nodules, back to vertical hair lines again, as either JZ, or G, were suitably displaced normally to their planes. To shift the fringes of any Non-reversed Spectrum Interferometry. 157 form into the middle of the wide slit image, a glass compen- sator in either } or b’ may be resorted to, or both JZ and G’ may be displaced together. Again, whereas the micrometric displacement of J/ produces a marked displacement of fringes within the strip in accordance with the equation, the micromet- ric displacement of G’ leaves the fringes stationary within the strip. Very remarkable results were obtained with compensators of glass plate. Placed in one or both beams and _ rotated around a vertical axis, they rotate the fringes. If however they are placed nearly normally in one beam, they produce no effect either of rotation or on the size of the fringes; but the grid is displaced bodily across the wide yellow slit image. Glass plates -2, -5°™ were used. It is not until the thickness of plate reaches 2™ that appreciable thinning of the interference fringes occurs when the plate is placed in one beam. In case of homogeneous light the Ives prism grating G” is not needed and much more light is available if the telescope is used directly. The strands of interferences being on a yellow ground are not very strong, Nevertheless a few measurements of ranges of displacement were made by moving both J (dis- placement e) and G’ (displacement /), alternately. The follow- ing values of e, h, and A tan 6’ were found, the film gratings having nearly the same constants : C= 6=19° 37’ —ieo™ 6'=20° 40’ jo tam G49" e and / tan @’ coincide as closely as may be expected, seeing that the fringes in neither case can be quite brought to vanish. 17. The same, continued. Duplicate fringes. Achromatic Sringes.—The occurrence of strands and apparently duplicated fringes has already been suggested in the preceding paragraph. In further experiments definite results were eventually obtained, with sunlight. These oceur in very great variety, but typical phases may be accentuated. In intermediate cases fine large strands occur. These pass into each other continuously; the manner does not admit of description. They are seen best in the principal focal plane and both sets are about equally strong. To obtain these fringes the adjustment was first carefully made with the sodium are. Thereupon the are was replaced by concentrated sunlight and fine fringes were recognized in the superposed spectra (longitudinal and transverse axes coin- ciding). These fine fringes were then enlarged both by rota- ting the grating G”’ (fig. 24) on its normal axis and readjusting JM in each case, and by adding trial compensators in the J/ or lV ie A glass plate 3 millimeters thick gave the best results. 158 C. Barus— Methods in Reversed and Rotation of the compensator in the first place moves the fringes as in interferometry, as does also the normal micromet- ric displacement of JZ. If this motion requires readjustment of J the range of displacement is curtailed and the corre- sponding change of phase appears. In the second place the compensator on rotation traces the contours of the curves by successively accentuating vaguer parts, as will presently be explained. The most remarkable results occurred on widening the slit. Supposing that large strands were visible in case of the fine slit, and that this was gradually widened until the slit width was half a millimeter or more; the strands were found to have coalesced in a way which defies description. In their place appeared a wide vertical strip of equidistant parallel crescents. The Fraunhofer lines had long vanished and the appearance of the spectrum was whitish and intense. The fringes in ques- tion may thus be termed achromatic. The strips appear quite regular through the breadth of the spectrum and its width may be one-third of the length of the spectrum. The fringes move with the normal displacement of J/ (interferometry) and the range is large (°5°™ without adjustment) provided J/ does not require readjustment by rotation. Simultaneously the strip is displaced longitudinally in the spectrum in the usual way. On closing the slit the ellipses break up into sharp strands again without offering a systematic clue as to the manner in which this is done. The strands usually trend more or less vertically with two sharp strong groups, flanked by one or more weak groups on each side. On removing the condenser, these crescents became more slender but much sharper, so that in spite of the diminished light they could be well seen. They were then found to be like the approximately confocal ellipses of displacement inter- ferometry, though not subject to the same laws. They embraced over one-third of the visibly overlapping (green yellow through red) spectra, terminating in very fine hair-lines on one side but coarse lines on the other. On opening the slit from a breadth of 0™ to about ‘1™™ the evolution was curious. With a very fine slit a relatively narrow strip of strong slant- ing lines was seen in the yellow. As the slit widened they developed curvature, adding the more slender complements of the ellipses on the red side, until this part of the spectrum was filled with confocal half ellipses having a transverse major axis. The range of displacement of J/is practically indefinite, depend- ing simply on the degree to which the spectra overlap. Three or four centimeters were tried. Both sides of the ellipses may be traversed by rotating the plate compensator, which suc- cessively accentuates (in a transverse strip) a definite part of Non-reversed Spectrum Interferometry. 159 their contours. In this way the thick apices or either of the hairlike lateral ends may be clearly brought out. To further study this result the grating G’* was successively rotated in small amounts on a normal axis with adjustment at J/. It was thus possible to find both the upper ends of the ellipses and their lower ends, as well as the central part. The con- focal ellipses are extremely eccentric with very turgid apices so that the central part (if in the spectrum) consists of trans- Fie. 25. Fic. 26. verse straight lines. Motion of JZ moves the fringes to and from the center where they originate or evanesce. The ellipses shift as a whole with J without changing form appreciably throughout the spectrum, but they move very slowly, quite differently in this respect from the round ellipses in displace- ment interferometry which are extremely sensitive to displace- ment of J/. In the present work it may take 5 or 10™ at to pass the ellipses quite through the spectrum. They are strong and fine in spite of the film gratings used. 18. Zhe same continued. Prismatic Adjustment.—The 60° prism has certain advantages in experiments like the present, particularly when non-reversed spectra are to be obtained. Fig. 25 is a device of this kind, in which P is the separating prism and P’ the collecting prism, the beam of white light Z from a collimator entering the flat face normally on the front side and issuing normally on the rear side at cand c’. Mand J are opaque mirrors parallel to each other, G a direct vision prism-grating. The telescope is at Z. The reflection may be either internal, as in the strong lines of fig. 25; or it may be external on silvered faces of the prisms * When nearly centered rotation of M about a horizontal axis is also suf- ficient to complete the centering of the ellipses. 160 C. Barus—Methods in Reversed and p and p’, the appurtenances being shown in dotted lines. In this case the separated rays, a, a’, b, b', are collected at c”, c’”, to be joined in the telescope at Z. The internal reflection being total, I made use of it for the following experiments. M and WV and FP’ are on micrometers with the screw in the directions normal to their faces. PP, JJ, V, P’ must all be adjustable. After preliminary measurement for equal dis- tances, the fringes were found. They were strong but fine, beginning with vertical hair lines and gradually rotating as they grew coarser till they rather abruptly vanished. The displace- ment of the J/ mirror did not exceed 06, nor the rotation 30°. The spectra being non-reversed, the fringes covered the whole field. One would naturally suppose that the abrupt evanescence of fringes was due to the escape of the beam of the edge of the prism P’, but this is not possible as the mirror J/ was traveling toward the rear. Furthermore the fore-and-aft motion of the prism /’ over several millimeters had scarcely any effect on the fringes. This is unexpected; for the rays, ¢, c', are com- pelled to approach or recede from each other by this motion. Finally the sodium doublets may be moved at some distance (many times their breadth) apart, without destroying the fringes. They are often most distinct when the JD lines are not superposed. The same is also true for the longitudinal axes, though to a less degree. To enlarge the fringes, the prism P’ may be rotated around a horizontal axis parallel to LZ. The fringes then also rotate, but the increase of size so obtained is usually not striking. Moreover no observable effect either on the size of fringes or on the range of displacement is produced by inserting compensators in one beam or both. A great variety of differ- ent adjustments showed a range of displacement at JZ, about the same (‘06™), whether the patch of light on the prism was wide or narrow. The range of fore and aft motion of /’ within which fringes are visible was °52°". They vanish quite abruptly when the light is near the edge of the prism, although both spectra are still strongly visible. When the light is nearer the base of the prism, they vanish more gradually. Definite strips of white light on both sides of the prism therefore cooperate to produce the fringes. The attempt to find a systematic method for enlarging the fringes failed, possibly because the prism angles were not quite identical. The striking contrast in the results obtained here in comparison with those of the preceding paragraph, although both methods are essentially the same, is noteworthy. It is for this reason that I thought it desirable to test the method in fig. 26 which accomplishes with a prism, what was Non-reversed Spectrum Interferometry. 161 done in my original experiments* with reversed spectra, by the aid of a grating. In the figure the incident beam of white light Z from a collimator strikes the 60° prism at its edge, and is then refracted into the paired pencils a, a’. Those are reflected normally by the opaque mirrors J/ and J, again refracted by P as each pencil nearly retraces its path. The return beams however are given a slightly upward trend, so as to impinge on the opaque mirror m (curved or plane). The rays reflected from 7, in such a way as to avoid the prism P, may be reunited in the focus /’ observed by the lens 7’ or (if parallel) collected by a telescope at Z. In view of the prism, the spectra are small and reversed, but may be brought to overlap at the red ends which are towards each other. The small dispersion makes it necessary to use a strong tele- scope if the Fraunhofer lines are to be visible and the VP lines separated. When the adjustment has been made symmetrically, a strong linear phenomena may be found not differing in appearance from the results obtained when a grating was used at P, fig. 26. When the mirror J/ is displaced, however, the fringes first appear in the form of multiple vertical hair lines, which grow coarser until but a single dark line flanked by a bright line is visible. With further displacement the phe- nomenon again vanishes in passing through multiple hair lines. An important result is the small range of displacement. This was found to be, between appearance and evanescence of fringes, about oe: 12'emi, thus scarcely larger than a millimeter, whereas in the case where a grating (D = 352 x 10-°cm.) was used in place of P, the range of displacement was of the order of 5 millimeters. Suppose that for low dispersion, the fringes may be regarded as extremely eccentric virtually linear ellipses, the lateral dis- tance between which very rapidly diminishes, so that (6¢ = 12) but 1 L2 aon x LOF - = 2000 can be seen by the given telescope. These lines would move behind the strip carrying interference fringes, as J/ is dis- placed. If now the dispersion is much increased, say from d0/dX = 2 X< 760 for the prisms to 2 x 2880 for the grating, the ellipses will be much less eccentric as a whole and their lines would have grown coarser, so that many more would be visible by the given optical system. As the dispersion is increased 2880/760 = 3°8 times, the range of displacement * This Journal, xl, p. 486, 1915. 162 C. Barus—Methods in Spectrum Interferometry. should increase similarly to "12 x 38 = 46. The plane ruled grating ( = 3852 X 107° cm.) was now again mounted in place of P and under good illumination the range ‘48° was found experimentally. This agrees very well with the esti- mated value. Moreover on close inspection it is discernible that the linear phenomenon really consists of extremely eccen- tric ellipses which in case of the best adjustment manifest the very sharp arrowlike forms. It also enters and vanishes in multilinear form, though the lines are not hair lines. Thus the assertion that increased uniformity of wave train accounts for the long range of displacement and visibility in case of the grating is not warranted. : [To BE CONTINUED. | Brown University, Providence, R. I. W. 7. Schaller—Identity of Hamlinite with Goyazite. 168 Art. XVI.—On the Ldentity of Hamlinite with Goyazite ; by Watpemar T. ScHALLER. Farrineton* has recently stated that the present evidence hardly seems sufficient for regarding the two minerals, goya- zite and hamlinite, as identical. Damour’s original formula for goyazite would then have to be accepted even though it “evidently needs confirmation ” as Farrington states. The resemblance of goyazite and hamlinite in their physical and optical properties is great enough to warrant the sugges- tion that their chemical composition is of the same type of formula, and since Damour’s calcium has been shown to be essentially strontium, the two minerals have the same qualita- tive composition, both being hydrous phosphates of aluminium and strontium. If their formulas are of the same type, it is most reasonable to consider them identical—until they are proven to be different. The table published} earlier by the writer showed the proba- ble identity of the two minerals goyazite and hamlinite. In order to emphasize the possible distinction between these two minerals, Farrington has published a table of differences between them ; there are, however, three errors in the six items listed in this table. The percentage of P,O, for hamlinite is wrongly given as 20°92 per cent. It should be 28-92 per cent. The suggestion was earlier made that the separation of P,O, from A1,O, was probably not accurate in Damour’s analysis of goyazite, so that the comparison between the sums of P,O,and A1,O, for the two minerals (65°53 and 62°87), is probably the most valid comparison of the analytical figures. Goyazite is stated by Farrington in his table to be ‘ Infusi- ble.” The original description statest : “ A la flamme du chalu- meau, il fond difficilement sur les bords des plus minces frag- ments.” Damour’s statement regarding goyazite, ‘il fond difficilement,” and Penfield’s statement for hamlinite, “‘ fuses about 4,” show a resemblance, not a difference. The table of Farrington states further that hamlinite is “slowly soluble in acids.” This statement, quoted in Dana’s System of Mineralogy, on page 762, must be somewhat modi- * Farrington, O. C., Studies of Brazilian favas, this Journal (4), xli, p. 358, 1916. + Schaller, W. T., The alunite-beudantite group, this Journal (4), xxxii, 009, 1911. Also, U. S. Geol. Survey Bull. 509, p. 70, 1912. ¢{ Damour, A., Note sur un nouveau phosphate d’alumine et de chaux, des terrains diamantiféres, Bull. Soc. Min., France, vol. xvii, 204, 1885. 164. W. 7. Schaller—Ildentity of Hamlinite with Goyazite. fied by Penfield’s later statement* that “hamlinite is almost insoluble in boiling dilute hydrochloric acid,’ and by Bow- man’s statementt that the “mineral is insoluble in hydro- chlorie acid.” By correcting these errors in Farrington’s table, the main supposed differences disappear and the list of similarities of the two minerals, as shown below, is sufficient to justify the conclusion that hamlinite is identical with goyazite—at least until their difference is proven. Similarities in properties of goyazite and hamlinite, supporting the conclu- sion that the two minerals are identical. Goyazite Hamlinite. Yellowish white. Colorless, yellowish, honey-yel- low, reddish brown. More or less transparent. Transparent. Uniaxial, positive. Uniaxial, positive. Cleavage, good, basal. Cleavage, good, basal. Hi 5: H = 4:5,4+. SiG, == 1326 S.G. = 3°228, 3°2, 3°159 —— ages 3°219 — 3°266. Fuses with difficulty. Fuses about 4. Not attacked by acids. Insoluble in HCl. Per cent Al,QO, + P,O, = 65°53. | Per cent Ai,O, + PO == 62a Per cent H, O (det. by loss on | Loss on ignition (hamlinite from ignition = == 16°67. | Switzerland) = 15°6 to 16-0 | per cent. Hydrous phosphate of alumin- Hydrous phosphate of alumin- ium and strontium. ium and strontium. * Penfield, S. L., On the chemical composition of hamlinite and its occur- rence with bertrandite at Oxford County, Maine, this Journal (4), iv, 313, 1897. + Bowman, H. L., On hamlinite from Biennenthal, Switzerland, Mineralog. Mag., xiv, 391, 1907. Chemistry and Physics. 165 SCIENTIFIC INTELLIGENCE. I. Cuermistry AND Pauysics. 1. The Occurrence of Free Carbon Monoxide in the ‘ Floaters” of Kelp.—Sxrtu C. Lanepon has made the interesting observation that the gas in the bulb and hollow stem of the large Pacific coast kelp, Wereocystis lwetkeana, contains carbon monoxide in quantities varying from 1 to 12 percent by volume. This gas has never before been found in the free state in a living plant. This kelp is a very large brown one. The author worked with specimens 85 feet long, and much longer ones have been reported. This enormous growth takes place in the short period of 10 or 15 weeks. Another interesting feature of this plant is the fact that while it contains about 92 per cent of water, more than one-fourth of the remaining 8 per cent consists of potassium chloride, and the commercial extraction of this valuable salt from the kelp has been extensively discussed. The hollow space in the large Specimens has a capacity of from 3 to 4 liters, and it has been found that the gas contained in this cavity is almost always under less than atmospheric pressure. Many of these pressures were determined, most of them ranging between 520 and 600™, while the normal atmospheric pressure is 760™™. The gas from nearly 1000 specimens was examined, and almost invariably carbon monoxide was present, the average amount being about 4 per cent. Carbon dioxide was either practically absent, or present in very small quantity, while oxygen varied from about 22 per cent, or about the same as in the atmosphere, down to about 15 or 16 per cent. It would seem that a mixture shown by one of the analyses to contain 12:2 per cent of carbon monoxide and 17:4 per cent of oxygen would be explosive when exposed to a flame, but the author makes no mention of such a circumstance. The samples of gas for analysis were carefully taken from the plants in their natural situation, sometimes over mercury, and sometimes by displacement of sea-water which had been saturated with the gas from other specimens. The qualitative tests for carbon monoxide in the gas were elaborate and convincing. Palladium chloride paper was black- ened, very dilute blood gave the characteristic change of color, the spectrum of blood treated with the kelp gas and ammonium sulphide showed the characteristic behavior of carbon monoxide, a guinea pig placed in a vessel through which the kelp gas was passing died in less than 10 minutes, while a canary bird died in about 15 seconds and a young chicken in about a minute. All these animals showed clear evidence from blood color or blood tests of carbon monoxide poisoning. The author suggests that the presence of carbon monoxide in the kelp may indicate an intermediate step in the reduction of carbon dioxide in connection with photosynthesis, a theory ad- 166 Scientifie Intelligence. vanced by Baeyer. He mentions a possibility also that the car- bon monoxide may be an accumulated waste product, or that it may be formed by processes of decay. He does not mention, however, an explanation that appears to be more plausible and interesting than the others, namely, that the carbon monoxide is produced by the plant as a poison to protect itself from such animals as might bore into it, make a habitation in its cavity or feed upon its substance. Since carbon monoxide is the simplest of organic poisons, it is perhaps reasonable to suppose that it might be utilized in this way by a simple organism like kelp, while higher plants produce alkaloids and other more complex poisons. It will be interesting to examine other plants of low orders for the presence of carbon monoxide. It appears that sea- weeds are generally very free from the attacks of marine animals, and it seems possible that we now have an explanation of this immunity. The further suggestion may be made that possibly the curiously large amount of potassium chloride in this kelp and also the iodine that occurs generally in these plants may be protective poisons.—Jour. Amer. Chem. Soc., xxxix, 149. He Tey 2. The Atomic Weight of Lead of Radioactive Origin.—The results of recent work by several independent investigators have shown with very little room for doubt that the metal derived from this source has a much lower atomic weight than ordinary lead. This conclusion is of such theoretical importance that THEODORE W. Ricwarps and CHarLEs Wapsworts, 3d, have extended the investigations previously made by Richards and Lembert upon the same subject. Their results entirely support the earlier con- clusion. Atomic weight determinations were made with ordinary lead and four samples of radioactive lead with the following results: Ore. Origin. Atomic wt. Galena (?) American ordinary lead 207°18 Carnotite Colorado, U.S. A. (?) 207°00 Carnotite Radium Hill, N. S. W. 206°34 Broggerite Moss, Norway 206'12 Cleveite Langesund, Norway 206°08 The last sample is the most carefully selected one, and is prob- ably most nearly free from ordinary lead. While it is possible that there may be two kinds of radioactive lead with different atomic weights, it seems more probable that the higher results from the carnotite ores are due entirely to the accidental admixture of ordinary lead. Careful spectroscopic examination showed no lines peculiar to the radioactive material. It was found that the magnitude of the radioactivity of the samples of lead seemed to bear no relation to the lowering of the atomic weight.—Jour. Amer. Chem. Soc., XXvili, 2613. H. L. W. 3. Engineering Chemistry ; by Tuomas B. STILLMAN. 8vVvo, pp. 748. Easton, Pa., 1916 (The Chemical Publishing Co.).—The Chemistry and Physics. 167 appearance of five editions of this book since the first issue in 1895 shows that it has been extensively used. The sub-title of the work describes it as “‘ A manual of chemical analysis for the use of students, chemists and engineers,” but it is to be observed that descriptions of many physical tests, copies of specifications and other topics not dealing strictly with chemical analysis are included. On the other hand, the book deals with only a restricted field of analytical chemistry, confining itself practically to such materials as are used in civil or municipal engineering. However, the book presents much that is useful in this field, and it is to be highly recommended to those who are interested in this kind of work. | Among the subjects most extensively treated are the proximate analysis of fuels, their colorimetry and their physical examination, the analysis of iron, steel, and a number of non-ferrous alloys, the analysis and physical tests of cements, concrete, clay, sand and building stones, the examination of asphalt and other bituminous road materials, of coal-tar lubricating oils, illuminating oils and fuel oils, soap analysis, varnish analysis, paint analysis, the chemi- cal and physical examination of paper, the analysis and treatment of boiler waters and potable waters, the analysis of flue gases, illuminating gases, etc., the manufacture of producer gas, water gas and acetylene, photometry and pyrometry. Many official methods are quoted and many interesting details of manufactur- ing operations are given. H. L. W. 4. Qualitative Analysis, by EK. H. 8. Barmey and HamitTon P. Capy. 8vo, pp. 294. Philadelphia, 1914 (P. Blakiston’s Son & Co. Price $1.50 net).—This is the eighth edition of a well- known laboratory guide, which gives an excellent course of in- struction in chemical analysis. The methods of qualitative separation and detection are well chosen and clearly presented. The title page states that the book is based upon the application of the theory of electrolytic dissociation and the law of mass action. There is a rather elaborate introduction dealing with the topics just mentioned, and the ionic nomenclature is used to an extreme extent throughout the practical part. There are some teachers who would prefer to lead up to the ionic theory by means of the facts encountered in qualitative analysis, rather than to attempt to explain the facts by means of a theory, but at present the lat- ter method appears to be popular. There is a generalization on page 32 which needs modification to correspond with all the facts : ‘“‘'The solubility of difficultly soluble salts of strong acids, how- ever, is not increased by the addition of an acid.” H. L. W. 5. X-Ray Wave-Lengths.—A valuable paper on the new branch of spectroscopy —the study of the wave-lengths of char- acteristic or fluorescent X-rays—has been recently written by Manne Sizepaun. The article begins with a list of 66 biblio- graphical references which indicates how rapidly the subject has grown in the course of a very few years. The author then describes the various types of bulbs and spectrometers which have proved most efficient in producing and analyzing the radia- 168 Scientific Intelligence. tions in question. The vacuum spectrometers perfected by Sieg- bahn and others are illustrated and explained in detail. This is followed by a discussion of the relations between the speeds of _ the exciting cathode rays and the excited characteristic X-rays. The rest of the article is devoted mainly to the relatively accu- rate wave-lengths determined experimentally by Siegbahn and his co-workers, full credit being also given to the more explora- tory and less accurate work of their predecessors. In the case of the A-group the wave-lengths cover the interval from 0°292, for neodymium, to A11°951 for sodium. Eight series of lines have been traced for the A-group. The wave- lengths belonging to the Z-group fall into 14 series extending from 4 0°596 for uranium to A 12°346 for zinc. For the elements of high atomic numbers,—gold to uranium,—the author has dis- covered a third group of lines which he calls the W-series. The paper closes with a very useful table containing all the accurately known wave-lengths, the number of which is 1729.— Jahrbuch d. Radioaktivitat u. Hlektronik, vol. xiii, pp. 296-341, Sept., 1916. 7 H. 8S. U. 6. General Physics; by Witiiam 8. FrRanKiIn and Barry MacNorr. Pp. vii, 604; 479 figures. New York, 1916 (McGraw-Hill Book Co.).—This book has the sub-title “ An Elementary Treatise on Natural Philosophy” and it is designed as a text-book for colleges and technical schools. The field cov- ered is divided into five parts: I Mechanics (pages 5-104), II Theory of Heat (107-177), III Electricity and Magnetism (181- 343), IV Theory of Light (347-484), and V Theory of Sound (487-528). Each part is preceded by a list of titles of selected reference-books together with comments on their nature and scope. But little space is devoted to statics and elasticity since these subjects have been presented at some length in two earlier volumes by the same authors. A list of 409 problems, for solu- tion by the student, constitutes Appendix A. The spirit and method of the calculus pervade the entire text. In this connection the authors say: ‘‘ The method of differential calculus is used quite freely in this elementary treatise on physics, but the authors believe that the use of this text does not depend upon the previous study of calculus by the student. Indeed the authors are convinced that the use of this book or its equivalent is a necessary preparation for the study of calculus; and the authors suggest that the student be required to turn again and again to the brief discussion of the methods of differential and integral calculus in Appendix B”. As a matter of fact, formal proofs depending upon theorems of the calculus are presented in small type and in such places as not to interfere, in the least, with the continuity of the main body of the text. The authors’ style is lucid and straightforward. Clearness of conception is facilitated in other branches of the subject by numerous mechanical analogies, presented both graphically and by the use of corresponding sentences in parallel columns. The definitions of units and the statements of fundamental laws are Chemistry and Physics. 169 characterized by accuracy and succinctness, and the entire text is as rigorous as possible for an elementary book. The diagrams are clear and to the point, italic and clarendon type are frequently used for emphasis, and the number of misprints is negligible. On the other hand, the innovations in terminology frequently introduced by the authors may not be pleasing to all readers. For example, the terms “spin-inertia,” “spin-velocity,” and “spin acceleration” are substituted for moment of inertia, angular velocity, and angular acceleration, respectively. The opinion is advanced (on page 285) that it is unsuggestive to speak of charg- ing an electric condenser but that it is extremely suggestive to speak of “squeezing” a condenser. The statement (p. 33) that “The limiting value of a is always represented by ay a must refer to the succeeding pages of the book since the symbol y does not seem to have become obsolete, especially with English writers. Nevertheless, as implied by most of the foregoing com- ments, the book, as a whole, seems to be a valuable contribution to the pedagogy of the subject. her Steue 7. Cosmical Evolution, Critical and Constructive. Second Edition ; by Evan McLennan. Pp. xxi, 490. Corvallis, 1916 (The Author).—The copy of this book submitted for review was -accompanied by a circular from which a few quotations will be made for the purpose of presenting the author’s point of view. “'This book comprises the principal part of the author’s life-work, continued over a period of about forty years.” “It contains undoubtedly the most complete and destructive criticism of the accepted fundamental views of physical science that has ever been published ; and it also contains an equally complete constructive theory.” ‘To all those who receive this book for review .... the author earnestly and finally appeals for an impartial and ade- quate treatment of it.” In the winter of 1914-15 the final manuscript was given a criti- cal examination by the heads of the physical departments of four of the principal colleges of the western United States. “ All of the results of this examination, for which permission to publish has been obtained, are given in Appendix A, together. with the author’s replies thereto.” The opinion of one of the physicists is summed up, on page 408, in the following words: ‘In reading over your first part I have not found one single argument that is valid. In some cases you have misunderstood the facts, in others you have drawn conclusions that are not warranted by the facts and in many cases you have quibbled over or misunderstood definitions of terms which are the most elementary foundation of physical science and which are verified experimentally thousands of times every year in laboratories.” The writer of the present notice has fully verified this adverse criticism, but has found it to be altogether too mild and considerate. In our. opinion the only use to which the text can be put is to furnish teachers of elementary logic with numerous simple illustrations of logical fallacies. 18h fe We Am. Jour. Sct.—Fourts Srries, Vou. XLIV, No, 204.—FrEsruary, 1917. 12 170 Scientific Inteilagence. 8. An Introduction to Astronomy. New and Revised Edi- tion; by Forest Ray Moutrton. Pp. xxi, 577. New York, 1916 (The Macmillan Co.).—The present volume kas been entirely rewritten. ‘As in the first edition, the aim has been to present the great subject of astronomy so that it can be easily compre- hended even by a person who has not had extensive scientific training.” Since it is impossible to even begin to do justice to this elegant book in a brief notice it seems desirable to suggest its scope by quoting the titles or sub-titles of the chapters, which are: I Preliminary Considerations. II The Shape of the Earth. The Mass of the Earth and the Condition of Its Interior. The Earth’s Atmosphere. III The Rotation of the Earth. The Revolution of the Harth. IV Reference Points and Lines. | V The Constellations. VI Time. VII The Moon. VIII The Law of Gravitation. Orbits, Dimensions, and Masses of the Planets. IX Mercury and Venus. Mars. Jupiter. Saturn. Uranus and Neptune. X Comets. Meteors. XI The Sun’s Heat. Spectrum Analysis. The Constitution of the Sun. XII General Considerations on Evolution. Data of Problem of: Evo- lution of Solar System. The Planetesimal Theory. Historical Cosmogonies, and XIII The Apparent Distribution of the Stars. Distances and Motions of the Stars. The Stars. The Nebula. With regard to details no pains seem to have been spared, either by the author or by the publishers, to make the volume as attractive and useful as possible. Of the 194 figures no less than 68 are photographic illustrations, those pertaining to the moon, the Pleiades, comets and nebule being especially beautiful. Five star maps have been incorporated to enable the reader to quickly recognize the constellations. Lists of problems (243) have been given at the ends of the principal divisions of the chapters. “They cannot be correctly answered without a real comprehen- sion of the principles which they involve, and in very many cases, especially in the later chapters, they lead to important supple- mentary results.” ‘The volume closes with both author and sub- ject indexes. Unquestionably this book is the most up to date and inspiring presentation of the subject that the writer of this notice has ever seen. H.Se s0b II. Grotogy anp MInErRALOGY. 1. Some Geophysical Observations at Burrinjuck; by Leo A. Cotton. Proc. Roy. Soc. N.S. Wales, vol. xlix, pp. 448-462, 1915.—Studies of the strength of the earth’s crust resulting in the contrasted doctrines of high rigidity and of isostasy have dealt largely with the validity of assumptions and the discussion ef theoretical conclusions. Experimental observations have been concerned chiefly with the crushing strength of rocks under actual and hypothetical conditions. The results of geophysical research have so far been unable to answer the question raised by geolo- Geology and Mineralogy. 171 gists, viz., do the adjustments of level initiate cycles of erosion and determine the locations of areas of denudation and of deposi- tion, or do adjustments of level result from shifting of load by streams and other transporting agents? In a problem involving a large number of unknown factors any new line of investigation is welcome, and for this reason the experiments being conducted in New South Wales assume high value. It was my good fortune to see the geologic features along a 40-mile stretch of the Murrumbidgee and to note the methods adopted for this line of investigations. ! The Burrinjuck dam on the Murrumbidgee River, now nearing completion, is a rival of the Roosevelt dam of Arizona, which it resembles in form and geological setting. The dam, 240 feet high, will form a lake 41 miles long, holding 33,000,000,000 cubic feet, 1,031,250,000 tons of water, equal in weight to a mass of sand and gravel 2 miles long, 1 mile wide, 322 feet deep. If the earth’s crust is sufficiently elastic or plastic to yield in response to a load of such weight and dimensions, it should be possible to record the fact with suitable instruments. Acting on the sugges- tion of Dr. W. G. Woolnough, three seismographs were obtained, two of the Rebeur-Ehlert type, previously used by Hecker and Schweydar in the investigation of earth tides, and one of the Zollner suspension type designed by Hecker. Careful considera- tion was given to the selection of sites by David and Cotton, geologists ; Father Pigot, seismologist ; and D. F. Campbell, resi- dent engineer. The geologic conditions were kept in mind and elaborate precautions were taken in the installation of instruments. Tunnels 60 to 80 feet long were excavated in rock and the pendu- lums placed in the innermost of three compartments, the second of which contains the recording apparatus. The instruments were first used to supplement geologic investigations on the normal stability of the earth’s crust at Burrinjuck. The pendu- lums are recording four types of movements: earthtides, which will be compared with those recorded on the instrument of the International Geodetic Association at Cobar, 360 miles from the coast (Burrinjuck is 125 miles from the sea) ; earthquakes ; fault movements ; slow deflections from the vertical. This last type of movements “‘may be related chiefly to the water loads or may be due to other causes, but it seems almost certain that the former cause is in operation.” The results of the experiments so far car- ried out look toward a conclusion that either elastic yielding or iso- static adjustment affects areas as small as a few square miles. Cotton raises the pertinent question as to whether the extremely contrasted views of Barrell and Hobbs on the one hand and of Hayford and Bowie on the other hand may not be reconciled on the basis of the stability of a given area prior to the removal or deposition of a load. ' Further reports on the Burrinjuck experiments will be awaited with interest, and it is highly desirable that similar investigations be undertaken within the United States. H. BE, G, 172 Scientific Intelligence. 2. Geology of Cincinnati and Vicinity ; by Nevin M. Fenne- MAN. Geological Survey of Ohio, Fourth Series, Bulletin No. 19, 1916, 207 pp., 59 figs., 12 pls.— Educators have frequently called attention to the dearth of books and articles on geology suitable for the elementary student and general reader. In the fields of botany, zoology, and even physics are many descriptive and explanatory treatises, and even periodicals which are widely read by the non-professional public, but most publications dealing with geology need the interpretative assistance of a teacher. One reason for the scarcity of suitable books on the earth sciences is the unusual difficulty of preparing a manuscript which combines authoritative statements with simplicity of writing. In this con- nection the “Geology of Cincinnati and Vicinity,” by N. M. Fenne- man, deserves high praise. It is prepared to meet an educational demand. It is well planned and clearly written, discards contro- versial matter and supplements the statement of hypothesis and fact by description and illustration of local features. The book may be studied with profit as a type of a local geologic report. H. E. G. 3. Field Geology ; by F. H. Lauer. 12mo. Pp. 508; figs. 409 in text. New York, 19i6 (McGraw-Hill Book Co.).—There has long been need of a work of this nature, which could be placed in the hands of students who are taking their first practi- cal field work in geology. The excellent little hand-book of the late Dr. C. W. Hayes is well known to professional geologists, but is of a rather technical character, written more especially to meet the demands of the U. 8. Geological Survey, and _ less adapted to the use of beginners. The present volume gives the student those criteria which, on the one hand, will enable him to understand what he sees in the field and on the other will suggest to him the data that should be searched for to gain a solution of his problem. ‘The scope of the work may be indicated by a men- tion of some of the topics treated : identification of rock features, characters of rock particles and pebbles, surface features of sedi- ments, structures of sedimentary rocks, field interpretation of sedimentary materials, features of igneous rocks, folds, faults, metamorphic rocks, mineral deposits, topographic forms and expression, topographic maps and profiles, geologic surveying, geologic maps, diagrams and sections, geologic computations, preparation of reports, etc. The work is clearly and simply writ- ten with avoidance of details and discussions, the writer having evidently his student audience in mind. It is published in a form and size convenient for transporting ; it is well printed and very fully illustrated by half-tones, sections and diagrams. ‘The only criticism which could be passed on its make-up, is that if a greater reduction in the reproduction of the drawings had been made, the illustrations would have been as serviceable and had a neater effect ; compare figs. 49 and 281 for example. : The book is commended for examination to all teachers who have courses in field geology. Ts Vics Geology and Mineralogy. 173 4, The Bundamental Principles of Petrology, translated from the German of EK. WrinscHENK by A. JOHANNSEN. 8vo. Pp. 214; 187 figs. in text, VI pls. New York, 1916 (McGraw-Hill Book Co.).—This translation of Grundzige der Gesteinskunde has been well carried out by Professor Johannsen, and in several ways the book is an improvement on the original work. As its name implies, it 1s not in its nature descriptive of the different kinds of rocks, but treats of such subjects as volcanism and the origin of igneous rocks, differentiation, rock-weathering, the nature of sediments, contact metamorphism, post-voleanic pro- cesses, regional metamorphism and jointing and rock-textures. These subjects are treated in a simple and yet condensed way so that the work, as intended, is suitable for students commencing work in petrology. Regarding matters concerning which there is yet no agreement it naturally, for the most part, presents the views of the German school of geologists. Professor Weinschenk him- self has worked largely in areas of regionally metamorphosed rocks, and thus in the subject of metamorphism he speaks with a certain authority and presents views of his own. The book may be usefully read by teachers and students who are not acquainted with it in the original text and in giving them this opportunity the translator has performed a commendable task. The volume is well printed and handsomely illustrated. L. v. P. 5. LP Oural du Nord; Le Bassin. des Riviéres Wagran et Kakwa, par Lovis Durarc et Marcuerire Trxanowitcu, (Mem. Soc. de Phys. et d’Hist. Nat. de Genéve, vol. xxxviil, fasc. 2, pp. 69-166. . Pls. 6-7, 1914).—In continuation of the geologica land petrographical researches being carried out on the northern part of the chain of the Ural Mts., by Professor Dupare and his stu- dent assistants, the present work is an interesting addition. There is no topographic map of this region and the field work is mostly of the nature of a reconnaissance, but the exploration of the area derives importance in connection with the occurrence of plati- num. Here again, as elsewhere in the Urals, occur great masses of gabbros associated with olivine and pyroxene rocks, similar types to those described in former publications by the senior author. The main part of the work is devoted to a detailed petrographic and chemical study of the rock types collected. Cent oe 6. Etude comparée des Gites Platiniferes de la Sierra de Ronda et de’ Oural; by Louis Duparc and AvuGUSTIN GROSSET. Mem. Soc. Phys. et Hist. Nat. de Genéve, xxxvili, 253, 1916.— Platinum has recently been found in connection with peridotite rocks that occur in the Sierra de Ronda, Province of Malaga, Spain. A petrographic study of the rocks of this locality shows, however, that the character of this occurrence is quite distinct from that of the classic locality in the Ural Mts. The conclusion is reached that the primary occurrences of platinum may be of the following three types: (1) In dunite rocks. Occurrences of this kind are the most common and richest ; this type is found in the Urals Mts. (2) In pyroxenites which are composed of a 174 Scientific Intelligence. monoclinic pyroxene, olivine and magnetite ; this is a less com- mon type and poorer in content of platinum. (3) In peridotites which contain both orthorhombic and monoclinic pyroxenes and commonly a brown spinel and chromite. They may be com- pletely changed to serpentine. ‘The occurrence in the Sierra - Ronda is of this type. So few examples of this sort are known that it is impossible to make any generalizations concerning their robable richness. W. E. F. 7. Etude Cristallographique et Optique Wun certain nombre de Minéraux des Pegmatites de Madagascar et de Mineraux de ? Oural ; by RenE-Cuar Es Sazor. A doctor’s thesis presented to the University of Geneva, 1914.—This paper gives the results of the study of a number of minerals from Madagascar and the Ural Mts. ‘Their crystals are figured and described, their optical constants and in a number of cases chemical analyses are given. The list of Madagascar minerals is as follows: from Ambatafotsikely, quartz, muscovite, monazite, columbite, euxenite, ampangabeite, stritiverite, hematite, spessartite ; from Ambositra, zircon; from Amampatsakana, mica; from Antson- gombato, tourmaline and apatite. The following minerals from the Urals Mts. were described: from Tokowaia, brookite, topaz, spessartite, hematite, rutile, quartz; from Syssert, muscovite, tourmaline. A detailed study is also given of the amphiboles in the diorite pegmatites of the platinum districts. WwW. E. F. 8. Lhe Geological History of Australian Flowering Plants ; by E. C. ANDREWs, pp. 171-232, September, 1916.—Exrrata. age 186, nine 14, for were contemplated, read are considered. 186, 21, “ Coprisma, read Coprosma. BEN FDO, ASE 12, 6: Holartie to Folarcite: 6 QOL UK MOO! Wal alee," “Melateuon: SV DOT ns Seo restocdae, “< Restiacee. «911, “ 8, % Leucondendron read Leucodendron. OBITUARY. Witiiam Exxis, formerly superintendent of the magnetic and meteorological department at the Royal Observatory, Greenwich, died on December 11 in his eighty-ninth year. Crement Rep, of the Geological Survey of Great Britain, well known for his work in paleobotany, died on December 10 at the age of sixty-three years. DanizL Otiver, professor of botany at University College, London, from 1861 to 1888, died on December 21 in his eighty- seventh year. A. M. WorruHINnGTOoN, professor of physics at the Naval Engi- neering College at Keyham from 1887 to 1909 and at the Royal Naval College at Greenwich from 1909 to 1911, died on December 5 at the age of sixty-four years. JOHN Wrigatson, professor of agriculture at the Royal Agricultural College, Cirencester, died on November 30 at the age of seventy-six years, £ Warns Narorat Science EstaslisHMENt 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. J-148. Price List of Rocks. Mineralogy: J-109. Blowpipe Collections. J-74. Meteor- ites. J-150. Collections. J-160. Fine specimens. Paleontology: J-134. Complete Trilobites. J-115. Collec- tions.. J-140. Restorations of Extinct Arthropods. Entomology: J-80. 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. 2 Taxidermy: J-138. Bird Skins. J-139. Mammal Skins. Human Anatomy: J-16. Skeletons and Models. General: J-155. List of Catalogues and Circulars. Ward’s Natural Science Establishment 84-102 College Ave., Rochester, N. Y., U.S. A. Pu blishers: WILLIAMS & NORGATE, 14 Henrietta Street, Covent Garden, London, W. C. “SCIENTIA” INTERNATIONAL REVIEW OF SCIENTIFIC SYNTHESIS. Jssued monthly (each number consisting of 100 to 120 pages). Editor: EUGENIO RIGNANO. “SCIENTIA’’ continues to realise its program of synthesis. It publishes articles which relate to the various branches of theoretic research, and are all of general in- terest; it thus enables its readers to keep themselves informed of the general course of the contemporary scientific movement. **SCIENTIA’”’ appeals to the cooperation of the most eminent scientific men of all countries. It has published articles by Messrs. Abbot (Washington) - Arrhenius (Stockholm)- Ashley (Birmingham) -Bechterew (Petrograd) - Bohlin (Stockholm)- Bonnesen (Kopenhagen) - Borel (Paris) - Bottazzi (Napoli) - Bragg (Leeds) -Bril-e louin (Paris) - Bruni (Padova) - Castelnuovo (Roma) -Caullery (Paris) -Chamberlin (Chicago)-Ciamician (Bologna) - Clark (New York) - Costantin (Paris) -Crommelin (Greenwich) - Daly (Cambridge, U. 8S. A.) - Darwin (Cambridge) - Delage (Paris) - De Martonne (Paris) - De Vries (Amsterdam) - Durkheim (Paris) - Eddington (Greenwich) - Edgeworth (Ozjford) - Emery (Bologna) - Enriques (Bologna) - Fabry (Marseilte)-Fisher (New-Haven, U. 8. A.)-Foa (Torino)-Fowler (London) - Firedericq (Liége)- Galeotti (apoli)- Golgi (Pavia) -Gregory (Glasgow)- Guignebert (Paris) - Janet (Paris) - Jespersen (Gentofte) - Kapteyn (Groningen) - Kidd (Oxford)-Langevin (Paris)-Lebedew (Moscou)-Lodge (Birmingham)-Loisy (Paris)- Lorentz (Haarlem) - Loria (Yorino)—Lowell (Flagstaff, U. S. A.)-Maunder (G'reen- wich) - Meillet (Paris) - Pareto (Lausanne) —- Peano (Torino) - Picard (Paris) - Poincare (Paris) - Puiseux (Paris) - Rabaud (Paris) - Righi (Bologna) - Rignano (Milano)=Russell (Cambridge)- Rutherford (Manchester)- Sayce (Oxford) - Schiapa= relli (Zilano) - Seligman (Vew York) - Sherrington (Liverpool) - Soddy (Glasgow)-= Svedberg (Upsala)- Tannery (Paris) - Turner (Ozford) - Vinogradoff (JZoscou)- Vol= terra (Roma)-Von Zeipel (Upsala) -Westermarck (Helsingfors)-Willey (Montreal, Canada) -Zeeman (Amsterdam) -Zeuthen (copenhagen), and more than a hundred others. ; ‘*SCIENTIA’’ publishes, at present, in the section dedicated to sociological articles, a series of studies on the present questions of an international character raised by the war. ‘*SCIENTIA”’ publishes its articles in the language of its authors, and joins to the principal text a supplement containing the French translations of all the articles that are not in French. (Write fora specimen number.) Annual Subscription: , Bi ist. BOs’ ince. Office: Via Aurelio Saffi, 11- MILAN (Italy). CONTENTS: Page Art. X.—The Water Content of Coal, with Some Ideas on the Genesis and Nature of Coal; by KE. Mack and G. A. Hipiwr to ei a a oe a eee 89 XI.—An Apparatus for Determining Freezing Point Lower- ing; by R. G. Van Name and “W. G. Brown 2550 =e XII.—The Sodium-Potassium Nephelites; by N. L. Bowen 115 XIII.— Pottsville Formations and Faunas of Arkansas and Oklahoma; by 1K. FL Matuur 222202. 2s ee 133 XIV.—A Study of Two So-called Halloysites from Gasca and Alabama; by P. A. vanprR MeEuLeEn .-._.__._-._ 140 XV.—Methods in Reversed and Non-reversed Spectrum Interferometry (continued); by C. Barus .--...-:.---- 145 XVI.—On the Identity of Hamlinite with Goyazite; by W. TO SCHALLEE <5 ee ees eee 163 SCIENTIFIC INTELLIGENCE. Chemistry and Physics.—Occurrence of Free Monoxide in the ‘‘ Floaters” Kelp, S. C. Lanepon, 160.—Atomic Weight of Lead of Radioactive Origin, T. W. RicHarps and C. Wapsworts, 38d: Engineering Chemistry, T. B. STILLMAN, 166.—Qualitative Analysis, E. H. S. Bainny and H. P. Capy: X-Ray Ware-Lengths, M. SimeBaun, 167.—General Physics, W. S. FRANKLIN and B. MacNorr, 168.—Cosmical Evolution, Critical and Constructive, Second Edition, E. McLennan, 169.—An Introduction to Astronomy, New and Revised Edition, F. R. Moutton, 170. Geology and Mineralogy.—Some Geophysical Observations at Burrinjuck, L. A. Cotron, 170.—Geology of Cincinnati and Vicinity, N. M. Fry-. NEMAN: Field Geology, F. H. Langer, 172.—The Fundamental Prin- ciples of Petrology, (E. Weinschenk), A. JoHANNSEN: L’Oural du Nord; Le Bassin des Rivieres Wagran et Kakwa, L. Duparc et M. Tikanowitcu: Etude Comparée des Gites Platiniféres de la Sierra de Ronda et de l’Oural, L. Duparc and A. Grosset, 173. —FEtude Cristallo-* graphique et Optique d’un certain nombre de Minéraux des Pegmatites de Madagascar et de Minéraux de |’Oural, R.-C. Sapot: The Geological His- tory of Australian Flowering Plants, Errata, E. C. ANDREws, 174. Obituary. —W.-Evuis: C. Rem: D. OnLtver: A. M. WortTHINGTON: J. WriaGuHtTson, 174. . | | | . | ' | | | __ “se ‘y he eee ay “Library, U. S. Nat. Museum. « VOL. XLII. MARCH, 1917. 4 Established by BENJAMIN SILLIMAN in 1818. 3 THE 3 AMERICAN lJ OURNAL OF SCIENCE. Epirorn: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or Camsrincez, PROFESSORS ADDISON E. VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT E. GREGORY | -anp HORACE S. UHLER, or New Haven, Proressor JOSEPH S. AMES, or Battimore, aos ~ Mr. J. S. DILLER, or Wasuineton. un50 | | Proressor HENRY S. WILLIAMS, or Iruaca, FOURTH SERIES oe VOL. XLUI—[WHOLE NUMBER, CXCIII}. ~ | No. 255—MARCH, 1917. Zz | NEW HAVEN, CONNECTICUT. 3 soe gee 1917. THE TUTTLE, MOREHOUSE: & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. Published monthly. Six dollars per year, in advance. $6.40 to countries in the _ Postal Union ; $6.25 to Canada. Single numbers 50 cents. ss Entered as ’second- class matter at the Post Office at New Haven, Conn., under the Act of March 3, 1879, LIST OF CHOICE SPECIMENS Wiluite and Achtaragdite, Wilui River, East Siberia. $5.00 to $20.00. PAS Vesuvius; Poland, Maine; Morelos, Mexico. $1.50 to $3.00. . .Wilkeite, Crestmore, Riverside Co., California. $2.00 to $4.00. Niccolite, Mansfeld, Thuringia. $1.50 to $4.00. Native iron, Cassel. $8.00 and $50.00. Pink beryl] with orthoclase and albite, Grotta d’Oggi, San Piero, Elba. $5.00 to $12.00. Pink beryl with orthoclase and albite, Grotta d’Oggi, San Piers: Elba Museum specimen, 8" x5”. $12.00. Tourmaline-pink beryl- albite-orthoclase, Grotta d’Oggi, San Piero, Elba, 6" x5", $18.00. ‘ Tourmaline with foresite and albite, Grotta d’Oggi, San Piero, Elba. $5.00, $8.00 and $10.00. Tourmaline, Elba, loose crystals. 50c. to $1.00. Diaphorite, Pribram, Bohemia. $12.00. Proustite, Freiberg, Saxony, $13.00; Chafiarcillo, Chile. $5.00 and $15.00. Argentopyrite-pyrargyrite-arsenic, Andreasberg, Hartz. $15.00. Cinnabar, Idria; New Almaden, California; Broken Hill, New South Wales. $1.00 to $5.00. Jordanite, Binnenthal, Switzerland. $1.50 and $3.00. Jordanite, massive, Silesia, Germany. $2.00. Millerite, Binnenthal, Sw itzerland. $2:00 and $3.00. Sartorite, Binnenthal, Switzerland. $3.00. Binnite, Binnenthal, Switzerland, $2.50, $3.00 and $10.00. Native Lead, Langban, Sweden, $3.00. : Natrolite, Aré, Langesundfjord, Norway. $8.00; very stout xls. (mew find). Wagnerite, loose xl., Werfen, Salzburg. $7.00. Stephanite in calcite, Andreasberg, Hartz, $5.00: Pribram, Bohemia. $3.75. > Erythrite, Schneeberg, Saxony. $2.00 and $4.00. Smaltite, Schneeberg, Saxony. $2.50 to $5.00. Mimetite, var. campyllite, Dry Gill, England. $4.00. Realgar, Kapnik, Hungary. $3.50 to $10.00. Realgar and lorendite, Allchar, Macedonia. $2.50. Minium, Tombstone, Arizona. $8.00. Josephinite, Del Norte Co., Oregon. $2.00 to $3.00. Awaruite, Smith River, California. 25c. to 7dc. Lawsonite, Marin Co., California. 75c. to $2.00. Eulytite, Schneeberg, Saxony. $2.00. Diaspore in dillnite, Dilln, Hungary. $3.00. baer ite, N atural Bridge, New York. $4.00; large crystals i in aie AM Wernerite, Norway. $2.50. Andalusite, Seliraintal, Tyrol. $5.00; large crystals in matrix 44" x 3”. ALBERT H. PETEREIT 81-83 Fulton St., New York City ‘Ge \ 4 i , a ; \ aR MAR 2 = [9 if ' J THE On: am . f er AMERICAN JOURNAL OF SCIENCE PROURTH SEHERIES.| oe Arr. XVIIL—A Method for the Determination of Dis- sociation Pressures of Sulphides,* and its Application to Covellite (CuS) and Pyrite (Fes,); by E. T. Atren and Rosert H. Lomparp. I. INTRODUCTION. Amone the sulphides there are a considerable number, 1f we include those of complex composition, which dissociate (lose sulphur) at accessible temperatures. In all such instances the dissociation pressure is of course an essential condition of formation of the sulphide. This fact was impressed upon us in a preliminary studyft of the sulphides of iron and copper. Very little work has been done on the subject; in fact only the dissociation pressure of covellite (cupric sulphide) seems to have been studied with any approach to accuracy, though some preliminary work has been attempted with pyrite.t In measuring the pressures of sulphur vapor, it is obvious that a mereury gauge would be out of the question. Three§ methods have been employed heretofore ; a dynamic method in which is determined the quantity of sulphur volatilized in a measured * This work is a portion of an investigation of the copper sulphide-ores undertaken by this laboratory in co-operation with Prof. L. C. Graton and colleagues of the Harvard Mining School. + Annual Report of the Director of the Geophysical Laboratory, 1914, 156 {Karl Schubert, Dissociation der einiger Oxyde, Karbonate, und Sul- phide. Dissertation, Berlin, 1909. Schubert made use of the Victor Meyer principle. See also Hempel and Schubert, Z. fiir Elektrochemie, xviii, 729, 1912. § A method for the determination of vapor densities at high temperatures published by G. E. Gibson might be used for the solution of this problem. The method depends on the elasticity of a quartz membrane, Unfortunately the apparatus appears to be unusually difficult to construct. See G. E. Gib- son, Dissertation, Breslau, 1911; also Proc. Roy. Soc. Edinburgh, xxxiii, 1, 1912. Am. Jour. Sct.—FourtH Series, Vou. XLIII, No. 255.—Marca, 1917. 13 176 Allen and Lombard—Determination of volume of nitrogen, the gas being passed over the sulphide while the latter is held at constant temperature; .a static method in which the sulphide is maintained at constant temperature while sulphur vapor at a known partial pressure in a mixture with nitrogen is passed over it, the sulphide losing or gaining in weight according as its dissociation pressure is greater or less than the sulphur vapor pressure ; and finally a static method in which the sulphur vapor pressure is directly measured by the very ingenious spiral quartz-glass gauge devised by F. M. G. Johnson.* The first two methods were used by Wasjuchnowa,t the lastt by Preunner and Brock- moller. We had prepared to use the spiral gauge in the investiga- tion of the stability relations of the copper-iron sulphides, but the outbreak of the European war about that time prevented the importation of quartz-glass gauges, and the construction of sufficiently sensitive instruments in this country proved im- possible. II. New Method for Dissociation Pressures. In this dilemma the present method was devised. It con- sists in balancing the dissociation pressure to be determined by a known vapor pressure of liquid sulphur. For this purpose an evacuated glass tube is required, having at each end a small bulb (fig. 1). One of these bulbs contains the sulphide, the other the liquid sulphur. Now while the sulphide is held at any desired temperature, the experimenter ascertains the temperature at which the sulphur must be heated in order that its vapor pressure may equal the dissociation pressure of the sulphide. It is obvious that the sulphur bulb must be the coolest part of the system, and that the method applies only to cases where sulphur is the only volatile product. In practice it is necessary to find by trial two temperatures for the sulphur, thus fixing two pressures, at one of which the sulphide loses sulphur, at the other remains unchanged ; or gains, if the dis- sociation product was originally chosen for experiment. At the first temperature it is evident that the vapor pressure of the sulphur is lower, at the second higher than the dissocia- tion pressure of the sulphide. Between these two pressures lies the dissociation pressure. In an actual determination, the interval between the two sulphur temperatures is narrowed down as far as practicable, the size of the interval at any point depending naturally on the steepness of the dissociation pressure curve. *Z. phys. Chem. lxi, 457, 1908. + Das Gleichgewicht Cupro-cuprisulfid. Dissertation, Berlin, 1909. {Z. phys. Chem., Ixxxi, 149, 1912. Preunner and Brockmédller used the spiral gauge as a zero instrument. Dissociation Pressures of Sulphides. Le This method suggested itself to us during a study of the copper-iron sulphides. At that time we were unaware of the fact that Wasjuchnowa had made use of the same principle, though jess conveniently applied, in a control method, the second mentioned above. She even conceived the idea of using an evacuated tube, though she never actually tried it. The method was used by her as a control only, most of her work having been done by the dynamic method. Apparatus.—The tube used in our measurements (see fig. 1) is made of Jena combustion glass where temperatures below about 675° are to be reached ; for higher temperatures quartz- BiG. ds | ) | D y Y é A es t | i = ae Fic. 1. Tube of glass (or quartz glass) for the determination of dis- sociation pressure. 1. Side view. 2. Top view. glass is employed. The apparatus is constructed as follows: Two glass tubes, about 1™ in inside diameter, are first sealed to either end of a stem of proper Jength and abont 5™™ in inside diameter. One of the tubes is then closed and blown into a small cylindrical bulb about 2° in length to hold the sulphur. To prevent the liquid sulphur from flowing out into the stem, the bend at Bis made. Two glass hooks to hold the thermocouples in place are now attached at E, and E,. A con- striction D is then made at a convenient distance from the open end of the tube after which about 0°5 g.* of pure sulphurt is dropped into the bulo A, melted and solidified. A small glass tube which contains 0:2 g. or more of powdered sulphide is now slipped into the bulb C. In some cases small lumps about 2™™ in diameter were substituted for the powder. LD is f *This quantity always insures a liquid phase under our working condi- 10NS. + The sulphur we used was three times distilled in vacuo. Since investiga- tions of the boiling point of sulphur have disclosed no material difference in the boiling points of different samples, and since our results make no pre- tension to high accuracy, the sulphur used in these experiments was not further investigated. of on O 178 Allen and Lombard—Determinat ee “AVTO OAL IT *SULIOAOD-Od Lg *XOLO[RO $o}s0qsV “wun punLy ‘ang dns 9 Je) WY) 99 G ‘Oplydyjns oy} ywoy 0} pos ‘| x OINSIOIA UOTJLVIOOSSTP JO UOTJVULULIOJoOp OY} UL pasn soovuAINy *Z “OTYT ees SEA ASE A> So BSS sant ashen sn Atas P| LOUDY EVEL UE VEU OL YRUU CULE GLE VOTRE VEL UT VLU AR AN UA UU WY I LEU LY URL WY WUE LULU UTE LEU 995 Fa ce NEALE SEER TOU ALATA RAE he AAA LEE truction used in furnace 1 was tried by us at ferable on account of is pre but the fire clay construction of furnace 2 D q co) rs) wr . 4S Rare Pes os o?Op on ie) ab ae ro ae 2 0) 275 salen 5 * mM ty M eS Dissociation Pressures of Sulphides. 179 then further constricted* to a thick-walled capillary. The open end of the tube, after being drawn down to proper size, cut off and smoothed in the flame, is attached to the vacuum pump by a piece of heavy-walled rubber tubing, and the pump is started. A little caution is required at first to prevent any sulphide powder from being sucked into the capillary. When the pressure is sutticiently reduced (to 1™™ or less), the sulphur is again melted by the free flame and the apparatus is carefully heated thr oughout its whole length to remove water vapor as completely as possible. While the pump is still in action, the tube is slowly sealed off at D. Lhe furnaces.—The above apparatus is heated by a pair of coaxial cylindrical furnaces which are attached to carriages in a horizontal position in such a way as to slide easily on a track (figs. 2 and 8). By this means the two furnaces may be easily slipped over the glass tube and brought tightly end to end when an experiment is to be made; or quickly drawn apart when the system requires to be chilled after sufficient reaction has taken place. The most important consideration in the con- struction of the furnaces is the maintenance in both of them of constant temperature ranges of sufficient length so that each bulb of the glass tube (fig. 1) shall be kept at a uniform tem- perature throughout, and so that a slight displacement of the tube from its proper position in the furnace can not disturb this uniformity. To fulfil this important condition it was found necessary to make the furnaces of considerable length and to use three separately controlled heating coils for each furnace. In our most satisfactory installation the furnace which serves to heat the sulphide is 46°" long, while the other, used for heating the sulphur, is 68° long. The two main heating coils are made of nichrome wire 1:3™™" in diameter and having a resistance of 0°75 ohms to the meter. The coils are wound on alundum tubes about 5™ external (4™ internal) diameter, 8 turns to the inch. The longer one has a total resistance cold of about 23 ohms, the shorter about 11°5 ohms. The auxiliary coils at each end of the furnaces are made either of nichrome tape, wound spirally into a flat disk with heavy asbestos paper as the insulator, or, for higher temperatures, they are constructed of flat circular disks of alundum with a spiral groove in one side; the wire is laid in the groove and cemented in by a paste of oround alun- dum and water, which is then dried and baked. This device was contrived by Mr. J. B. Ferguson of this laboratory. For these auxiliary coils, wires or tapes about 7 meters in length and having a total resistance of 5 ohms were found adequate. * Tf the tube is not partially constricted before the sulphide is dropped in, oxidation is to be feared because of the hotter flame and longer time required to draw down the wider tube. 180 Allen and Lombard— Determination of x The current in each of the six furnace coils is separately con- trolled by a sliding resistance. Ours possessed a resistance of about 30 ohms and a carrying capacity of 5 amperes. With this arrangement the currents could be regulated to about ‘01 ampere. The coils were used either in 55 volt or 110 volt cir- cuit according to the desired temperature. An important aes: Fic. 3. Apparatus used in the determination of the dissociation pressures of sulphides. detail of the furnace construction, which should be mentioned, is a pair of asbestos plugs, A, B, fig. 2, which afford the neces- sary thermal insulation between the two furnaces. These plugs fit the alundum tubes which form the working spaces of the furnaces. They are about 2°5™ thick. Each “plug i is pierced in the center with a hole which admits the stem of the glass apparatus (fig. 1) easily, and each is split into halves so that it Temperature in Microvolts. 3300 Dissociation Pressures of Sulphides. 181 may be readily set in position when the tube is put into the furnace (fig. 2). Thus the constant temperature portions of the two furnaces may be kept several hundred degrees apart. In these furnaces it is possible to maintain temperature regions of from 4°" to 12° in length, over which the variation is not more than 0°5°. This requires frequent temperature readings throughout the entire length of the apparatus at inter- Fic. 4. 40 36 32 98 24 poe AG 12 Centimeters from outside end of furnace. Fic. 4. Curves showing temperature distribution in the furnaces. I, I. Curve for sulphur furnace. i ii | + sulphide furnace. vals of about 3°" and frequent adjustment of the currents as the temperature readings may indicate—a process frequently consuming four or five days. A curve plotted from data of this kind is given in fig. 4. Such curves were worked out at temperature intervals of about 25°. It is obviously essen- tial that no point on these curves between the positions of 406° Temperature—C° 182 Allen and Lombard—Determination of bulbs A and O (see fig. 1) shall be lower than the temperature of bulb A containing the sulphur. The furnace ends must of course fit well together or a cold region will be found at the joint. Calibration of thermocouples.— All the thermocouples used in these experiments were of platinum-rhodium, and their readings were carefully taken at the following melting points :* tin 231°9°, cadmium 320°9°, zinc 419-4,° sodium sulphate 885°, silver 960°2°, copper 1082°8°. In some instances the reading was also taken at the melting point of antimony 630:0°. Manipulation.—W hen the furnaces are ready for an experi- ment they are drawn apart on their carriages, the split plugs are removed and the thermocouples pushed through their por- celain jackets till the junction of each projects a few inches beyond the inner ends of the furnaces. The thermocouple in furnace 1 is then slipped over the hook E, of the glass tube and so adjusted that the thermojunction is brought close alongside of the sulphide powder. The glass tube is now sup- ported and guided by one operator while the other draws it carefully into the furnace by means of the thermocouple, until the glass chamber enclosing the sulphide is within the region of constant temperature. The split plug B (fig. 2) is now placed in position. Furnace 2 is pushed up close to the bulb A (fig. 1) so that the second thermocouple may be attached to the hook E,; the plug A is set in place, and while the two thermocouples are held taut the two furnaces are brought care- fully together so that the tube is not displaced. Although the whole operation requires but a few minutes, the temperature in the middle of each furnace naturally falls 10° or 15°, but it rises rapidly again and soon returns to its former value. After a period of several hours, during which the temperature is carefully maintained constant, the furnaces are quickly drawn apart, the plugs removed by forceps, the thermocouples detached and the tube withdrawn. This operation generally requires less than a minute, and, since the tube is very light and its heat capacity small, its cooling is very rapid. Vapor pressures of sulphur.— We used in our experiments values for the vapor pressure of sulphur obtained by other investigators, namely, Ruff and Graf, Matthies, and Bodenstein. The determinations of Ruff and Graft which extend from 0° to 211°3° were made by the dynamic method; the sulphur was volatilized in a stream of hydrogen. Bodenstein{ made five determinations by observing directly the boiling point of sul- * The nitrogen thermometer from zine to palladium, Arthur L. Day and Robert B. Sosman, this Journal (4), xxix, 93, 1910. + Z. anorg. Chem., lviii, 209, 1908. { Z. phys. Chem., xxx, 118, 1899. Dissociation Pressures of Sulphides. 183 phur under measured pressures. His determinations extend from 874° to 444:8°. Some years later Matthies* supple- mented these by measurements extending from 210°2° to 379-4°, using the same method as Bodenstein. Earlier determinations by Regnaultt extended to higher temperatures (450° to 570°), but as they are beyond the range of our experiments they are not quoted. For the convenience of the reader the data of Ruff and Graf, Matthies, and Bodenstein are tabulated below. TABLE I, The vapor pressures of sulphur as determined by other observers. Observer Ruff and Graf W. Matthies Max Bodenstein p in mm. of é' p in mm. p in mm. of éyoyOr mercury fin Cy of mercury tin C° mercury 49°7 0°00034 210°2 1°35f 374° 240° 18°3). "0023 216°7 2a ON 393° 336° 89°0 "0057 222°4 3°20 410° 443° 99°3 "0089 230°6 4°48 427°0 580° 104°0 "0115 234°4 5°54 444°8 764°5 110°8 °0200 237°3 6°5 114°5 "0285 241°8 8°45 123°8 "0535 265°0 20°5 131°9 O81 306°5 53°5 132°2 "079 341°7 105°5 133°1 ‘088 352°5 133°0 141°0 °13] 363°0 176°0 147°0 “192 379°4 250°1 157°0 *332 1G62:0 "403 172°0 502.0 189°5 138 AleleS 3°14 Ill. Zhe Dissociation Pressure of Sulphides. Just how widely applicable the above method is to the determination of dissociation pressures can not yet be predicted, as too little is known about the stability of the various sul- phides, but we can see no impediment to its use at any tem- peratures where quartz glass will hold gas tight (1100° to 1200°). * Physik. Zeit., vii, 395, 1906. + Mem. de l’Academie, xxvi, 339, 1862. + Matthies states that his results below 4 mm. pressure are unreliable. 184 Bs and Lombard Determination of We have used quartz glass in a number of measurements, and have found no special difficulty with it. Doubt less some sul- phides like cuprous sulphide do not decompose appreciably within this temperature range, and some volatilize unchanged. Others like FeS,, CuS, OuFeS,, Cu,FeS,, PtS, and presumably those of many other heavy metals dissociate at considerably lower temperatures and are within the range of the method. Hig 5: Pressure—mm. of Mercury. ety Paes Bee oehns ERRERERRRERSE SS wee RRR RERORS =: CREB eae ee ie Ws ee i PCC eis feelers ieee ae ee ee eee a eee Y= | CEPEE EEE Eee test to ae BRERA REREEREROR ARR EORe BRERA RREMSRERERERP Ale See eee PERERA SS oe tr aaeceeeoeeeeeo BeRBERAERRARERSS BREESE SESBERERRASe BRR REREREES RARER BRERERERRSARERERRSESSRee Ss et Te he a aad ee Daa eee ee eee ceeececce— 400 420 440 460 480 Temperature—C°. 500 Fic. 5. Dissociation pressure curve of covellite (CuS). + + Authors’ results. © © Wasjuchnowa’s results. % & Preunner and Brockmoller’s results. 1. Dissociation of covellite ( Cu8). In each particular case where the method is applied, there must of course be some means of ascertaining whether the sul- Dissociation Pressures of Sulphides. 185 phide. has gained or lost sulphur or has remained unchanged. This may always be determined by an analysis of the product and perhaps, with sufficient care in handling, it might be deter- mined by weighing the sulphide before and after the experl- ment, though this latter method so far has not been tried. In special cases the question may be settled by the change in some physical property. Thus the dissociation of covellite is most readily detected by an inspection of the color. Covellite is deep blue, while chalcocite, its dissociation product, is dark grey. Small lumps of the mineral about 2™™ in diameter were found very convenient to handle and on such surfaces as they presented the colors could be easily distinguished. Such lumps of covellite are readily penetrated by sulphur vapor under these conditions, for in a number of cases the lumps were removed from the apparatus after experiment, broken open and exam- ined with a lens, and the change was nearly always complete. A pure synthetic covellite was used in some of the experi- ments; in others chaleocite either natural or synthetic. The natural chaleocite, from Butte, Montana,* showed the follow- ing composition : Cal. for Cu.S Op Ae ene 79°67 79°85 eee tee.) MOOSE POLS jG Ql a TA he LORE cee we! 09 ae 100°06 100°00 Frequently lumps of both chalcocite and covellite were used in the same experiment but the resulting products could never be distinguished from one another except by the shape of the lumps. The results follow: TABLE II. Dissociation pressures of covellite from 400° to 490° Corresponding pres- Mean dissoci- Temperature Temperature surelimits of the ation pressure Estimated of sulphide limitsofsul- sulphurinmm.of inmm.of error in in C° phur in C° mercury mercury mm. 400° iss) Sills 12 1°5 + 1° 410° 198 —210 2— 3°4 2°7 + 07 433°7 246 —251 10—12°5 diel ap’ = og 450° 280 —284 30-32 31°0 Se) us 460° 306 -315 50-60 55'0 az S- 468° 334 -337 90—96 93° Se 3° 475° 363 -—365 166-174 7,0" Se 04 482° 389 -391 314-326 320° se (0 485° 401 -—408 386— 399 393° aay 490° 417°5-420 501— 520 510° + 10° * Analyzed by Eugen Posnjak. 186 Allen and Lombard—Determination of tee | These results are plotted in fig. 5. They are expressed tol- erably well by the well-known equation: log p = — + Blog T+C. By substituting the observed values of p at 410°, 460° and 485° respectively, the following values of the constants were found: A = —96397°514, B = +356°43227, C = —1150- 98605. From these values, p at the other temperatures was calculated. The calculated and observed values of p are com- pared in the following table: TABLE III. Comparison of observed and calculated values of p, the dissociation pressure of covellite Abs. temp. Temp. C” p observed p cal. diff. 763 490° 510° 600° —90° 758 485° 393° 393° 0° 755 482° 320° 306° +14: 748 475° 170° ee — 3° 741 468° 93° 100° — 7 733 460° 55° 55° 0: 723 450° 31° 27° + 4° Om 433°7 7) iis? + 1°5 683 410° Bey) Dall O; 673 400° 1°5 1°8 — 0'3 Though the formula serves fairly well for the purposes of interpolation it is our belief that a graphic interpolation gives more accurate results. Results of other observers. In Table IV our results are com- pared with those of other observers. Wasjuchnowa’s original results were not stated in terms of pressure ; only the mass of sulphur, volatilized at a measured temperature in 1 liter of nitrogen under standard conditions, is given. As the results are probably not accessible to every reader we quote them below in their original form. The partial pressure of the sulphur anor is given by the formula 7 mM ] y'32 Pen where m is the mass of sulphur volatilized in 1 liter of nitrogen at the given temperature and the pressure p, and v is the aver- age number of atoms in the molecule of sulphur under the Temperature —_— Results of Preunner and tilts of all obs eovellite ( [ VECO ue CuS) i (ei rn we ers on the dissociation pressure of Dissociation pressures in mm. of mercury Brockméller determined Results of Wasjuch- Results by Pin ‘C° by the spiral quartz nowa determined by the authors’ ENG gauge the dynamic method method 393°4 eke 0°6 Neel 400° ena ae ee 1°5 405°8 Lee 1°6 Bw 410° rere gh Oe 421°5 ids 5°3 eae 425°2 gees 0) Ee 433°7 Beat Eee Wey 436°2 Rohe 14:0 Tas 444°5 ive 21°8 te 447-4 sean 29°0 Epp 450° 80. (34)* 31 460° acl Be etd 55 467°3 isotipede 110 mee a 468° Seat ‘ete 93 470° 200° (145) Ses gies 475° 250° (195) 170 480° 313° (255) et ieg 482° Hee ee 320 483°8 neat 295 Dbges Sr ares a 9 485° oe oer 393 490° Beat ee 510 500° 980° (560) es ae 502°3 ae 595 Jam TABER, V\. pressure of covellite (CuS). Wasjuchnowa’s original results on the dissociation 2 ~Temperatur des Cus 393°4 405°8 421°5 436°2 447°4 467°3 483°8 502°3 T, abs 666°4 678°8 694°5 709°2 720°4 740°3 756°8 775°3 3 Schwefel menge pro Dampftdruck des Schwefels 11. Stickstoff g. 0°0054 0°0131 0°0405 0°1220 0°2884 1°290 5274 30°380 uber dem CuS berechnet f. Ss.s, bei 444°5°C. Schwefel in g pro 1 1. * Results in ( ) are interpolated by Preunner and Brockmoller. 188 Allen and Lombard—Determination of same conditions. Wasjuchnowa had at her disposal only one value of v (at the boiling point of sulphur) and was, therefore, able to calculate only one of her results in terms of pressure. Preunner and Brockmdller* by the aid of their data on the vapor density of sulphur calculated a part of her results (447°— 502°) in terms of p by a method which they describe. In a similar way we completed the series. The results are given in Table V. Wasjuchnowa also gives one other result by a differ- ent method,t viz. Schwefel in g tiberfuhrt Temperatur des Schwefels Temperatur des CuS von 11. N bei 0° C es T. abs. c° T, abs. und 760™™ 216°5 489°5 425°2 698°2 0°0559 If the mass of the sulphur is caleulated as above in terms of — p, the value 7™™ is obtained which we give in the table. If, however, we take the vapor pressure of sulphur from Ruff & Graf’s results we get, at 216°5° by a short extrapolation, 4™™. The discrepancy is unaccounted for. In the lower part of the curve our results agree surprisingly well with those of Was- juchnowa,t while both are much lower than those of Preunner and Brockmdoller. The suspicion naturally arises that their higher results have been due to water vapor in their apparatus or in the sulphide or possibly free sulphur in the latter, but an examination of their procedure seems to set the suspicion at rest.§ It is certain that their temperatures were not very uniform; variations of 5° occurred along the dissociation chamber of their apparatus, but even this would not account for the great differences between their results and ours. They admit that their pressure measurements varied considerably (Die Messungen wichen erheblich voneinander ab), but they do not account for the variation. ‘The small temperature in- terval within which the dissociation of cupric sulphide reverses itself in the lower part of our curve, and the care with which the temperature measurements were performed give us special confidence in that portion of it. 2. Dissociation pressure of pyrite. The mineral used in these experiments was a very pure variety of pyrite of the kind used in wireless telegraphy, which is said to come from Colorado. No impurities could be eer- tainly detected in 1 g. though doubtful traces of silica and * Z. phys. Chem., Ixxxi, 150, 1912. + Das Gleichgewicht Cupro-cuprisulfid, Dissertation, Berlin, 1909, p. 20. ¢{ W. used for the two points obtained at 484° and 502° a different method for condensing the sulphur. § Z. phys. Chem., lxxxi, 152, 1912. Dissociation Pressures of Sulphides. 189 heavy metals were found. Both sulphur and iron were deter- mined. Found Cal. Fe= 46°71 46°56 S= 53:29 53°44 SiO. = trace ? Heavy metals of H,S group = trace? 100°00 100°00 Pyrite does not seem so readily penetrated by sulphur vapor as covellite even at a much higher temperature. Covellite has a perfect cleavage, and lumps of it are always filled with seams while pyrite shows a smooth conchoidal fracture. For this reason, probably, pyrite dissociates much more slowly than covellite. Even after four or five hours heating the 100 mesh pyrite powder used in the experiments was only partially changed. The dissociated portion, 1. e. the pyrrhotite, amounted to about one-third to one-half of the original pyrite. It was of a somewhat darker tint and crystallized in mossy aggregations. The magnetic test for pyrrholtite—The difference in mag- netic susceptibility between pyrite and pyrrhotite affords a fairly sensitive method for separating them. An ordinary horseshoe magnet however is unsuitable. We employed a strong electro-magnet for the purpose. The glass apparatus containing the sulphide need not be opened for the preliminary magnetic test. After cooling, the bulb, C (fig. 1), is brought close to the poles of the magnet. Unless the quantity of pyrrhotite is small, some particles will show attraction very plainly. If the test is negative, it is necessary to open the bulb, remove and triturate the sulphide in an agate mortar and test it again. The magnet will now in some cases remove a little pyrrhotite. As a control on the results obtained with the magnet, a number of the products were also analyzed. The analyses were all carefully done in a platinum dish, the iron being twice precipitated with ammonium hydroxide which was prepared by saturating, with ammonia gas, water contained in a gold vessel. The following table shows the agreement between the magnetic tests and the analyses. The percentage of iron in pure pyrite is 46°56. All the magnetic products contained more iron than this. (See Table VI.) The non-magnetic products which were analyzed gave 46°90, 46°73, 46°99, 47°39. The first three results are as close to pyrite as could be expected. The last is the only real dis- 190 Allen and Lombard—Determination of TABLE VI. Comparison of magnetic behavior with composition. | Pres- Analysis Tem- sure of Tem- | pera- | sul- pera- |tureof| phur | Magnetic Portion Weight | Weight | Per- tureof| sul- jinmm.| behavior analyzed of sul- of cent- pyrite | phur | of . phide | Fe.0O3 | age of in C°./in C°.| mer- taken | found | iron cury : magnetic 9 s ° ° ° 610 50 12 | magnetic portion 0°1043 |0°0735 | 49°29 610 | 236 6 He a "1040-| -0788 | 53:00 665 | 380 | 959 | P°™” . | the whole | :2030 | -1361 | 46:90 magnetic 665 | 378 | 243 | magnetic | M28R°HC | -jo48 | -oga5 | 55-07 portion 680 | 418 | 505 os the whole | ‘0754 | :0514 | 48°43 680 | 418 | 505 rs | ef "2145 1446 | 47°16 680 | 411 | 453 c magnetic’ 1953 | 1068 (eae portion 680 | 424 | 554 | 7°" | the whole | -2244 | -1499 | 46-73 magnetic 680 | 421 | 530 | oe “a °2047 ‘1375 | 46599 595 | 209 | 38 /| magnetic | M#88RCUS | .ox90 | -0557 | 61-94 portion 595 | 228 4} HOD’ .- | the whole | -1011 | 0685 | 47:39 | | | magnetic | | crepancy in the table, but even here the result involves an error of less than a milligram. For a rigorous comparison of the methods, more material should of course have been taken for analysis, but small quantities of pyrite were used in the pressure experiments because in that way a uniform tempera- ture for the whole charge was more certainly attained. Dissociation pressures._-The dissociation pressures estab- lished by the aid of the magnetic method are tabulated in Table VII and plotted in fig. 6. These results are expressed by the equation: log p = — 7 + Blog T + C, with about the same degree of accuracy, as are those with covellite. The values of log p at 595°, 645° and 672° were chosen for the computation of the constants. The values found were: A = + 191942°61, B = — 434°1950%75, INssociation Pressures of Sulphides. 191 C = + 149756707. A comparison of the values of p com- puted from the constants, with the values observed are given in Table VIII. Dissociation pressures of pyrite from 575° to 680°.* TABLE VII. Tempera- ture of sulphide im: ©° 680 672 665 655 645 635 625 610 595 575 Tem pera- ture limits of sulphur ime 418-421 393-395 378-380 363-364 341-342 314-316 288-291 250-255 209-228 167-1838 #7) Correspond- ing pressure limits of sulphur in mm. of mercury 505-530 337-349 243-259 166-170 105-108 60-62 35-37°5 12-15°0 3°0— 4:0 (oa eI Mean dissociation pressure in mm. of mercury 518 343 251 168 106°5 61:0 36'S 13°5 3°5 0°75 Estimated error in mm. sey 2 -L oo i i Oe Oe) He HE He HE He HE HEH 0°25 *The upper pressure limit was the lowest pressure actually found at which the product was entirely non-magnetic ; the lower pressure limit was the highest actually found at which magnetic substance was detected. TABLE VIII. Comparison of observed and calculated values of p, the dissociation pressure of pyrite. Abs. temp. Temp. C° p observed 953 680 518 945 672 343 938 665 251 928 655 168 918 645 106°5 908 635 61:0 898 625 36°3 883 610 13°5 868 595 3°5 848 575 0°75 p cal. 447 343 264 0°0 0°25 In a former paper from this laboratory+ it was stated that with hydrogen sulphide gas at 1 atmosphere pressure, pyrite and pyrrhotite appeared to be in equilibrium at 565°-575°. By an extrapolation of the curve of Preunner and Schupp,t the par- + Allen, Crenshaw and Johnston, The mineral sulphides of iron, this Jour- nal, xxxiii, pp. 203-204, 1912. tZ. phys. Chem., lxviii, 161, 1909. Am. Jour. Sct.—Fourts Series, Vou. XLIII, No. 255.—Marcs, 1917. 14 Pressure—mm. of Mercury. 192 ‘Allen and Lombard— Determination of tial pressure of sulphur i in hydrogen sulphide at this temperature was estimated to be about 5"™. The free hydrogen very likely has an influence on this equilibrium, for in pure sulphur vapor we now find the dissociation pressure of pyrite to be less than 1™™ at the above temperature. In the same investigation a strong absorption of heat was found in pyrite between 665° Fic. 6. 600 Enea eRe Rae eee eee (ee ee fe a Ea rE 500 ees 400 EERE alee RRR se. s00 ft J = CCE pe SI Se se Bel a Oe a De ea eae emia ne ale Pi oer 575 595 615 635 655 675 Temperature—C°. Fic. 6. Dissociation pressure curve of pyrite (FeS2). and 685° which was attributed to the fact that the sulphur pressure had reached 1 atmosphere. This is confirmed by the present work. A short extrapolation of the curve plotted in tig. 6 indicates that a pressure of 76°" would be reached about 683°. One atmosphere pressure is evidently reached only at the wpper end of this absorption interval (665° -685° ) rather than the lower as we then supposed. Bae Ane eee Dissociation Pressures of Sulphides. 193 Heat of dissociation of covellite and pyrite.—The two cases of dissociation which we have studied are unusually compli- cated. Not only does the sulphur vapor change in density with change of temperature, but the solid products, pyrrhotite and covellite, vary in composition. The validity of the Van’t Hoff equation : | log”: — log?) = Q am ane Po ae Ve ee, may be doubted here, but assuming that it applies, the experi- mental data are not sufficiently accurate for the calculation of satisfactory values of Q. An error of 1° in the temperature of the sulphur causes an error in p of 8-10™ in certain parts of the curve (450-500™") which occasion a very large error in the calculated value of Q. Small errors in the temperature interval T’,—T, also cause large errors in the value of Q. IV. Advantages and Disadvantages of the Method. 1. The above method for the determination of dissociation pressures has the important advantage of approaching the equilibrium from both directions so that, if the temperature measurements are accurately made, the course of the curve within the limits set should be certain. It will be noted that the chilling of the glass tube as a consequence of its light weight and the method of manipulation, is very quickly accom- plished. It is possible that at high pressures there might be some reversal of the reaction in this operation, but nothing of the kind has been observed in our experiments. 2. Hither the sulphide or its dissociation product may be used and the presence of free sulphur would have no effect, advantages not possessed by the spiral gauge method. The influence of moisture or other foreign gases would probably be less serious than in other methods. 3. It is unnecessary to wait till the sulphur vapor has come to equilibrium with the sulphide; all that is needed is to know with certainty the direction in which the reaction is proceeding under given conditions. With the spiral gauge method it would of course be necessary to a correct measurement that the sulphide should be held at constant temperature until equili- brium was reached. This necessity would cause much trouble if the reaction moved very slowly, because the temperatures require constant watching. The dynamic method could not be used at all in such a case, for it depends on the assumption that the partial pressure of the sulphur in the stream of indit- ferent gas is always equal to the dissociation pressure.* This * Wasjuchnowa proved that equilibrium under her working conditions was quickly established between covellite, chalcocite and sulphur vapor. 194 Allen and Lombard—Determination of advantage is quite important. The two mineral sulphides covellite and pyrite reach their equilibrium pressures very slowly under the conditions described in this paper. For example, 1 g. pyrite was crushed to a powder which passed a screen of 100 meshes to the inch. After 5-6 hours heating at a temperature of 672° and a pressure of 320"™ of sulphur vapor, therefore very close to the dissociation pressure curve, about 10% remained unchanged while the rest contained 49°6% iron and had therefore lost only 3% sulphur. A portion of this dissociation product was heated a second time under simi- lar conditions. Analysis showed that it now contained 57% iron. It had therefore lost 75% more sulphur in the second heating. COovellite also was found to move slowly to equili- brium, though not so slowly as pyrite. 4. The apparatus required is easy to construct, is not fragile, and is inexpensive. 7 5. The most serious disadvantage of this method is the time which it involves; it being necessary to make a number of experiments for the determination of every point on the curve. The adjustment of the furnace temperatures is also tedious, but it is difficult to see how this could be avoided in any method, for uniformity in temperature is a prime neces- sity in any solution of the problem. 6. The method is not adapted to the measurement of pres- sures of much more than an atmosphere, since the vapor pressure curve of sulphur in this region becomes so steep that a small error in temperature causes a comparatively large error m the pressure. Other uses of the method.—In principle, of course, this method for dissociation pressures need not be confined to the sulphides ; any compound which dissociates at accessible tem- peratures with the formation of a single volatile product which condenses at accessible temperatures could be investigated in this way. A further proviso is that glass (or quartz glass) should stand the chemical action of the vapor sufficiently well. The method also affords a convenient means for the synthesis of most* sulphides and can be employed to advantage in cer- - tain instances where dissociation gives trouble. More import- ant than this, the method promises to be a valuable instrument in ‘the investigation of complex sulphides which can not be synthesized by simple fusion in an indifferent gas. Summary. 1. A new method has been devised for the determination of dissociation pressures at comparatively high temperatures in * Some sulphides like marcasite could probably not be made in the dry way. Dissociation Pressures of Sulphides. 195 cases where mercury gauges can not be used. It is intended especially for sulphides. It depends in principle upon balanc- ing the dissociation pressure of the sulphide by the vapor pres- sure of sulphur at a known temperature; the pressure is not directly measured. The method applies to other compounds than sulphides provided there is a single volatile dissociation product which does not attack glass (or quartz glass) and which condenses at accessible temperatures. The method can not be used above about 1100° to 1200°. 2. By this method the dissociation pressure curves of covel- lite (CuS) and pyrite (FeS,) have been determined from about {= to 5OO™™, 3. The chief advantage of the method is that the equilibrium is approached from both directions and the experimenter is therefore not liable to be deceived by false equilibria. The method has the disadvantage of being slow and is inaccurate at pressures much above an atmosphere. 4. The method was devised as an instrument for the inves- tigation of complex sulphide systems where the dissociation pressure is a factor in stability that can not be neglected. There seems to be no reason why it should not find a broader application to other systems of similar characteristics. 5. The method also supphes a convenient means for the synthesis of sulphides the dissociation of which causes difficulty. Geophysical Laboratory, Carnegie Institution of Washington, Washington, D. C., November 22, 1916. 196 PP. EB. Raymond—Beecher’s Classification of Tritobites. Art. XVIII.——Beecher’s Classification of Trilobites, after Twenty Years; by Percy E. Raymonnp. In the February and March numbers of this Journal for 1897 appeared the two parts of Professor C. E. Beecher’s short paper, on an * Outline of a Natural Classification of the Trilo- bites.” The classification there proposed, modified rather by curtailment than expansion, appeared again in the Eastman edition of Zittel’s Textbook of Paleontology in 1900. On this occasion, just twenty vears later, it seems fitting to inquire how Beecher’s work has stood the test of actual use and I wish also to present. what follows as a slight tribute to the memory of my revered teacher. Beecher himself regarded this classitfica- tion only as an outline, and often expressed the wish that he himself or some other person could find the time to elaborate the classification in the same manner that Professor Schuchert had elaborated his earlier outline scheme for the Brachiopoda. Professor Beecher’s untimely death just seven years after the first part of his “classification” was printed prevented his own return to the subject, and the studies and contributions of other writers on trilobites during these twenty years have dealt usually rather with single species, genera or families than with the subject as a whole. Numerous objections to the scheme proposed by Beecher have appeared, from Pompeckj, in 1898 to Swinnerton in 1915, the objections varying in strength from Wood’s* sweeping remark that “the only classification of Trilobites which can be adopted is a division into families” to Swinnerton’s statement that the majority of new trilobites found since the publication of the classification “fit into it without difficulty and prove that to a large extent it is conceived on a sound basis. SS) vo .& 4 a be | ole o 6 Ww On es S15 ® "ote =| ~ © ON mae 2 230 A. P. Honess—Kiching Figures of Beryl. this is only a slight modification and may represent a different stage of development, or a potassium solvent shghtly different from that used by the writer. The figures produced by the writer also resemble natural etchings found on the second- order pyramid of a light green, transparent beryl crystal from Topsham, Maine. The figures occurring on this form are symmetrical to a vertical plane. The second-order pyramid was not well-developed, con- sequently the figures obtained are rather indefinite but resem- ble very closely those produced by Traube on the same form. They are elongated oval forms (diagram E), extending = |i) RVR “| PEL) bx < |X Dia.A Dia.B parallel to the crystalline edge 1120/1121, widest in the center and tapering at each end. Four planes comprise the figures, the two larger planes meeting in the bottom of the pit and ex- tending almost the entire length; at the ends, right and left, a small face descends rather abruptly to the bottom. The figures are also symmetrical to a vertical plane, which accords with the symmetry of the type. Thus the four forms etched serve conclusively to illustrate the holohedral character of beryl. While the etch figures produced by the separate alkalic hydroxides are distinctly different, those produced by a one to one mixture of sodinm hydroxide and potassium hydroxide are strikingly similar to those of the pure sodium fusion. Solution. was stopped after six or seven seconds, in order to prevent intergrowth, hence a slight difference might be ex- - pected, as the sodium hydroxide figures, and also the potassium A. P. Honess—Etching Figures of Beryl. 231 figures, were immersed for fifteen seconds. Nevertheless, large, well-defined etchings were obtained, which may be easily compared with those previously described. These figures, as before mentioned, do not resemble the potassium figures to as great a degree as they do those produced by sodium hydroxide, and a comparison will eventually evolve into slight modifica- tions of the sodium forms. In the first place a detailed study of the sodiuni-potassium etchings (fig. 7) reveals four slightly eurved bounding lines instead of two, as described in the sodium figures. Also these figures are deeper with a larger plane lying parallel to the prism face; in most of the larger figures the basal groove so prominent in the sodium figures is noticeably absent. These characteristics reveal the presence of the potassium molecule, in the mixture, as they are quite com- mon in the forms produced by that fusion. It is possible that they are shorter and thicker than either of the others. But, on the whole, the shape of the figures and their orientation resemble very closely the sodium figures. Other etchings ap- pear on the same face representing various stages of growth, and even these bear much resemblance to the sodium forms. They are not so deep and possess a large plane parallel to the unit prism 1010; oftentimes they are so shallow as to be barely visible under the microscope, and still they are about the size of mature figures. Etchings of this kind have little detail, revealing practically nothing to the observer, except a broad flat plane, which gradually diminishes in size as the figure deepens, due to the development of the lateral bounding faces. In the mature figure, these lateral faces are much more prominent than the plane forming the base of the pit. Occa- sionally among the smaller figures, well-defined pits occur, which are bounded by four triangular faces, the apices of which meet in a common point, at the center of the figure (see fig. 8). These smaller forms appear to be identical with natnral etchings on Brazilian beryl crystals described by Kohlmann.* There is, then, a very noticeable resemblance in the two types of etch figures; those produced by the pure sodium hydroxide and those produced by the one to one mix- ture of sodium and potassium hydroxides; and, although the figures of different size and shape occur on the same face, which merely indicates a variation in conditions of etchings, each form reveals very clearly the symmetry of the face upon which it occurs, and the several fundamental forms etched serve conclusively to illustrate the holohedral character of beryl. Natural Kichings. Prisms.—The writer, in his investigation of the beryl crystals from some of the more important localities, has been «3 JL, ©p 232 A. P. Honess—KHitching Figures of Beryl. successful in establishing the symmetry of the mineral by means of several different natural etchings hitherto undescribed. While but three or four different forms have been etched on the crystals present, these are, in every case, so distinet and so well defined as to reveal very clearly the symmetry of the face upon which they occur. The first to be described is a light green crystal almost trans- parent, with the forms 1010, 1120, and 0001 well developed, collected near Hiddenite, North Carolina (diagram A, fig. 13). Fie. 14. 1010 1010 1010 j : 1010 / \ | PES Ue \ | \ // y | | i | \/ \V/ { | gS | | <> ee) IN if A \ | } || / | Dia.C-I Dia.C-I The unit prism 1010 contains many small pits, more or less quadrilateral, extending across the face, with their long axes at right angles toc. The figures are bounded by straight and curved lines, the smaller forms having rounded terminations. The quadrilateral outline of the larger figures is more apparent as they are composed of upper and lower faces (diagram A), and occasionally right and left faces, very rarely a plane parallel to the prism face. Figures of various shapes are represented upon a single face, indicating various develop- mental stages of growth. The figures are symmetrical to one vertical and one horizontal plane. The second-order prism 1120 is a very narrow face and con- tains but few distinct forms. ‘These are perfect spindle-shaped pits elongated parallel to the vertical axis of the crystal, and resemble very closely the natural figures occurring on the 1120 form of Brazilian beryls, as described by Kohlmann.* The six-sided forms described by Kohlmann as primitive figures are absent on the North Carolina erystal. Base.—On the base are beautifui, regular, hexagonal pits, with the sides parallel to the edge (0001/1010 (see 1 and 2, * Tc: A. P. Honess—Kkitching Figures of Beryl. 233 diagram I’, fig. 15); some are terminated by a basal plane, others are pyramidal with the six bounding planes lying approximately in the plane of the first-order “pyramid ; both forms are symmetrical to six vertical planes, which conforms to the symmetry of the type. Another form of natural etch figure occurs on a beautiful emerald from Musa Valley, near Bogota, United States of Columbia (diagram B, fig. 13). Three forms are etched. real: 121 a Prism.—The unit prism 1010 is beautifully pitted with quadrilateral figures, many of them being perfect squares with four sloping planes; others are slightly elongated at right angles to the vertical axis of the erystal, but all angles are ninety degrees, as in the more regular type. The right and left faces of the figures le in the prism zone, the upper and lower faces lie in the zone of the first-order pyramid and the base. The position as regards the intercept could not be accurately determined as these pits would not yield casts of sufficient size for measurement. The figures are very inter- esting from a genetic point of view, as they are identical] in form with the artificial figures produced by the writer on the unit prism of a North Carolina beryl, by sodium hydroxide ; they are perfect duplications at this state of growth. This is to be noted, however: if the artificial figures are allowed to develop to a more mature stage, they assume an elongated ap- pearance exactly as do the natural etchings on the emerald from Musa Valley, South America, but they differ in orienta- tion; the former have their long axes parallel to ¢, the latter 234 A. P. Honess—KEitching Figures of Beryl. at right angles. This seems rather suggestive, for may we not assume a solvent rich in soda, and is it not probable that the beryl was associated with rocks and minerals with a high sodium content? At least the form of the figures would indi- cate this. The etchings are symmetrical to two planes. Pyramid.—The second-order pyramid 1121, so well de- veloped on this erystal, reveals one or two etched faces. The figures are too small to permit of detailed study, but they ap- pear as top-shaped forms, with the point downward and divided symmetrically by a vertical groove (diagram B, fig. 13). They are, therefore, symmetrical to a vertical plane. Base.—The fioures on the base are similar to the natural figures on the North Carolina crystals, with the exception of a trigonal form (see 3, diagram F, fig. 15). The base, the second-order pyramid and the unit prism, then, reveal the symmetry of the type, as illustrated by natural etchings. As a usual thing the etch figures are alike upon the same form in the same locality, but erystals from Miask, Russia, reveal the presence of two different solvents. The change in form and orientation may be due to increase in concentration or temperature, possibly both, but very evidently there has been a decided change in the nature of the solution, for the difference in the figures is not only orientation, but size and shape as well. For comparison two crystals were examined and described; these crystals are very clear and a greenish yellow, the surfaces being very bright and in every way suitable for the study of the etchings : they are of the same size and similarly developed with 1010 as the dominant form, and upon this face the figures appear. Upon the one crystal most of the figures are hexagonal, due to the replacement of the acute angle of the rhomb by a small face lying in the prism zone ; the larger angle is approximately a hundred and_ thirty-five degrees (diagram C-I, fig. 14). Other figures occurring on the same face are very shallow and perfectly diamond-shape in outline. The pit is so shallow that the four lateral planes are reduced to mere lines. The base of the pit is very large and also diamond-shaped. The long axis extends at right angles to the prism edges. The figures of the second crystal are diamond-shaped ; measuring one and one-half mm. wide and three im. long. The long axes of the figures are parallel to the prism edge, the reverse of the eee of the figures on the first crystal. Occasionally the obtuse angle is replaced by a curved lateral boundary, producing a hexagonal form (diagram O-II). Although these crystals are from the same locality this would indicate two different solutions. As to the nature of the solvents, the writer las not been able to arrive at any con- A. P. Honess—Eitching Figures of Beryl. 235 clusions, although several attempts were made to duplicate the figures with laboratory preparations. Perhaps the most elaborate figures occur on a beautiful light green beryl from Topsham, Maine. The erystal possesses three of the fundamental forms, the unit prism, the base and the pyramid of the second order, and all faces are distinctly etched. Prism.—Upon the prism three stages of development are manifest (figs. 9 and 10). The primitive form is the shallow diamond-shaped pit, very simple and distinct, with very often a small pit in the center, just discernible in some figures and quite well developed in others. The pit, having reached an advanced stage, completely fills the diamond form, producing a second stage in development. The mature figure possesses a deep groove at right angles to the prism edge and passes the entire length of the figure. These two planes as they ascend quite suddenly diverge at a much larger angle and intersect the surface in a curved line and in some cases a straight line. This gives the figures an hexagonal appearance, elongated at right angles to ©; the acute angles at the ends of the pit measure approximately forty-five degrees, and are formed by the intersection of two small faces, lying in the zone of the second-order pyramid and the first-order prism. | Base.—The base reveals two different figures, one the out- growth of the other; one is hexagonal, the other twelve-sided and both regular with the edges of the hexagonal form par- allel to 0001/1120. (See 4 and 5, diagram F.) The dihex- agonal figures (fig. 12) ultimately become hexagonal as was observed in the ease of the artificial figures produced by sodium hydroxide on beryl, where the figures were composed of so many small faces as to appear circular, but which gradually became hexagonal as solution continued. This well-defined figure (fig. 12) possesses twelve triangular faces meeting in a common point. All figures on the base are symmetrical to six vertical planes. Pyramid.—Upon the pyramid of the second order are three or four well-defined triangular figures; they are isosceles triangles with the large angle turned upward and the base parallel to the edge 1121/0001. The larger forms often ap- pear pentagonal (diagram D, fig. 15) and all are symmetrical to a vertical plane. The artificial etchings produced by the alkalie fusions are represented in diagram E, where the relation of the figures to the crystalline edges may be observed. Forms on 1011 and 1121 are produced by potassium hydroxide and on 1010 and 0001 by sodium hydroxide. 236 A. P. Honess—Etching Figures of Beryl. Diagram G (fig. 15) represents the natural etchings found on a light green beryl from Mt. Antero, Chaffee County, Colorado. The figures are very large and simple, considerably intergrown, but occasional individuals may be readily detected with the unaided eye. The more common type of etching is represented by 1 and 2 of diagram G. In every figure there is present the two long faces lying in the prism zone, which either intersect at the base of the figure, or extend downward, intersecting the third long face, forming the bottom of the pit. Very often the larger figures possess narrow depressed areas extending lengthwise the pit and which, when observed in number, give the crystal face a striated appearance, parallel to C. The small face at each end may or may not be present. The figures are symmetrical to two planes at right angles. A few well-defined natural etchings, which were observed on the base of a ight green beryl, from Mursinka, Siberia, are shown in fig. 11. Hence the various forms, naturally etched, conform to the symmetry of the type. And from the foregoing investigation, it is readily observed, that the etch figure, be it produced in nature or in the laboratory by solutions of various constituents, concentrations and temperatures, the result is always an accurate outward reflection of the interior structure of the crystal. Princeton University, Princeton, N. J. J. M. Blake—Plotting Crystal Zones on the Sphere. 237 Art. XXI.—Plotiing Crystal Zones on the Sphere; by Joon M. Brake. (Article 4.) In the early sixties the writer made use of plotting on the sphere in the solution of certain crystal problems. Some pre- liminary trials were made with a nine-inch wooden slate-coated sphere, and the results were so encouraging that a hollow cast zine sphere was made of about twelve inches diameter. This sphere was accurately turned in a special lathe, and was then lay A Fie. 1. J. M. Blake’s Plotting Sphere. coated with a liquid slating compound, and its spherical figure was perfected by recoating the low portions and then grinding the surface with a pumice stone mould of the same curvature. A hard, smooth, true surface was thus secured. It was pro- vided with a graduated equatorial ring. On this ring a right- angled spherical triangle could be rested while plotting. The base of the sphere rotates on rollers set with the axes radial. The sphere can be turned independently of the base by slackening three leather-tipped centering screws; and it 238 J. M. Blake—Plotting Crystal Zones on the Sphere. rests in a shallow leather-lined cup which turns on a central adjustable screw point. The main value of the sphere when used in connection with crystal work is dependent upon its adaptability for the reten- tion of data relating to plane positions. It is obvious that plotting on the sphere cannot be performed with as great accu- racy aS upon paper; but if we refer to the measurements that have already been made with the goniometer, and keep the plotting work corrected, we can attain sufficient exactness to enable us to perform valuable work having erystal description for its object. This work can be done by such aid with econ- omy of time and labor. In fact, we can perform work that would hardly be attempted if dependent on the methods that have so long been in most frequent use. In order to secure the greatest degree of accuracy en a given crystal will yield, we make use of the best average zone tangent spaces that can be obtained from a measurement of the whole crystal. It will follow from this that the greatest num- ber of planes we can develop on a salt, or the greater number of trustworthy reflections we can obtain from a mineral, will tend to a more accurate determination of the length of the crystal axes. With these few hinis relating to the hoped-for, eventual, exact determination of crystal laws and constants, we will turn our attention to some practical methods of collecting and pre- senting data in available form for future generalization and mathematical treatment. In so doing we will endeavor to make it apparent, that graphic and mechanical methods can, with advantage, be substituted fora great mass of the work com- monly undertaken by algebraic and analytical methods. These algebraic attempts are many times founded on uncertain meas- urements, and mathematical precision cannot, therefore, be expected on such a basis. By the methods here advocated, the matter of collecting useful data is much facilitated, and the work can be undertaken by a larger number of observers. ; If we examine a tangent- -plane projection of the normals made from a suitable view-point, we will havea series of points arranged in parallel rows. Another set of rows having a different spacing length may cross the first set through the center at right angles, as in the orthorhombic forms, or they may pass obliquely at one side of the center, as we find with the oblique systems. In addition to these rows of points, which are the intersec- tions of the normals with the projection plane, there are planes whose normals do not pierce this plane, but lie parallel to it. These may be called the prismatic planes. These latter have their part in the growth and shaping of the crystal, and they J. M. Blake— Plotting Crystal Zones on the Sphere. 239 conform to the same laws. When we deviate from this best view-point, the rows converge to some point in the distance. The planes of a fully developed erystal may be regarded as forming a mutually dependent whole. We may find certain planes undeveloped on one crystal which may be present on another crystal identical in composition; and also, a crystal may present a perfect microscopic edge where there should be a truncation by a frequently occurring plane. We will assume that each species has a limited number of planes, all of which may be developed under favoring conditions. By growing a polished sphere of a salt to a proper degree, reflections may be obtained from its surface for all developable © planes. Such a sphere, when grown, has distinctive markings brought out on its surface. With the same end in view, a corner, or an edge of a crystal, can be rounded and then grown. Fig. 2. Ss | 4 Cc \ \ (aed c 2 © © OGD oO * / Ke © oe ~ A ° (0) cece 2 ran 0) ° ae Go 0 O Oo oO —O am e O-— 12) S ROB SR o 7 ea = fo) fe A S Or ; Hog e 1 2 3 In the latter case, we have some of the original planes left to aid in locating any less frequently occurring planes that may be developed by the treatment. The system of measuring crystals by complete zones was originally suggested by the serious need of such a plan for use in the ready plotting of the planes on the sphere. The expedi- ent of attaching the crystal by a short piece of lead wire, while measuring, was proposed in 1866 by the writer. This wire support will allow of making a complete zone reading for each adjustment. The Gnomonic Projection.— We would here draw attention to the value of the gnomonic projection of crystal planes, as an important first step in the study of crystals. An objection has been made that some of the tangents are liable to become too seriously extended, and hence, a very general resort has been Am. Jour. Scir.—Fourts Sertss, Vou, XLIII, No, 255.—Marcg, 1917, 240 JS. M. Blake—Plotting Crystal Zones on the Sphere. made to the stereographic projection which involves the use of the tangent of the half angle. By such a substitution we lose some important advantages which can follow the use of the gnomonic projection. The objection alluded to can be over- come in the case of the orthorhombic and oblique species by taking care in selecting the proper projecting point. There is a certain position on a crystal which gives the most compact projection of the planes. This position will furnish the least complicated set of indices. A projection made at right angles to this, will probably show the planes less compact but still within reasonable bounds. A third position at right angles to the other two is apt to give a very widespread pro- jection of the planes. This third position is to be avoided as a rule. If the drawing were made from this third view-point some of the planes would be seriously foreshortened. Fig. 3 gives three projections of humite, type I, an ortho- rhombie species. It shows a remarkable segregation of planes in projection 1 at right angles to the ¢ axis. The projection shown in 2 is also made at right angles to the ¢ axis and is also at right angles to the 1 position. The projection on the ¢ plane given in 3 is made at right angles to the 1 and 2 posi- tions. It is, on the contrary, very diffuse, and for this reason the drawing 3 has been made to include only one quadrant. ° The radiating short lines in 1 and 2 represent the normals of the prism planes according as we adopt one or the other posi- tion. The humite series of minerals are unique in having forms bounded by a multiplicity of planes. Oblique crystals may give several projections in which the planes are not widely scattered. Fig. 4 represents a triclinic erystal of albite. It is selected to show an inclined crystal in which the projection can be made from several view-points without wide-spreading the plane positions. The upper cut represents a crystal laid flat on the 6 plane and the numbered lines pointing towards its center give the direction of each pro- jection. The middle figures show the crystal in several posi- tions while being rotated on the 6 axis. The favored positions are selected with reference to giving an edge-on position to certain planes. The lower figure shows the several projections corresponding to the different positions. Radial lines at the margins indicate what are temporarily the prismatic zones. The triclinic character of the crystal is shown in measurable quantity on the original drawings which were made to a two inch radius, but is not plainly shown on the greatly reduced scale. Plotting the Planes on the Sphere._-The planes of a single selected zone are first plotted on a great circle drawn on the sphere with a pencil guided by the equatorial ring. If we have a crystal with two rectangular axes, a second great circle can be drawn at once at right angles to the first circle at the J. M. Blake—Plotting Crystal Zones on the Sphere. 241 intersection of the two zones. Preliminary zone plots on paper are useful as a guide in this work. The cross zones can be added in turn, and we eventually secure all the planes by this Fie. 3. y a! ae BY ts “ asi 4 , 1 ? ; ‘ 5 |. f i ’ br imanee a I 2 3 process. In ease of a triclinic crystal, it may be required to triangulate with the dividers to attain a first approximation in locating the second zone. The balance of the cross zones can then take their proper places. The next step is to obtain a record of the position of all the planes that occur in one hemisphere. The reading of the equatorial circle will be from 0 to 360°. The reading of the vertical leg of the triangle will be from 0 at the point of con- tact of the tangent plane at the top of the sphere down to 90° where the triangle rests on the equatorial ring. We shall require a record of the horizontal or equatorial reading, and of the corresponding vertical angle. It will 242 J. M. Blake—Plotting Crystal Zones on the Sphere. facilitate this plotting if we prepare a suitable drawing board. This need not exceed twenty inches square. A hard-wood peg is inserted in its center and is cut off flush with the surface of the board. From this center a circle is drawn, and is divided into degrees. The graduation marks should extend sufficiently outward so as not to be covered by the square of the drawing paper. There should be two sets of numbers in reverse order both on the ring and on the board. The graduation marks on the board correspond to the divisions on the ring of the sphere. We next mark on a paper strip a series of natural tangents up to eighty degrees. This scale of tangents can be one eal- culated for a two-inch radius. A shorter radius is sometimes used, but requires more exactness in drawing. We pin this paper strip by its zero point to the center of the drawing board. The graduated edge of the strip should be radial. A bit of paper is pasted to the strip to receive the pin through the zero point of the scale as this will bring about the radial position of the scale edge. We can now go on and plot the position of the normal in- tersection points on paper, and we thus make the gnomonic projection. The equatorial zone will appear only as radial marks. The radius of the tangent scale can be shown by a circle drawn about the center. The normal lines used in projecting start from the center of the sphere. When we use the projection for our purposes we look down upon it from the outside as we would upon a crystal. If we also project the back planes, these should appear as though we looked through the crystal and should be in reverse. When we have made our tangent plane projection from the proper view-point, the indices of the planes can be obtained by inspection without calculating them. It will be gathered from these articles that there has been a constant aim throughout to penetrate an almost untrodden field, which we know instinctively must lie just beyond our reach. It has also been the aim to improve the means for securing as reliable data as possible to be used in making further advances in the desired direction. The present article treats of .a means of making the gnomonic projection, and gives a few illustrations of its application to crystal work. There remain at least two other methods to be described. These latter are needed to place the whole series on a complete and practical working basis. We should then be able to describe a crystal with little dependence upon algebraic methods, and the results of our work should pave the way for really effective mathematical treatment in the future. New Haven, Conn., Jan. 1917. Richardson— Note on the Age of the Scranton Coal. 248 Arr. XXII.—Wote on the Age of the Scranton Coal, Denver Basin, Colorado ;* by G. B. Ricwarpson. In the course of a reconnaissance in the Denver Basin, Colo- rado, in 1910-11, I traced, over a considerable area on the plains east of Castle Rock, a zone of coal beds carrying fossil leaves which Knowlton considers to be of post-Laramie age. On stratigraphic evidence I correlated this coal zone with the Scranton coals, which a number of years ago were mined in a shaft 20 miles east of Denver, and which, ever since the publi- cation of the Monograph on the Denver Basin,t+ in 1896, have been referred to the Laramie formation. This discrepancy in age assignment prompted the desire that more field work be done before publication of the results, but the opportunity not having arisen it seems desirable, considering that the Denver Basin is the type area of the much-discussed Laramie forma- tion, to publish the evidence now: The coal zone referred to occurs between 900 and 1200 feet above the base of the Laramie formation. It crops out over a considerable area and has been prospected in a number of places, especially in the vicinity of Calhan and Fondis. Only here and there the coal is pure. enough to warrant opening country banks, such as the Purdon mine in sec. 27, T.11S%., R. 61 W., and the Moseby mine in sec. 18, T. 13 8., R. 62 W. The coal is a low grade sub-bituminous variety. I have traced this coal zone to within 15 miles of Scranton, in sec. 16, T.3S., R. 65 W., and although actual connection with the Scranton coals by following the outcrop is impossible, because of the cover of later deposits, it is evident from the field relations that the Scranton beds are in the zone traced. The following list of fossil leaves from this coal zone were collected by the writer and C. W. Cooke in 1910 at the locali- ties mentioned : Sec. 30, T. 98., R. 60 W.; secs. 7 and 33, T. pom Glew secs. 2,0). It §., Kh. OL W.; sec..18, T.:13.8., R. 61 W. List of fossil leaves from upper coal zone, Denver Basin ; by F. H. Knowrron. Pteris undulata Lx. Ficus spectabilis Lx. Laurus socialis Lx. Platanus Haydenii Newb. * Published by permission of the Director, U. S. Geological Survey. + Emmons, S. F., Cross, Whitman, and Eldridge, G. H., Geology of the Denver Basin in Colorado, U. S. Geol. Survey Mon. 27, pp. 373-3875, 1896. 244 Lichardson—WNote on the Age of the Scranton Coal. Platanus rhomboidalis Lx. Platanus raynoldsi1 Newb. Populus nebrascensis Newb. Vitis olriki Heer Cissus lobatus-crenata Lx. Fraxinus eocenica Lx. Sequoia longsdorfii (Brgt.) Heer Lygodium kaulfussi | Carya antiquorum Lx. Salix augusta Al. Br. Knowlton reports that these leaves are of post-Laramie (Denver) age. They indicate, therefore, a reassignment of the age of the Scranton coal and the consequent modification of the constitution and delimitation of the Laramie formation in the type area. This conclusion is corroborated by the record of a well sunk in search of oil near Sable Station on the Union Pacific Rail- road in sec, 24, T. 3 8., R. 67 W., about midway between Denver and Scranton. ‘The record was obtained by Willis T. Lee who, at my suggestion, visited the Denver Basin in 1915 and kindly turned his results over to me. Unfortunately the drillers did not keep the critical part of the record in writing but the results were confirmed independently by two men con- nected with the work. It is reported that a bed of con- glomerate 50 feet thick, thought to be the Arapahoe, was encountered in the well about 350 feet below the Scranton coal and about 700 feet above a coal zone thought to mark the lower part of the Laramie formation. Because the record is from memory it should not be given too much credence, never- theless it is of interest as supporting the evidence of the fossil leaves that the Scranton coal is of post-Laramie age. Chemistry and Physics. 245 SCTENTIFIC INTELLIGENCE. I. Cyemistry AND Puysics. 1. Reminiscences.—A privately printed pamphlet of 21 pages has recently appeared, giving a memorandum of the remarks of Dr. C. F. CHANDLER made at a dinner of the “Society of Gas Lighting ” in New York City. As these reminiscences deal to a large extent with experiences and history concerning household illumination, beginning at a period when the candle was the prin- cipal source of artificial light, an abstract of this particularly interesting part of the address is given here. Dr. Chandler spent his boyhood in New Bedford, where he became familiar with the whale-oil industry. At that time three or four hundred whale-ships were sent out from that place on voyages of from 23 to 3% years to all parts of the world. The oil of the sperm whale was used extensively for illumination at, that time and sold as high as $1.75 per gallon. The oil from the “right ” whale, the one yielding “ whalebone,” was much cheaper, but it had a property of gumming up the lamps. Therefore, “ Camphene,” which was rectified spirits of turpentine, was intro- duced as a comparatively cheap substitute. This gave such a naturally smoky flame that it could be used only in lamps pro- vided with chimneys. In order to modify this product so that it could be used in open wick lamps, it was mixed with a suitable proportion of alcohol to make “ Burning Fluid.” Both the cam- phene and the burning fluid were so volatile as to be danger- ously inflammable, and they were the cause of many accidents, but there was no way to make these materials safe. This was the condition of artificial illumination when Professor Chandler went to Germany to study in 1854. The next year in Berlin he saw “Coal-oil” used in lamps with chimneys. This product was manufactured from Boghead coal from Scotland by distillation. Two or three years later coal-oil factories were established in America, from Portland, Maine, to Wilmington, Delaware, in which imported Boghead mineral was used. Other materials were soon found also, such as Breckenride coal, Nova Scotia Albertite and West Virginia Grahamite, for which factories were started at or near the localities where they occurred. One of these factories named its product “ Kerosene,” and this name was afterwards given a much wider application. The demand for coal oil increased with great rapidity. It was furnished for about 50 cents a gallon and gave a beautiful light. The coal-oil industry had hardly been established fully when attention was directed to ‘‘ Petroleum” or “ Rock Oil” which was found floating on pools of water in Western Pennsylvania and Virginia, and which had been used for some time for external application in the treatment of rheumatism. Some one in New Haven had seen a small specimen of this oil in the mineral collec- 246 Scientific Intelligence. tion of Yale College, so he became interested in it, went to Oil Creek in Pennsylvania and brought back a sample of the oil. This was examined by Professor Benjamin Silliman, the younger, who pronounced it to be a superior quality of crude oil, substan- tially identical with crude coal oil, or the kerosene of that time. The New Haven man, whose name is not given, organized the first Petroleum Company, and the collection of this oil for the pur- pose of illumination was started. Col. Drake, the superintendent of this company, was the man who first drilled a well for petro- leum and he thus found oil in 1859. Dr. Chandler gives an account of his important work in exam- ining samples of dangerous kerosene which had been adulterated with the much cheaper naphtha. This work, and his publication of the results, brought about a proper inspection of the product, so that accidents from it became rare. He mentions the fact that as a boy he saw the first gas works built at New Bedford, but he states that illuminating gas had been in use in some of the larger cities for 25 years or more. H. L. W. 2. The Removal of Barium from Brines Used in the Manu- facture of Salt.—W. W. SKkInNER and W. F. Baucumaw have succeeded in removing all but traces of the barium in the brine used at a large salt works by the simple expedient of adding a proper amount of sodium sulphate (30 per cent in excess of the theoretical amount) in the form of “salt cake” to the brine before evaporation. It was necessary to add a little lime also to neutralize the free acid in the salt cake. The occurrence of barium in the brines of the Ohio Valley District of West Virginia and Ohio has long been known. These brines are peculiar in being free from carbonates and sulphates. An analysis of one of them in parts per thousand is as follows: KCl = 0°61, NaCl =: 63:95, CaCl, = 15:95, SrCl, = 0:21, BaGin— 0°79, MgCl, = 5:28, Fe(HCO,), (?) = 0°08. The lower grades of salt produced from these brines have been found .to contain as much as 5 per cent and sometimes 10, 16 and 18 per cent of barium chloride, and the poisoning of animals appears to have happened frequently by the use of such products. After the application of the new process the barium in all the grades of salt is so low that there is no danger in its use.—J. Indust. and Eng. Chem., ix, 18. H. L. W. 3. Ammonium Chloride as a Food for Yeast.—It is stated by Cuarues H. Horrman that when this salt is added at the rate of 5 lb. to 1000 lbs. of flour in mixing dough for bread there is a saving of 30 per cent in the quantity of yeast required for baking. When a little calcium sulphate and potassium bromate are used in addition to the ammonium chloride a saving of 50 per cent in yeast results. Upon investigating the subject carefully Dr. Hoffman has found that the ammonium chloride disappears entirely during the fermentation of the dough and that it is utilized by the yeast in forming new cells and is evidently converted by the yeast into Chemistry and Physics. 24-7 albuminous matter. The saving in yeast is evidently due to the increased growth under the influence of a suitable food.—Jour. Ind. and Eng. Chem., 1x, 148. H.'L. W. 4. General Chemistry ; by Hamitron P. Capy. 12mo,. pp. 522. New York, 1916 (McGraw-Hill Book Company, Inc.).— This is one of the ‘ International Chemical Series.” It is based upon the autbor’s earlier “Inorganic Chemistry,” being ‘‘ some- thing of an abridgement and much of a simplification ” of the latter. It gives a very satisfactory presentation of chemical facts with a suitable amount of theory introduced in such a manner as to explain the facts as they are brought forward. Comparatively little attention is paid to the details of chemical experiments, so that it is to be presumed that the book is expected to be used in connection with another book, a guide to laboratory work. | H. L. W. 5. Generalized Relativity and Gravitation Theory.—In a number of papers published in several scientific journals during the past few years Einstein has developed a generalization of the original principle of relativity and has deduced certain results which seem to be of great importance. Since the mathematical analysis pertains to four-dimensional “space” and is extremely complex it will only be possible to give in this place a slight sug- gestion of these very original investigations. The fundamental assumption consists in the ‘hypothesis of equivalence.” Some idea of this conception may be obtained from the following considerations. Suppose an observer at rest in @ gravitational field notices that all bodies fall with the same, constant acceleration. According to the non-generalized theory this state of affairs will be equivalent to the case where there is no gravitational field with the system of bodies at rest but with the observer moving in such a manner as to cause the bodies to appear to him to possess the same acceleration as they seemed to have when accelerated in the gravitational field. Because of the changes in the units of time and length in the two cases, the acceleration of the observer relative to the bodies in the second set of circumstances will not necessarily be equal numerically to the acceleration of the bodies in the first or gravitational field. Hinstein’s equivalence hypothesis consists in assuming that the two cases are not only equivalent for purely mechanical phe- nomena but that they are absolutely equivalent in all respects,— electrodynamically, ete. Another noteworthy feature of Kin- stein’s theory is that it involves only the Gaussian constant and the velocity of light in the free ether. No other quantities having physical dimensions are introduced. Attention may now be directed to certain practical deductions from the general theory. In the first place, it can be shown that a ray of light will be deflected by a large mass just as if the luminiferous vibrations were endowed with ordinary momentum. Accordingly, when light from a star passes close to the sun and then proceeds in such a manner as to be aimed at the center of 248 Scientifie Intelligence. the earth at the instant when the light strikes the observer’s eye, the star will appear to be displaced from its true position on the celestial sphere for the reason that in passing through the sun’s gravitational field the ray will be bent out of the line in which it had previously been moving. The theoretical angular displace- ment only amounts to 0”’°83 which is probably. too small to be observed. A second special case relates to the solar spectrum. As a con- sequence of the fact that the rates at which synchronous clocks will run depends upon the potential of the gravitational field in which they are immersed, it follows that solar spectral lines should appear to an observer on the earth to be displaced toward longer wave-lengths. The magnitude of the shift is given by the approximate equation (v, — v)/v, = 2 X 10~° where y, denotes the frequency of the light vibrations on leaving the sun, and v stands for the frequency on reaching the earth. Displacements of this magnitude and direction have been observed by L. E. Jewell and others, but, unfortunately for the present, theory, it has not been possible to eliminate other causes which might have produced the effects observed ; increase of pressure, for example. The only quantity of appreciable size for which the theory seems to account perfectly is the secular motion of Mercury’s perihelion. If this be true, and not a coincidence, then Kinstein has overcome a difficulty which has completely baffled all astrono- mers. A certain angle associated with the orbit of Mercury has the values +118"°00 and +118"°58 according to observation and to Ejinstein’s theory, respectively. The agreement is perfectly satisfactory since the probable error of the experimental datum is +0°:43. On the other hand, according to Newcomb, the differ- ence between observation and theory amounts to as much as +8"48 (instead of —0":58 + 0”°43) when the computations are based on the laws of Newtonian mechanics. H. S. U. 6. Ball Lightning.—The absolute reliability of the following interesting observations is vouched for by M. KE. Marntas in a short article on the subject. On the fifteenth of last April, at about 18" 20™, a flash of lightning burned out the telegraphic apparatus associated with the observatory on the Puy de Dome. One observer, who happened to be looking toward the south, noticed that the discharge first assumed the form of a ball of fire with a soft, delicate outline, then became spheroidal with the greatest diameter horizontal, and finally exploded, throwing out tongues of fire in all directions. At 18" 30™ and 18" 50™ two similar electrical discharges, which behaved in precisely the same manner as the first, were seen by all of the people at the signal station. The balls of fire made their appearance suddenly in the fog and at the same point in the air. ‘They seemed to be very close, in fact, at the level of the second story of the dwelling house near the ruins of the temple of Mercury. The balls appeared to remain sensibly stationary for two or three seconds before bursting. Their color was primarily white with a slight Geology. 249 tint of mauve. ‘The explosions sounded like the sharp crack of a whip and the associated luminosity was sufficiently intense and extended to light up the interior of the observatory. The appar- ent diameter was a little less than that of the moon. No other strokes of lightning occurred during the entire day.— Annales de Phys., vol. v, pp. 365, 366 ; May-June, 1916. H. 8. U. 7. A Laboratory Course of Practical Electricity ; by Mav- RICE J. ARCHBOLD. Pp. ix, 222. New York, 1916 (The Mac- millan Co.).—This manual may be used in connection with any text-book and the experimental work suggested is supposed to be supplemented by lectures and demonstrations. Each page is essentially a skeleton of a laboratory report so arranged that the student has to supply the missing words in sentences, enter numerical data in ruled blank tables, and draw graphs on cross- ruled areas. Wiring diagrams for all of the experiments are given in the appendix. The number of exercises for each of the first, second, third, and fourth semesters is 27, 25, 24 and 22, respec- tively. To make the work practical, as many of the tests as possible are given with commercial apparatus. The experiments are carefully graded from very elementary qualitative manipula- tion of magnets to more advanced work, such as : “ Methods of Connecting Polyphase Transformers,” etc. The book is bound in the loose-leaf manner so that the instructor may change the order of the experiments. EROS We 8. The National Physical Laboratory. Report for the Year 1915-16. Pp. 80. Teddington, 1916 (W. F. Parrotr).—‘ The work of the Laboratory during the past year has been greatly affected by circumstances arising out of the war. A considerable part of the ordinary research work has been in abeyance ; in its place a large number of special investigations have been under- taken for Government Departments, and the testing work for Gov- ernment Departments has also very greatly increased. There has been a reduction in other test work, and the Laboratory has thus been almost entirely employed with Government work.” For these reasons the report is of much Jess general scientific interest than usual. In the appendix is given a description of the new aeronautics buildings which were constructed in a very short time. The text is followed by large diagrams of the new struc- tures in plan and elevation, together with photographs of the ex- teriors both of the buildings and of the wind-channels. HS: 0. II. Gronocy. 1. Investigations of Gravity and Tsostasy ; by Witiiam Bowtg, Chief of Division of Geodesy, U. 8. Coast and Geodetic Survey. Special Publication No. 40, U. 8. Coast and Geod. Surv. Pp. 196, page figs. 1-9, in pocket 10-18. Washington, 1917.— This publication extends and elaborates the lines of investigation begun in previous publications. In 1912 Bowie published the results of observations at 124 gravity stations in the United 259 Scientific Intelligence. States. The present volume raises this number to 219 and adds 42 stations for southern Canada. The number of stations is thus a little more than doubled. The principal facts for 73 stations in India and 40 other stations are added, all reduced to the new method of determining the gravity anomalies. This large amount of data is subjected to analysis and several hypotheses of the nature of isostasy are tested by it. As is familiar to those who have followed the previous publi- cations, the hypothesis has been introduced that every topo- graphic feature is balanced by a corresponding variation in density in a column reaching to a uniform level,—the level of complete isostatic compensation. Thus every unit column con- tains the same mass. Furthermore, the compensation is assumed to be uniformly distributed down to the level of complete com- pensation, but there to abruptly terminate. The data of geodesy shows that such a hypothesis of perfect, uniformly distributed, and local isostasy 1s much nearer the truth than a hypothesis of no isostasy, but it is obvious to geologists that this hypothesis must be a very much simplified and generalized picture of the truth. The new gravity stations show, as seen in fig. 11, a greater complexity of the map of lines of equal anomaly. A number of new localities of high anomaly have been discovered, and, in general, where the stations are most abundant the anomaly gradients are steepest. The question of local versus regional isostasy is important. The hypothesis of local isostasy satisfies the geodetic data about as well as that of regional isostasy up to some radial distance between 59 and 167 kilometers ; but the geologic evidence favors regional isostasy to as great a limit as is compatible with the geodetic data and would favor a wide regional distribution for a part of it, a narrower regional distribution for another part. The problem of the depth of compensation is also tried out by a number of solutions, with the result that Bowie states the best mean for the depth of the level of complete compensation under the hypothesis of uniform distribution to be 96 kilometers. He believes that future values derived from much more extensive data will fall between 80 and 130 kilometers. ese 2. The Prodromus of Nicolaus Steno’s Dissertation Concerning a Solid Body Enclosed by Process of Nature Within a Solid. An English version with an introduction and explanatory notes by Joun Garretr WinTER, University of Michigan ; with a foreword by Wittram H. Hoxsps, University of Michigan. Pp. v, 124, pls. vii. Macmillan, 1916. This is Part 7, Vol. XI. Humanistic Series, University of Michigan Studies, and is one of the “contributions to the history of Science.”—Nicolaus Steno was born in Copenhagen, Jan. 10, 1638 and belongs thus to the generation which began to lay the foundations of modern science. In 1665 he was in Florence and in 1669 the Prodromus was published. It was to be the introduction to a larger work which was however never written. The work is remarkable in its Geology. 251 record of geological observations and the sound-induction from them to principles concerning the nature of strata, fossils, crystals, and mountains. Where one solid object is contained within another, as a fossil shell, or a crystal, within a rock; how did the internal object come to exist and what was the nature of the surroundings at the time of origin? Thus the title covers the field of geology. The work antedates the real rise of geology by more than a century and is of importance to those interested in the history of science: if the time had been ripe it might well have led on to a fruitful germination. JU Bi 8. Atlantic Slope Arcas; by PEart G. SHELDON. Paleon- tographica Americana, Vol. I, No. 1, 101 pp., 16 pls. Harris Company, Ithaca, 1916.—All interested in the Tertiary Paleon- tology of North America have welcomed the useful and well illus- trated ‘“ Bulletins of American Paleontology ” which have been issued from time to time under the auspices of Professor G. D. Harris of Cornell University. The time has now come when pub- lication in quarto form of more monographic papers may advan- tageously be undertaken. The editor of the Bulletins now offers the first installment of such a series under the name of “ Palzeonto- graphica Americana, Vol. I, No.1.” This comprises a paper of 100 pages, with 16 plates quarto, on the Tertiary and recent species of the genus Arca from the Atlantic slope, with references to Cretace- ous species and to others from the Antillean area, by Pearl G. Sheldon. This includes descriptions of five new forms. The excellent quality of the plates deserves special mention, consider- ing the poor quality of many recent paleontological plates, especially of the European Tertiaries. It is to be hoped that the enterprise of Professor Harris will be so well sustained by those interested in this subject that the series may be continued indefinitely, as it is certain that the riches of our Tertiary horizons will afford material for many years to come. W. H. DALL. 4. Iowa Geological Survey, Annual Report, 1914, with accompanying papers. GEORGE F. Kay, State Geologist. Pp. XXlll, 627, with 72 plates and 53 text figures, 1916.—Besides the administrative report and one on the mineral production of the state by the director of the Survey, this volume contains: (1) _ The iron deposits near Waukon, by J. V. Howell ; (2) Pleistocene history of Iowa River valley, by M. M. Leighton ; (3) Trilobites from the Maquoketa beds of Fayette County, by A. W. Slocum ; (4) The origin of dolomites, by F. M. Van Tuyl (a summation of this paper appeared in this Journal for September 1916) ; (5) Physical features and geologic history of Des Moines valley, by J. H. Lees. ens 5. New York State Museum, Twelfth Report of the Director. Bull. 187, 1916, pp. 192.—In this report are set forth the activi- ties of Director Joun M. Ciarke of the Museum and the State Geological Survey, of the State Botanist, Entomologist and Archeologist, and of the Zoology Division. Great progress has 252 _ Serentifie Intellagence. been made in the exhibition rooms of the new Education Build- ing, where striking installations may be seen of Iroquois wam- pums (here reproduced in two plates), seven life-sized groups representing the life of the Iroquois Indians (seven plates), and a testimonial case to former State Botanist Charles H. Peck, showing models of edible and- poisonous fungi (four plates.) Three plates illustrate the report of progress in Paleontology, and four that in Zoology. The scientific papers accompanying the report are: (1) Landslides in unconsolidated sediments; and (2) Albany molding sand, both by D. H. Newlands; (3) On the genus Urasterella, by G. H. Hudson; and (4) Ancient water levels of the Crown Point embayment, by E. E. Barker. c. s. 6. Review of the Geology of Texas; by J. A. Uppen, C. L. BakeER, and Emit Boss. Bull. of the Univ. of Texas, 44, 1916, pp. Xi, 164, 1 pl., 10 figs.—Only 10,612 square miles out of 262,398 square miles constituting the State of Texas, are covered by detailed geologic maps. Reconnaissance maps are available for 114,066 square miles and exploratory maps for 53,7385. For 87,367 square miles, or one-third of the State, no maps have been published. Likewise the detailed and reconnaissance reports on widely scattered areas, issued by various organizations, have not given a satisfactory view of the geology of the State and have been insufficient for use in comparative studies. The Review of the Geology of Texas, including a geologic map of the State, is, therefore, welcome. Under the heads Physiography, Geology, Geologic History, and Kconomic Mineral Products existing knowl- edge is classified, summarized, and briefly discussed. The student of Texas geology is given his bearings. Those who have had occasion to construct State geologic maps and to prepare compre- hensive but condensed accounts of the geology of a large and diversified area will appreciate the time and thought involved in evaluating and correlating the work of various authors and in filling gaps in the record. The Director of the Texas Bureau of Economic Geology and Technology is to be congratulated on the successful completion of a most useful piece of work. 4. ©. G. 7. Annual Proyress Report of the Geological Survey of Western Australia for the Year 1915; by A. Gtsp Marrianp. Geol. Survey Western Australia, 1916, pp. 44, 1 map.— Western Australia is the Nevada of the Southern Hemisphere. It owes its growth and development to its mines and receives its revenue chiefly from this source. It is but natural, therefore, that its Geological Survey should be almost wholly occupied with eco- nomic problems. During 1915 investigations were continued in the Murchison, Coolgardie, Yilgarn, and other gold fields ; the building stones of the State were classified and described ; lime- stones and foraminiferal sands were studied ; possible sources of petroleum were examined and magnesite deposits were mapped. H. E. G. 8. Economic Geology ; by Huinricu Ries, A.M., Ph.D. Fourth Edition. Rewritten, xx+856 pages, 291 figures, 75 plates. New York, 1916 (John Wiley & Sons. Price $4.00). Miscellaneous Intelligence. 253 —Prof. Ries’ new edition of his well-known text book, like its predecessors, 1s divided into two separate parts, “ Non-metallics ” and “ Ore Deposits.” Greater prominence is given to the former as it occupies the first part of the volume and 429 pages are devoted to it, while 329 pages are assigned to ore deposits. The book is much larger than the former editions and embraces a greater number of Canadian occurrences so that it is now essen- tially a text book of Economic Geology of the United States and Canada. The illustrations are profuse, and considerable dis- crimination has been used in their choice and arrangement. The text matter is supplemented by full statistical tables and by a complete and up to date bibliography at the end of each chapter. The volume is essentially a compilation, but the numerous references, the complete bibliography and the illustrations taken from widely scattered sources attest to the far-reaching nature of the compilation. Personal experience of many occurrences, shown by many of the author’s own photographs from scattered locali- ties, fills in details ordinarily lacking in most compilations and the part dealing with clays is a contribution. Part I is the most able and comprehensive treatment of non- metallics that has yet appeared and it will serve as the standard text on that subject. In Part II the author has added much new material and made many changes in the mode of treatment and classification of the subjects. The illustrations, tables and bibliography are commendable and the volume indicates discrimination and judgment on the part of the author in the selection of the subject matter. The book is the only one of its kind that covers the whole field of geology and will be of great value to teacher and student. On the other hand an introductory chapter, connecting together the two entirely separate parts of the book and pointing out their differences and relative importance, would enhance its value. Also, it is to be regretted that so much of the book is given over to descriptive material. This is of unquestionable value, but the volume might serve a wider field as a text if more space were devoted to the principles and theory of Economic Geology. This applies particularly to Part Il. While the treatment of Part II leaves much to be desired the book as a whole is a valuable addition to the literature of economic geology and is bound to have a wide use as a text book in the United States and Canada. | A. M. BATEMAN. Ill. MiscennAnrovus Screntiric INTELLIGENCE. 1. Annual Report of the Superintendent, United States Coast and Geodetic Survey, K. Luster Jones, to the Secretary of Com- merce, W.C. REepFiE nD, for the fiscal year ending June 30, 1916. Pp. 1643; 53 illustrations. Washington, 1916.—This annual report recently issued gives the usual detailed statement (Part III), illustrated by many charts, of the field and office work accomplished by the Survey during the past year. Facts in 254 _ . Serentifie Lntelligence. regard to the organization of the Bureau, and its immediate needs, in order that the work should go on efficiently, are discussed in Parts I and II. It is stated that important changes in the organ- ization became effective in October, 1915, and further changes were made in July, 1916 ; gratifying results as regards output have already been noted. “The needs of the Bureau, however, are many and important. It is now inadequately housed, in part in a building erected as a dwelling (the former home of Gen. Ben- jamin Butler) and in part in another structure originally built for a hotel. Considering the many lines of work carried on, not only administrative, but including a printing plant, a lithographic plaut, a machine shop, etc., the conditions for effective work as well as for the safety of important documents are most unfavor- able. The situation now existing is presented to the public in a telling fashion by a series of photographs, showing how inade- quately and inconveniently the charts and other materials are housed and with how great loss of economy the work must go on. Hardly less important than a new building are the needs of the Survey for expanding and perfecting its hydrographic and geodetic work. This subject is alluded to at some length in the notice of February, 1915, (see pp. 225-227). An important memoir recently issued by The Survey is noticed on an earlier page (p. 249). 2. The Fundamentals of Psychology ; by W. B. Pirisspury Pp. ix, 562; 92 figs. New York, 1916 (The Macmillan Com- ' pany).—This is a text-book, designed for relatively mature col- lege classes, covering the range of topics usually deemed essential in a general introduction to psychology. Among such topics may be mentioned sensation, perception, attention, memory, reason, feeling and emotion, and voluntary activity. While not pro- found, the book will serve to put even a general reader in touch with the trend and many of the results of scientific psychological sage age ee ROSWELL P. ANGIER. 3. Mechanisms of Character Formation: an Introduction to Psychoanalysis ; by Wittiam A. WuitE. Pp. 342. New York, 1916 (The Macmillan Company). —An introduction to what is commonly termed ‘Freudian’ psychology,—Freud, the Viennese psychiatrist, being the first to employ its unique method * psycho- analysis.’ This psychology aims to be “ humanistic” ; avoiding metaphysics and physiology alike, it formulates certain broad principles underlying behavior necessary to real appreciation of human beings, “especially as the priest and the physician knows them.” Abnormal mental phenomena, mythology, dreams, etc. are the chief sources of material. The author presents uncriti- cally much which requires further substantiation. RICHARD M. ELLIOTT. Warns Natura Science EstaBiisHMent -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. J-148. Price List of Rocks. Mineralogy: J-109. Blowpipe Collections. J-74. Meteor- ites, J-150. Collections. J-160. Fine specimens. 2 Paleontology: J-134. Complete Trilobites. J-115. Collee- 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-188. Bird Skins. J-139. Mammal Skins. - Human Anatomy: J-16. Skeletons and Models General: J-155. List of Catalogues and Circulars. Wards Natural Science Establishment 84-102 College Ave., Rochester, N. Y., U.S. A. Publishers: WILLIAMS & NORGATE, 14 Henrietta Street, Covent Garden, London, W. C. “SCIENTIA” INTERNATIONAL REVIEW OF SCIENTIFIC SYNTHESIS. Jsswed monthly (each “number consisting of 100 to 120 pages). Editor: EUGENIO RIGNANO. “SCIENTIA”’ continues to realise its program of synthesis. It publishes articles which relate to the various branches of theoretic research, and are all of general in- terest; it thus enables its readers to keep themselves informed of the general course of the contemporary scientific movement. **SCIENTIA’’ appeals to the codperation of the most eminent scientific men of all countries. It has published articles by Messrs. Abbot (Washington) - Arrhenius (Stockholm)- Ashley (Birmingham) - Bechterew (Petrograd) - Bohlin (Stockholm)- Bonnesen ( Kopenhagen) - Borel (Paris) —-Bottazzi (Napoli) - Bragg (Leeds) -Brile - louin (Paris) - Bruni (Padova)-Castelnuovo (Roma) - Caullery (Paris) -Chamberlin (Chicago)—Ciamician (Bologna) - Clark (New York) - Costantin (Paris)-Crommelin (Greenwich) — Daly (Cambridge, U. S. A.) - Darwin (Cambridge) - Delage (Paris) - De Martonne (Paris) - De Vries (Amsterdam) - Durkheim (Paris) - Eddington (Greenwich) - Edgeworth (Oxford) - Emery (Bologna) - Enriques (Bologna) - Fabry (Marseille)-Fisher (New-Haven, U. 8S. A.)- Foa (Torino)-Fowler (ZLondon)- Fredericq (Liége)-Galeotti (Vapoli)- Golgi (Pavia)-Gregory (Glasgow)- Guignebert (Paris) - Janet (Paris) - Jespersen (Gentofte) - Kapteyn (Groningen) - Kidd (Oxford)-Langevin (Paris)-Lebedew (Moscou)-Lodge (Birmingham)-Loisy (Paris)- Lorentz (Haarlem) -Loria (Torino)-Lewell (Flagstaff, U. S. A.)-Maunder (Green- wich) - Meillet (Paris) - Pareto (Lausanne) - Peano (Torino) - Picard (Paris) - Poincare (Paris) - Puiseux (Paris) - Rabaud (Paris) - Righi (Bologna) - Rignano (Milano)—Russell (Cambridge)- Rutherford (Manchester)- Sayce (Oxford) -Schiapa= relli (JZilano) - Seligman (New York) - Sherrington (Liverpool) - Soddy (Glasgow)- Svedberg (Upsala) -— Tannery (Paris)- Turner (Ozford) - Vinogradoff (JZoscou) - Vol= terra (Roma)- Von Zeipel (Upsala) -Westermarck (Helsingjfors)-Willey (Montreal, Canada) -Zeeman (Amsterdam) - Zeuthen (Kopenhagen), and more than a hundred others. ‘“SCIENTIA’’ publishes, at present, in the section dedicated to sociological articles, a series of studies on the present questions of an international character raised by the war. ** SCIENTIA’’ publishes its articles in the language of its authors, and joins to the principal text a supplement containing the French translations of all the articles that are notin French. (Write fora specimen number.) Annual Subscription: 24 sh. post free. Office: Via Aurelio Saffi, 11- MILAN (Italy). CONTENTS Art. XXIII.—Lava Flow from Mauna Loa, 1916 ; by T. A, J AGGAR, DPI. lS oS ee es eet XXIV.—The Formation of Salt Crystals from a Hot Satu- rated Solution; by E. Tatum Lone _2___._- 22-222 XX V.—An Upper Cretaceous Fulgur; by B. WapE..--.-- XXVI.—A Middle Eocene Member of the “Sea Drift” ; by He VWWio Sey ie see anor : -2e se - eee wee ee ee ee ee er ee eK Ht Page 293 XX VII-—Correlation of the Mississippian of Ohio and Penn-_ sylvania;: by W. A. VeRWIBERE. 2.2 2232-5 oe XXVIIL—A New Labyrinthodont from the Triassic of Pennsylvania; by W. J. SINCLAUS: 5. - 5) esse XXIX.—On the Calcium aca in Meteoric Stones; by Go PMERRILL. 22 ook So ee XXX. ee of Prac by E. V. SHannon...-- SCLENTIFIC INTELLIGENCE. 319 322 325 Chemistry and Physics—The Penfield Test for Carbon, W. G. MixtTER and F,. L. Haicu, 327.—New Volumetric Method for the Determination of Cobalt, W. D. EneLte and R. G. Gustavson, 328.—Determination of Molybdenum by Potassium Iodate, G. S. Jamirson: Fixation of Nitrogen, J. E. BucHEer: Volatilization of Potash from Cement Materials, EK, ANDERSON and R. J. NESTELL, 329.—Dispersion and the Size of Molecules of Hydro- gen, Oxygen, and Nitrogen, L. SInBerstern, 330.—Lubrication of Resist- ance-Box Plugs, J. J. MANLEY, 331.—Unipolar Induction, E. H. KeEnnarp: Model Drawing, C. O. Wricut and W. A. Rupp, 382. —Teaching of Arith- metic, P. KLAPPER, 333. Geology and Mineralogy—American Fossil Cycads, Volume II, Taxonomy, G. R. WIELAND, 833.—New Zealand Geological Survey, P. G. Morean, 330.—The Gold Belt South of Southern Cross, T. BLATCHFORD, etc., 336.— Relationship of the Tetracoralla to the Hexacoralla, W. 1. Ropinson: A new genus and species of the Thecidiinae, etc., J. A. THomson: Tertiary formations of Western Washington, C. E. WEAvER: Paleontologic centri- butions from the New York State Museum, R. RUEDEMANN, 387.—Great Eruptions of Sakura-jimi in 1914, B. Koro: Synantectic Minerals and Related Phenomena, J. J. SEDERHOLM, 338.—Bibliography of Australian Mineralogy, C. ANDERSON: Rings, G. F. Kunz: Economie Geology, H. Ries: The Geological History of Australian Flowering Plants, E. C. ANDREWS, 399. Miscellaneous Scientific Intelligence—Indians of Cuzco and the Apurimac, H. B. Ferris, 3389.—Carnegie Institution of Washington, R. 5S. WooDWARD, 340.—Observatory Publications: Publications of the Museum of the Brooklyn Institute of Arts and Sciences: Tables and Other Data for Engi- neers and Business Men, C. E. Frrris, 342. Obituary—W. BEEBE: T. PurpiE: N. H. J. Mruume, 342. ea / an i # e " “ , +! . & ie =. < oe a — —- aA —s VOL. XLIII: APRIL, 1917. Established by BENJAMIN SILLIMAN in 1818. THE AM HRICAN JOURNAL OF SCIENCE. Epirorn: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW ann WM. M. DAVIS, or Camprinae, Proressors ADDISON E. VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT E. GREGORY AND HORACE S. UHLER, or New Haven, Proressor HENRY S. WILLIAMS, or Irnaca, Proressor JOSEPH S. AMES, or Battimore, “Mr. J. S. DILLER, or Wasuinerton. qnsonian Ins# (ES \ FOURTH SERIES |“ APR 3-'j9 Af om ty, VOL. XLITI—[WHOLE NUMBER GR@]I1). PS - No. 256—APRIL, 1917. NEW HAVEN, CONNECTICUT. BAO. Leyes Be THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. SE LT ST TI LT ES RIE STE ETOCS CT TI TE SEE TDI LI POAT I EE TEE EE 2 __ Published monthly. Six dollars per year, in advance. $6.40 to countries in the - Postal Union ; $6.25 to Canada. Single numbers 50 cents. 4 _ Entered as ’second- class matter at the Post Office at New Haven, Conn., under the Act of March 3, 1879. LIST OF CHOICE SPECIMENS Miargyrite, Pyrite, Zacetecas, Mexico. $15. Franckeite, Poopo, Bolivia. $2. Ludwigite, Magnetite, Dognaczka, Hungary. $8. Dyscr asite on Arsenic, Andreasberg, Harz. $2.5 Argentite, Freibergite, Freiberg, Saxony. $4. Rathite, Pyrite, on dolomite, Binn, Canton Wallis, Switzerland. $2. Rathite, Ducktown, Tennessee. $3, Calcite, stalactitic, copper-stained, Cochise Co., Arizona. 44” x 3%" x 3’, very showy specimen, green, blue and white. $8. Apatite, amethystine, Mt. Apatite, Maine. $38. to $10. Apatite, Ehrenfriedersdorf, Saxony. $1.50 to $5.00. Apatite, pink, with Fluorite, Ehrenfriedersdorf, SAXONY. $2. to $4. Vesuvianite, Templeton, Canada. $6.50. Tourmaline, green, Ramona Co., California, group of three crystals 2” in diameter each. §8. Galena, Wellington Mine, Breckenridge, Colo., specimen 4” x 3}; crystals ttof’. $4. Ouvarovite, var green Garnet, Orford Co. , Quebec, Canada. $2; to $0. Sternbergite, Freiberg, Saxony. $3. to $5. Sternbergite, Proustite, ‘Freiberg, Saxony. $0. Goethite, var. Pribramite on Calcite, Pribram, Bohemia. $1. to $4. Herderite, Poland, Maine. Group of twinned crystals in matrix. $2. to $8. Chaleosiderite, West Phoenix Mine, Cornwall, England, very fine, $2. to $6. Iimenite, Arendal, Norway. $4. Harmotome, Calcite, Strontian, Scotland. $2. to $9. Autunite, Schwarzenberg, Saxony. $1. to $3. Lollingite, Schwarzenberg, Saxony. $1.50 to $5. Greenockite, Schwarzenberg, Saxony. $1. to $1.50. Cordierite, Barule, Hungary. $2. to $3. Baddeleyite, Minas Geraes, Brazil. "$2. to $3. Chalcocite, Redruth, Cornwall, England. $1.50 to $3. Gismondite, Phillipsite, Vallerano, Campagna Romana, Italy. $1. to $3. Pharmacosiderite, Cornwall, England. $1. to $3. Mercury (native) in slate, Idria, Austria. $1. to $2. Laurionite, Laurion, Greece. $1.50 to $5. Cabrerite, Laurion, Greece. $2. to $4. Uraninite, Gummite, Joachimsthal, Bohemia. $1. to $2. Anatase, Switzerland. $1. to $8. Pucherite, Schneeberg, Saxony. $3. to $5. Wulfenite, Bleiberg, Austria. $1.50 to $3. : Wulfenite, Mies, Bohemia. $1.50 to $4. Tennantite, Cornwall, England. 50c. to $2. Steinmannite, Pribram, Bohemia. $1. to $3. ALBERT H. PETEREIT 81-83 Fulton St:, New York City TRE [FOURTH SERIES. | oe Art. X XIII.—Lava Flow from Mauna Loa, 1916 ; by T. A. JAGGAR, Jr: Iy accordance with published expectations, the outbreak of Mauna Loa voleano, Hawaii, from the flank of the mountain, relieving the congested fluids within by effervescent ejection of them through gas pressure, took place May 19-31, 1916.* The summit crater Mokuaweoweo did not appear to partake in the activity but one report from an exploring party which visited the floor of the crater in the suinmer of 1916 indicates that there may be some changes there. The account of the erup- tion here given wil! treat of the sequence of events at both Kilauea and Mauna Loa, 1915-16. Kilauea 1914-16. The accompanying diagram (fig. 1) shows the fluctuation in level of lava in Halemaumanu pit, the inner summit crater of Kilauea voleano, from the time of low level of the spring of 1914 to November 10, 1916. The black vertical lines show the height of the lava surface on the dates indicated, measured as depression below the rim of the pit in feet (left column of figures), the rim of Halemauman standing 3700 feet (1128 m.) above mean sea-level as recorded by government bench marks. The elevations above sea-level are shown in the right hand column of figures. The periods of premonitory seismic spasms, believed to originate on Mauna Loa, are shown in the upper part of the diagram, as well as the gas and lava outbreaks of * This Journal, Feb., 1915, Pp. 167: Dec., 1915, p. 621 ; , Boney Honolulu Chamber of Commerce, 1912, ‘*‘ The Cross of Hawaii,” p. 12 ; Report Hawai- ian Volcano Observatory, Jan.-Mar. , 1912, Boston, p. Pi ‘Ball Hawn. Voi. @bs'y. vole 1, -No. 4, April, 1915, p. "39 : Science Conspectus, Boston, vol. v, No. 4, 1915, p. 98. Am. Jour. Sct.—Fourts Series, Vou. XLIII, No. 256.—Aprit, 1917. 18 Sig! Aon “30 uvosuvpP’ va 000¢ ool 00¢o Sores 00ve ooss oog£ ag 0/8 NON IE 81 Ly Fnac Ne ine NAP 2 213 CZK AYW citi na ii 7 jee} SANT. Oy WRIA er IA Ju VEN WNW) eee cw a alc Be U6 #II01AON!] L530 “Ld “did 02 = os Lt} +4 o£ o0z JSaggar—Lava Flow from Mauna Loa, 1916. 257 that voleano. In each case the local earthquake spasms occurred at the beginning of the pronounced subsidence of Kilauea lava after a temporary spurt of rising. A greater earthquake spasm of September, 1915, accompanied a more pronounced rise and fall at Kilauea. The period September to November, 1914, exhibited high seismicity, which ceased when the summit crater flows of Mauna Loa poured forth freely in December.* The rising of the Kilauea column spurted synchronously with the term of the Mauna Loa outbreak. Excepting for the July-September rise of 1915, the Kilauea column remained well below the 400-foot (122 m.) mark in the pit for the year February, 1915, to February, 1916, and then it climbed rapidly and a sudden spurt in May, 1916, culminated this rising spell, apparently in coordination with the new outbreak of Mauna Loa, and its swarm of earthquakes. Thereupon the lava of Kilauea sank suddenly with cataclysmal effect, accompanied by a localized earthquake spasm, and immediately recovered, rising with increased volume and rapidity until the present time (Novem- ber, 1916), except for a sluggish month in September. A general tendency to culmination at equinox and solstice is evident throughout the chart, and this habit has proved useful in guiding expectation, as for instance at the present time, when it is reasonable to look for a turning point in the great rise about December, 1916. The writer surmises, however, that a gas pressure control, with release and recovery depen- dent on accumulation and resistance in an adjusted conduit system, is effective in inducing rhythm in volcanic mechanism, as he pointed out for Mont Pele in 1902, when, however, he supposed the gas to be water vapor.t The luni-solar stresses, superposed on the more dominant long-term trends inherent in the ever-rising gas and lava, probably act as trigger in opening and closing fissures, but the two curves respectively of summit culminations and sudden depressions at intervals of about eight months shown on this plat imply increasing maxima without respect to any obvious astronomic period. If this tendency led to a fourth crisis of like interval we might expect Halemau- mau to overflow before February, 1917,{ and thereafter to col- lapse still more profoundly than in June of 1916. But sucha * This Journal, Feb., 1915, p. 167; Dec., 1915, p. 621; and May, 1916, p. 383. { Science, Nov. 28, 1902, p. 871. See also Gilbert, ‘‘ Harthquake fore- casts.” Science, Jan. 22, 1909, p. 121. t On Feb. 1, 1917, the lava was still rising and less than 50 feet (15m.) below the rim of the pit. (Author, Feb. 2, 1917.) Fie. 1. Diagram showing fiuctuation of level of lava in Halemaumau, in relation to seismic and volcanic activities of Mauna Loa, 1914-1916. 258 Jaggar—Lava Flow from Mauna Loa, 1916. speculation is futile without more extended records for com- parison. It is evident from this chart that the lava of Kilauea was slowly rising from April, 1914, to May, 1916, that it exhibited a spurt upward and a decline synchronous with each of the outbreaks of Mauna Loa 1914-15, November—January, and 1916, May—June, and that just at the close of the latter out- break the Kilauea lava column sank suddenly to profound depths. It also appears that after this time the Kilauea lava column recovered with extraordinary rapidity, so that within five months it regained all that was lost by the subsidence and continued to rise still higher. In volume of lava this rising of 1916 amounted to an inflow of approximately 14,567,700 cubic meters from June 7 to November 16 inclusive (practically a cylindrical mass 335 meters in diameter, for the pit was left after the June subsidence with walls nearly vertical). This condition of verticality was quite different from the funnel shape of the vent as observed by the writer in 1909 and from 1912 tor 1916 ) Earthquakes 1915-16. The seismic prelude of the 1914 outburst of Mauna Loa has been deseribed.t The seismographs of the Hawaiian Voleano Observatory registered local shocks per month in 1915 as fol- lows: 1915 Jan. 22 shocks July 28 shocks Bebz—-9 Ss UG 2210) eee Marn2om oc Sept. 1385 —< APE Toe Octin2sr a. May 32 ‘“ INO so ielts ieee Jumer sO. 3° Decin2orae. It will be seen from this table that from nine to seventeen local earthquakes are commonly recorded per month at the Whitney Laboratory,t that near solstice and equinox in 1915 the number tended to increase, and that the approach, eulmina- tion and recession of the autumnal equinox, allowing for a short lag, coincided with an extraordinary swarm of local quakings which led us to watch Mauna Loa attentively. At the same time the lava of Halemaumau sank suddenly. The phase intervals of the seismograms indicated origins up to sixty miles (97 km.), and while some of the shocks were of * Report Hawaiian Volcano Observatory, Jan.—Mar.,1912. Boston. Fig- ures 13, 25, 27. + This Journal, Dec., 1915, p. 623; Bull. Seis. Soc. Amer., Mar., 1915, p. 39. } This Journal, Dec., 1915, p. 628. JSaggar— Lava Flow from Mauna Loa, 1916. 259 Fig. 2. u aah og tte ae aertk Ay {EB | i at L tiele fae a § 10 15 20 25 5101520. 25 9 10 15 20 95 5 10 15 20 25 APRIL MAY JUNE UIEY. Fig. 2. ber of local earthquakes per day registered at Hawaiian Voleano Observatory Diagram showing by height of vertical lines approximate num- from April to July, 1916. The May crisis corresponds with Mauna Loa eruption, that in June with Halemaumau subsidence. 260 JSaggar— Lava Flow from Mauna Loa, 1916. much nearer origin and felt sharply three miles (5 km.) north- east of Kilauea crater, the larger number were propagated from origins as far as the Mauna Loa rifts, This September crisis of 1915 deserves close scrutiny in view of the double Mauna Loa—Kilauea culmination, accompanied with earthquakes and intense volcanic vibration, which hap- pened eight months later. A similar seismic spasm, premoni- tory to, and two months before, the 1914 outbreak, occurred in September, 1914, when twenty-nine shocks were felt at Kapapala on the southeast flank of Mauna Loa from Septem- ber 27 to September 28 and theseismologist of the Observatory reported* that “‘ without question this forty-eight hours... . was a period more active seismically than any interval of like duration since the establishment of this station.” A similar but milder seismic spasm had occurred during the early days of July, 1914 (fig. 1). Coérdinately with each of these premoni- tory seismic events of 1914 there was a spurt of rather grad- ual rising of Kilauea lava followed by relatively sudden subsi- dence. The earthquakes, however, showed origins more distant than Halemaumau and were believed to arise on the Mauna Loa rifts. The 1915 spasm as instrumentally recorded at the Observa- tory was described as followst by H. O. Wood, seismologist : “Tn the following six days (after September 16, 1915) sixteen well-marked local earthquakes were registered. “None of these were felt close by the Observatory, but several, one in partic- ular, were felt at Hilo, Kapapala and on the Hamakua coast. “Most of these, while distinctly of local origin, were from more distant sources than the average of local shocks. Besides these there were thirteen, more or less, seismic disturbances grading in character and energy from well-marked earthyuake wave- groups to those usually considered as voleanic-vibration wave groups. All these were of very small amplitude; and there is no genetic distinction between an earthquake record and a voleanic-vibration record. The former is a group of earth- waves with a distinct indication of fore-phases. Though Mauna Loa shows no sign from the Observatory, the possi- bility that these disturbances point to renewed action there cannot be overlooked. “The week ending with September 29, 1915, has been marked by extraordinary local seismic activity—though none of the earthquakes registered have been felt at the Observa- tory or at the Voleano House. However, within three miles (5 km.) in the direction of Hilo several have been felt, partic- ularly on September 26, some quite sharply. * Bull. Hawaiian Vole. Obs’y, Nov., 1914, p. 134. + Bull. Hawaiian Vole. Obs’y, Sept., 1915, pp. 100 and 103. JSaggar—Lava Flow from Mauna Loa, 1916. 261 ‘Disturbances were registered during the twenty-four hour periods, ending in the forenoon of the dates specified, as follows : . September 23, three well-marked shocks and about nine other disturbances ; September 24, one well-marked shock and about five others ; September 25, two well-marked shocks and one other ; September 26, the day of the crisis, forty-four well- marked shocks, thirty-seven definite shocks of lesser amplitude and eighty-three other wave-groups, some of which almost surely were minor earthquakes; September 27, ten well- marked shocks and twelve others, more or less ; September 28, ten well-marked shocks and three others; and September 29, one well-marked shock and eight other disturbances. The total for the week, therefore, is 108 earthquakes about half of which were of very small amplitude and none of large ampli- tude, and 121 other definite wave-groups, many of which probably were minute earthquakes and many, doubtless, only groups of volcanic vibrations.” It will be seen from this description that the total of all local seismic disturbances instrumentally recorded at Kilauea, greater and smaller, for the fortnight ending September 29, 1915, was 258. The 138 for the month in the foregoing table refers only to positive earthquakes. The earthquakes of the nine days September 17 to 25 were especially strongly felt in the Mauna Loa axis (districts Hilo, Hamakua and Kau), while those from September 26 to 29 were sharply felt in the Kilauea axis (district Puna). The Weather Bureau reports* that eight shocks, September 17, 18, 19 and 25, 1915, were reported from Kohala, Hamakua and Hilo districts, especially severe the 19th and 25th at Ookala and Honomu. These districts lie on the north side of Mauna Kea in the line of the long axis of Mauna Loa which extended would divide the island of Hawaii into almost equal halves. In southern Hawaii no earthquakes were reported from Weather Bureau stations for September, and none whatever for the date of the instrumental crisis September 26. Thus it would appear that the actual equinox period Sep- tember 19-25, 1915, was marked by a few strongly felt earth- quakes in the Mauna Loa axis; that this was followed by a spasm of trembling partly felt on Kilauea mountain, very intense on September 26 and declining thereafter, and exhib- iting earthquake origins locally remote from Halemaumau ; and that a great crisis of subsiding lava in the Halemaumau pit took place at the average rate of sixteen feet per day dur- ing the six days September 22-28, with a sudden acceleration in the subsidence September 25-26 (fig. 1). In view of the record of a year before, September-November, 1914, and the * Climatolog. Data, Hawaii Section, Sept., 1915, Honoluiu. 262 Jaggar—Lava Flow from Mauna Loa, 1916. similar seismo-voleani¢ crisis first on Mauna Loa and then on Kilauea in May-June, 1916, it seems probable that the Sep- tember disturbances of 1915 indicated a spasm of rising lava- foam splitting its way upward in the edifice of Mauna Loa, attaining release of gas pressure in some unobserved irruption subterranean or submarine, and so temporarily impoverishing Kilauea as to induce subsidence in its lesser edifice. Even an outward eruption of fume from the Mauna Loa rifts enduring several hours might take place at any time in the year during the long cloudy spells, and pass entirely unperceived except as a seismic event. Complete observation of Mauna Loa will remain impossible until there are permanent observers housed above the cloud zone. In 1916 the earthquake frequency increased by leaps and bounds with the outbreak of Mauna Loa and during the sub- sidence cataclysm of Halemaumau. The numbers im the fol- lowing table for 1916 are only approximate as the seismograms have not yet been measured : 1916. Jan. 16 shocks before Jan. 15. June 418 shocks Feb. Rec anne recorded because ot | July. 07am Mar. storm damage to laboratory. Aug. “Issa Apr. 14 shocks. Dept. = ltommme May 454 “6 Oct: 23S That there was seismic activity codrdinate with the volcanic events cannot be questioned (fig. 2). Whether the deeper earthquake rift movements were cause or effect may well be considered a subject of controversy. There is no reasonable doubt that the epicentral tract of the earthquakes of May 19-28, 1916, lay along the southern rift zone of Mauna Loa in the district of Kau. (See map of Hawaii, this Journal, Dec. 1915, p. 622;) nor that the earthquakes of June 5-7, 1916, centered about Halemaumau. In proof of these places being the respective seismic centers may be cited the fact that the first series was felt principally at Hilea and Papa, on opposite sides of the place of lava out- flow and that the second series was felt most strongly at Kilauea. The Weather Bureau reports* that felt earthquakes on the island of Hawaii for 1916 were as follows : January 1916. 6 earthquakes in north Hawaii. In south Hawaii :— AleWilee teh, G0) 12 ae Hilea (SE. base Mauna Loa) 10, 5:20 p. m., Glenwood (NE. slope Kilauea) 21, 4:20 P. M., ’ Pahala (SE. base Mauna Ha) 21, 5:00 Pp. u., Hilea *Climatolog. Data, Hawai Section, Jan.-July, 1916, Honolulu, U. 8. WW. “Be Jaggar—Lava Flow from Mauna Loa, 1916. 263 Kebruary 1916. No earthquakes reported. March 1916. 2 earthquakes in north Hawaii. In south Hawaii :— Mar. 27, 6:00 P. m., Hilea April 1916. 4 earthquakes in north Hawaii. (1 heavy, 28th) In south Hawaii :— Apr. 28, 7:30 a. m., Glenwood 66 GG 6 Hilea SC OE G6 Kealakekua (W. base Mauna Loa) re a6 El olmalion Sia os 2 May 1916. No earthquakes reported from north Hawaii. In south Hawaii :— May 20, numerous earthquakes in Hilea. Ml 6 ce (15 3 (1 June 1916. 2 earthquakes in north Hawaii. In south Hawaii :— June 12, Hilea (severe) “ Glenwood “« Hilo 24, Hilea (severe) July 1916. 2 earthquakes in north Hawaii. In south Hawaii :— July 2, 5:25 a. m., Glenwood 11, 10:00 p. m., Hilea (severe) ss Glenwood 12, Pahala 21, 7:55 a. M., Glenwood The absence of complete records for May is covered in the Weather Bureau published data by a blanket statement of unusual voleanic activity. There are unfortunately no Weather Bureau stations on the southwest slope of Mauna Loa. The earthquakes from May 20 to 30 were generally felt in Kona aud Kau. The earthquake of June 12 was felt generally over the island of Hawaii just three weeks after the Mauna Loa outbreak as was the earthquake of April 28, three weeks before the outbreak. It will be observed that Hilea, which bore the brunt of seismic activity of the eruption, also reported 264 JSaggar—Lava Flow from Mauna Loa, 1916. X more earthquakes throughout the seven months than any other station and greater severity in the shocks felt during and after the lava flows. Mrs. Z. V. de la Nux, living at Hilea, was so impressed with earthquake swarms of Saturday and Sunday, May 20 and 21, that she noted the times of the stronger shocks in the after- noon and evening of both days as follows : May 20, 1916. May 21, 1916. WAS DENG 2:45 P. M. 1:35 P. M. (strong) SAO men. AO ego ee AT OVE Dison tee AS yamiece 2S Oe s ANON; = 00 SOE es A Ac yOu lyfe anos A296. 48 5:18 « 4:30 « HV De Sins os GAO Ses eg) 7:28 “ (strong) 5340s S200 ae BaD wee Sis0Siaec G20 aE 82210) Seo Oma 2s) 9S LAO -& 9:09 < Dios ese O59 Biase 9°32. “é One Maeda, a Japanese living in the Mohokea hills (a large ancient crater) five miles (8 km.) nearer the lava source than Hilea, felt more and stronger earthquakes than the in- habitants of Hilea. He reported thirty strong shocks in three hours on the evening of May 20. On the other hand, neither Waiohinu nor Pahala, respectively eight and six miles from Hilea, on either side of it, and both on the Mauna Loa flank, felt any such number of shocks as were reported from Hilea. Waiohinu, an equal distance from the lava source with Hilea, reported on the evening of May 20 “one or two shocks in the iast two days.” An overseer in the upper Waiohinu sugar fields, however, reported several earthquakes felt in the open on the afternoon of May 20. Pahala for May 20 reported ‘an earthquake this morning.” The middle Kona district on the west flank of Mauna Loa reported two earthquakes on the morning of the 20th. It was evident, therefore, that the posi- tion of Hilea on a line at right angles to the lava rift opposite the point of emission, was peculiarly sensitive to the jarring of JSaggar—Lava Flow from Mauna Loa, 1916. 265 the outbreak aud that the felt earthquakes did not travel far from the source at the time of actual rapid rise of lava near the surface. Cursory inspection of the seismograms of the Mauna Loa eruption period of 1916 shows May 19, morning, continuous trembling. 19-20, trembling followed by a shock. 20-21, trembling followed by several shocks. 21-22, some trembling, but increase of definite shocks. 22-23, many distinct shocks of large amplitude. 23-31, earthquakes continued, generally decreasing in amplitude, except for occasional ones of very large amplitude. One of these appears May 25-26 and another May 30-31. June 1-3, Four ordinary earthquakes. The larger amplitude shocks of the closing days of the erup- tion were felt especially strongly on the Mauna Loa slopes. The writer felt a sharp shock at Waiohinu about 1:50 a. m., May 27. Mr. Wood was on the mountain at an old cone east of the lava source on May 30, when* “at about 8:45 Pp. m., a short sharp earthquake occurred, plainly felt by all three of us sitting or reclining on the cinders in the reéntrant of the cone. This shock was felt sharply at Waiohinu and at Kapapala. At Hilea it was felt as the strongest shock of the entire series connected with this eruption. Within less than a minute after the shock there occurred a spasm of greatly increased action at the southern active lava vent, with the jetting of lumps of lava and a great increase in the flow. This quickly declined to normal.” This observation is of interest in showing the immediate kinship of the lava-gushing and the earthquakes. The above cited sequence in the seismograms shows that the earthquakes waxed from a continuous jarring to large numbers of moderate shocks in four days, and thereafter for eight days there was fairly regular decrease in numbers of shocks and increase in their occasional intensity. The approximate numbers of earth- quakes per day are shown diagrammatically in fig. 2; the column of figures on the right of diagram signifies numbers of earthquakes and larger wave groups; the length of the ver- tical line corresponds to number of such shocks counted on the seismograms for each twenty-four hour period following the morning of the date indicated. The swarm of local earthquakes which accompanied the sud- den drainage of the lava from Halemaumau, June 4 to 7, 1916, increased in number to about double the maximum registered * Bull. Hawaiian Vole. Obs’y, June, 1916, p. 54. 266 Jaggar—Lava Flow from Mauna Loa, 1916. during the May crisis in a similar period of four days from June 3 to 6 and declined more abruptly, the greatest number of shocks being recorded and also felt during ‘the twenty-four hour period following the morning of June 5. This happened simultaneously with the most rapid fall of the laval” Whe sequence is shown in fig. 2, and the following notes were made by the writer from casual inspection of the seismograms: June 1-2, two shocks. 3, two shocks. -4, eight very small shocks. 5, 69 shocks of various amplitudes. 6, about 217 shocks ; but from noon to 6 p. m. June 5 increasingly, a steady trembling followed by steady decline. 6-7, 43 definite shocks, one of large amplitude but not larger than some of June 5. , 25 definite shocks, some rather strong. , 12 very weak shocks. 0, 9 very small shocks and one very strong one. 10-11, 13 feeble shocks. 11-12, 4 shocks, one a considerable earthquake followed by continuous trembling for three hours. This second seismic spasm was described by the seismolo- gist* as comprised wholly of local shocks of feeble character felt near Kilauea crater but attracting no attention at Hilo or Pahala. None was stronger than III of the Rossi-Forel scale. They proceeded from origins nearer to the Observatory than the shocks of the May spasm. Volcanic vibration exhibited larger amplitude than usual during the week of this June crisis and the tilting of the ground as registered seismometrically changed from northeastward to southwestward about the time of the great subsidence. Explosion of May 19, 1916. The first recorded symptoms of actual outbreak on Mauna Loa in 1916 were earthquakes recorded in a swarm by the seismo- craphs during the very early morning hours of May 19, and felt, in part, as smart shocks at the southern end of the island of Hawaii. Glow and fume, also, above the mountain were reported seen before daybreak from a vessel. N othing was seen from the vicinity of the Observatory until 7 a. m., when a white cumulus rose above the profile of Mauna Loa south of the summit region and developed two pronounced jets of fume and a vortex ring. At 8 a. m. the two jets were united into one and the column was 20,000 feet (6100 m.) high above its base (ules o) sueluhe higher fume jet bore approximately S. * Bull. Hawn. Vole. Obs’y, June, 1916, p. 49. Jaggar-—Lava Flow from Mauna Loa, 1916. 267 Ine, &). Fic. 3. Fume jet over southwest rift of Mauna Loa, 8 a. m., May 19, 1916. Looking S. 84° W., taken from Voleano Observatory at Kilauea erater by H. O. Wood. 268 JSaggar—Lava Flow from Mauna Loa, 1916. 82° 30’ W., and the lower one approximately 8. 85° 30’ W., from the Obser vatory , indicating an explosion about the 11, 000-foot (3360 m.) contour of Mauna Loa along a N.E.-S.W. rift over a distance approximately one and a half miles long (2-4 km.) and a few miles to the southwest of the crater Mokuaweoweo. This preliminary explosion of vapor reached its maximum force in a little over an hour, and by 8:15 a. mM. was declining so that by 10 a. m. the stem ‘of the column had disappeared. The fume made a milky bluish cirrus which spread over the sky and was visible above the lower rain cumulus until night- fall, developing a ripple pattern. The color was quite distinet from that of ordinary clouds and showed iridescence in places. Some lava was ejected from this high vent according to the reports of cattlemen, anda line of whitish discoloration along the high rift at the site of this outbreak was seen and photo- oraphed from a distauce of about ten miles (16 km.) south by an Observatory party in July. This was probably a spatter of pumiceous lava adjacent to the fissure. A little fume was seen at sundown May 19 and 21 in thin wisps near the place of first outbreak. Miss Paris in upper Kaawaloa, the middle Kona district, reported seeing in the morning of May 20, a “ pillar of smoke” high up the southern Mauna Loa slope. Local earthquakes diminished on the 19th after the first storm of vibrations, increased on the 20th in frequency and amplitude, and thereafter were registered by hundreds as recorded above. Lava Outflow May 21, 1916. At about 11 p. um. May 21, liquid basaltic lava spouted up in fountains along the main southern fissure system of Mauna Loa north of the well-known cone called Puu o Keokeo in the middle of the southern lobe of the mountain at an elevation higher than 6600 feet (2000 m.) above sea-level. As is usual with Hawaiian lava, the outwelling was quiet and produced no sufficient noise or quaking to awaken inhabitants of the coast- wise plantations nine to fifteen miles (15 to 24 km.) away. The glow on the sky, however, became very brilliant and was first seen from the Observatory in a direction 8. 66° W. During the night, the writer accompanied by Mr. H. O. Wood, seismologist, motored from Kilauea by the government road to Kau-Kona boundary making a partial cireuit of the center of outflow east, south and southwest at a distance from sixteen to eleven miles (26 to 17 km.) of Pun o Keokeo. A column of fume illumined red from below was seen at appar- ently about the same distance inland everywhere from Waio- hinu westward looking north across the lava flows of 1868, 1887 and 1907. A brilliant radiance extended southeast from JSaggar—Lava Flow from Mauna Loa, 1916. 269 the fume column showing that the first extension of the lava from the vent was eastward, all being dark to the west of it. It was a clear, calm night, and there was not the slightest noise or perceptible quaking of the ground. We were obliged to awaken the residents of Hutchinson Plantation and Waiohinu who were unaware of the eruption although the sky above them was brilliant with a ruddy glare. At 2 a. mM. May 22, W. Vredenburg, manager of Kahuku Ranch, rode inland between the 1868 and 1887 flows and after passing the koa forest, came to the open country of the upland and saw two flows in rapid motion headed respectively south- east and south-southeast at a distance approximately ten miles (16 km.) north of the road. At daylight May 22 looking northwest from the Kahuku Ranch gate, we saw the main fume column extending itself in a series of puffs developing to the right or northward. This may have been the beginning of the second flow which poured westward during the next twelve hours. This western, or Kona flow, passed through the land called Honomalino, a cattle ranch,‘and its progress southwestward was mostly accomplished during this day, for no glow in that direction was seen before daylight, whereas on the following night most of the illumina- tion was on that side. Distribution of 1916 Lava. To make clear the general arrangement of these flows of 1916, there is here reproduced (fig. 4) a tracing from a rough sketch map made by a government surveyor of the Territory of Hawaii after the eruption showing known trigonometric stations. As shown on Baldwin’s map of Hawaii,* the 1907 lava flow originated in a fissure extending some miles above Puu o Keokeo, but the main outpouring of that year was from an opening below Puu o Keokeo, which perhaps remained open after the upper fissure had closed so that it came to be considered the 1907 source. Fig. 4 shows only this lower 1907 orifice. The 1916 lava welled up cracks extending three and a half miles (5°6 km.), about N. 10° E. from Puu o Keokeo, the zone of outpouring being about thirty feet (9 m.) wide, and bending more to the northeast above. There is more lava to the north and east than the map shows. The elevation of this source rift les from 6500 to 7500 feet (2000 to 2300 in.) above sea-level. Along it about ten new cones were developed in 1916 of lava and vari-colored cinders, and the upper mile (1°6 km.) is solfatarie with much sulphur deposit and no lava except as spatter. The lava flow eastward in Kahuku poured from cones for about one mile north of Puu o Kecokeo, and “locy cits. psio22: Lava Flow from Mauna Loa, 1916. 270 Saggar the flow westward from cones still farther north, the line of cones being nearly straight (fig. 5). | Comparing the map of Hawaii (this Journal, Dec., 1915), it will be seen that the new flows straddled the northward exten- sion of the 1907 flow, coming out of the same general zone of cracks, and forked southeast and southwest in the same fashion as the 1907 flow had done, but from a higher point on the mountain, and with a more extensive spread of the lava to ID(s, 44 MHAWAl TERRITORY SURVEY WALTER E.WHLL SORFEYOR i LOCATION SIAUNA LOA FLO OF 13216 : APualehua KAU-KowA, ~farale — 6033.5 Seal il Afizu Ahinui Agere 7 Surrey “By 3, RANARANO} * 3967.84. NS . SUV a 491E Loula Aku Hoomah 6347. 6636.4 QAChiclele 2H 6oT0 ft. ——— oi. SS A thuanuw Thus ft. Anois poste Fic. 4. Sketch map 1916 lava flows, showing general arrangement fissure eruption north from Puu o Keokeo. Map furnished by courtesy Territorial Surveyor of Hawaii. the east. This was due to the presence of a flattening of the mountain surface near Puu o Keokeo and it was probably this flattening and the expenditure of the flood in two branches thus early which partly saved the ranches and prevented the flow from reaching the government road. There is great need of a good topographic map of Mauna Loa for the benefit of ranches during such a crisis. The flood of lava which spread eastward during the first night was the widest and probably the thickest portion of the 1916 basalt and this area has tongues eastward not shown on the map. JSaggar— Lava Flow from Mauna Loa, 1916. 271 An Observatory party skirted the edge of this eastern flow on May 30 and “ encountered many thin narrow tongues (from five to eight or ten feet (from 1°5 to 3m.) in depth and from fifty to two hundred yards (46 to 180 m.) in width) radiating to the southeast and east. These departed from the main stream at higher and higher points. Wherever junctions were seen, the departures of these minor branches appeared capricious—1. e. no evidences of local damming or pooling were seen. Though much thinner and less massive than the main stream, these were still fuming, and in varying degrees the air above them Wig, A. Fie. 5. Line of heapings at fissure-source, 1916; lava flows looking north from Puu o Keokeo. Eastward flow in foreground. Photo. Miss Tulloch, May 26, 1916. was in a state of shimmer from heat. The emanation of the fumes, however, furnished a more reliable indication of their course than the heat-disturbed air above them.’’* Journal of the Eruption May 22-31, 1916. The following narrative is quoted from the writer’s account prepared on June 1:+ “ During the day, May 22, the accumulated lavas of the upland were discharging southwestward from a vent about two miles (3°2 km.) above Puuo Keokeo. As no glow was seen west of the * Bull. Hawaiian Vole. Obs’y, vol. iv, No. 6, p. 56, June, 1916. + Bull. Hawaiian Vole. Obs’y, vol. iv, No. 6, pp. 40-43, June, 1916. Am. Jour. Sct.—FourtsH Series, Vou, XLIIT, No. 256.—ApriL, 1917, 18) 272 Saggar—Lava Flow from Mauna Loa, 1916. vent before 6 A. m., this southwest flow across Honomaline Ranch must have progressed at least six miles (9°6 km.) in twelve hours, for at 6 Pp. M. it was within three miles (4:8 km.) of the road at Honomalino. The settlement here lies in a valley and the flow was headed directly for the houses. The timber could be seen burning from the road and the flowing lava, advancing down a ten degree slope, from the housetops. Detonations from explod- ing gas, and the crash of falling trees, could be heard. The homes were temporarily evacuated but fortunately the flow stopped at this point, its front resting on a hillside forested with lehua, kukui and guava. This flow was rugged aa eight to twelve feet (3 to 4 m.) in thickness and a quarter of a mile (400 m.) wide along its lower course. | At eight o’clock on this night, Monday, May 22, the fiery radi- ance above the fountain head and the flows probably reached its greatest brilhaney. Unlike the night before, the glow was now all to the west of the column of fountain fumes which shot up in purling volutes on the right spreading to the west in a high cloudy haze, which was lighted from below a bright red from the glowing river of melt in Honomalino. On the left a great pur- plish swirl of smoke from the burning forest rose and curled over to the east. ‘There was darkness over the Kahuku flow, showing that for the time the lavas on that side were sluggish and con- gealing. During the night May 22-23 there was revival in Kahuku. Messrs. Waldron and Liardy found the front of the first flow nearly motionless at altitude 5,500 feet (1700 m.) by aneroid, At 1 a. M. May 23, an arm of this flow more southerly in trend devel- oped near the source and for two hours advanced brilliantly. Then it became sluggish, cooled down and stopped. About day- light the western glow and the fountaining at the source had both diminished. On May 23 motion within the Honomalino flow diminished rapidly. In the morning flowing lava could be seen and trees were still being overturned. At 3 Pp. m. high ohia trees were still standing in the midst of the flow, their trunks blazing, a few blocks were tumbling at the front but no pasty lava was to be seen. The side of the flow resembled a heap of burning anthra- cite coal, with flames playing through the incandescent blocks, and the only odors were a strong smell of burnt charcoal and occasional coal-gas (fig. 6). At 7 ». M. the glow seen from the south indicated distinctly a revival of the southeast flow toward Kahuku. All was dark west of the fume column. ‘There was a tapering line of reflected light on the clouds trending south-southeast, and in the same direction along the ground below could be seen a low straight-topped film of smoke, highest near the. source and lowest at the lava front. Over this smoke could be seen the radiant glow from the lava. May 24. On this date the writer inspected the Kahuku flow eight miles (12°8 km.) from the road and in the region between the JSuggar— Lava Flow from Mauna Loa, 1916. 273 1868 and the 1887 flows. At about seven miles (11°3 km.) there is a broad rolling paddock with sparse koa. Above this is a steeper slope of koa forest, down which the ancient aa flow, Pele o Iki, had cut a swath. Above this forest is an extensive bushy upland consisting of old pahoehoe and aa which stretch out as far as the eye can reach toward distant cones and on the west to the rugged ridges of the 1887 flow. On reaching the upper edge of the koa slope, it was found that the fronts of the new flow had moved down more than a mile Fie. 6. e Fic. 6. Front of Honomalino lava flow in forest, May 23,1916. Photo. Colville. (1°6 km.) since the previous morning and were now close to the forest. On riding up the west side of the flow, which was in process of overriding Pele o Iki, we came to the moving lava advancing in billows and cascades of red hot viscous liquid and overflowing the already cooled aa surface of the previous day. The moving liquid lay in streams from 100 to 300 feet (30 to 90 m.) in width and one of these only 100 yards (91 m.) from us was cours- ing down the main lobe of the previous day, driving in front of it a fan-shaped pool of dark red color witha sugary surface which advanced at the rate of perhaps a half a mile (800 m.) per hour. The cascades above carried huge rafts and bowlders of a ten to thirty feet (3 to 10m.) in diameter, and these blocks moved along majestically in a procession, about as fast as a man might walk. 274 Jaggar—Lava Flow from Mauna Loa, 1916. The advancing flood became deflected to the west and tumbled over the bank of congealed lava into the vegetation at our feet not fifty yards (46m.) away. The liquid would developa cindery surface and then congeal into what appeared to be a glowing mass of coals. The only odor was of the carbon gases and these were strong and oppressive, particularly as we were directly to leeward and a fresh breeze was blowing from the northeast. The heat was intense, and occasionally fierce small whirlwinds developed at the edge of the hot flood and moved along the border with a whistling noise, carrying up clouds of grit and smoke. Out in the molten river gigantic bubbles would occasionally burst with a thud but there were no fountains as in Halemanmau. Cascade beyond cascade could be seen up stream vanishing into the hot vibrant fume of the distance and I was much impressed by the piling-up effect of this motion, the advance being always by over- flow. ‘This is in marked contrast to the usual mechanism of pahoehoe flow where the new floods push out from beneath in tongues which break through the skin, or else a crust breaks up and founders in the rising melt. No crusts whatever were seen forming on this a@@. There were moving blocks and there was the red hot flood. When it cooled, it appeared to cool as a unit and the new floods overflowed it. There were one or two patches of roundish shape in the chilled part of the flow which resembled pahoehoe. These were probably blocks transported from far above where near the vent the flow actually was pahoehoe. The front of the overriding flow was advancing in the direction S.15° E. The under-flow was in two lobes, one leaf-shaped and headed in this direction and one more to the east pushing for- ward very slowly at a rate estimated sixty feet (18 m.) per hour. Between 2 p, m. and 7p. M., according to Messrs. Bonesteel and Moses, the fan-like advance pool which we had seen overran the leaf-shaped lobe and coursed down the koa slope carrying the forest before it, following the western margin of Pele o Iki. This became the main Kahuku front and flowed a mile farther, stopping on the paddock about seven miles (11 km.) from the road. The flood would surround the trees, burn them and uproot them, and in some cases the koas were seen to be carried along upright like blazing torches before they toppled over and were engulfed. At the front the flow was only a few hundred feet wide. An interesting mechanism which occurred whenever the flow poured over old cavernous pahoehoe was furnished by gas explo- sions in caves. ‘This was probably coal gas mixed with air. Detonations were frequently heard, and the writer saw the dust from one of these explosions a few feet away, puff up through cracks above a cavern. The old rock was very hot, and it was evident that some of the hot gases and perhaps the new lava had penetrated a subterranean tube. Mr. L. A. Thurston in the same region on the following day found several of these places where the whole roof of the cavern had blown off and large blocks of rock were strewn aboat several feet from the orifice. Jaggar—Lava Flow from Mauna Loa, 1916. 275 The noise of the lava itself in motion was not great. It was a splashing sound, and where the pasty lava dripped the quickly cooled droplets fell with a tinkle. The bubbles bursting made thudding noises, but there was no sustained explosive fountaining such as the writer had imagined. In general the advancing floods were unexpectedly noiseless. The horses advanced fearlessly to within two hundred feet of where the liquid lava to windward was actually cascading on to the grass and bushes, and showed as complete indifference as they would to a camp fire. In the for- est, of course, there were the noises of falling timber and crack- ling flames. The various estimates of rate of motion of lava flows which different observers reported of these flows were based usually on observed progress of the runways compared to a man’s pace. But the runways are no measure of the movement of the lobe as a whole, nor 1s the movement of a lobe a measure of the advance of the whole arm of lava which may be slowly pushing down a grade. The flows progress appareatly by impulses of overflood- ing into hollows, the central vent above feeding a great pooled area. Somewhere the margin of this area finds outlets. From what the writer saw he would judge five miles (8 km.) an hour to be a fast rate for the rush of the runways, and half a mile (800 m.) an hour for average general progress of a flow as a whole during its most rapid advance on a steep slope. ‘This progress appears to be not gradual, but in spasms, with intervals of many hours of stagnation. At midnight May 24, the profile north from the Kahuku gate showed a column of dark smoke thought to be burning forest, bright flames at two points shooting up on its right, and strong radiance both east and west of the smoke column. The smoke and flames were at the front of the Kahuku flow, and so nearer the observer, while the radiance was from the floods of molten stuff in the distance. R. MceWayne about this time reported for the head of the Honomalino flow a cone about the 7000-foot (2300m.) level ejecting with a roar high lava fountains and blocks of rock. There were two main vents in action, both above Puu o Keokeo and in a line with it, the higher feeding several tongues in motion toward Kona, the lower supplying the Kahuku flows. Vredenburg left the Kahuku flow at 2 a. wm. May 25. The main lobe had then advanced three quarters of a mile in twelve hours, and the eastern lobe had pushed forward somewhat along the east margin of Pele o Iki. The main front was stagnant and three hundred feet (92 m.) in width, widening above. The main direction was about.8.15° E. During this day the main flow advanced a quarter of a mile (400 m.) farther, there was some activity and burning forest on the front of the east lobe, and to the westward glowing lava could be seen in motion overriding the flows already cooled. 276 Jaggar— Lava flow from Mauna Loa, 1916. At 11 p.m. May 25 the mountain profile showed a red general glow behind a screen of cloud, with rapid rising white cumulus over the fountain head. : May 26. S RS 8 DQ a = ies) ‘S 8 £ = Ss | SD = S Q — toby Taped 1 SOU a E.. T. Long—Fformation of Salt Crystals. 291 hot brine. During the first step of the eo ale a second glass tube was run through the stopper (X) in the right hand side neck of B without rubber tube connection, at the outer end of which, however, there was a connection on which another stopcock was placed. During the second step, instead of the glass tube, a common rubber finger was used. It was punctured at the tip and a small glass tube inserted, while the open end was stretched over the same rubber stopper, xt The most effective temperature to bring about deposition from the hot brine was found to be between 70° and 75° C. Below that, there was not enough NaCl absorbed for an appreciable deposition, and above 75° recrystallization went on in Bb. As enough hydrostatic pressure could not be obtained by raising A so that the passage through the stopper - at X could be kept open, it had to be frequently cleared by a steel rod with a right-angle turn run through the top neck of B. This stopper was also removed to replenish the supply of salt, but otherwise B was kept air-tight to prevent crystallization by evaporation. After five or six days the hot brine became sufiiciently saturated to begin the formation of erystals at C, the first point where crystallization took place, and therefore the point where it was always farthest advanced. At the end of a few weeks the crystals had grown so that they completely filled the tube and stopped the flow entirely. Figure 2 shows the conditions at this stage of the experiment. The hydrostatic pressure was increased but no seepage could be forced through. This run was repeated several times under slightly different conditions, but the result was always the same even though the time varied somewhat. The second step was then started: to make the crystals grow against pressure, in this case a rubber finger. This was so successful that after a month, one crystal expanded the rubber wall of its prison so far as to puncture it. The tiny hole was cemented over and effectually closed, but in a few days another crystal went through at a different point, this time making a much larger hole. On page 550 of Mr. Taber’s article he says, “ The tendency to form erystals is much stronger in some substances than in others, but it.is never so strong as to cause growth on a face which is not in contact with a supersaturated s solution, and even if a growing surface is in contact with a supersaturated solution, the relative rate of growth is chiefly controlled by the rapidity with which the material for growth is made avail- able.” The results of the two series of experiments thus indepen- dently worked out would therefore seem to establish without 292 £. T> Long—Lormation of Salt Crystals. question, not only that a hot saturated brine will deposit salt when cooled, but that the developing crystals in growing exert a lateral pressure sufficient to permit continued growth even against opposing external forces. Some of the stages of growth of the crystals in this experi- ment are given in figs. 83-6. Fig. 3 shows the collapsed finger Fie. 3. Hines Mie, 0. Fig. 6. with only the liquid running through. Later the erystals began to expand the rubber, and at the end of the first run, the connection with the stopper sprang a leak; fig. 4 shows the condition of the second run, after the first puncture and before the second one, which ended the experiment. Fig. 5 is an enlargement of 3, and 6 an enlargement of 4. B. Wade— Upper Cretaceous Fulgur. 293 Arr. XXV.—An Upper Cretaceous Fulgur ;* by Bruck Wank. Tue family Fulguride was proposed by Grabau and Shimert in 1909. It includes a well-defined group of pyriform gas- tropods that are very common in the Tertiary and Recent of North America. The group was first recognized by Fischert in 1887 and given the rank of subfamily under the Tur- binellidee, and its unity was again pointed out by Cossmann§$ in 1901. Since the assigning of the group to the rank of family in 1909 the name Fulguride has been used by some paleontologists to include Zudicla, Pyropsis, Perissolax, Busycon, Strepsidura, Levifusus and Pyrifusus. Of these genera Pyropsis, Pyrifusus and Perissolax are especially profuse in the Upper Cretaceous of North America. Zudtvcla, Levifusus and Strepsedura are most abundant in the Eocene. Busycon, commonly known as Fulgur, and sometimes Sycotypus in part, is very abundant in the later Tertiary and vecerin: The Fulgurs are especially interesting on account of their limited geographic range which is restricted to the eastern United States. This distribution is explained by the fact that the animal is deprived of an active free-swimming larval stage by the loss of the velum before the young emerge from the egg-capsule. The geological range of Fulgurs is also of interest. The most common and important Pliocene to Recent Fulgurs are: Busycon carica and Busycon canaliculatum. Among the forms ranging from Miocene to Recent Busycon perversum is important. Busycon excavatum is found from the Miocene to Pliocene. Busycon pyriformis, Busycon _imeile and Busycon coronatum are limited to the Miocene. The above-named species are less than half the number of known forms, but include the more important that are well- established. Busycon (spiniger var.!) tampaense Dall, Busy- con spiniger var. nodulatum Conrad and Busycon stellatum Dall are common in the Orthaulax pugnax zone of the so- called Oligocene of Florida. The present knowledge of the Fulgers older than the Oligo- cene seems to be in a chaotic state. There has been a con- fusion of both species and genera of the fulguroid forms in the Eocene. Grabau| has questioned the existence of any Fulgurs * Published with the permission of Dr. A. H. Purdue, State Geologist of Tennessee. + Grabau and Shimer, North American Index fossils, vol. i, pp. 764-772. { Fischer, P., Manuel "de Conchyliologie, Paris, 1887, p. 61 8. use ane M., Hssais de Paleoconchologie Omanares, Paris, 1901, Liv. . 61. | Ceapael A. W., American Naturalist, vol. xxxvii, pp. 515-539, 1903, 294 B. Wade— Upper Cretaceous Fulgur. older than the Oligocene. Dall,* however, points out that the genus had assumed its essential shell features before the close of the Eocene and that the group took its rise within the Eocene. The fulguroids are abundant in the Eocene and are represented by a number of forms whose generic relationships have not been well-established and have been variously assigned. The more important of these are “Busycon spiniger”’ Conrad, Fulgur triservalis Whitfield, Pulgur? dallianum Harris, Levifusus dalei Harris, Levifusus Blakei Conrad, Levifusus trabeatus Conrad, Bulbifusus eornatus Conrad, ete. Which of these are true Busyeons has not been established. The Eocene Fulgurs have small, thin shells and less bulbous pro- toconchs in contrast with the ponderous conchs which con- stitute one of the most conspicuous elements of the later Ter- tiary. This has led to the conclusion that the Eocene species were the most primitive forms and that the genus evolved in that period. However, the recent discovery, in the Upper Cretaceous of McNairy County, Tennessee, of a typical Busy- con or Fulgur would seem to show that the genus was well- differentiated long before Eocene time. Further collecting and study of well-preserved Cretaceous and Eocene material may throw much light on the relations of this most interesting family of North American late Mesozoic and Tertiary gas- tropods. There appear to be two species of Busycon in the Upper Cretaceous of Tennessee, the larger and more perfect of which forms the basis of the present note. For the sake of complete- ness the generic synonymy is given followed by a description of the new form. Family Busyconipa Genus Busycon Bolten. Busycon Bolten, Mus. Boltenianum, p. 149, 1798, First species. Fulgur carica Montfort. Fulgur Montfort, Conch., vol. ii, p. 503, and figure, 1810. Type, Fi carica var. eleceans Montfort. Sycopsis Conrad, Amer. Jour. Conch., vol. 111, p. 184, 1867. Sycopsis (Browne) Gill, ibid., vol. ili, p. 147, 1867. Fulgur Fischer, Manuel de Conchyliologie, Paris, p. 620, 1887. Fulgur Dall, ‘Trans. Wagner Free Inst. Sci., Philadelphia, vol. ii1, pt. 1 pps L09- Tis. 1890: Fulgur Cossmann, M., Essais de Paleoconchologie Comparee, Paris; pp. 76-77, 1901. Fulgur and Sycotypus Grabau and Shimer, Index Fossils of North America, vol. i, pp. 767-770, 1903. Busycon Dall, Fauna of the Orthaulax pugnax zone, Bull. 90, U.S. Nat. Mus; p: 66, 1905: * Dall, W. H., Trans. Wagner Free Inst. Sci., Philadelphia, vol. iii, pt. 1, p. 109, 1890. B. Wade— Upper Cretaceous Fulgur. 295 This genus was. assigned to the subfamily Fulgurine in the Turbinellide by Fischer in 1887. In 1890 Dall gave a dis- cussion of the known species at that time and assigned the genus to the family Fasciolariidee. In 1901 Cossmann published probably the most careful generic description to be found in the literature on the genus Busycon and assigned it to the sub- family Fulgurine. In 1903 Grabau and Shimer very fittingly applied the family name Fulguride. In 1915 Dall called attention to the ruling of the International Committee on Bigs Ie Mire 2: Fies. 1, 2. Back and front views of Busycon (Protobusycon) cretaceum, Sp. nov. Nomenclature and revived the old Bolten name Busycon, but referred the genus to the family Buccinide. The present paper extends the range of the genus by including a descrip- tion of an Upper Cretaceous species, and emphasizes the desirability of recognizing the family rank of the Fulguride for the fulguroid group. In accordance with the accepted rules of nomenclature this should be the family Busyconide, as Busycon has replaced Fulgur as the type genus. The new species described below differs from its congeners sufficiently to form the type of a new subgenus for which the name Protobusycon is proposed. 296 BL. Wade— Upper Cretaceous Fulgur. Busycon ( Protobusycon) cretaceum, new species. Description.Shell rather small for the genus; outline typically fulguroid; spire low, less than one-fourth the entire length of the shell; earlier volutions broadly rounded, later whorls broadly and somewhat obliquely shouldered,—the peri- phery falling about two-thirds of the distance from the posterior to the anterior sutures; apex obtuse; protoconch broken away in the type,—the scar large; the remaining four in number, increasing rapidly in size to a much inflated body; external sculpture inconspicuous and rather irregular; axial sculpture restricted to a series of low, sub-spinose protuberances crown- ing the shoulder keel, ten in number upon the last whorl of the spire, horizontally elongated and irregular in size and spac- ing upon the final half turn of the ultima; a second obscure keel outlining the base of the body, obsolete toward the aper- ture,—the keel beset with four, or possibly five, rudimentary spines; incrementals vigorous and crowded toward the aper- ture, especially upon the shoulder; spiral sculpture obscure and irregular,—the liree approximately ten in number on the medial portion of the body of the type, tending to alternate in size, minutely crenulated by the incrementals; surface of spire so badly decorticated that the character of the finer sculpture can not be determined ; line of demarkation between the base of the body and the pillar outlined by a shallow sulcus; in- crementals very sharply folded along the sulcus, the fold directed toward the aperture, and terminating as a slight pro- jection at the margin of the labrum ; posterior portion of whorl closely appressed ; suture inconspicuous; aperture pyriform, feebly sulcate at the posterior commissure, terminating anteriorly in a long, open canal; outer lip broadly arcuate, notched at the shoulder, the incrementals produced into a series of varix-like spines; labrum feebly insinuated also at the base of the body, directly in front of the basal suleus ; inner margin of the aperture quite strongly excavated ; parietal wall widely and heavily glazed ; columella smooth, sinuous; anterior canal broad, slightly recurved, probably feebly emarginate at the anterior extremity. Dimensions.—Altitude, 63°2"™; length of aperture includ- ing canal, 50™™; maximum diameter, 35°2™”. This very interesting species is represented in the Coon Creek collection by the single specimen figured which, aside from the loss of the protoconch, is well-preserved. In shape and elevation of the body, angle of the shoulder, sub-spinose angula- tion of the posterior portion of the body and in the general aspect, it is curiously similar to Hulgur carica (Gmelin) so abundant in the recent faunas. Theimpressed line at the base B. Wade— Upper Cretaceous Fulgur. 297 of the body does not appear on any of the later Tertiary and Recent representatives of the genus and may possibly be ex- plained as an inherited character from a more primitive type with an abruptly constricted body whorl such as that of Pyropsis and Tudicla. Much more probably, however, the basal suleus and marginal notch are in some way analogous to the more or less well-defined band and marginal notch used by many of the recent groups, notably in Strombus, for the ex- trusion of the eye-stalks. Although the sulcus is peculiar to Protobusycon, the abrupt basal constriction is shared by some of the Eocene members of the group, and by Busycon stellatum* Dall of the Florida Oligocene. Occurrence.—Rirtny Formation: Dave Weeks Place, on Coon Creek, McNairy County, Tennessee. (Collected by the writer.) Geological Laboratory, The Johns Hopkins University. * Dall, W. H., Molluscan Fauna of the Orthaulax Pugnax zone, U. S. Nat. Mus. Bull. 90, p. 67, pl. 10, figs. 7,9, 1915. 298 Berry—Middle Eocene Member of the “ Sea Drift.” Art. XX VI.—A Middle Hocene Member of the “Sea Drift”; by Epwarp W. Berry. Stupies made during the past few years have brought to light extensive fossil floras in southeastern North America which range In age from the lower Eocene to the Pliocene. Practically all of these are coastal floras and the bulk of the plants discovered belong to the strand flora and comprise many plants whose fruits or seeds are normal constituents of the sea drift and are distributed mainly by ocean currents. Among these are such striking forms as the Nipa palm, Dodonea, Sapindus, Sophora, Rhizophora, Conocarpus, Avicennia, ete. These Tertiary floras, partially published,* indicate pro- gressively warmer climates commencing with the lower Eocene and culminating in the lower Oligocene, at which latter time tropical climates prevailed throughout this region. Some years ago I received from Mr. Otto Veatch a striking seed collected by him from the middle Eocene of western Georgia. This has remained undescribed until now, but as it represents such a well-defined type of seed of the tropical or sub-tropical sea drift and a type not heretofore represented in the fossil state by seeds, it seems worthy of a brief note. It is referred to the genus Carapa of the family Meliaceze and may be characterized as follows: Carapa xylocarpoides sp. nov. Seed of large size, somewhat trapezoidal or pyramidal in outline, tapering toward the hilum, rounded distad. Length about 3°5™, maximum width about 3°5™, thickness about 6™™. The lateral margins are rounded. The distal margins tend to be somewhat angular and there is a more or less pronounced angular ridge on the proximal face of the seed. The outline and the degree of rounding or angulation and the variability are the results of mutual pressure of the seeds of a head, and the range of variation among existing forms is considerable. The texture is ligneous and the seeds obviously formed part of the middle Eocene sea drift and are contained in marine sediments. This striking form is unquestionably referable to the modern genus Carapa Aublet or to the allied genus Xylocarpus Konig and Jussieu, the latter often made to include two existing oriental species frequently referred to Carapa, especially in the older literature. The differences between the two genera are * Berry, E. W., Lower Eocene Floras of Southeastern North America, U.S. Geol. Survey, Prof. Paper 91, 1916. Berry— Middle Eocene Member of the “ Sea Drift.” 299 i uce mae 300 Berry—Middle Hocene Member of the “ Sea Drift.” those of floral structure, degree of buoyancy of the seeds, and manner of dehiscence of the fruit, characters upon which the fossil sheds but little light. The seed coat of Carapa is woody while that of Xylocarpus is corky and consequently more buoyant and better adapted for dispersal by ocean currents. In this feature the fossil seed seems more like those of Xylo- carpus. In form Xylocarpus seeds are somewhat more regularly pyramidal than those of Carapa and in this respect also the fossil is more like Xylocarpus, especially the oriental mangrove AXylocarpus obovatus, a seed of which is shown in fig. 10. Hotere in considering an American Tertiary form and recalling that Carapa is much the better known generic term even for the oriental mangrove and other modern species of both Carapa and Xylocarpus, and that Carapa is represented by foliage in the lower Eocene (Wilcox) flora of this general region,* it seems advisable to adopt the generic term Carapa for the fossil, since it can hardly be more closely related to the oriental mangrove AXylocarpus obovata Blume or the oriental beach plant Xylocarpus moluccensis Lamarck than to the half dozen existing species of the American and west African tropics. The absence of the more or less massive seeds of the various existing species in our larger herbaria is my excuse for not making more detailed comparisons between them and the present fossil form. The locality where the fossil was collected is 24 miles east of Fort Gaines, Clay County, Georgia, and it came from the marine clays of the middle Eocene (Claiborne). The fossil is shown natural size in fig. la and a somewhat larger seed of the oriental Xylocarpus obovata is shown in fig. 16. Johns Hopkins University, Baltimore. * Berry, op. cit., p. 253, pl. 59, fig, 4. W. A. Verwiebe—Correlation of the Mississippian, etc. 301 Arr. XX VII.— Correlation of the Mississippian of Ohio and Pennsylvania ; by Wattrer A. Verwirse, Ohio State Uni- versity, Columbus, Ohio. Ever since the time when it became possible to discuss geo- logic systems in this country in formational detail, the problem of the correlation of Devonian and Carboniferous formations in northeastern Ohio and northwestern Pennsylvania loomed large in geologic literature. It was not a simple matter to trace the Devonian formations of New York into northwestern Pennsylvania, nor was it an easy task to define the continua- tion of the lower Carboniferous formations of eastern Penn- sylvania, in western Pennsylvania. Both of these problems, however, have been solved to a satisfactory degree. In the west the excellent work of Orton, Newberry, and Prosser soon brought order out of the profusion of shales and sandstones which compose the Devonian and lower Carbonifer- ous of Ohio. By some strange coincidence, however, the line dividing this state from Pennsylvania seemed also to form an insuperable barrier between the formations in both. Many of our most capable stratigraphers have attempted to correlate the Devonian in these two areas, but it must be confessed, with unsatisfactory results. The writer does not presume to think that he has found the key to this riddle, but after spending two seasons in the field he has reason to hope that some of his findings may contribute a little in that direction. The area covered in ‘this survey extends approximately from the Cuyahoga River in Ohio to the Allegheny River in Pennsylvania and from Lake Erie south to the parallel of 41° north latitude. It thus includes roughly the northeastern part of Ohio and the northwestern part of Pennsylvania. In a former article published in this magazine* the correlation of the Berea formation was treated in some detail as the basal member of the Mississippian. In this article the remaining formations will be taken up. Formations involved in this discussion. Sharon conglomerate; This is the basal formation of the Pottsville series of the Pennsylvanian system. It was used merely as a key horizon and was not studied to any further extent in connection with this work. Mississippian formations of Pennsylvania. Shenango Shale: This formation was named by I. C. White. It is recognizable as a stratigraphic unit only over a limited extent of territory since it merges imperceptibly with the * This Journal, vol. xlii, pp. 48 to 58, July, 1916. 302 W. A. Verwiebe—Correlation of the Shenango sandstone toward the east and is largely absent because of erosion toward the west. It is generally a bluish gray, argillaceous shale though locally also very sandy. Its maximum thickness is perhaps 50 feet, but, on account of the unconformity at the top, usually less than this is found. Shenango Sandstone: As a key horizon to the stratigraphy of western Pennsylvania this is a most important formation. It received its name from the Shenango River of Crawford LE ey ils ly wiht pb ee hon i) HON Ah ‘h)! Fie. 1. Sections showing the character and thickness of the Shenango formation. The map on p. 312 (fig. 6) inserted shows location of sections. and Mercer counties. In lithologic character it is unique. Always coarse and quartzitic, it is marked as well by the great amount of iron present in eoncretionary form, in secretions, and in veinlets. The solution and redeposition of this sub- stance has given it a deep brown color throughout, but espe- cially on exposed surfaces. In thickness it is very uniform, ranging gradually from 15 feet in the west to about 40 feet in the east. Farther east and south of the region studied it merges more and more closely with the overlying shale and has been named as a unit the Burgoon formation.* In fig. 1 are platted sections to show the nature and thick- * Butts, Chas., U. S. G. 8. Folio No, 115, p, 9. Mississippian of Ohio and Pennsylvania. 303 ness of the Shenango sandstone and the Shenango shale over- lying it. Since the top of the latter is the locus of an uncon- formity, the thickness is variable. A feature well brought out, however, is the gradual thickening of the sandstone at the expense of the shale as they are traced eastward. Toward the west both shale and sandstone are cut out by the unconformity, but reappear in the longitude of Cleveland as the uppermost part of the Royalton formation. The formation representing the aang shale and sandstone in central Ohio is the Logan. Bie, 2: me mitre yh AUN ULE Fie. 2, Sections showing the Meadville and Sharpsville formations. Meadville formation: The county seat of Crawford County, Pa., is Meadville. It suggested the name of this formation to Dr. White. Consisting essentially of bluish gray sandy shales, it would appear at first not to be of any great value for the stratigraphic geologist. However, two layers at definite hori- zons are of such striking lithology that they serve as excellent guides. They are essentially limestones, but because of their high silica content are very hard and compact and possess a conchoidal mode of fracture. Pieces of these layers can there- fore readily be recognized in streams even though the parent rock may be covered. In fig. 2 sections have been platted showing the character and thickness of this formation together with the Sharpsville Am. Jour. Sc1.—FourtH Srerizes, Vou. XLIII, No. 256.—Aprit, 1917. 21 3 304 W. A. Verwiebe—Correlation of the below. The two limestones are somewhat exaggerated in order to make them more prominent. ‘They are usually six inches to two feet thick but may be less and are sometimes absent. A third limestone of exactly the same nature as the Meadville limestones is found about the middle of the Sharpsville. It will be noted that along the state line between Ohio and Pennsylvania the Meadville already begins to show a con- siderable proportion of thin sandstones. These increase in number toward the east. But toward the west the opposite takes place and the Royalton formation, which represents it toward its upper part, has only a small percentage of sandstone layers. Sharpsville formation: This formation was named after the village of that name which lies a few miles north of Sharon on the Shenango River. It consists largely of sandstone, though shale layers may occur; furthermore muddy sediment is so thoroughly mixed with the sand grains that the sandstone is very impure. Frequently the formation is divided into two subequal parts by a limestone layer like those in the Meadville formation, which is on the average one foot in thickness. In tracing the formation toward the east it was found more and more difficult to separate it from the Orangeville below. Toward the west it becomes more shaly and finally also loses its identity in the Cuyahoga terrane. It is distinctly recog- nizable as far west as the Warren (Ohio) region, but becomes indefinite beyond that. Orangeville formation: Lying just beyond the border of Pennsylvania in Trumbull County, Ohio is Orangeville. Here a series of soft, argillaceous, bluish gray shales are exposed to which I. C. White gave the name Orangeville. They are interbedded to a slight extent with thin sandstones. Iron is present in the form of marcasite concretions and the solution of these causes the formation to present a rather rusty outerop. When followed toward the east the sandy layers become more abundant and in the region of the Allegheny River form the largest part of the formation. Toward the west the opposite holds true, also the layers toward the base become more carbo- naceous, blacker, tougher, and more fissile. This phase in its typical development is called the Sunbury shale. Whenever the Orangeville is encountered it is found to contain an abundant though not varied fauna which consists of the follow- ing species : Lingula melie, Hall Lingula membranacea, Winchell Discina newberryi, Hall Discina pleurites, Meek Mississippian of Ohio and Pennsylvania. 305 Fig. 8 shows that the thickness increases noticeably from east to west, also that the formation has a definite sandstone horizon toward the west which separates the Brecksville from the Sunbury shale. This is the Aurora sandstone. Berea formation; This formation is considered by far the most important of the Mississippian formations in the territory surveyed. In Pennsylvania it is subdivided into the Cusse- wago sandstone, Cussewago shale and Corry sandstone, of HIiGs or Fic. 3. Sections showing the Orangeville formation. which the first two are unimportant because they are only local in extent. The Berea is to a large exteut the key horizon to the stratigraphy not only of the district under discussion but also of the region to the south. Therefore much time and care were spent in tracing it accurately. All described exposures were visited and many new sections involving its nature or position were made. A careful search for fossils was also instituted, but none were found. I.C. White, who has perhaps done more detailed work in this region than any other geolo- gist, mentions finding fossils only at two localities.* One is at Corry, where, he states, the formation contains a “ few ill-pre- served fossils.” The other is “ east of the county line (Craw- * Second Geol. Surv. Pa., vol. Q*, p. 92. 306 W. A. Verwiebe—Correlation of the ford), in Warren County on the road to Enterprise,” where “ it is finely exposed and very fossiliferous near its base.” This last-named place was visited by the writer and it was found that the exposure showed not the Corry sandstone, but a sand- stone located stratigraphically about 180 feet lower, the first Venango sandstone. In the volume mentioned above White has described numerous sections involving the Corry sandstone, but nowhere does he mention the occurrence of fossils, although that is his custom in the case of other formations. It appears, therefore, that the Corry must be considered essentially an unfossiliferous sandstone. This conclusion does not seem to be in harmony with the findings of Dr. Girty. In a paper recently published,* he aims to show that the Bedford formation of Ohio is Devonian inage. The Bedford is admittedly found beneath the Berea and in order to strengthen the case the fauna of the latter is adduced and an attempt made to show that it is Carboniferous. Inasmuch as the Berea is practically barren in Ohio, its corre- late the Corry is analyzed faunally. No sections are given and the only clue to guide the reader in deciding where the fossils were found is the following statement : “The ‘ Corr 5a horizon carrying the fauna can be traced eastward to Cobham’s hill just east of Warren, where it comes in immediately above what has been called the ‘sub-Olean conglomerate’ (Knapp formation), in the short interval which separates that formation from the Olean conglomerate.” This locality (Cobham’s hill) was visited by the writer, and it was found that the fossilifer- ous horizon corresponds stratigraphically to the Venango first sandstone. This can betraced very easily along the Allegheny River toward tke south in the numerous rock cuts along the railroad, and is found to underlie the Berea or Corry by an interval of about 180 feet. It is very easy to see, however, how these two sandstones might be confused. Both are very similar lithologically and both are overlain by bluish shales, the Orangeville and the Riceville respectively. One important guide to their separation is a layer of very calcareous sandstone at the very base of the Corry. This resembles the Meadville limestone in every detail, breaks with a conchoidal fracture, and is usually very hard and covered with a ferruginous crust. In addition the base of the Orangeville is generally quite fos- siliferons, the characteristic genera being Lingula and Discina. This is not true of the Riceville formation above the first Venango sandstone. If it should be found that a confusion of sandstones has occurred here then a peculiar anomalous situation arises. Since the Venango sandstone lies considerably below the Corry, the * Ann. N. Y. Acad. Sci., vol. xxii, p. 295-319, Nov., 1912. Mississippian of Ohio and Pennsylvania. 307 equivalent of the Berea, then it must also lie stratigraphically beneath the Bedford. Thus we should have a formation carry- ing fossils which show a distinct affinity with Carboniferous species below another, the fauna of which is distinctively Devonian. To escape from the enigma three possibilities sug- gest themselves ; either the Bedford must be considered as Carboniferous in age or the ‘Corry’ as Devonian, or, the fauna of one or the other is not sufficiently conclusive. Hie. 4. 2B Fie. 4. Sections showing the Cuyahoga formation. Mississippian formations of Ohio. Royalton formation:—This formation was named by Dr. Prosser from Royalton township in Cuyahoga County, Ohio.* It consists largely of shales both arenaceous and argillaceous, which are generally bluish or bluish-gray in color. Inter- bedded are thin, impure sandstones many of which are full of Sptirophyton markings. It includes the Sharpsville and Mead- ville formations of Pennsylvania and is equivalent to the upper part of the Cuyahoga of central Ohio. Also it is quite probable that it includes the Shenango of Pennsylvania and the Logan of centra] Ohio. However, since the unconformity between the Mississippian and Pennsylvanian has cut down lower in northern Ohio than in central Ohio, it is only in favored spots that we may expect to find the Royalton thick enough to include the stratigraphic horizon of the Logan. * Geol. Sur. Ohio, Bull. 15, p. 493, 1912. 308 W. A. Verwiebe—Correlation of the Such a locality as this is found two miles east of Strongsville, Ohio, along Willow brook. This section is No. 107 in fig. 5 and shows 172 feet of Royalton exposed with about 60 feet covered, a total of 232 feet. Brecksville :—This name was applied by Prosser to the upper and largest part of the terrane which is equivalent to the Or angeville of Pennsylvania. The other members are the Aurora sandstone and Sunbury shale. The formation consists Rico: Fic. 5. Sections showing the variation in the Mississippian from west to east. of shale generally dark in color and prevailingly argillaceous. In thickness it varies from about 70 to 115 feet. Aurora sandstone :—This is a rather local member of the Orangeville terrane in Ohio. It was named from Aurora creek in the northwestern part of Portage County, Ohio.* Here it is a blue, fine-grained sandstone six feet thick with some thin shale partings. Elsewhere it is frequently a single unit and of considerable value as a stratigraphic horizon. It is clearly recognizable as far east as Warren, Ohio, but beyond that it soon loses its identity. Sunbury shale :—This is a familiar term to a stratigraphic geologist. The name was first applied by Prof. Hicks (1878) and has since been discussed a good deal in connection with * Geol. Sur. Ohio, Bull. 15, p. 211, 1912. Mississippian of Ohio and Pennsylvania. 309 Ohio geology. The black, rather tough, strongly Jaminated shales to which it is applied are, though thin, a well marked lithologie unit. When traced from Ohio into Pennsylvania this formation loses its distinctive character and becomes an inseparable portion of the Orangeville shale. ‘The black color is noticeable as far east as the longitude of Linesville in Craw- ford County, but the hard compact texture is perceptible some- what farther east. The fauna is also characteristic and is the same as that of the Orangeville formation. Mississippian as a whole. The correlation of the formations enumerated has been given in connection with each. In conclusion a few words may be added regarding the Mississippian as a whole. Figure 5 shows a number of typical sections across the region under discussion which will indicate the variation in thickness and character from west to east. The first section (107) has a considerable portion covered. As explained elsewhere, this interval most probably consists of Royalton shales. The great irregularity in thickness is explained largely by the unconformity between the Mississippian and Pennsylvanian systems, still there is evi- dence of considerable thinning toward the east. In sections 31, 47, and 74 the Sharon comes to rest upon formations below the Shenango sandstone or its equivalent, whereas this as well as the overlying shale forms part of the sections farther east. Another prominent feature is the tendency to increasing coarseness toward the east. All forrmations show this, so that the whole system changes from a shaly aspect in the west to a sandstone and conglomerate facies in the east (along the Alle- gheny River). Still farther east the whole series becomes the Pocono formation. The following table gives a summary of the formations and their correlation in different areas : Mississippian Pennsylvania Ohio east west north central ( Mauch Chunk : Greenbrier ! f Shenango shale | : (Shores) Shenango sandstone | oe 8 ee MW ee ee + Royalton agaist sors i Meadville | | | “~DSVi | Pocono 4 Ppavile J . Cuyahoga Brecksville | Orangeville Aurora Sunbury Sunbury | ie | ete | Cussewago shale & + Berea Berea [ [ sandstone 310 W. A. Verwiebe—Correlation of the Unconformity between Devonian and Mississippian. It will be noted that in the above table the Berea has been considered the basal formation of the Mississippian system. A good deal has been written to determine the proper place to put the dividing line between the Devonian and Mississippian and a few words of discussion on this subject may therefore not seem out of place. In the article by Dr. Girty discussed above some excellent reasons are given for drawing the line at the base of the Berea. The fact that the Berea inaugurates a new series of rocks as a basal conglomerate (though it is most commonly a coarse sandstone), secondly the fact that the under- lying rocks are extensively eroded giving rise to a disconform- ity, are strong arguments in favor of this view. The impor- tance of this disconformity is not admitted by all geologists. Professor Cushing, for instance, holds that it is merely con- temporaneous erosion. Ina paper published in the Bulletin of the Geological Society of America® he states certain condi- tions which should be present in order to indicate a gap of sufficient length of time to entitle the break to be considered of diastrophic importance. The first is that oscillation upon which an unconformity depends must be accompanied by some warping. He finds that nowhere does the Berea rest upon any formation but the Bedford (ranging in thickness from 50 to 100 feet) and assumes that therefore no warping has taken place. This reasoning is no doubt very good. Still, would it not be perfectly possible that a negative movement of the strand line involving a narrow strip of continent, say one hun- dred miles wide and extending roughly east and west, might be caused by a slight lowering of the sea level ? Assuming that the shore line during the Bedford age lay to the north approx- | imately within the present confines of Lake Erie and that the shales and sandstones of that age were being accumulated in a shallow sea gently sloping toward the south, does it seem unrea- sonable to suppose that the strand line shifted slowly south until it reached approximately the latitude of central Ohio ? Such a shifting would explain very nicely the fact that the Bedford shale increases in thickness toward the south and that ’ the sandstones are present in the south and missing toward the north. The Berea formation then represents the coarse phase of deposition which should accompany a sea transgressing the region again toward the north. And again we should expect that the Berea would show a less thickness in central Ohio and a greater thickness in northern Ohio, which an analysis of the facts proves to be true. If we accept this condition of affairs * Vol. xxvi, p. 205-216, June 15, 1915. Diastrophic Importance of the Un- conformity at the Base of the Berea Grit in Ohio. Mississippian of Ohio and Pennsylvania. 311 it will not be necessary to have any warping in the region affected by the unconformity. On the other hand, it has not yet been proved that the Berea nowhere cuts down below the Bedford. In fact Mr. Burrows* cites one example in Lorain County, where a channel 175 feet wide and extending down considerably into the Cleve- land shale (which underlies the Bedford) was carved in the interval between Bedford and Berea times. It is quite likely that further search will reveal others that penetrate beneath the Bedford. Even if we disregard this evidence, however, and look at the matter from another angle, we may say that the sub-Carbonifer- ous period witnessed the inception of the epeirogenic move- ments which culminated in the formation of the coal-swamps of the upper Carboniferous in eastern United States. The Laurentian shield has always been a positive segment of the continent and it does not seem improbable that one of its pre- liminary movements brought a section three hundred miles wide into relief without appreciable warping. Furthermore, the coal measures offer abundant testimony of periods of emergence of slight relief, so that we may assume that northern Ohio also emerged but slightly and that therefore the agents of erosion consumed a great deal of time in cutting through the sandstones of the Bedford and into the red, blue, and black shales beneath. Another bit of evidence from another direction may not seem amiss. Sufficient well records are now available to enable one to trace the Berea from its outcrop in central and southern Ohio across the eastern part of the state well into Pennsylvania. In a similar way the writer has traced the Corry into southern Pennsylvania from its outcrop in north- western Pennsylvania.t By this method it will be found to correlate pretty definitely with the base of the Pocono. Since this formation is universally admitted to represent the base of the Mississippian in the east, it should indicate that its corre- late, the Berea, also occupies the same horizon. It must be admitted that the problem of finding the divid- ing line in Ohio is a difficult one and that the evidence at present available is insufficient to warrant a clear-cut conclu- sion; still, the writer is of the opinion that further evidence will place it at the base of the Berea. Summary. The correlation of the Devonian and Mississippian forma- tions in northwestern Pennsylvania and northeastern Ohio is a * Burroughs, W. G., Berea Sandstone in Eroded Cleveland Shale, Jour. of Geol., vol. xxii, p. 766, 1914. + This Journal, vol. xlii, p. 51, 1916. 312 W. A. Verwiebe— Correlation of the problem which has not been satisfactorily solved up to the present. In this paper an attempt is made to correlate the Mississippian formations in these two areas, preliminary to a discussion of the Devonian. Also an effort is made to defend the proposition that the Berea is the basal formation of the Mississippian. New Sections bearing on this Discussion. The following hitherto unpublished sections were made by the writer in the territory under discussion, and, since they Fie. 6. vLe NEW YORK FX | aus | Mead ville 3a. 6Q PENNSYLVANIX Fic. 6. Map showing the location of sections platted in figs. 1 to 5. throw new light on the stratigraphy are deemed worthy of publication : la. Sharon Section. Thickness No. Feet. 6. Sharon :—Sandstone, coarse, friable, white quartz grains, but weathering buff pebbles scattered thr oughout j in IONSES. 2. os eh ee ree por ae cae ae ee ae 18 5. Shenango:—Sandstone, grayish white, rather loose tex- tured, numerous 1ronespots 222222 2k 4 ee 4. Shale, bluish, largely argillaceous, with many sandstone layers, some of which are buff colored. .._-----.--- 40 3. Sandstone, white, quartzitic, turns reddish brown on weathering Da 5 oa ek RE Ep a oe 7 EN eee 10 2. Covered to quarry of Sharon Clay Products Company 25 1. Meadville :—Sandstone and shale, mostly hard, sandy and blue imcolor.c is: s2c222eee Oe eee 34 Mississippian of Ohio and Pennsylvania. 313 This section was secured partly in the quarry of the Sharon Clay Products Co., above. Both are Ieriad about one mile west of Sharon. 38a. Buchanan Ravine. and partly in an abandoned quarry just KS Wo do OO = Cre oO = ke bo ko Thickness No. Feet. 29. Meadville. Shale, bluish, argillaceous ; some sandstones MORE MMCOM COC tee es os oe ue ee eee 9 28. Mead. upper Limestone. Hard, blue, siliceous limerock in four layers, fossils abundant,-also pebbles. --_- - - -- 1? 27. Meadville lower. Shale, bluish, argillaceous, some sand- SCTE) Sa ae ee ee A) pis eee MR yee ee nee Se Sears 15 Jeo tale and sandstones: tsi zckoes- jeock eee ceediccas 1 Zoe SANOS hone laweGeees 63 24o0) ea do, See ees Zoe piiulevandisamdstomesiac) 225. sos) sik. le eke it 23. Mead. middle Limestone. Sameas No. 28 ; 2 layers. .- 22, Sharpsville. Sandstone, flaggy, brownish in color--__.- 1 21. Sandstone, fine-grained, bluish grey, micaceous --.-.- -- 1 20. Mead. lower Limestone. Same as No. 28 ; 2 layers.--- 19. Sharpsville Lower. Sandstone, massive layer ---.----- 18. Sandstone, thin layers from a fraction of an inch to 6 or 7 inches thiekeven-bedded. 2 22. noe 20 17. Orangeville. Shale, blue and grey, soft, ar gillaceous, McdL ers) CO ayRuUsty, COlOn 2i2)1.5 40) sero. ee dee 8 hee Sangstone: dull chocolate coloric2s 2e)i)o: yosbs 22. fe 15. Shale yas above; butemore sandy 2-022 22 Ja ele 7 if) Shale--as) above: itossiliferousi. 4:22- 20 eed =| 19 Beer Veber slant pein ae Sake y ee aL 11 12. Shale, same as No. 17, interbedded with thin, hard MICACeOUS Sanastones, TOSSIIS 25.252 25. 2 2 ede 23 11. Sandstone, dull brownish grey, with black carbonaceous SURE AILS SS Oe ss eee Senrie B ee ge eee 3 10. Sandstone, hard, blue, calcareous and ferruginous ; con- tains a great amount of marcasite_.__..--.-------- s 9. Shale, blue, weathers rusty on outcrop.-_------ Te 3 53 Sy oandstone, dull oray; fine-oramed esse. 2.2L S222... Oe ie uae bide arovllaceous ave. 222 eso elles e ek 1b 6. Sandstone, blue, rather fine-grained._-..-_------.--- 1 SS GNCR CUR naar ne tween Oe SE eS ee te 4. Corry. Sandstone, coarse, rather friable, massive, fer- ruginous, as shown by deep crust of weathering ---- ee OC cerca meme aren NINE Dea OSS ol 2 ae Ase NOM meen ce ME eR A Soe ee 1. Cussewago. Shale, blue, argillaceous ; thin sandstones, TEAL UGS SINS) acs sea aloes, a Ne a 25 This section was secured a few miles south of Meadville on the farm of Mr. David Buchanan. It is just opposite the - small station of Buchanan on the Erie R. R. The section is a 314 W. A. Verwiebe—Correlation of the valuable one, inasmuch as it shows the formations from the upper Meadville down through the Corry and well into the Cussewago, in great detail. The Corry is here shown to have a thickness of 13 feet and 8 inches and the Orangeville of 81 feet and 10 inches, while the Sharpsville is 55 feet 11 inches and the lower Meadville 40 feet and 8 inches thick. 6a. Shaws School House Section. Thickness No. Feet. 15. Shenango Sandstone. White quartz rock, weathering buff and brown, massive layers, iron concretions.... 11 14. Meadville upper. Largely thin sandstones with some argillaceous.shales interbedded _ 22/22. 224 722 S22 eeeee 13. Mead. upper Limestone. Blue, siliceous, conchoidal. - - 3 12. Mead. lower. Shale, bluish, with numerous sandy layers 26 11. Sharpsville upper. Sandstone, brownish gray, irregu- lar layers’ 222200832 2 ORO Ahab. en 25 10. Mead. lower Limestone. Fine development, breaks up into several layers ; much iron, concretionary - - --- - - ies 9. Sharpsville lower. Sandstone, irregular layers, dull chocolate color, lagey. 2 51222 Weer Blt eae 22 6... Limestonen oo c2ccc ec Ue ee ee ee 3 7. Covered:2. te 20 2 ie. Ser. Bu eee ee 54 6. Sandstone, very thin, bluish grey, separated by a few thin-zones of clay shalewi2s Bila su a0) 1S ee eer 5. Orangeville. Shale, rather arenaceous; upper part is bluish black in color and more argillaceous; middle largely occupied by thin sandstone ._.....--_---.--- 24 4. Shale, blue, argillaceous; thin sandstone ..--.------- 254 a. Covered. 200. peacoat PRG 342 ye eee ae ee 18 2, Sandstone, ash gray, hard, fine-grained_ _.:_.22-) 22289 3 1. Shale, typical, rusty outerop.-_- 22.2 322 3 aa This section was measured in the ravine which passes the Shaws schoolhouse. Shaws is a small station on the Erie rail- road about eight miles south of Meadville, Pennsylvania. 182. Torpedo Section. Thickness No. Feet. 12. Conoquenessing. Sandstone, very massive, but poorly cemented and friable; composed of white quartz grains which weather buff on exposure. Scattered through the mass are seams and lenses of small pebbles, especially near the bottom; many seams of iron or rather bands of quartz highly charged with iron also occur. Interesting are the clay dikes, some twenty feet high, which fill the joint planes in the TOCK... 2: / BolT. ok wish wal es a aot ene eee eee Mississippian of Ohio and Pennsylwania. 315 Thickness No. . Feet 11. Sharon. Shale, black, fissile, tough, somewhat argilla- ceous In spots, carbonaceous and ferruginous, This seems to be disconformable under the Conoquenes- Sn ar ees ol PO Pe oe ee Pa oa ga I 10. Covered (section continued on road to south)._.._--- 54 9. Sandstone, hard, close grained, pebbles._-_...------- 3 Bee OCMC cen (oN Oe oe ee ea 43 7. Shenango. Sandstone, shaly, drab; some clay shale... 24 6. Sandstone, white quartz rock, weathering deep to buff and brown, massive, ferriferous.___...._.......... 12 5. Covered, with occasional shale or sandstone. -__-.__-- 270 4, Venango. Shale, blue, and thin sandstone...-._._... 12 3. Sandstone, single layer, weather flaggy, fossils.._.... 34 Dig SIDING eh pa a 2 wae L ok ae aig ene ie n me pe nu ean G2) We Oovered to railroad level.) 8. i The above succession of rocks may be seen by following the road from the works of the Warren Silica Co. up to the quarries from which the raw material is secured. Torpedo is a station on the D. A. V. & P. R. R. between Titusville and Warren in Warren County, Pa. Rea. Hyde Run Section. Thickness No. Feet. 15. Meadville. Sandstone, rather coarse, brown color; layeksr ane taithymassives ! ole e sae ake 21k 18 14. Mead, upper Limestone. Bluish, siliceous, conchoidal 14 13. Shale and thin sandstone, partly covered._--.------- 52 12. Mead. lower Limestone. ‘Same as above...-.-------- 3 11. Sharpsville. Shale and thin sandstone, bluish grey and LORCA AMS eS ee Ns a MON TO OVvened mi ta eae LOT ee Soe 80 9. Orangeville. Shale, hiss and drab, soft, argillaceous, mire seamed: byinonese eM aoe e en ies 45 8. Corry. Sandstone, thin-bedded toward top, shaly in center and massive below (lowest layer is 21 inches thick); color is buff throughout and the rock is of fine OX C ME ae ee ae Ber I Lee 14 Cussewago. Shale, blue, argillaceous__.--.-...------- 4 Cussewago Limestone. Very typical, a blue siliceous lhmerock which also carries much iron. On weather- ing the lime is dissolved, leaving behind a crust of deeply stained ferruginous sand. Here the limestone at OCCURSMIMMEMORlAVeS oS Meet se kl be 5. Shale, soft, argillaceous, bluish; lithologically very munchies tier Oraneevilles 292 726 ell 44 Ame CONROE etme hu et. AD 3. ficeville. Shale and thin sandstone.....------------ 8 9. Sandstone, thin-bedded and shaly._.....----------- 12 1. Covered to Thompson’s Run. 316 W. A. Verwiebe—Correlation of the A few miles northwest of Titusville is the village of Hyde- town, through which flows Thompson’s Run. A branch of this run just above the village is called Hyde Run. Along this creek the section was seen. 24a. Run on Simmons Farm. Thickness No. Feet. 8. Shenango. Sandstone, medium to coarse grain, brown 15 7. Shale and thin layers of sandstone -_--_-2_. (222 13 6. ‘Covered Sree ceiaes 2. 8 5eee ee eee ane 33 5. Meadville. Sandstone, brownish, flaggy, rather coarse... 5 4. Shale and sandstone in thin layers, largely covered__- 110 3. Orangeville. Shale, blue, argillaceous -.__- .--2 = =) es 2. “Goverédae 2 ki sS ke ee ee ae ae ee er 1. Corry. Sandstone, massive, buff, hard, close grained. Middle section shaly and upper part less massive than:thelowetso.' 222 foes 34 ea 15 In the section above it is more than likely that the Meadville formation begins with No. 5. Im that case the Shenango sandstone is concealed in the 33-foot interval (No. 6). The general topography of the stream also seems to bear this out. The part of the section including No. 2 to 5 was obtained by following the road leading to Thompson’s Mills. The Corry may be seen directly below the bridge over Thompson’s Run, about one-half mile below Kerr’s Mills, which is a few miles north of Titusville in Crawford County, Pa. 25a. Wolfkill Hollow Section. Thickness No. Feet. 24. Shenango Sandstone. Massive, white quartz sand, coarse and rather friable, iron concretions___._.--.-- 1 23. Meadville. Sandstone, weathered to a deep brown color; layers fairly -«massives®.2624..2 527 ee 40 22. Mead. upper Limestone. Blue, hard, conchoidal frac- ture, ierruginous crust of decayed rock._.-...---.- 14 21) Sandstone, . coarse; brownish ._.. .... 2-22. ee 4 20. Shale, thin, even- -bedded, eray, sandy 2.2 5..cksaeee 4 19: Sandstone, ‘irregular layer, coarse, greenish..___..-.- $ 18. Shale, same as Wor B0i nce cans de ob geen 17. Sandstone; layers iup-to-3 am! thiek ei: Si sa) ee 2 16. Shale and sandstone, partly covered...-_:-2.2:4-25- 23 15. Mead. tower -Limesione.- take Nose? See 4 14. Sharpsville. Shale, bluish gray, weathering to deep brown, thin, sandstones interbedded__._.---------- 10 Mississippian of Ohio and Pennsylvania. 317 Thickness No. Feet. 11. Sandstone, massive, close-grained, buff.--..-_----.-. Z Mees ialenisorts-avab:) fossils...) te. ot ee eee a 8 Wee camascone. Hike No. Tl. 2 0o ooo oe) Clee eee cole le + 8. Sandstone, irregular, flaggy, some shale.......-.---- 9 actrees Covered’) lS Ue ee eee ete 4a 6. Sandstone, buff, not massive, thin layers._-.._------ 5 5. shale and sandstone, bluish grey ...--_.._.......--. 2 Paimestones blue, fairty: hard’... ol2> 622 yo Pe 4. =a phale-and sandstones, bluish, fossils... _.-..__2 = 22 8 2 CU Wig ee oe Be Pe ahi a ee 1. Shale and sandstones to railroad tracks_............ 8 Wolfkill Hollow is a deep ravine with all the earmarks of an interglacial valley. It empties into Oil Creek just below the station called Miller Farm, on the Pennsylvania railroad between Titusville and Oil City, Pa. The section shows that in this region the upper Meadville is a coarse, brown sandstone and so similar to the Shenango sandstone that it becomes dif- ficult to separate them. The Corry, unfortunately, is not exposed, but is quite probably concealed in the covered inter- val No. 7 26a. Sprott Ravine Section. Thickness No. Feet. 15. Shenango Sandstone. Coarse white. quartz rock, massive, layers up to three feet thick; on weathering SELMER Sea CM GUO eo Ob eee kes Os ea ny 14 Peles lie caretllaceons. 258. 6 ok 2 See Sano scone MRE UN Gt hos 0 ee Lk. 24 12. Meadville. Sandstone, deeply weathered to a brown color, layers are thin, mostly less than 3”: the rock is GF medium, cram and micaceous_...- 122... _-_..---- 13 11. Sandstone like No. 12 but more massive.._...------- 254 10. Mead. upper Limestone. Blue, hard, in two layers, dees rounded acsvusgualo ole AS ee 13 9. Sandstone, as above, but more flaggy ._....--.------ 63 Se eecuale, bine, arellaceousoe. «82-022. ose 7. Sharpsville. Shale and thin sandstones, mostly Segeete es sc een nner Sai 2 22 6. Mead. lower Limestone. Same as limestone above... 4 5. Covered to outcrop at junction with Hammond Run. 88 4. Orangeville. Sandstone, buff color, hard but rather Pi Pl, Lbs eee ee Oe he 3 Se oiaelue ang drab. sandy. 222252)... 93 2. Sandstone, oie massive layer, buil color: —..--.--..-- lz 1. Shale, chocolate color, largely argillaceous--_---.----- 4 This section was made in a small ravine crossing the road just above the farmhouse of W. H. Sprott, two miles southeast 318 W.A. Verwiebe—Correlation of the Mississippian, ete. of Titusville and leading into Hammond Run below. It also shows: the character of the upper Meadville in excellent fashion. But for the presence of the upper limestone it might well be mistaken for the Shenango sandstone. 29a. Markley Quarry Section. Thickness No. Feet. 10. Shenango Shale. Shale, bluish grey, somewhat argilla- G@OUS22< SO +2 See Be a eee 2 9. Sandstone, brownish... 235. sa 5 13 8. Shale, sandy2. 2225-8655: 234 Ae ee 7. Sandstone, coarse, ferruginous, whitish grey--.--._-_-- - 6..;: Shale,-blue, clayeye_ S225. Soe Ao 52 5. Shenango Sandstone. Sandstone, somewhat flagegy, more massive toward the top, whitish grey, mica- GOOUS 225 29h. UY eS ee Eee ee ee eee 26 4, .Same, single layer: .2)6 5.534 225 25+ Bi, ,Cowercdh sss 3 oho a eat 2. Corry. Sandstone, white quartz rock weathering to buff, massive, layers irregular ~_2> 2222 .. ==) 53 eee 4 1. Riceville. Shale, bluish grey, partly argillaceous with , interbedded sandstones to R. R. level__-----_- .--- 38 Just south of Titusville, Pa., is the abandoned quarry of O. W. Markley. It was worked for the sandstone of the Shenango formation. The section then continues down to the Pennsylvania railroad tracks nearby. The importance of this section is that it shows in great detail the character of the Shenango shale and underlying sandstone. A study of it will show that the term Shenango Shale is a misnomer, since it consists largely of sandstone and that the term Shenango should cover both phases in other parts of the district as well as this. List of Sections used in Figures 1-5. The following sections are taken from Bull. 15 of the Geol. Survey of Ohio: No. 16, page 72; No. 31, p. 143; 38, p. 158; 47, p. 194; 74, p. 316; 79, p. 336; 82, p. 353; 86, p. 363; 89, p. 380; 90, p. 382; 92, p. 399; 107, p. 489. The following were published in vol. xliii of this Journal, July, 1916: No. 13@ on page 56; 21a, p. 56; 22a, p. 57. The following may be found in vol. Q4, Second Pa. Geol. Sur- vey, 1881: 6a, p. 194; 8w, p. 1825 llw, p. 1395 120, p. 173. No. 2B was taken from the general section published on page 193 of the Report of the Top. and Geol. Surv. Comm. of Pa. for 1906-1908. The following will be found published in this article: la, p- 312; 3a, p. 313; 6a, p. 314; 18a, p. 314; 24a, p. 316; 25a, p- 316; 29a, above. W. J. Sinclair—A New Labyrinthodont. 319 Art. XX VIII—A New Labyrinthodont from the Triassic of Pennsylvania ; by W. J. Stvcratr. Remains of vertebrates from the rocks of the Newark group are so rare that every additional specimen obtained from them cannot fail to be of interest. The material which forms the subject of the following note was collected by Professor Gilbert Van Ingen and the writer from the locally basal beds of the Norristown shale at Holicong, Bucks County, Pennsylvania. It comprises the front part of the left half of the lower jaw of a large Stegocephalian, No. 12302 Princeton University Geological Museum, comparable in size with A/astodonsaurus, and apparently pertaining to an undescribed genus and species for which the name Calamops paludosus is proposed. Two large pieces, evidently belonging to the back of the jaw, do not make contact with the tooth-bearing portion of the ramus and have been omitted in the drawing (fig. 1). They aresome- what fragmentary and so heavily covered with matrix that little or nothing can be made of them. The jaw was found in a red shale abounding in impressions of large horse-tail rushes which are observabie on the surface of the hard siliceous capsule investing the bone (whence Calamops, “reed-face”’). This capsule is of variable thickness up to half an inch or so, is apparently siliceous, of radiate structure, and is applied so closely to the outer surface of the bone, and has become so much a part of it, that its removal has proved impracticable, especially as the bone-substance is - quite soft and spongy, owing to partial solution. The portion of the jaw preserved is straight lengthwise, except toward the forward end where it curves inward toward the symphysis. Here the matrix is broken off and bone tissue exposed below the base of one, or perhaps two teeth. The upper margin of the ramus carries a series of seventeen large teeth withont sockets. Perhaps an additional small tooth or two occurs in front of the first large tooth in the series. The matrix has been chipped off the tips of several of the teeth giving some rather imperfect cross sections. These have been supplemented by a thin section of the penultimate tooth, cut transversely. Apparently the teeth are without enamel, with smooth or but slightly grooved crowns, circular in cross section at the base and, in the case of some of them, with the crown slightly flattened laterally toward the tip, with acute anterior and posterior edges. In thin section, cut transversely to the crown, radial prolongations of the pulp cavity are seen toward which numerous fine dentine tubules seem to converge from Am. Jour, Sct.—FourtH Series, Vou. XLIII, No. 256.—Aprin, 1917, 2 3 2 0 Hie. 1, 1 Museum, left half of 1ca ty Geolog three-eighths natural size. iversl Princeton Un de, inner si men, No. 12802, the ype speci mandible from Calamops paludosus, t Bre: lk W. SJ. Sinclair—A New Labyrinthodont. 321 the outer wall of the tooth. Labyrinthine infoldings of cement have not been made out. The radiating pulp canals are especially well shown on the tenth tooth im series, counting forward from the back of the jaw, where an attempt at remov- ing the exceedingly hard matrix resulted in chipping off the outer wall and exposing the ends of the pulp canals which ap- pear as a series of vertical grooves. The anterior teeth are somewhat variable in height, but none seem to be specially enlarged unless the tooth supported on the base exposed at the broken symphysial margin might have beenso. The last three teeth are much shorter than the rest. Although their tips are broken off, their bases are considerably higher than those of the teeth in front owing to a sudden increase in depth of the jaw. Externally, the teeth anterior to these three are sup- ported by a flange from the jaw, only their tips projecting above it. Owing to the thick capsule of hard material adhering to the bone, none of the jaw elements can be made out, which renders difficult a comparison with forms hitherto known. On the outer side of the jaw, the matrix has an irregular hummocky surface suggestive of pittings in the bone. The latter is flat behind and slightly convex in front in vertical section. In- | ternally and toward the back of the jaw, a wedge-shaped depression is inclosed between a strongly convex ridge above and a lower ridge below. ‘This depression fades out anteriorly until below the eleventh tooth, counting forward from behind, the jaw is strongly convex in vertical section. Its lower mar- gin is broadly concave upward, increasing in curvature toward the symphysis. Calamops is larger than any labyrinthodont hitherto described from the Newark group and is the largest amphibian thus far reported from the Triassic of North America. The fragment figured has an extreme length, measured in a straight line, of 446 millimeters ; at its narrowest part below the four- teenth tooth, counting forward from the last, the depth of the jaw is 55 millimeters; measured vertically below the base of the last tooth it is 118 millimeters. Owing to the heavy coat- ing of matrix, not even approximate figures for the thickness can be given. The longest tooth rises 48 millimeters above the inner, upper margin of the jaw and is about 16 milli- meters in diameter anteroposteriorly at the base. - Princeton University, Department of Geology, January 1917. 322 G. P. Merrill—Calcium Phosphate in Meteoric Stones. Art. XXIX.—QOn the Calewum Phosphate m Meteoric Stones ; by Grorcr P. Merritz, Head Curator, Department of Geol- ogy, United States National Museum. [Published by permission of the Secretary of the Smithsonian Institution.] . Unper the caption On the monticellite-like mineral in mete-- orites, | have elsewhere* described a previously unrecognized form of calcium phosphate occurring among the silicate con- stituents of stony meteorites. Examination of a large number of thin sections of other stones than those there mentioned has shown the occurrence of this mineral to be so widespread as to warrant a brief note devoted exclusively to it. The first noted occurrence of a mineral phosphate in a meteor- ite was that of C. U. Shepard, who in his description of the Richmond, Virginia, stone mentions the occurrence of a minute, but macroscopically visible, yellow mineral, particles of which reacted for phosphorus and calcium and which he interpreted as indicative of the mineral apatite. In this he was essentially right, but singularly enough the discovery remained unverified almost to the present time (see my paper), nor was apatite again reported until Berwertht found it among the silicate con- stituents ot the Kodaikanal iron, and Tschermakt in the stone _of Angra dos Reis. Singularly enough, too, the phosphorus reported in chemical analyses of meteorites is in the majority of cases relegated to the metallic constituent, a matter to which attention was called in my paper on the minor constituents of meteorites published by the National Academy of Sciences.§ The cause of this is not difficult to determine, as noted in the following descriptions, in which I have repeated, in part, mat- ter given in previous papers, adding the results of more recent observations, particularly in slides of the New Concord and Waconda stones. The New Concord stone, a veined intermediate chondrite according to Brezina’s classification, fell in 1860. It has been the subject of much study, and need be referred to here only in connection with its mineral and chemical composition. Dr. Smith’s analyses| showed the stone to consist of 10°7 per cent nickel-iron and 89°37 per cent “earthy mineral,” the nickel- iron yielding 0-012 per cent of phosphorus, but no mention is made of this same constituent in the “earthy” portion. An examination of the stone in thin sections, a method not avail- able in Smith’s time, reveals the presence of numerous areas of the phosphatic mineral 1 have elsewhere described and _ pro- * Proc, Nat. Acad. Sci., vol. i, p. 302, 1915. + Min. Petr. Mitt., vol. xxv, p. 188, 1906. t{ Ibid. xxviii, 110, 1909. § Memoirs, vol. xiv, p. 27, 1916. In the 42 selected analyses there tabu- lated, the phosphorus is given as P in 24 cases and as P.O; in the remaining 14. | This Journal, vol. xxxi, 1861. G. P. Merrill—Caleium Phosphate in Meteoric Stones. 323 visionally referred to francolite. Three of these occurrences in a single slide (No. 62-a) are figured herewith, the actual diameters being from 0°5 to 1:5"™, The mineral is in all cases colorless, without crystal outline (its form being controlled by that of the interstices in which it occurs), no definite cleavage, a complete lack of pleochroism, polarizing if at all only in light and dark colors, sometimes almost completely isotropic, and giving at best poor interference figures indicative of its biaxial nature. The isotropic sections are so colorless and lacking relief that on hasty inspection such might be mistaken for holes in the slide, a glass, or even the problematic maskelynite. Once t seen, however, they are readily Ld Ree recognizable. In the three forms : here figured the position of maxi- mum extinction is shown by the small black cross. In the case of the larger form, the black brushes ot the hyperbola emanating from a point outside the field of view, indicative of its biaxial character, are readily obtained. The phos- phatic nature of the mineral has been determined beyond question by the usual microchemical method. An attempt at a quantitative deter- mination was made, but with unsat- isfactory results. Studies of thin sections of the Waconda, Kansas, stone, a rede- scription of which is in process of preparation, reveal the presence of the phosphate even where it can not be determined microscopically. It was found that when the surface of an uncovered slide was treated for but a few minutes with a dilute acid the solution obtained would react for both calcium and phosphorus, and the slide when again placed under the microscope be found to contain numerous minute and irregular cavities left by the dissolving away of the mineral. A quantitative analysis shows the presence of 0:26% P,O, in the stony portion, or 0°23% in the bulk or mass analysis. Unmistakable evidences of the presence of the phosphate I have thus far found in the stones of Alfianello, Bath (South Dakota), Bluff, Dhurmsala, Estherville, Farmington, Felix, Homestead, Indarch, Knyahinya, Mocs, Plainview, Pultusk, Quenggouk, Rich Mountain, and Waconda. 324 G. P. Merrill—Calcium Phosphate in Meteorie Stones. These results agree in all respects with those given In my previous paper, and show with seeming conclusiveness that a calcium phosphate is a very general if not universal constituent of meteoric stones, and that, further, it differs from normal apatite in its optical and other physical characteristics, which may be summed up as below, a repetition in part of what I have previously stated : Occurrence sporadic, without crystal form, very brittle, colorless ; cleavage for the most part lacking though sometimes impertect and interrupted, showing angles of 60° and 120°; optically biaxial and positive (4), birefringence weak, less than 0-005, refractive indices a=1°623+0:002 and y=1'627+0:005 ; no pleochroism and often undulatory extinction, polarizing in light and dark colors, sometimes almost isotropic ; easily soluble in cold dilute nitric acid and less so in hydrochloric, giving solutions reacting for calcium and phosphorus. The mineral is further distinguished from normal apatite in that it is a product of the last rather than the first stages of consolidation.* Just what meaning and how much importance is to be attached to this different habit of the phosphate in the history of a meteor- ite it is yet too early to say. Obviously it bespeaks conditions governing crystallization unlike those which prevailed during the consolidation of terrestrial rocks. For the present it would seem best to designate this member of the apatite group by the name francolite, with which it most nearly agrees,t as was done in the paper mentioned above. It may be added that its common presence in optically recog- nizable quantities suggests the advisability of exercising greater care in the determination of phosphorus and calcium in chemi- cal analyses of meteoric stones. *In his description of the meteorite of Angra dos Reis, Tschermak men- tions the occurrence of abundant small colorless granules, without cleavage, weakly doubly refracting, uniaxial and negative. These he identified as apatite, a conclusion borne out by the chemieal composition of the silicate portion of the stone. The mineral I have described, which apparently is quite similar, is, however, biaxial. The possibility of the apparent biaxial interference figures being those of a uniaxial mineral cut parallel with an optic axis was considered, but deemed wholly improbable from the fact that in not one of the many sections examined was I able to find a uniaxial figure. It seems improbable that among so large a number should not be found at least one, did such exist. + See Schaller, Bull. 509, U. S. Geol. Surv., p. 91, 1912. i. V. Shannon— Crystals of Pyromorphite. 325 Art. XX X.— Crystals of Pyromorphite; by Hart V. SHANNON. 1. Pyromorphite from the Caledonia Mine, Coeur d’ Alene District, Idaho. _ CrrtTain specimens showing hexagonal crystals of some secondary lead mineral were collected by the author in February, 1915, in the Caledonia Mine near Wardner, Idaho. Recent study has shown this mineral to be pyromorphite but; although the crystals are quite simple and show no new forms, the habit and mode of occurrence are somewhat unusual. Occurrence.—The Caledonia vein consists in its unaltered portion of a rather large body of pure, fine-grained, argentif- erous galena with smaller amounts of chalcopyrite and tetrahe- drite. From.the surface downward for some 800 feet the vein maintains a rather high dip, but on the 900-foot level it flattens out into an almost horizontal position and forms a ‘ blanket’ ore-body of considerable size and richness. Below this level the vein is traceable as a gouge-filled fissure which has not been shown to contain any commercial ore. The inclosing rock con- sists of a much crushed white quartzite belonging to the Burke division of the Belt Series of sediments of Algonkian age. No igneous rocks are known from the vicinity. The zones of oxidation and secondary enrichment were well defined in the portion of the vein above the 700-foot level. In the oxidized ore, pyromorphite, of the pale green color common in the Coeur d’Alene District, occurred in crystal masses of extra- ordinary size and beauty, associated with cerussite, massicot, bindheimite, and native silver. The material here described is totally unlike the green pyromorphite in appearance and was_ not noted in the oxidized nor in the secondary sulphide ore. The crystals commonly occur deposited in cracks in masses of unaltered galena or in the wall-rock adjacent to such masses of galena, from the 500-foot to the 900-foot level. There is no other secondary lead mineral associated with the pyromorphite which has here apparently formed directly from the galena without the intermediate compounds, anglesite and cerussite. Pyromorphite in this district is a characteristic mineral in the extreme upper portion of the oxidized zone. Ordinary green pyromorphite has not been found here at a greater depth than 300 feet below the surface. No specimen is known, from the district, showing green pyromorphite associated with valena. It is of importance to note that the mineral described below occurs in some abundance in the lowest ore opened, 900 feet, vertically, below the surface. 326 E. V. Shannon— Crystals of Pyromorphite. Description.—The mineral occurs commonly in crusts of minute erystals coating cracks in galena or as larger individuals in cracks in quartzite. The color ranges from faintly pink to colorless in the smallest crystals to quite deep grayish violet in some of the larger ones. In size they range from microscopic to an occasional length of 1:5°™, the larger being those in the wall-rocks. Those over 5™™ in length are commonly nearly opaque with curved prism faces and brush-like terminations. The luster in the smaller crystals is adamantine while that of the larger opaque crystals is resinous. Quite commonly the crystals are attached by a prism face and are then doubly terminated, the form being essentially like the accompanying figure 1 with the length several times the diameter. At times Fic. 2. these small prisms greatly resemble quartz crystals, the resem- blance being heightened by an unequal development of the pyramid faces which gives them a rhombic aspect. As no suitable material was at hand when it was decided to describe the occurrence, a specimen of the type material was borrowed from Col. W. A. Roebling. The crystals measured were taken from this specimen. While promising in appearance these crystals are somewhat dull and give very poor signals. The forms noted were the pyramid a (1011) and the prism m (1010). No other planes have been observed. The best erystal, measured on the reflecting goniometer, gave the value va a’’’= 80°50’. This is sufficiently close to the caleulated angle (80°44’) to identify the form. The presence of lead, chlorine, and phosphoric acid were proven by qualitative methods, as were the absence of vanadium, calcium and arsenic. Col. Roebling reports that a preliminary EE. V. Shannon—Crystals of Pyromorphite. 327 analysis on insufficient material gave approximately lead 75 per cent and chlorine 3 per cent. The composition thus corresponds to ordinary pyromorphite. 2. Pyromorphite from Broken Hill, New South Wales. For purposes of comparison Col. Roebling very kindly loaned the author a recently acquired specimen of pyromorphite from Broken Hill. The specimen consists of an irregular mass of iron oxides completely encrusted with glittering microscopic erystals of gray pyromorphite. Measurement on the reflecting goniometer showed that the dominant forms present were the prism @ (1120) and the pyramid s (1121) of the second order and the prism mm (1010) of the first order with a small base and an occasional pyramid of the first order. The habit is thus quite unlike the Idaho material. Figure 2 shows the appearance of these crystals. | SCIENTIFIC INTELLIGENCE. I. CHEmIstRY AND FPuysics. 1. The Penfield Test for Carbon.—W. G. Mixter and F. L. Haicu have made an interesting examination of the delicacy of this method which was originally devised for testing for carbon and carbonates in minerals. ‘They observe that the method does not appear to be in the literature of analytical chemistry. It is novel only in the very delicate way in which the late Professor S. L. Penfield applied well-known reactions. He fused the sub- stance to be tested with lead chromate in a small horizontal hard glass tube, closed at the heated end, and containing near the open end a small drop of barium hydroxide solution. The appearance of a film of barium carbonate indicates the presence of carbon in the substance. If no film is seen when the fusion is effected, the open end of the tube may be closed with the finger to keep out carbon dioxide from the air and the tube may be removed from the flame so that the result may be more carefully observed. The method has been used by Professor Mixter in the Sheffield Chemical Laboratory for a number of years, especially for test- ing metals for carbon. It is so delicate that lead chromate which has been exposed to the air in preparation will react for carbon, but the reagent may be freed from carbon by heating it below its sintering point in an atmosphere of oxygen. Special precautions are necessary also to clean thoroughly the glass tubes used for the experiments, and to protect them from dust. 328 _ Scventifie Intelligence. In order to determine the delicacy of the method a mixture of very finely powdered silicon carbide (“carborundum’”’) and aluminium oxide, prepared by igniting the hydroxide in oxygen, was made, containing 9'990 g. of alumina and 0°01 g. of the car- bide. The mixture was heated for an hour in oxygen to burn out all carbon compounds except silicon carbide, and it was care- fully protected from atmospheric dust. Experiments were then made by fusing varying quantities of the mixture with lead chromate in sealed tubes containing in each case a drop of saturated barium hydroxide solution in a small depression blown near the open end, and provided at this end with an attached protective tube containing solid potassium hydroxide. The following results were obtained with different -weights of the mixture : 100 mg. gave an abundant white film. 20 mg. gave a distinct white film. 14 mg. gave a slight but distinct white film. 10 mg. gave a very faint white film. 8 mg. gave a doubtful result. 4 mg. gave absolutely no result. 1°5 mg. gave absolutely no result. Ten milligrams of the mixture was the smallest amount that gave an unquestionable reaction. ‘This corresponds to 0:01 mg. of silicon carbide containing 0-003 mg. of carbon. Undoubtedly smaller amounts of carbon may be detected by using a minute drop of solution of barium hydroxide and observing the result with a microscope. In doubtful cases a blank test should be made for comparison.—Jour. Amer. Chem. Soc., xxxix, 374. Hi, Us ava 2. A New Volumetric Method for the Determination of Cobalt.—W. D. Ener and R. G. Gustavson have devised a method for the determination of cobalt when nickel is present, which according to their test-analyses gives excellent results. Metals of the copper and iron groups as well. as manganese must be first removed by the usual methods, and while zinc, the alkali, and the alkali-earth metals may be present, the solution must contain nothing that will liberate iodine in acid solution. To about 100° of the solution containing about 5° of dilute sulphuric acid in excess, 1 or 2 g. of solid sodium perborate are added, and after this has dissolved an excess of sodium hydroxide is added and the liquid is boiled for 10 minutes. The mixture is then cooled to room temperature, and after 1 g. of potassium iodide has been added it is acidified with dilute sulphuric acid, and finally after the precipitate has dissolved the free iodine is titrated with sodium thiosulphate solution with starch as an indicator. The thiosulphate solution may be standardized by means of a pnre cobalt compound or by means of potassium dichromate. In the latter case, as the cobalt is oxidized to the trivalent condition, one dichromate is equivalent to six cobalts.— Jour. Indust. and Eng. Chem., viii, No. 10. H. L. W. (2) Chemistry and Physics. 829 a On the Determination of Molybdenum by Potassium Todate.—GuorcE 8. Jamieson of the Sheftield Scientific School has applied the general method of titration of L. W. Andrews to the determination of molybdenum. He recommends that the solution of a molybdate strongly acidified with bydrochloric acid, and heated to about 50° C., should be passed slowly through a rather long and warm “ Jones reductor ” of granulated and amal- gamated zinc, the outlet of which leads to the bottom of a glass- stoppered bottle placed in cold water and containing iodine mono- chloride in hydrochloric acid, as well as a little chloroform for use as an indicator. Under these conditions the molybdic acid is reduced to Mo,O, in the reductor and is readily oxidized to Mo,O, by the iodine monochloride with the liberation of iodine. The liquid is then titrated, in the presence of at least 10 per cent of actual hydrochloric acid, with a standard solution of potassium iodate until the chloroform loses its color, A good end-reaction was obtained at this point, but it was observed that the oxidation goes further to MoO, in the course of about 2 days. The test analyses show good results, but they all deal with small quantities of molybdenum, the largest of which is 0:0627 g. It would be interesting to know if the process can be easily applied to the ammonium phosphomolybdate precipitate.—Jour. Amer. Chem. SOC., XXX1x, 246. H. L. W. 4, The Fixation of Nitrogen.—Joun E. Bucur has made the observation that a process depending upon the formation of potassium cyanide by the action of atmospheric nitrogen upon a heated mixture of carbon, potassium carbonate and finely divided metallic iron, described as long ago as 1839 by Lewis Thompson, is the basis of a-practically important method for the fixation of nitrogen. It appears that numerous attempts to utilize the method or to confirm its application in those early times failed to be successful on account of omitting the use of the finely divided iron which acts as a catalyzer and is of the greatest importance. The author has found that at moderate temperatures of 920° C., or lower it is possible to obtain large yields of sodium cyanide from mixtures of sodium carbonate, carbon and finely divided iron upon passing nitrogen gas. It is even possible to start with oxide of iron in the place of the metal and to use producer gas under proper conditions in place of pure atmospheric nitrogen. The author’s description of his many experiments with their results leads to the opinion that the process is a promising one for commercial application. An interesting experiment was the distillation of the sodium cyanide from the mass in which it was formed by heating in vacuo.—Jour. Indust. and Eng. Chem., 1x, 2338. H. L. W. 5. The Volatilization of Potash from Cement Materials.—In connection with the possibility of volatilizing potash from silicates during the process of manufacturing Portland cement and the collection of the volatilized material by means of the electrical precipitation of the dust from the furnace gases, H. 330 _ Setentific Intelligence. ANDERSON and R. J. Nesrextt have made a long series of ex- periments with a number of actual cement materials. They have found that it is possible to volatilize all the potash from cement material provided the temperature, time, and volume of gases are sufficient. The lower point of temperature for this volatilization is 1100° C., and the rate increases rapidly with the temperature. The presence of chlorides, particularly calcium chloride, increases the rapidity, while sulphates decrease it. Sodium is driven off nearly as easily as the potassium. No conclusions are given as to the applicability of this operation as a practical source of potassium saits.—Jour. Indust. and Eng. Chem., ix, 2538. H. L. W. 6. Dispersion and the Size of Molecules vf Hydrogen, Oxygen, and Nitrogen.—From theoretical considerations, it has been recently shown by L. SinsEeRsTEIn that the molecular refrac- tivity VV of an isotrupic substance the molecules of which consist of two equal atoms may be expressed by the formula N= (2/3) N,[2 + (1—oM,)"]. MM, denotes the atomic refrac- tivity, and o = a/(27f*) ; where a = 3m,/1:008 = 4°88 X 10-” gram, and / symbolizes the mutual distance of the “centers” of the two atoms composing each molecule, that is, of the positions of equilibrium of the dispersive particles within their atoms. This formula reduces to the usual relation V = 2, when £#is so large as to make oJ, negligible as compared with unity, and it owes its generality to the fact that in its derivation the mutual influence of the atoms in one molecule was taken into account. When X,/A is small enough to admit of the omission of terms involving the fourth and higher powers of this ratio the equations of refractivity reduce to the very simple form N=6+9/N and N=), + 9/r. and A, denote respectively the wave-length of an incident radiation in the visible spectrum and the free wave-length in the extreme ultra-violet characteristic of each atom when undisturbed by its neighbors. 0, and g are explicit functions of certain properties of the atoms, such as r,, 0, the mass m of a dispersive particle, and the charge e of such a particle. 6 and g, which are called respectively the “ refraction- coefficient ” and the “dispersion-coeftficient,” are amenable to direct experimental determination. Under the conditions of ap- proximation considered, 6 and g are dependent upon }, g, and / (through oc) in the following manner: 5 = (2/3) 6,[2 + (1 —oh)~'| and g = (2/3) q[2 + (1—o},)~*]. . The denominator 1— ob, being a fraction, the dispersion will show a greater departure from additivity than the refraction, which is a well-known feature of this class of phenomena. The problem before us may now be stated: “ Given the molec- ular refractivity WV of the diatomic substance, that is, given its coefficients 0, g, find the atomic coefiicients 0), g), and therefore the atomic refractivity JV, and also the interatomic distance & involved in o.” Obviously, another relation between J, and g 1s necessary for the solution of the last two equations. Silberstein Chemistry and Physics. 331 makes the very plausible assumption that 6,’/g, = xe where « is the electronic value of 0,’/g, (« == 183 X 10° cm/gram) and « is the smallest integer compatible with the physical conditions. This may be interpreted as meaning that an atomic resonator con- sists of x electrons or that each atom contains x dispersive electrons. The practical application of the preceding theoretical outline to the case of hydrogen will now be considered. The indices of re- fraction for the radiations 4,, D, H,, and HT, for hydrogen at, 0° C, and 76°™ pressure, are 10001387, 1:0001392, 1:0001406, and . 1:0001412. Hence, the well-known formula WV = (p’ — 1) (pw? + 2) Md™ gives 2°074, 2:082, 2°103, and 2°112 for the cor- responding molecular refractivities. (Molecular weight of /, is M = 2-016, density = d = 8'9873 X 10-° grm/cm’.) These values of WV satisfy the equation WV = 2:0445 + 1:279 x 107-"*/X*_ very Salisiactorily (A in em.).. Hence 6 = 2°0445, g = 1°279 xX 10 ™. It is also found that «=1, so that ‘we shall attribute one dispersive electron to each hydrogen atom.” The equation b/g) = 183 X 10° when combined with the two equations con- necting 6 and g with }, gq, and o leads directly to ob, = 0°4932, i 110..G¢, ==: 3250.< 10” ¢.e.s. units. Accordingly the atomic refractivity of hydrogen is expressed by VV, = 0°7719 + 0°326 x 10°-"/A*. The “free” wave-length 2» of a hydrogen atom = 649°'4 A. The central distance between two atoms in a hydrogen molecule = A = 1-067 x 10-*cm. For oxygen x = 2, Ape 622°6 A, and A =1-265 xX 10-°em. Yor nitrogen x = 3, X,. = 570°5 A, and A = 1:493 x 10-*cm. In conclusion, attention should be called to the facts that (1) x agrees with the valency of the elements, (ii) A, falls near the limit of Lyman’s extension of the Schumann region so-called, and (iii) the values of & agree more closely with the molecular semidiameters derived from data on viscosity than these data agree with the radii obtained from other conse:uences of the kinetic theory of gases, such as Boyle’s law, heat conduction, and diffusion. In fact, the viscosity values of FR are given respectively as 1:024 x 10~*, 1°405 x 10~*,. and 1-448 x 10-*° for hydrogen, oxygen, and nitrogen.— Phil. Mag., XXX11, p. 215, February, 1917. THs, Se. Ua 7. Lubrication of Resistance-Box Plugs.—In making aecu- rate comparisons of electrical resistances, by the aid of plug or dial boxes, it is important to have the surfaces of contact of the plugs and lugs maintain sensibly constant resistances. The fact that a satisfactory approximation to this condition may be made by lubricating box-plugs and switches has been known for at least fourteen years, since it has been the custom at the National Physical Laboratory (England) to use petroleum or paraffin for this purpose during the interval of time specified. Since, how- ever, the scheme does not seem to be generally known and as quantitative data are apparently lacking, it may not be super- fluous to call attention to some of the results recently obtained and published by J. J. Mantey. 332 . NScrentifie Intelligence. A certain box, fitted with 9 mine was tested 10 times, by Carey Foster’s method, under each of the following conditions, the plugs being removed between the tests : (a) plugs ina bath of Geryk pump oil, (6) plugs smeared with Geryk pump oil, (c) plugs smeared with medicinal paraffin oil, (d) plugs smeared with “ pure” vaseline, and (é€) plugs and lugs clean and unlubri- cated. The mean contact resistances were found to be respec- tively 0°00135, 0°00118, 0:00124, 0°00116, and 0°00119 ohms. The corresponding extreme variations from these means are given as oo 20, + 18°6, + 19°3, 4 8°6, and + 72°2 per cent.” (Similar results were obtained with three other boxes having 17 plugs each, and with a box of 4 dials. The conclusions are that the usual process of cleaning the surfaces of contact with emery- paper or otherwise and using the plugs dry is very unsatisfactory, and that the scheme of smearing the plugs with pure vaseline is the most convenient and reliable piece eure to follow.— Phil. Mag., XKxM, p21 1,’ Mebruary, 1917. H. 8. U. 8. Unipolar Induction.—The work of E. H. Kennarp should be added to the list of investigations which have dealt with the vexed problem of unipolar induction and its bearing on the mov- ing-line hypothesis, on the existence and state of the luminiferous ether, etc. For lack of space, a description of the apparatus used and an outline of the associated theory will be omitted. By rotating a cylindrical condenser inside a magnetized coaxial sole- noid it was found that the condenser became charged in accord- ance with the theory of Lorentz. Rotation of the solenoid alone had no effect, as had previously been observed by Barnett. The author also says: “The disproof of the moving-line theory is - thus completed ; electromagnetic induction depends in part upon ~ absolute rotation in the mechanical sense. Analysis in terms of electrons seems to make necessary the existence of a stationary aether in order to explain the observed effect; so that the phe- nomenon seems to present difficulties for those relativists who reject the aether.” The following by-product of the work merits attention for pedagogical reasons. Some text-books state that the effect of rotating the armature of a dynamo is the same as that of rotating the field magnets in the opposite direction. This statement must be understood to refer to the total electromotive force produced. On the other hand, the electromotive force in the first case is developed almost entirely in the longitudinal parts of the drum winding, while in the second case a large fraction of it is pro- duced in the radial parts, and the distribufion of electrification on the armature will be different in the two cases.— Phil. Magq., XXXlll, p. 179, February, 1917. He Sate 9. Model Drawing; by C. Ocravius Wrigut and W. ARTHUR Rupp. Pp. xviii, 246; 300 figures. Cambridge, 1916 (University Press and G. P. Putnam’s Sons).—The authors have found that students become much more proficient in the art of model drawing when the usual apparatus of simple geometrical Geology and Mineralogy. 333 models is supplemented by numerous practical exercises selected from typical architectural forms than when the subject is pre- sented in a stereotyped Euclidean manner. Accordingly, although the volume is not intended to be an architectural text- book, it contains an undercurrent of suggestion of the historical devélopment of architecture which should add interest to the theory of correct drawing, appeal to the esthetic sense, and act as an incentive to original work. The text is divided into two parts, the first of which deals primarily with geometrical constructions (4 chapters) and the second (4 chapters) pertains mainly to drawing in perspective. The following salient features also deserve notice. No measur- ing points are employed. ‘The authors believe that the perspec- tive treatment of the circle is quite new, and they have found that pupils can use the method with ease and advantage.” Com- parisons of representations on changing picture planes are given, and their relative advantages are discussed. Frequent oppor- tunities are afforded for sketching from memory, and free use of tracing-paper in ubservation work and of clay in modelling is in- dicated. The line diagrams and other illustrations are excellent, the pages make a pleasing impression on the eye, and the subject is presented in a very attractive manner. H. S. U. _ 10. The Teaching of Arithmetic ; by Paut Kuiarper. Pp. vii, 387; 51 figures. New York, 1916 (D. Appleton and Co.).—This volume is an outgrowth of a course of. lectures given to teachers in the elementary schools. The book is not a text on the subject of arithmetic, but it is a manual of method of teaching arithmetic. The early chapters are devoted to a critical study of the values of arithmetic, of the principles governing the organization of the course of study, and of the pyschology underlying sound methods in arithmetic. The later chapters set forth methods rather than the method of teaching each of the important branches of arithmetic. The index is preceded by two appendixes, of which the first is a classified list of the titles of reference books on the subject, and the second contains standard tests for classes of different grades. Although the text refers primarily to arithmetic it merits the attention of all teachers of elementary subjects because the underlying principles are of general applica- bility and since the material is presented in a very clear and thorough manner. H. Ss. U. II. Grotogy AND MINERALOGY. 1. American Fossil Cycads, Volume II, Taxonomy, by G. R. Wirexianp. Issued by the Carnegie Institution of Washington as Publication 34, Volume II, on July 28, 1916.—In this, Pro- fessor Wieland’s second, volume on American Fossil Cycads, are embodied the results of his labors for the last ten years, reports of which have appeared from time to time chiefly in this Journal. 334 . Seientifie Intelligence. Like its sister volume, the present one is splendidly and gener- ously illustrated and its beautiful quarto pages are a delight. Taxonomy is the special subject treated, and though Wieland recognizes the flimsy basis* of external morphology upon which Cycadeoid species have been created, yet he does not quarrel with the ‘“ cast-makers ” of species but considers it “ of the great- est importance to present structural details as rapidly as those can be assembled, and with as little specific rearrangement as pos- sible.” So thoroughly has he presented these details, that con- siderable headway has been made in the reduction of species, possibly also in the reduction of genera. For example, Wieland has shown that the score or so of species of Cycadella can prob- ably all be included in four, and that even the genus Cycadellat can well be merged into Cycadeoidea. Though Wieland has no quarrel with the makers of species, yet he has one complaint which must appeal to every scientist as a very valid one. In some cases “‘ type specimens ” are hoarded intact in museums and only their external appearance is known, it being considered a sort of sacrilege to have their structure determined by section- ing! This most unenlightened view of the purpose of a museum retards scientific advance. It is a new phase of the old struggle between the systematist and the modern botanist, in which the hand of the strateographic geologist can be clearly discerned. In the end such museums will suffer by their narrow policy, for as time goes on, and the structure in related species become known, © the value of the type ones will become less and less. The National Museum at Washington and some European museums have set a good example, and it will no doubt not be long until their lead will be followed. Wieland further considers that extensive field work, opening up of quarries like those for Dino- saurs, affords the only means for an adequate study of fossil plants. Aside from the taxonomic results of his work the chief points of interest lie in the elaboration of the woody-structure of the stem and in connection with the fruiting habit and the character of the staminate disk. A most remarkable new species (C. dartonz) which bore 500- 600 well-preserved fruiting strobili, one in practically each leaf axil, and all uniformly mature, illustrates the occurrence of a culminating fruiting period (monocarpy) in the Cycadeoidea. Wieland had found traces of such a condition, but in this case the evidence is practically conclusive. The similar habit in the *< BELL REEF 2} MALIMA 1? 4 !Qp'anpi ; AVEA \L 8 a9? WATHE ay “oy x en sere ee eset teen, NUKU THIKOMBIA Rf THIKOMBIA Ll LAV « Fie. 1. The Exploring group and adjacent Reefs and Islands. Black=elevated limestone. Fie. 2. Fig. 2. Island of Lakemba. Black=limestone. White=andesite. Lined=coastal flat. 346 W. G. Foye— Geology of the Lau Islands. of the upper curve. Here, near the center of the island, is situated its highest peak, Koro Mbasanga, which rises 930 feet. North of this peak the country slopes down in wide, spoon- shaped valleys to a limestone platfor m, having a broken surface which forms the northern end of the island. The average elevation of the platform is 350 feet. It swings southward on the western side of Koro Mbasanga and terminates just west of the peak, wedging out along the shore. Southward from this peak a broken ridge forms a serrate backbone to the island. The map shows hills 500 and 700 feet high along this ridge, but the usual altitude is but 300 or 400 feet. The hills are low and rounded. At the southern end of the island the low ridge dips south- ward beneath a small area of limestone standing at an elevation of about 300 feet. The peak, Koro Mbasanga, has an amphi- theater-like depression on its eastern side. The southern limb of the ridge inclosing the amphitheater was followed south- ward and it was found that lavas from this peak had over- flowed an area of limestone and baked it red. The limestone was a foraminiferal variety containing little or no coral. Just north of Koro Mbasanga a similar limestone, but silver-grey in color, was found in patches overlying an eroded voleanic sur- face. Here it included bits of the underlying rock and formed an undoubted basal layer. From the facts just recorded it was inferred that an eroded surface of volcanic rocks had been submerged and overlain unconformably by limestone. The evidence was inconclusive, however, as to whether the limestone had been elevated and eroded before the extrusion of the voleanic rocks of the second period. This question was answered by exposures in the small island of Andivathi, lying off the northwestern coast of Vanua Mbalavu within the upper curve of the question mark. The larger part of Andivathi is composed of andesite agglom- erate and ash dipping 30° E. The western beds, representing the lower members of the series, are of coarse agglomerate with vesicular blocks of lava a foot or more in diameter. The upper beds become increasingly finer and show many small fragments of pumice. The total thickness of these beds is about 200 feet. The eastern side of Andivathi is composed of elevated coraliferous limestone. Both ash and limestone are cut by a network of basic dikes, the largest observed being 20 feet in width. It is deeply lateritized and outcrops as a long, well-defined depression in the limestone into which it was in- truded. The limestone stands up in nearly perpendicular walls on either side of the depression which was once filled with the dike, indicating that in this climate the igneous rock weathers more rapidly than the sedimentary. On the N.E. side of W. G. Foye— Geology of the Lau Islands. B47 Andivathi ash beds rest on an irregular surface of limestone and dip S.E. toward the eroded and nearly vertical walls of limestone, forming the coast of Vanua Mbalavu less than a hundred yards away. The irregularity of the surface beneath the ash makes it apparent that the limestone was elevated and eroded before the volcanic rocks of the second period were extruded. The central ridge south of Koro Mbasanga is more deeply lateritized than the rocks of the highest peak, and bits of silici- fied coral often occur in the laterite. The writer inferred that the corals were residuals of a limestone cover which once man- tled the voleanic rocks. The andesites, lateritized to a depth of 20 feet or more, strengthen this inference since it is apparent that they are of a very different age than the rocks of Koro Mbasanga. Tsoi Eee SII pS) . * Oy o['| 2 4 | ay] g Sih Ss] & ~ wv wv | o \ | o{{| ¥ “3 \s| 3 =i ™ a = a ee era Peahtend b= = G % > } ex = ere: | a) w 5 ee vo i LF ) } ae 2 o) al pe 5 2 | | os = S U 53 / yn 1 ce a) sere wn She o-; t t is o o a\ell ge > g ~ 5 oe oc | ce 8 vo ,¥ eels c Ole = © | ee of fe ) Si] ° c } 2 no 3 oil) ) fo) ia Sores * = eal 3 ‘= > 2) S x vi] = °o iS a & t v | = v Ss Sil = v | x = uv Le) N ° ” 3 3 aS a) Bs -FL29 + v O ~ na) ) fry ° Uv << S oF ip) S ,ooski 70021 ,009 Fic. 2. A—section from shore at Toronto, across Island to the Filter basin ; see map, fig. 1. (From Toronto Water Supply Report.) B—section outward from Victoria Park. (Chipman & Powers.) C—section outward from Searboro Bluffs. (Chipman & Powers.) 356 Spencer—Origin and Age of the Ontario Shore-Line. Island, the outer margin of which is 10,800 feet from the shore (section A, fig. 2). At the shore, only a few feet of sand cover the basement shales, but at the outer edge of the island variable beds of sand with clay (180 feet) rest on transi- tion and clay beds to a depth of 230 feet, overlying the base- ment shales. Their origin is complex. Only the upper 30 ~ feet are composed of coarse sands and gravels, washed as are beach materials. These alone can be regarded as belonging to the Ontario beach, and their thickness corresponds to the maximum depth of the bay behind the island. Below are 10 feet of transition beds with some clay, which limit the water- bearing strata. Beneath, are 100 feet of variable beds of very fine sand and clayey sands, not well-washed. These rest upon very fine, closely packed sand without clay (40 feet), beneath which is a transition bed of clay nodules in sand (10 feet) over- lying 40 feet of blue clay beds covering the basement shales,* The fine clayey sands below the coarse beach sands are the covered post-Iroquois delta deposits of the Don River. The Toronto beach, in front of Ashbridge Bay and Toronto Harbour, is a barrier nearly 6 miles long, leaving the main lake shore near the Woodbine (see map, fig. 1). By comparing the various maps of it, made since 1788, the beach is seen to have had a rapid westward growth—-the waves building up a succession of spurs and hooks, with intervening ponds and marshes, until they coalesced (see fig. 3). Early in the Nine- teenth Century the western end of the Harbour was open. Here a spur, 4,000 feet long, was developed between 1841 and 1882 (forming Hanlan’s bay; see B, fig. 3). In 1854, the west- ern end of the Harbour was being further silted up so as to | occasion concern. In 1891, and again in 1912, piers were con- structed for keeping open the deepened westeril channel, now otherwise closed. The materials for the westward orowth of the barrier have been mostly obtained by robbing the sonthern shore of the beach. About 1850, the waves forced a passage between the lake and the bay (B, fie. 3). This was further opened in 1852, 3 and *5, when a steamer passed through. ‘The gap was closed in 1856 owing to low water, but it became permanent in 1857. By 1882 the beach had been washed away for a length of 1,800 feet and a depth of 5 feet, leaving the site of the tavern of 1841 some hundreds of feet outward in the lake, as also the triangulation station of 1818. Thus, Toronto Island became separated from the peninsula at dates mentioned. No con- siderable amount of drift from the east re-entered the gap. In 1883 the channel through it was deepened by a canal and the *This geological section was obtained from borings at the Filter basin, near the southwestern point of the barrier beach. Spencer—Origin and Age of the Ontario Shore-Line, 357 Fie. 3. Fic. 8. Maps of Toronto Harbour, showing development of the Island, the barrier broken by the waves, and gap again closed artificially and by redeposits due to protective works (after Chipman). A, B, C. 358 Spencer—Origin and Age of the Ontario Shore-Line. remainder of the gap closed by erib-work and piling, and the deepened canal is now protected by transverse piers nearly half a mile long. The sands and gravels were originally derived from the remains of the drifts of Scarboro Bluffs, which upon being undermined and wave-washed were current-borne westward. However, these features were insufficient to explain the origin of Toronto Island until soundings and borings had revealed the character of the underlying delta deposits hitherto unknown. The source of these is found in the excavation of the Don valley* (behind the barrier to Toronto and Ashbridge bays (see map) which, below a height of 190 feet above the lake, has been excavated since the Iroquois epoch. The delta basement (with the closely packed very fine sand which may be inter-Glacial) is more than six times the thickness of the Ontario beach itself. The now buried delta was accumulated while the lake sur- face was rising from its lowest level (some 200 feet) to that of Mill-Haven stage, which is now represented by the ‘“ bottom” of the present valley, before the waters fell again to the exist- ing shore, when the beach materials from Scarboro Bluffs were brought by the waves to cover the delta mass. While elsewhere the lake deepens gradually to 75-100 feet or more, the slope of its floor is only 17 feet in a distance of 1,600 feet at the southwestern angle of the Island, but im- mediately beyoud it descends abruptly to 69 feet (in 1,400 feet) and to 150 feet just beyond (section A, fig. 2). Such is in strong contrast with the gentle gradient elsewhere. This changing slope also shows the approximate limit of wave action to 20 feet as appears farther east. No features below 20-30 feet can be correlated with the Ontario Beach. The materials of both the older delta and the newer beach have been dumped into the eastern side of the Humber Embay- ment, producing the peculiar configuration of the southwestern extremity of the Island. But the Humber is unlike the Don Valley, the former being of inter-Glacial age, with a later mantle but without a post-Iroquois delta as has the Don. Age of Lake Ontario and of the Saint Lawrence River. Investigations about Toronto Island alone furnish no clue of the age of the Ontario Beach ; but from the encroachments of the lake upon Scarboro Bluffs, for over 9 miles between Victoria Park and Port Union, and from the topography of the *Such materials as were removed may be seen at the Don brick yard ; till 3 feet ; stratified sand with grit and layers of clay 20-20 feet ; laminated drab clay (sandy) akout 25 feet; clay with fine sandy iaminations 35-40 feet ; sand coarser 6 feet ; fine sandy clay 40 feet covered with sand. Spencer—Origin and Age of the Ontario Shore-Line. region the approximate age of the Ontario Beach is determinable. Between these two points, the en- croachment of the lake has been measured along 18 lines codrdinate to the lake shore. These points* are indicated on the map (fig. 4). \ The recession of the two terminal points has been unimportant (re- spectively 8 and 3 feet in 50 years), but between them the mean lake encroachment has been 2 feet a \ year. The total recession of Scarboro \ bluffs, since the establishment of the Ontario Beach, is determinable from the topography. While the bluffs reach to 850 feet above the lake, the Iroquois terrace skirts their face except for less than half a mile, where it has been under- ’p mined by the waves (see fig. 4). “a, This shows that the post-lroquois 2 surface, removed by the lake, rose less than 200 feet above it, and its slope should correspond to that of the country between the Iroquois Beach and the low shore, just west of Victoria Park, a distance of 4,000—-4,400 feet. Such would in- dicate the original extension of the a Ontario shore in front of Searboro [| ~ Blufts. , = This location is confirmed by the 3 restoration of the natural shore-line in front of the indentation in the bluffs (see figs. 1 and 4), where their truncation represents the recession of the original Ontario shore. Combining the mean rate of recession of 2 feet a year with 5 the amount of encroachment deter- ee mined, the age of the Ontario 2% Beach is found to be approximately 2 2,000 years. * My. F.M. Passmore marked by 18 corner- _ a stones the position of the bluffs in 1882-3. Messrs. Speight and Van Nostrand re-sur- Fie. 4. Map showing Scar- veyed them in 1912 (for the Toronto Water boro Bluffs, Iroquois Beach Supply). By comparison of these surveys and later encroachments on the mean recession has been found. the shore. Am. Jour. Sct.—FourtH Series, Vou. XLIII, No, 257.—May, 1917. 25 Gee AS Port Union # 18 2%, Scarborough Plateau 4 Bo Dutch church Bluffs | G Wipe ie TUNA WeeaM Metta and Terrace Lake ", “Aas iin PUA a Imiles Hunt Club Recession points measured |- 18 359 360 Spencer—Origin and Age of the Ontario Shore-Line. But the age of the lake itself is somewhat older—its surface having been lowered a few feet, while the St. Lawrence was deepening its channel, but there is no direct measurement derivable from the St. Lawrence River itself. Researches, relating to the great accession of water to the Niagara River, show that the terrestrial tiltmg about the north- eastern angle of Lake Huron occurred as late as some 3,500 years ago. This earth movement extending to the St. Law- rence River was that which gave birth to the great modern river itself. No appreciable deformation has since occurred. Consequently, Lake Ontario is found to be some 3,500 years old. The difference between this figure and 2,000 years for the age of the Ontario Beach seems ample for the lowering of the lake level from the Mill Haven to the present shore-line. New fesults—Lake Ontario and the St. Lawrence River date only from the time when the discharge was first confined to the river channel, since lowered some 15 feet by current scour. The former earth movements of some 3,500 years ago have disappeared in the higher and present Ontario shore-lines. The age of the Ontario Beach is satisfactorily determined at about 2,000 years, while the age of the lake itself, and that of the St. Lawrence River, is 3,500 years or perhaps slightly less, as based upon the data so far discovered, without any great probable variation indicated. Speculations relating to the Age of Lake Ontario and Post- Glacial Time. To close this study and not analyze the speculations would leave the subject in confusion, as few have the available data or opportunities for investigation. Furthermore, a discussion introduces other important features. A notable attempt to find the time value for the Ontario Beach has been made by Prof. A. P. Coleman,* who based his estimate on beach accumulations artificially obstructed by the canal piers crossing the beach at Toronto Island (see fig. 3); and according to his assumed rate of deposit and the gross quantity of sand under Toronto Island he concluded that the age of the Ontario shore is 8,000 years. Plausible as this method might seem, such assumptions are contrary to the observations and mechanies of wave action, as well as the mass of sand, being derived from one source. In transportation, the wave-action varies as the sixth power of velocity. Reduce the velocity one-half and the transporta- tion power is reduced to one sixty-fourth. Thus, a slight obstruction will cause a deposition of the wave-carried sand, * Estimate of Post-Glacial Time, by A. P. Coleman. Trans. International Geol. Congress, for 1913, pp. 455-472. Spencer—Origin and Age of the Ontario Shore-Line. 361 not merely that drifting along the shore, but also that washed by the waves from greater depths. An illustration of new-formed land is seen at Atlantic City, where groins or piers were built outward to check the force of the waves encroaching upon the coast. But the groins led to the deposition of sand between them. by extending these and adding others a strip of new land was acquired in 40 years for a length of three-fourths of a mile, with breadth reaching to one-fourth of a mile, curiously leaving the lighthouse in- land. During high water of 1908, the lake waves washed away the shore road around Humber Bay. Here also protec- tive groins were built (see map, fig. 1), and already the inter- spaces are being filled with beach material. Without artificial impediments, a barrier beach 3 miles long was formed at Asseagua Light, on the coast of Maryland, in 40 years; and another spit or barrier, 4,000 feet long, was developed on the western side of Toronto Harbour, between 1841-1882 (see 0, fig. 3) and now completely separates the bay from the lake, while the southern side of the beach is being washed away. After the formation of the gap through the Toronto barrier, no considerable deposit of beach material from the east was carried through the opening into the quieter waters of Toronto Bay ; and only when the gap was artificially closed in 1883 and further obstructed by the canal piers, did the sand and gravel accumulate here, although such were naturally closing the western end of the harbour. These were the deposits which Prof. Coleman unfortunately used as a basis for his speculations of geological time. The above examples illustrate the effects of artificial impediments upon wave action and natural changes of current, showing the worthlessness of any deductions as to time that can be derived from them. Further- more they are contrary to the laws of mechanics. Prof. Cole- man also failed to distinguish the underlying clayey sands of the Don delta from the overlying beach materials of the On- tario shore, thereby greatly lengthening his conjectures as to geological time. We are indebted to Prof. Coleman for calling our attention to the not-well-known surveys of the recession of Scarboro bluffs, which had been made for the Toronto Water Supply. By his selection of the figures, rejecting measurements of the greater erosion, while retaining those of reduced amount (which were compensatory in the variable conditions), he assumes 1°62 feet as the mean rate of recession in place of 2 feet per annum, which latter is the mean of 16 measurements between the points of practically no encroachment. 362 Spencer—Origin and Age of the Ontario EOP In the recession of Scarboro blufts the author, just mentioned, seeks to confirm his estimate of the age of the Ontario Beach derived from Toronto Island. Such requires a knowledge of the amount of the wave encroachment since the birth of the Ontario Beach. From the gentle slope of the lake floor to a depth of 100 feet, at a point 18,000 feet outward, where it declines more rapidly, he concluded that the Ontario beach had an amplitude of 100 (in place of 20) feet and had risen this amount, but from the soundings he could with equal value have selected 8 ,000 or 18,000 feet. From his assumptions here he again gets 8,000 years as the age of the Ontario Beach. Using “he other flour es, Just mentioned, a student could just as well obtain an age of 5,000 or 11,000 years. Any hypothesis based on such variables can not be accepted. But more, he includes in the Ontario Beach epoch the time of the water rising 100 feet to its present level, thus unconsciously embracing with the time of the Ontario Beach a long period before the begin- ning of that stage, which further throws confusion into his hypothesis. To this disjointed speculation of the age of the Ontario Beach (8,000 years) the same author assigns, without offering any evidence, an equal time for the Iroquois Beach, and then nae another 8,000 years (which he himself recognizes as only ** ouess ’”) for the time inter vening between the Ontario aud ee beaches—thus making 24,000 (or perhaps 27,000) years as the duration of post-Glacial time.* Buta further con- tradiction appears, for the professor has shown elsewhere that Lake Iroquois was a Glacial lake, consequently, his 8,000 years as the age of the Iroquois Beach must be taken away from his post-Glacial time, leaving 16,000 years. Such @ priori philoso- phy leaves a suspicion that its author had some specnlation to support, but the analyses of the data show that a confusion is thereby thrown into the problem of Geological time, when he had within his grasp the material for a lasting scientific contribution of great value; and if the confusion be not ex- punged such must lead to the retardation of scientific research. Washington, D. C. *From resemblances between the Iroquois and Ontario beaches their relative duration might be inferred within an error of 100 per cent; but the time required for the work performed between these epochs is not compar- able to either, for during this intervening period the lake level fell more than 500 feet, with pauses, and rose again 200 feet with the partial sub- mergence of some of the terraces. During this intervening period great denudation was effected, as shown by the large valley excavated by the Don. The varied work performed at this time was vastly greater than that of building either the Iroquois or the Ontario Beach. W. H. Twenhofl—Granite Bowlders. 363 Arr. XX XIII.— Granite Bowlders in (?) the Pennsylvanian Strata of Kansas ; by W. H. TwEnwoF Et. Introduction. From time to time granite has been reported from deep wells of different parts of east Kansas. On investigation many of the reported granites were found to be other kinds of rocks, or without basis in fact. In a few instances final judgment had to be reserved. The 1915 boom in oil develop- ment together with very conspicuous cases of granites reported from two wells near Zeandale, a small village about eight miles east of Manhattan, Kansas, brought so many inquiries to the Kansas University Geological Survey, and the possibility of encountering granites in drilling was having such a depreci- ative effect on further development, that Professor EK. Haworth, the State Geologist, carefully investigated all of the reported cases of granite in order to definitely ascertain, if it were pos- sible, whether such actually had been reached in any of the deep wells.* , In each of the Zeandale wells granite was reported at a strat- igraphic level in the upper part of the interval between the Oread and Iola limestones. Jn one it was struck at nine hun- dred and fifty-eight feet and drilling was continued to one thousand and ninety-three feet without penetrating the granite, although curiously, the log of the well showed that thirty-two feet from the top of the granite the drill had passed through a twelve-inch bed of shale. In the other well granite was struck at nine hundred and forty-five feet and continued drilling encountered no lithic change. A careful examination of the cuttings which are said to have come from the wells showed unquestionable fragments of unweathered granite. Haworth’s conclusions were to the effect that the granitic rocks which were derived from these two wells, and which perhaps had also been found in others, are probably firmly cemented sediments which came from rocks of a granitic character and, by way of illustration, he called attention to a rock of Tertiary age which occurs in Phillips County, Kansas, and is composed of firmly cemented granitic gravels derived from the Rocky Mountains. This rock resembles granite so closely that it has been quarried and used as paving blocks under that name. During the field season of 1916, while examining territory in Wilson and Woodson counties in the interest of the Fre- donia Gas Company, the present writer found granite bowlders which it is fairly certain came from Pennsylvanian strata. * Haworth, Bull. 2, Kansas Univ. Geol. Sur., 1915. 364 W. Hf. Twenhofel—Granite Bowlders in (?) The locality is in Eminence Township, Woodson Oounty, about eight miles almost due south of Yates Center and one mile west and a half mile south of the Missouri Pacific Rail- way station, Rose, the locality being on the very headwaters of a small tributary of the Verdigris River, into which it empties about twelve miles to the southwest. The bowlders lie on the northern edge of a low hill which has been determined by a small anticlinal structure. ” is supposed to have a greater amplitude: than “@ ~~ ; alias curve ‘“‘c” is supposed to have a smaller amplitude than “6”. A careful examination of the available temperature curves shows that curve “A” may be discarded. On the contrary curves “ B” and “C” must both be taken into consideration. As a typical example I will cite the temperature curve for Warsaw. The utilized daily means are those of the years 1826-1880.+ The following diagram (fig. 3) gives an interpretation of the detailed curve. | The variation is composed of phases 1, 2, 3 and 4 belonging to four concordant curves. The steps occur approximately at the following dates: Jan. 25, March 9, April 22, Oct. 2, Nov. 29, and Dee. 16. Then during the summer, on the contrary, from June 9th till July 28th, we have practically a straight lme of mean tem- * H. Renou, Annales Bur. Centr. Met. France, 1887, i, p. B. 195. +R. Gautier et H. Duaime, Arch. Sc. Phys. (4), xv, 545, 1903. +t Jan Kowalczyk, Pam. Fizjogr., Warsaw, 1881. C Mean Annual Temperature Variation. 405 peratures increasing from 17°6° to 193°C. This line 6 is evidently discordant : it cuts all the others. The summer maxi- mum is 8° below the maximum which would be observed if curve 4 were fully developed. The summer temperature may, therefore, be considered abnormally low. Now, between 6 and 4 we observe the fragments of a curve 5 extending from May 29d to June 8th and July 29th to Aug. 26th. “This curve intersects all the others: its amplitude is evidently much smaller than the amplitude of curves 1-4. Although the temperature curve for Warsaw may be con- sidered a most typical example of the normal anomalies of the annual temperature variation, the breaks being too well accentuated to be ascribed to chance circumstances, it is useful to cite a few other examples in order to show how, gradually, we pass from one type of variation to another. Hellmann has published* a detailed curve of the daily means derived from the observations made in Berlin during the years 1848-1907. Referring to the diagram for Warsaw (fig. 3), the Berlin curve displays the fragments 1 to 4,5 is missing and 6 is a well-developed curve extending from June 13th to Sept. 21st, with a maximum on July 22d. From March 18th to April 7th the increase in temperature is abnormally rapid so that 3 is discordant with 2 and 4. The drop of temperature between 4 and 6 (June 6—-11)is 15° C. This is perhaps the most characteristic feature of the curve. Besides, 4 and 6are discordant. Phase 6 hasa smaller amplitude than 4. Evidently if many curves were available it would be interest- ing to follow the progressive change from station to station. That these anomalies do not occur simultaneously at different * Preuss. Met. Inst., Abh.; v. 3, No. 6, Berlin, 1910. 406 H. Arctowski—Normal Anomalies of the stations but gradually propagate from place to place is a well established fact, at least for some of the temperature depres- sions of the spring. The curves for Arcachon, Greenwich, Gerlin, Lemberg, Penza and Wologda, published by Almstedt,* may serve as examples. Without adopting in their integrity the ideas expressed long avo by Dove,t let me suppose now that the temperature curve for Melbourne, or the curve for Warsaw, or the curves for other places, exhibiting similar anomalies, express the result of an antagonism between maritime and continental climates, respectively characterized by a very small and a very large annual amplitude. On the ocean, west of the coasts of France, the difference between the mean temperatures of the warmest and the coldest day of the year is certainly less than 10°C. In Paris it is 18°7°, in Warsaw 24°7°, in Barnaoul it is 41°8° and the observa- tions of Nertchinsk give 49°9° C. I have traced the curves for Barnaoul (means of the observations made from 18388 to 1882) and Nertchinsk (1839-1881)t and have found, to my great astonishment, that from March 15th to November 1st these two curves are practically identical, whereas during the winter months they differ very greatly one from the other. The winter in Nertechinsk is very much colder than in Barnaoul. The temperatures are: Barnaoul Nertchinsk October 31 —= 98° == FOS January 2 —18°3° — 30°7° Mareh 15 = Mes? —= 181° The maxima of the summer are: Barnaoul 20°2° on July the 4th Nertchinsk QD ie SES TER eae Othe The temperature curve of Barnaoul belongs therefore to two climates. During the months of April to October it belongs to the excessive continental climate of Nertchinsk, while in the winter it belongs to a Jess continental climate. : The following diagram (fig. 4) gives the daily means of tem- perature in Nertchinsk for the months of May and June. The diagram shows that the increase of temperature pro- ceeds by steps. Going up, the curve slides down here and there and goes up again. The. total depression of the yearly amplitude, due to these steps, may be estimated at 17°. Admit- * Loe. cit. + Abh. K. Akad. Wiss. Berlin, 1856, p. 121. t Repert. f. Meteor., Suppl. 3, St. Petersbourg, 1886. Mean Annual Temperature Variation. 407 ting this estimate and supposing that a curve of 67° (=50+17) amplitude would represent the variation corresponding to the coefficient of perfect atmospheric transparency* at the latitude of Nertchinsk, the lowest temperature of January must be con- sidered 6°5° too high and the highest July temperature 11:5° too low, since the extreme daily mean temperatures should be — 87° and + 380°. Thus it may be that, in this case, the decrease in amplitude is in close connection with the annual variation of atmospheric moisture. During the summer months, the greater amount of aqueous vapor diminishes the coefficient of atmospheric (thermal) transparency very much more than during the winter months, and so it is evident that the summer temperatures differ more,from what they should be than the temperatures observed during the winter. This leads to the question whether the steps of the Nertchinsk temperature curve, and perhaps also the steps of the curves of several other stations, are not partially due to a rhythmical transport of atmospheric moisture. * A. Angot, Ann. Bur, Centr. Met. France, 1883, i, pr Baie. Am. Jour. Sci.—FourtsH SERIES, Vou. XLIII, No. 257.—May, 1917. 28 408 H. Arctowski— Normal Anomalies of the In the ascending part of the curve of Barnaoul the depres- sions following the crests generally precede by two to four days the corresponding details of the curve of Nertchinsk. The inflexions of the isotherms, characteristic for these changes, progress therefore across Siberia from the W. towards the E. But, as a result of the international balloon ascents of May 13th, 1897, Hergesell has shown that the typical decrease of temperature observed then in Central Europe was very much more accentuated at high altitudes than it was near the ground.* Similar observations have been made since. In consequence, the inflexion of the isotherms must be more pronounced at an altitude of 10,000-m. than it is at the surface of the earth’s crust. After each step of the ascending temperature curve, the entire air mass above the station where the step has been observed is changed. The work of the progressive heating of the ground by solar radiation and the heating of the air mass above, by convection currents, must, to a certain extent, be begun anew, and probably under different conditions. Evidently, to reach ‘a definite conclusion it would be nec- essary to study the records of individual years and the weather maps as well. But the real difficulty, and at the same time the great interest of the study of these normal anomalies of the mean annual temperature variation, resides in the fact that we have to deal with a phenomenon showing intimate relationship between very far distant stations. On the opposite side of the world, in Baltimore,t the tem- perature crests of February 22d, March 10th, April 14th and May 10th, there observed, belong also to the curves of Barnaoul and Nertchinsk. In Barnaoul they occur: Febru- ary 18, March 9, April 21 and May 10. In Nertchinsk: Feb- ruary 29, March 12, April 20 (and May 12). Alr eady Dovet noticed the fact that the anomaly occurring during the month of May is noticeable in the records collected . in Arctic America and Greenland. On the other hand, accord- ing to R. C. Mossman,$ the cold period of May is well pro- nounced in Argentina and Chile, north of 40°S. lat., and it was also observed at the winter quarters of the “ Discovery ” in 1902 and 1903, at Cape Adare in 1899, at the South Ork- neys during all the years of observation from 1903 to 1908, with the exception of 1906. Mossman remarks that thus the temperature anomaly of May is a bipolar phenomenon and he adds that the curve of * Meteor. Zeit., xvii, 1, 1900. sh O. L. Fassig: The climate and weather of Baltimore, pl. 3, Baltimore, 1907. t Loc. cit., p. 162. S$ Symon’s Met. Mag., xliv, 1, 1909. Mean Annual Temperature Variation. 409 mean atmospheric pressure at the South Orkneys, for each day of the year, bears a close resemblance to that of Edinburgh. Forcibly, therefore, we reach the conclusion that in a com- parative study of the anomalies of the annual temperature variation, Teisserenc de Bort’s conception of the great centers of action of atmospheric circulation will find an extensive application ; because, although at present it would be prema- ture to try to explain why it is that some changes of phase may occur simultaneously in Arctic and Antarctic regions, or in North America and Siberia, it seems impossible to conceive such correlations without supposing some relationship with the exchange of pressure between the seasonal and permanent cen- ters of action. New York City, November 2, 1916. Se Lion PP Le “ENTE EEL GEN. OE. I. CuHeEmistry AND PHysIcs. 1. Attempt to Separate the Isotopic Forms of Lead by Frac- tional Recrystallization.—Although the complete inseparability of isotopes by chemical means has been frequently asserted, THEopoRE W. Ricuarps and Norris F. Hartt have deemed it desirable to test the possibility of such a separation still further and they have made a very elaborate attempt to do this with the ~ radioactive lead from Australian carnotite. The low atomic weight of this sample of lead, previously determined by Richards and Wadsworth, indicated that it contained three or four parts of the isotope (usually assumed to be radium G) to one of lead, and its B-ray activity showed that it contained sufficient radium D to serve as a basis for testing the possible separation of this isotope from the inactive varieties. About 1 kilogram of this radioactive lead was converted into the nitrate, and this salt was subjected to systematic fractional crystallization by cooling the boiling, nearly saturated aqueous solutions, and carrying the crystals and mother liquors in opposite directions in the series. About 1000 separate crystallizations were thus made, where the end-products were combined when they became too small for convenient work, and a series of 18 members was finally obtained. Four of these members at the “‘ more soluble” end and six at the “less soluble ” end were united in each case, and these products were carefully purified, converted into chloride, and subjected to atomic weight 410 Scientific Intelligence. determination by finding the relation of the lead chloride to the silver required to combine with the chlorine, according to the highly refined methods in use in Professor Richards’ laboratory. Two analyses of the chloride of ordinary lead had given 207:187 and 207°186 for the atomic weight. The results with the end fractions of the isotopic lead were as follows: | ‘* More soluble ” ‘“ Less soluble ” 206°426 206°406 206°409. 206°422 206°431 206°399 Av. 206°422 206°409 These atomic weights agree so closely that it is obvious that no separation, or at most not more than an exceedingly small separa- tion, of the isotopes was effected by the elaborate fractional crys- tallization. The end-products of the fractional crystallization, after purification, were tested as to their radioactivity in order to find if there had been any separation of “radium D” which is supposed to have an atomic weight of 210, and is supposed to be present in this lead in the proportion of about one part in ten million. The results agreed within the experimental error of 1 per cent, so that no evidence of any concentration was found.— Jour. Amer. Chem. Soc., xxxix, 531. Hi; We 2. Manganese in Soils—MaxwE.u O. Jounson of the Hawaii Agricultural Experiment Station has devised a remedy for a serious local difficulty in the culture of pineapples. The chief pineapple district of the Islands lies on the island of Oahu in a region where the water supply is insufficient for the growth of sugar cane. In this district, in distinction from the usual red soils, there occur various areas of dark or black manganiferous soils where, as has been known for 10 or 12 years, pineapples make a very poor growth. The leaves of the plants gradually become yellow, and the plants often die or finally produce very small fruit of inferior quality. As these dark soils, where the yellowing of pineapples occurs, aggregate from 6,000 to 10,000 acres of the lowest lying, most accessible and most easily culti- vated land of the region, the matter is of considerable economic importance. Several years ago this trouble was attributed by Kelly, of the same experiment station, to the presence of an abnormal amount of manganese in the black soils. He found that one of the black soils contained about 6 per cent of manga- nese oxide, while a red one contained only about 0°3 per cent of it. He analyzed also the ashes of a number of plants grown on the black soil, some of which were affected by the presence of manganese while others were not affected by it. From a study of the ash analyses Johnson came to the conclusion that the . affected plants were abnormally low in iron, and that the pres- ence of manganese causes a suppression of the assimilation of iron Chemistry and Physics. 411 by the plants. Acting upon this suggestion, a practical method . of controlling the trouble was found. Fields of yellow, unhealthy plants when sprayed with solutions of iron salts became green in a very short time. The effect on the fruit was most remarkable. The small, red, stunted pineapples, when sprayed, showed decided improvement within a week, becoming normally dark green, and commencing a vigorous growth within two or three weeks. When iron was applied to one side only of the unripe fruit, that side became green first and made such growth as to distort the shape of the fruit, but later the iron appeared to be distributed and the fruit became fairly symmetrical when ripe. Practically all the pineapple plantations have now applied spraying with fer- rous sulphate solution on the manganiferous soils with apparently perfect success in profitable cultivation.—/our. Indust. and Eng. Chem., 1x, 47. H. L. W. 3. Preparation of Sulphurous Acid.—It is stated by Epwarp Harr that the cheapest and most convenient method for prepar- ing small amounts of sulphur dioxide in the laboratory consists in warming fuming sulphuric acid containing 30 per cent of SO, with sulphur in the form of lumps. The sulphur dissolves form- ing a blue solution from which, on warming, SO, is given off mixed with some SO,. If sulpburic acid is objectionable the resulting solution may be warmed and the gas again absorbed. The evolution of SO, ceases when the SO, has been acted upon and the sulphur melts. The sulphuric acid remaining contains only a small amount of dissolved sulphur and is fit for most uses.—/our. Amer, Chem. Soc., xxxix, 376. H. L, W. 4, Gas Chemists’ Handbook, Compiled by Technical Commit- tee, Sub-Committee on Chemical Tests, THE AmeERiIcAN Gas IN- STITUTE. 8vo, pp. 354. New York, 1916 (The American Gas Institute, New York City).—This compilation has been made by a committee of eleven prominent gas works chemists, of whom C. C. Tutwiler is chairman, under the editorship of A. F. Kun- berger. It is a very useful book, not only for chemists connected with the gas industry for whose use it is particularly intended, but also for other analysts who may desire information in regard to the special methods used by these chemists. The book has a wider scope than would perhaps be expected, as it deals not only with the examination of the raw materials and the products of gas manufacture, but also with the tests of many miscellaneous materials, such as boiler-waters, paints, fire-clays, lubricating oils, solders, and other alloys, lime, cement, and both ordinary and special steels. The very important subject of sampling is well presented in connection with the various materials to be examined, and the book as a whole may be regarded as an excellent treatise, giving very full and clear directions of well-selected methods of examination and analysis. H. L. W. 5. Union of Glass in Optical Contact by noe Treatment. — In many optical and spectroscopic investigations much annoyance and incomplete success arise from the fact that it seems to be 412 Scientific Intelligence. impossible to obtain a suitable cement for fastening together the transparent parts of the various types of cells required. Cements . cause trouble in, at least, three different ways. (a) They deform or strain glass which has been figured plane or spherical to the highest degree of accuracy. (b) No one cement is immune to chemical action when all kinds of liquids at various temperatures come in contact with it. (c) It is very difficult, if not practically impossible, to cement optically plane plates to a separating piece so as to obtain perfect parallelism between the surfaces. Certain experiments recently performed, in the research laboratories of Adam Hilger, by R. G. Parker and A. J. DALLADay give excel- lent promise of overcoming all of the disadvantages pertaining to the use of cements. The basic idea consists in heating the pieces of glass— while under pressure and in optical contact—until they unite and form one single object. The necessary conditions of perfect figuring and absolute clean- liness present no insurmountable difficulties. The problem there- fore resolves itself into a careful study of the proper thermal treatment. The temperature for complete union must not only be below the melting point of the particular kind of glass under investigation, but it must also be inferior to the annealing tem- perature, for, even at the lower temperature sight non-uniform stress will give rise to permanent deformation of the optically per- fect surfaces. By keeping two pieces of the same kind of glass in optical contact, for about one hour, (under suitably applied pressure) at a temperature 60° C. or 70°C. below the annealing temperature, it was found that complete union, without deforma- tion or appreciable strain, resulted. This significant result was deduced from the empirical formula Teg established by the investigators. They found that S;=S,,A %, where S,; and S;, are the ‘“ mobilities” at the temperatures T and T, respectively,—measured by the rate of deformation of glass under constant stress,—and A and N are constants for the special sort of glass concerned. The authors say: “The quantity N is comparatively small, so that for quite small values of T—T,, T— T the index * would become equal to 1, 2, 3, ... If, for ex- ample, we take N= 8, A= 2, and compare the values of the mobility of the particular glass at temperatures of T = 560° C. and T, = 552° C., 544°, 536° successively, we find the respective values of the mobility to be Sr 1 rr -90° —— = 2’ when: T’ = 552° C. Tj = 2* = 4 when T, = 544° C. == 2218 when t= S367. That is, the glass becomes twice as “soft” for every rise of temperature of 8°C. Therefore, at 60° or 70°C. below the Chemistry and Physics. 413 annealing temperature a glass has become sufficiently hard to withstand very great local pressure for short periods without sensible deformation.” By taking suitable precautions in regulating the temperature of an electric furnace the experimenters have succeeded in making perfect cells of the type involved in the Rayleigh interference- refractometer. The inner opposing walls were as plane and parallel as the outer surfaces of the cell windows. The parts of a polarimeter tube of soda-lime glass with end windows of plate glass were readily united at a temperature near 470°C. For obvious reasons, success was not attained in attempting to unify an object-glass the component parts of which were made of ordinary crown and flint glass. Although the problem of con- structing ultra-violet absorption cells of fused silica has not been completely solved, the results thus far obtained give promise of ultimate success.— Phil. Mag., xxxili, p. 276, March, 1917. Hy 8s B: 6. The Flame Spectrum of Iron.—A promising method for differentiating and classifying the lines of complicated spectra has been recently developed, and applied to the case of iron, by G. A. HemsatecH. The special features of the apparatus employed consist in the ‘“‘ electric sprayer’ and the gas burners. The most essential parts of the sprayer may be described as follows: An inverted bell-jar (height 10°5 in., diameter 5 in.) is closed at the lower, narrow end by a rubber stopper and at the upper, wide end by a flanged wooden disk. supplementary spectrum.” For details concerning the structure of the iron spectrum refer- ence must be made to the original article.-— Phil. Mag., xxxiil,_ p. 1, January, 1917. H. 5. U. 7. The Nature of Matter and Electricity ; by Danizt F. Comstock and Lronarp T. Trotanp. Pp. xxii, 203, with 11 plates. New York, 1917 (D. Van Nostrand Co.).—“ ‘This book attempts to give in broad, schematic form the conception of the structure of the material universe which has developed in the minds of modern students of physical science.” The text is divided into two parts, the first of which is based on a series of Chemistry and Physics. 415 articles contributed in 1911 by Comstock to ‘Science Con- spectus.” The material of Part I, which is presented in an extremely elementary manner, has been amplified and brought up to date by the original author Part IT, from the pen of Troland, is virtually an appendix and it consists of fifty-six sections relat- ing to specific problems and details appropriately omitted in the rapid survey of the entire field made in Part I. This field covers the following subjects: the properties and behavior of atoms and molecules; the nature of heat, Brownian movement, absolute zero, etc.; the electron and its behavior; electrons, chemical action, and light; electrons and magnetism ; radio-activity; the structure of the atom; atomic numbers, the quantum hypothesis, radiation, and X-rays; atoms and life. The material of Part I is presented in such a pleasing style that the writer of this notice was constrained to read every sen- tence in it. Frequent reference to Part IL showed that it has also been written with care and that it constitutes an almost indis- pensable supplement to the earlier pages, in the sense that it con- tains just the right things to help satisfy the desire for more information so skilfully excited in Part I. The full-page plates are unusually good and no pains seem to have been spared by both authors and publishers to make the volume as attractive as possible. fH SSR 8. Hlectric and Magnetic Measurements ; by CHarLtEs Mar- quis Smiru. Pp. xii, 373, with 171 figures. New York, 1917 (The Macmillan Co.).—The degree of advancement of this text may be inferred from the facts that the equivalent of one year of general physics and some knowledge of the calculus are presup- posed. Although the ground covered by the fifty-six laboratory exercises does not seem to afford anything especially novel, never- theless the book possesses several valuable features. For illustra- tion, much space is devoted to the general theory upon which the experiments are based, special attention being paid to definitions and to the precision of measurements. In other words, the text has been evolved from a course of lectures as well as from labora- tory notes. Another salient feature is that most of the experi- ments are described in such a way as not to require particular types of apparatus. The exceptional cases involve apparatus which is well known and generally available. A few theoreti- cal problems, for solution by the student, are given at the ends of the earlier chapters. ‘The index is preceded by an appendix which contains an outline of the absolute measurement of current and of resistance, a table of constants, and a list of standard reference books. Because of the clearness and generality of the explana- tions the volume merits the serious attention of all who are engaged in conducting laboratory courses in pure as well as in applied physics. He 8. Uz 9. Recreations in Mathematics ; by H. E. Licks. Pp. v, 155. New York, 1917 (D. Van Nostrand Co.).—“ The object of this book is to afford recreation for an idle hour and to excite the 416 Scientific Intelligence. interest of young students in further mathematical inquiries.” The material selected is of such a heterogeneous character as to be susceptible only of the following very broad classification: arithme- tic, algebra, geometry, trigonometry, analytic geometry, calculus, astronomy and the calendar, mechanics, and physics. The num- ber of “recreations” is 180. These include puzzles, paradoxes, fallacies, historical memoranda, and serious remarks introductory to the study of the more advanced subjects. Taken as a whole the book should be found both interesting and instructive by those to whom it isaddressed. In two articles, however, the author makes statements which are either open to very serious objection or are altogether incorrect. In section 137 he concludes that mechanics ought to be omitted from courses in physics because the teacher of pure science makes use of dynes and poundals and since “ . . . no apparatus for measuring forces in such units has ever been made or used.” Article 138 contains several statements which are unqualifiedly wrong. For example, the unit of acceleration is given as ‘one unit of length per sec- ond.” With regard to the dimensional equation | #| = [ZL 7] he writes : “The student tries hard to comprehend this, but finds it impossible, for he knows that force is not ML./T° and he knows that there is no way to measure a force except by the number of units of force which it contains. The truth of the matter is that the equation / = m/f is not true.” Hy 8;/ Ue II. Grontogy AND MINERALOGY. 1. The Andes of Southern Peru. by Isatan Bowman, Director of the American Geographical Society. Pp. xi, 336, with 204 figures and 7 topographic sheets. New York, 1916 (Henry Holt and Company).—The attractively printed and beautifully illustrated ‘“ Andes of Southern Peru” is essentially the record of a reconnaissance along the seventy-third meridian, between par- allels of latitude 12° and 16° 37’ south. The book consists of two parts. In Part I, Dr. Bowman gives a well-composed picture of the life of the people, in which the factor of environment is emphasized. After reading the chapters on The Canyons of the Urubamba, The Rubber Forests, The Forest Indians, The Country of the Shepherds, and The Border Valleys of the Eastern Andes, one feels that he knows both the Indians and the Spanish planters. Two chapters on climate record and discuss the sig- nificance of observations made by the Yale Peruvian Expedi- tions, the Harvard Observatory, the Peruvian Government, and the Geographical Society of Lima. The Physiography of the Peruvian Andes is treated under the headings: The Peruvian Landscape, The Western Andes, The Eastern Andes, The Coastal Terraces, Physiographic and Geo- logic Development, and Glacial Features. The most significant Geology and Mineralogy. 417 topographic elements of southern Peru are: 1. A system of high level slopes of gentle gradient and great areal extent developed on rocks of various types and attitudes, deeply covered by soil and standing at an elevation 4,000 to 5,000 feet above the level at which they were formed; 2. deep, steep-walled, narrow-floored canyons sunk below the plateau surface; 3. lofty, residual, highly sculptured mountains; 4. voleanic cones and lava plateaus of the western Cordilleras; 5. glacial features at valley heads; 6. deep alluvial fill in valley bottoms, now in process of dissection. The system of high level slopes is substantially the surface of denuda- tion elsewhere called the Inca peneplain (Gregory, H. E., A geo- logic reconnaissance of the Cuzco Valley, Peru: this Journal, vol. xli, pp. 1 to 100). Dr. Bowman finds that along the seventy- third meridian a stage of maturity or post-maturity, but not peneplanation, has been reached and considers the well-marked peneplaned areas in the Cuzco region and in Bolivia as remnants of an older land surface. It appears that the extent of the Inca peneplain and its relation to other features must await further field work. The canyons sunk below the level of the Peruvian plateau are represented by the Urubamba (pp. 8 to 21) and the mountains rising above that surface by the Cordillera Vilcapampa (pp. 202 to 224). A separate chapter is devoted to glacial features. The strictly geologic results of this geographic reconnaissance are appropriately meager. Geologic maps and measured sections are lacking, and conclusions drawn regarding geologic history necessarily rest on scant field evidence and are subject to radical revision. The oldest sediments recognized along the seventy- third meridian are unfossiliferous slates and shales, provisionally assigned to the Silurian, and underlain by early Paleozoic (?) schists. ‘The Upper Pennsylvanian is widely represented and fossils were obtained at Huascatay, Pampaconas and Pongo de Mainique. Fossils from 2,000 feet of limestone at Cotahuasi were determined by Professor Schuchert as Lower Cretaceous. The physical history of the Tertiary of the Coastal region is pre- sented in a new form. ‘The lacustrine deposits of the Cuzco and other intermontane basins, also considered Tertiary, are probably late Pliocene or Pleistocene (Eaton, G. F., Vertebrate fossils from Ayusbamba, Peru: this Journal, vol. xxxvii, 1914, pp. 141 to 154, and papers by the present writer: this Journal, vol. xxxvii, 1914, pp. 125 to 140, and vol. xli, 1916, pp. 78 to 85). When one realizes the advancement of geographic knowledge of the Andes represented by this book and notes that it has upwards of 40 smaller companions resulting from the devotion of Professor Hiram Bingham to pioneer scientific work, he can but regret that the work of the Yale Peruvian Expeditions has been suspended, ‘The topographic maps of Kai Hendriksen and A. H. Bumstead alone justify the expenditure of time and money. | H. EB. G 2. Mount Rainier, A Record of Exploration, edited by Epmonp 8. Meany. Pp. xi, 325,17 pls. New York, 1916 (The 418 Scientific Intelligence. Macmillan Company).—Professor Meany’s book on Mount Rainier bears the evidence of accuracy of statement and judicious selec- tion of material. The discovery of the mountain by Captain Van- couver in 1792, its first approach by Doctor Tolmie in 1833, the first recorded trip through Naches Pass by Lieutenant Johnson in 1841, the attempted ascent of the mountain by Lientenant Kautz in 1857, the successful ascent by General Stevens and by 8. F. Hmmons in 1870 are recorded in the explorers’ own words. Pro- fessor Russell’s account of explorations of the glaciers of Mount Rainier in 1896 is reproduced from the Annual Report of the United States Geological Survey. Chapters on Glaciers of Mount Rainier, by F. E. Matthes ; on The Rocks of Mount Rainier, by George Otis Smith ; and on The Flora of Mount Rainier, by Pro- fessor Piper, are authoritative scientific descriptions. Accounts of Indian legends and an annoted list of place names are interest- ing and useful features. H. E. G. 3. Publications of the United States Geological Survey, GEORGE Oris Suiru, Director.—Publications of the U.S. Geological Sur- vey recently received are given in the following list (continued from vol. xli, pp. 440, 441): . Topocraruic ATLAs.—Seventy-two sheets. Foutos.—No. 200. Galena—Elizabeth Folio, Illinois-lowa; by E. W. SHaw and A. C. Trowsrivcr. Surveyed in cooperation with the Geological Survey of Illinois. Pp. 13; with 2 pls. topography, 2 colored pls. areal geology, 1 pl.of 13 half-tone ilus- trations. No. 208. Colorado Springs Folio, Colorado; by Grores I. Finuay. Pp. 15; 3 pp. sections, 2 pls. topography, 2 pls. areal geology, 2 pls. half-tone illustrations. PROFESSIONAL PapeErs.—Shorter Contributions to General Geology. Chapter L, The Pliocene Citronelle Formation of the Gulf Coastal Plain and its Flora. Papers by Gzorcr C. Matson and Epwarp W. Berry. Pp. iv, 167-204; pls. XX XII-XLVIL, 3 figs. O. Relations of the Embar and Chugwater Formations in Central Wyoming; by D. Date Conpir. Pp. ii, 263-270; 3 pls. 2 figs. P. 1. Stratigraphy of a part of the Chaco River Valley, New Mexico; by Crypz M. Baver. Pp. ii, 271-278; pls. LXIV-LXXI, 1 fig. Bu.ietins.—No. 631. The Yukon—Koyukuk Region, Alaska; by Henry M. Eakin. Pp. 88; 10 pls, 2 figs. No. 640, 641. Contributions to Economie Geology, 1919. 640, Part dT. KI... 641, -Part I ;G, a: No. 642. Mineral Resources of Alaska—Report on progress of investigations in 1915; by AtFrep H. Brooks and others. Pp. 279, x; 11 pls., 14 figs. No. 648 Notes on some Mining Districts in Eastern Nevada; by James M. Hitt. Pp. 213; 6 pls., 18 figs. | Nos. 637, 639, 643, 646. Spirit Leveling, R. B. MarsHatt, Chief Geographer. No.637. Texas, 1896-1915. Pp. 254; 1 pl. Geology and Mineralogy. 419 No. 639. Mississippi, 1901-1915. Pp. 80; 1 pl. No. 643. South Dakota, 1896-1915. Pp. 100; 1 pl. No. 646. North Carolina, 1896-1914. Pp. 71; 1 pl. Warer Suprry Parers.—Surface Water Supply of Ve United States, 1914. N.C. Grover, Chief Hydraulic Engineer. No. 381. Part I. North Atlantic Slope Drainage Basin. Pp. 195; Xxxvii; 2 pls. No. 382. Part II. South Atlantic and Eastern Gulf of Mexico Basins. Pp. 66, xxx; 2 pls. No. 393. Part XJI. North Pacific Drainage Basins, B. Snake River Basin. Pp. 248; 2 pls. No. 400. Contributions to the Hydrology of the United States, 1916. Chapters B, C, D. The Thirty- seventh Annual Report of the Director for the Year ending June 30, 1916, is announced as having been pub- lished under date of January 1, 1917; it has not as yet been received at this office. 4. West Virginia Geological Survey, I. C. Wuuirr, State . Geologist. Jefferson, Berkeley, and Morgan Counties; by G. P. GRIMSLEY, Assistant Geologist. Pp. xxvi, 644; 3 maps in Atlas (under separate cover), 37 pls., 20 figs.—The area, covered by this detailed report, is especially noteworthy because it contains vast deposits of the purest limestones in the country ; also ‘Immense deposits of pure dolomites and vast quantities of glass- sands; also many other valuable and interesting mineral deposits, including clays, road materials, iron ores, etc. The price of the volume, including case of maps (delivery charges paid), is $2.50; extra copies of geologic map, $1.00 each, and of the topogr aphic map, 50 cents each. 5. The Inorganic Constituents of Marine Invertebrates ; by Frank W. CrarKEand WaLtTER C. WHEELER. Prof. Paper 102, U.S. Geol. Survey, 1917, 56 pp.—This memoir, interesting alike to zoologists, paleontologists, and geologists, presents and dis- cusses 250 new analyses, along with many old ones, of calcareous algae and all of the chief shell-bearing marine invertebrates. It is shown that the foraminifers, aleyonarians, echinoderms, bryo- zoans, and crustaceans use in their skeletal structures from 5 to 25 per cent of magnesium carbonate, and that the quantity varies with the temperature of the water. The greatest amount is found in the warm waters and in association with calcite, whereas the aragonitic structures are essentially non-magnesian. C. 8. 6. A Synopsis of American EHarly Tertiary Cheilostome Bryozoa ; by Ferpinanp Canu and Ray 8. Basster. Bull. 96, U.S. National Museum, 1917, 87 pp., 6 pls.—In this very import- ant synopsis relating to the classification of American Cenozoic Cheilostomata are also described 50 new genera and 42 new species. For four years the authors have been developing these bryozoans, which now number nearly 500 species from the Eocene and Oligocene. Eventually all will be described in an extensive monograph now in preparation. 0. 8. 420 Scientific Intelligence. 7. Note on Goyazite ; by Ouiver C. Farrineron (communi- cated).—The writer desires to express his thanks to Dr. W. T. Schaller for his kindness in pointing out an error in the writer’s quotation of the percentage of P,O, m hamlinite.* It is worth noting, however, that the corrected figure makes the difference in the percentage of P,O, in goyazite and hamlinite much greater than the value which the writer used. Corrected, the relation - between goyazite and hamlinite is : Goyazite Hamlinite . Percent. O22. eee ew eee 14°87 |. 28°92 This discrepancy seems too great to allow the two to be consid- ered the same mineral, especially as no inaccuracy in Damour’s determination has been proved. Moreover the statement of Damour iu regard to goyazite that “it fuses with difficulty on the edges of the smallest fragments” does not describe a fusibility of 4, Again, Hussakt does not assert from his examination of Damour’s voyazite that there was no calcium present, but simply that it possessed a ‘“ very strong” strontium content in compari- son to that of calcium, (einen im Vergleich zum Kalkgehalt sehr starken Strontiumgehalt besitzt”). The entire absence of cal- cium is especially noted by Penfield{ as a feature of hamlinite, and Bowman found no calcium in the hamlinite which he analyzed.§ In view of these differences the identity of goyazite and hamlinite cannot be said to be yet proved. 8. Hlements of Mineralogy, Crystallography and Blowpipe Analysis; by A. J. Moses and C. L. Parsons. 5th edition. Pp. 631, 575 figs. New York, 1916 (Van Nostrand Co.).—This is a new edition of a well known text-book. The changes in this edition, which have involved the addition of almost two hundred pages, have been chiefly as follows: The description of new species and economic groups, a more detailed discussion of the occurrence and genesis of minerals, an enlarged chapter on the optical properties of minerals and new determinative tables. An interesting addition to the latter is the inclusion of optical tests on fragments of the crushed mineral. ‘These are given in addi- tion to the ordinary chemical and physical tests. The book covers a very wide field and its method of treatment is necessarily very concise. At times one is inclined to wonder if this condensa- tion, admirably as it is done, does not make some portions of the subject too difficult for the ordinary student. The book is excel-— lently printed and illustrated. W.E. F. 9. The Optical Character of Sulphatic Cancrinite ; a Correc- tion; by E. 8. Larsen (communicated).—In the table in the middle of page 333 of volume xli of this Journal (October, 1916), the optical characters of sulphatic cancrinite, cancrinite, and natrodavyne were inadvertently stated incorrectly but are stated correctly in the text. Sulphatic cancrinite and cancrinite are optically negative while natrodavyne is optically positive. * This Journal (4) xliii, p. 163, 1916. + Tsch. Mitth., xxv, p. 340, 1906. ¢ This Journal (4) iv, p. 314, 1897. § Min. Mag., xiv, p. 392, 1907. Miscellaneous Intellagence. 421 ITI. Misce,rutaAnrovus Screntiric INTELLIGENCE. 1. Morphology of Invertebrate Types; by ALEXANDER PETRUNKEVITCH. Pp. xili, 263, with 50 figures. New York, 1916 (The Macmillan Company).—This book is essentially a lab- oratory guide to the dissection of the principal types of inverte- brate animals. It differs in important respects from the other books in this field, particularly in insisting upon a more thorough and detailed study of invertebrate anatomy than has hitherto been possible for any except the most advanced students. To this end the book gives explicit directions for the complete dis- section and study of all the organ systems of the animals included instead of a few superficial structures. Each chapter contains a brief summary of the anatomical features of the species to be studied. ‘This descriptive part explains the technical terms employed and is accompanied by a general diagram of the ani- mal’s anatomy. It is to be read by the student before he begins the laboratory exercise. ‘There are also brief directions for obtaining the materials required and for preparing them for study. Wie Ra C. 2, Microbiology: A Text-book of Microorganisms General and Applied ; edited by CuarLes E. MarsHaty. Second edition, revised and enlarged. Pp. xxiv, 900, with 186 figures. Phila- delphia, 1917 (P. Blakiston’s Son & Co.).—As in the first edition of this widely adopted text-book, the subject matter is arranged in three distinct parts: I, Morphology and Culture of Microorganisms, includes a general and systematic summary of the structure and classification of molds, yeasts, bacteria and protozoa, with a brief account of the invisible microorganisms. Part II, Physiology of Microorganisms, contains chapters on nutrition, products of metabolism, mechanism of metabolism, and the influence of mois- ture, temperature, light, electricity and other physical and chemi- cal conditions on the life of the organisms. Part UI, Applied Microbiology, is subdivided into chapters on the economic import- ance of the microorganisms of soil, water, sewage, milk and milk products, and special industries, together with the microbial dis- eases of plants, insects, domestic animals and man, andthe methods of eontrol of infectious diseases. The work is in reality more than a text-book ; it is a reliable and authoritative summary of the entire knowledge of the micro- organisms of economic importance. It is the combined effort of many specialists, for in addition to the able work of the editor, the different features of the subject have been contributed by twenty-five of our foremost microbiologists. Wie Be. 3. Growth in Length: Embryological Essays ; by RicHarp AsSsHETON. Pp. xi, 104, with 42 figures. Cambridge, 1916 (University Press).—The first part of this little book contains three lectures on the formation of the embryo in the different groups of vertebrates, supporting the theory that the chordates 429 | Scientific Intelligence. have originated directly from a coelenterate ancestor rather than from the annelids, arthropods, or other highly specialized inver- tebrates. The second part consists of a reprint entitled “ The Geometrical Kelation of the Nuclei in an Invaginating Gastrula (e. g. Amphioxus) considered in connection with Cell Rhythm and Driesch’s Conception of Entelechy.” 'The early develop- mental stages may be explained by assuming the presence of a vital attractive force acting between the adjacent cells of the embryo. The book was prepared for the press by the author’s widow. W. BG. 4. The Respiratory Hxchange of Animals and Man; by Aveust Kroeu. Pp. vii, 1738. London, 1916 (Longmans, Green & Co.).—This is a satisfactory attempt to “trace out a course through the ocean of literature” that has accumulated since Lavoisier’s classic researches that first indicated more than a century ago something of the true nature of the respiratory exchange. The reputation of the author, reader in zoophysiology at the University of Copenhagen, as an expert in this field of science gives adequate assurance of intelligent treatment. The topics reviewed are grouped under the following headings: the physio- logical significance of the exchange of oxygen and carbon dioxide; methods for measuring the respiratory exchange ; the exchange of nitrogen, hydrogen, methane, ammonia, and other gases of minor importance ; the standard metabolism of the organism ; defini- tion and determination; the influence of internal factors upon the standard metabolism ; the influence of chemical factors upon the respiratory exchange ; the influence of physical factors upon the respiratory exchange ; the variations in standard metabolism dur- ing the life cycle of the individual ; the respiratory exchange in different animals. The volume is a fitting companion to the others in this series of monographs on biochemistry. _—_L. B. M. 5. Memoirs of the Queensland Museum, Vol. V.,; edited by the Director, R. Hamiyn-Harris. Pp. 234; 25 pls., 21 figs. Issued July 10, 1916. Brisbane (Anthony J. Cumming, Govern- ment Printer).—This new volume from the Queensland Museum is chiefly devoted, as were the earlier issues, to articles on natural history, especially zoology. Some fifteen of these appear, well illustrated by twenty-five plates. The opening paper by the Director, associated with Mr. Frank Smith, discusses fish poison- ing and poisons employed by the aborigines of Queensland, where the practice of stupefying and killing fish by this means has been long in vogue. Queensland fishes are discussed at length by J. Douglas Ogilby and A. R. McCulloch ; the two gentlemen together give an extended review of Australian therapons, while Mr. Ogilby continues his studies of the edible fish of the colony. (Parts iv-ix). Papers on Australian fish scales and on Queens- land bees are contributed by Professor T. D. A. Cockerell of the University of Colorado. Warp’s Naturat 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. J-148. Price List of Rocks. Mineralogy: J-109. Blowpipe Collections. J-74. Meteor- ites, J-150. Collections. J-160. Fine specimens. 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-185. Bacteria Slides. Taxidermy: J-188. Bird Skins. J-139. Mammal Skins. Human Anatomy: J-16. Skeletons and Models. General: J-155. List of Catalogues and Circulars. Ward’s Natural Science Establishment 84-102 College Ave., Rochester, N. Y., U.S. A. The American Journal of Science ESTABLISHED BY BENJAMIN SILLIMAN IN 1818. CONTRIBUTORS should send their articles two months before the time of issuing the number for which they are intended. The title of communications and the names of authors must be fully given. Notice is always to be given when com- munications offered have been, or are to be, published also in other Journals. Thirty separate copies of each article will be furnished to the author free of cost and without previous notice from him. They will be provided with a plain cover (but with reference to volume and year). If the author orders separate copies, they will be understood to be in addition to the thirty mentioned above, and he will receive a bill for the extra expense involved, as also for that of a printed cover (with title, etc.), when this is specially ordered. These charges will conform to the following schedule; but will be increased if there are plates. No. Copies. 50 100 200 300 500 B pages... _--| $2.25 $2.75 $3.50 $4.95 $5.50 AG ee ee eet a | 850 4.25 5.25 6.25 8.00 | Bee Voy ie 4.95 5.15 7.00 8.25 10.50 Cavemen rr $1.00 $1.25 $1.75 $2.25 | $3.00 (2s During the Paper Famine the above prices may be somewhat increased (e. g., from 5 to 10 p. c.). CONTENTS. x Page Art. XXXI.—The Geology of the Lau Islands; by W. G. WOVE S22 o8 SO a eae ee ae XX XII.—Origin and Age of the Ontario Shore-line,—Birth of the Modern Saint Lawrence River; by J. W. SpENcER 351 XX XIII.—Granite Bowlders in (?) the Pennsylvanian Strata of Kansas; by W: H: Twunnormn-.-<**). 3) 2 eae XXXIV.—An Oligocene Camel, Poébrotherium andersoni nesp.3* by ke dh: AERO RS 22 See een ere 2 38a XX XV.—Sand Fusions from Gun Cotton $ by C. E. Mae 389 XXXVI.—Electrolytic Analysis with Small Platinum Elec- trodes; by F, A. Goocu and M. KopayAsui__--..-.--- 391 XXXVII.—Crystal Drawing and Modeling; by J. M. Brake 397 XXXVIII.—Normal Anomalies of the Mean Annual a perature Variation; by H: ARcTOWSKI - 222 > S223 - 402 SCIENTIFIC INTELLIGENCE. : ‘ Chemistry and Physics— Attempt to Separate the Isotopic Forms of Lead by Fractional Recrystallization, T. W. RicHarps and N. F. Hatt, 409.— Manganese in Soils, M. O. Jounson, 410.— Preparation of Sulphurous Acid, E, Hart: Gas Chemists’ Handbook: Union of Glass in Optical Contact by Heat Treatment, R. G. PARKER and A. J. Datuapay, 411.—Flame Spec- trum of Iron, G. A. HEMSALECH, 413.—Nature of Matter and Electricity, D. F. Comstock and L. T. TrRotanp, 414.—Electric and Magnetic Measure- ments, C. M. SmirH: Recreations in Mathematics, H. E. Licxs, 415. Geology and Mineralogy—The Andes of Southern Peru, I. Bowman, 416.— Mount Rainier, a Record of Exploration, 417.—Publications of the United - States Geological Survey,~G. O. Smit, 418.—West Virginia Geological Survey, I. C. Waite: Inorganic Constituents of Marine Invertebrates, F. W. Cuarke and W. C. WHEELER: Synopsis of American Early Tertiary Cheilo- stome Bryozoa, F. Canu and R.S. Bassuier, 419.—Note on Goyazite, O. C. FARRINGTON: Elements of Mineralogy, Crystallography and Blowpipe Analysis, A. J. Moses and C. L. Parsons: Optical Character of Sulphatic Cancrinite, E. S. Larsen, 420. Miscellaneous Scientific Intelligence—Morphology of Invertebrate Types, A. PETRUNKEVITCH: Microbiology—Text-book of Microorganisms General and Applied: Growth in Length: Embryological Essays, R. ASSHETON, 421.—Respiratory Exchange of Animals and Man, A. KrocH: Memoirs of the Queensland Museum, Vol. V, R. HamytyN-HarRris, 422. ig Rey ogres ane a as iy ia ot gt Se - Noy" poe “aS on eS ee eS | - * ee ee ~ » #4 a ee A ee Lf Ba eee sibrary, U. at. Museum. VOU. XLII. | | JUNE, 1917. ue ee” tan Sr: Established by BENJAMIN SILLIMAN in 1818. THE ~ ‘JOURNAL OF SCIENCE Epirorn: EDWARD S. DANA. AMERICAN = ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or Camsruinge, Proressors ADDISON E. VERRILL, HORACE L. WELLS, LOUIS V. PIRSSON, HERBERT EF. GREGORY anp HORACE S. UHLER, or New Haven, Proressor HENRY S. WILLIAMS, or IrHaca, Proressor JOSEPH S. AMES, or Battiwore, Mr. J. S. DILLER, or Wasuinerton. FOURTH SERIES VOL. XLIII-[WHOLE NUMBER, CXCIII}. No. 258—JUNE, 1917. ns JUN 2-197 ss WV. é NEW HAVEN, CONNECTICUT; 3) yyseo™ WITH PLATE II. Loy | THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET, a Published monthly. Six dollars per ‘year, in advance. $6.40 to countries in the ~ Postal Union ; $6.25 to Canada. Single numbers 50 cents. _ Entered as second-class matter at the Post Office at New Haven, Conn., under the Act of March 3, 1879. . 2" LIST OF CHOICE SPECIMENS. Stibnite, Schinokawa, Province of Igo, Japan. Beautiful specimen 7 x 04"; over one dozen crystalsembedded. $15... — Benitoite-neptunite, San Benito Co., Culifornia. 6x 44”. .Good many bright crystals of benitoite and neptunite embedded ; very showy. $18 Ilmenite in serpentine, Snarum, Norway. 4x3": dozen crystals , y ; 7; embedded. $12. Copper, xlized, Bisbee, Arizona. 34x38”. $65. Dioptase, Khirgese Steppes, Siberia. 4% x 22"; number of large crystals of fine color embedded. $38. Stolzite on limonite, Broken Hill, N. S. W. 44x 3"; face covered with brilliant brownish red crystals. $18. Stolzite on limonite, Broken Hill, N. S. W. 24x 12”; specimen is covered with brownish red crystals, very choice. $15. Crocoite, Tasmania. 3x 3"; whole face is covered with fine brilliant erystals of rich red color. $10. Crocoite, Beresov, Ural.. 32x 2". $5. Crocoite, Beresov, Ural. 24x ee $4. Realgar, Allchar, Macedonia. 2} x14". $3.50. Linarite, California. Very brilliant azure blue crystals; 2x14". $4. Gersdorfiite, Steinbach, Germany. Very fine sae derelonc’ crys- tals; 24 x2". $7.50. ere Cripple Creek, Colorado. 3x 12"; whole face covered with large brilliant crystals. $8. Calaverite, Cripple Creek, Colorado. 24x14"; very rich, face covered with calaverite and shows veins of calaverite. Reticulated gold on quartz, Silver Peak Mine, California. 3x 2". $7.50. Silver, xlized, Batopilas, Mexico. 24x 14"; 2 faces covered with bril- liant crystals. $8. Wire silver, Zacatecas, Mexico. 2x13". $8. Tellurium, Colorado. 2x 13"; very rich, partly orystallized. $3. Sphene, Albrum Paes, Switzerland. 4} x8": good many twin xls. . embedded. $5. Epidote, Canton Uri, Switzerland. 34x 14"; number of transparent gem xls. embedded and one large xl. 14” long. $12. Epidote, Canton Uri, Switzerland. 38} x 2"; single crystal, doubly ter- minated, translucent. $7.50. Aquamarine, Siberia. 2” long, #" diameter; doubly terminated, abso- lutely clear gem xl, $22.50. I have just received one of the finest lots of phlogopite ever found at Quebec and Ottawa, Canada. About forty specimens, loose crystals from 2" to 8” and matrix specimens from 3” to 5". Very bright and showy, some associated with blue apatite and some with muscovite intergrown. Price 10c to $4. ALBERT H. PETEREIT 81-83 Fulton St., New York City THE AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES. | oe Art. XX XIX.—The Geology of Pigeon Point, Minnesota ; Reeinatp A. Dary, Harvard University. CONTENTS. Introduction. Revised map. Main intrusive a sill. Tilting after injection of the sill. Stratiform structure of the sill. Segregation of red rock through gas tension. Ribbon injections. Time relations of the magmatic phases. Differentiation necessarily postulated. orgie Pete MW, Nature of the magma differentiated. % . mest ti i al psd Kvidences of assimilation. Red-rock shells surrounding xenoliths. Magmatic stoping. Capacity of the original magma for assimilation and contact fusion. Assimilation in the sill conduit. Summary on the origin of the red rock. Lntroduction.—During the summer of 1916 the writer, in company with Professor Charles Palache of Harvard Univer- sity, spent six days in a field examination of the intrusive body which has become celebrated through the able memoir by Bay- ley, entitled “ The Eruptive and Sedimentary Rocks of Pigeon Point, Minnesota, and Their Contact Phenomena.”* With the kind assistance and personal guidance of Professor F. F. Grout of the University of Minnesota, the point was found to be easily accessible. Leaving the regular steamer from Duluth to Port Arthur at Grand Portage, six miles west of the point, the party there boarded the motor-boat of Mr. P. Gagnon and were soon comfortably camped in Little Portage Bay, an excel- lent center for the study of the whole eruptive area, especially *W.S. Bayley, Bulletin 109, United States Geological Survey, 1893. Am. Jour. Sci1.—FourtH SERIES, Vou. XLIII, No. 258.—Junz, 1917. 29 Pa Aone LL , £9 73 SUN Qe Old - , yA 494 RP. A. Daly—Geology of Pigeon Point, Minnesota. because transportation by the motor-boats of resident fishermen could be secured. The objects of the visit were to secure personal acquaintance with the rocks so admirably described by Bayley, and to collect structural and petrographical material bearing on the origin of the local “red rock,” which represents one of the main prob- lems of Minnesota geology and, yet more significantly, falls under the general question regarding the origin of the world’s granites. | The red rock has been called “ quartz keratophyre” and also “soda granite.” It is a highly micropegmatitic or granophyric rock, chemically very similar to many common granites, but locally it has a chemical composition much like a quartz sye- nite. The red rock passes gradually into “ intermediate rock,” that varies, chemically, from a monzonite-like type to a diorite- like type. The intermediate rock, always in small volume, in turn grades into gabbro. The latter is much the largest com- ponent of the main Pigeon Point intrusive. It has olivine-rich and olivine-free phases. Though possessing a somewhat low content of alkalies, the gabbro has nearly normal characters. For the full petrography of the eruptive and sedimentary rocks of the point, the reader is referred to Bayley’s memoir. Short as the field study was, it permitted a revision of the geological map and structure sections published in Plate XVI of Bayley’s paper. A map showing necessary changes in the lines of contact is given in the accompanying fig. 1. So far as the outcrops allow a decision, the main intrusive appears to be a sill cutting the local Animikie sedimentaries, rather than a dike, as stated by Bayley. The red rock is concluded to be a gravitative differentiate from the magma which filled the sill chamber. As Bayley and others proved long ago, the origin of that magma is a question involving the possibility of the fusion of the Animikie quartzites and metargillites by primary gabbroid magma. While Bayley showed the probable reality of this melting-up, its absolute proof, here as so commonly at eruptive contacts, is obscured by the processes of differentia- tion. A special novelty discovered is the existence of minute igne- ous bodies which are hereafter described under the name “Jinear or ribbon injections.” Professors Palache and Grout have greatly aided in the col- lection of facts and in clarifying the discussion, but neither should be held responsible for the theoretical views hereafter expressed. Revised Map.—Pigeon Point is a peninsula stretching east- ward from the main shore of Lake Superior, 150 statute miles (240 kilometers) northeast of Duluth. The peninsula is six R. A. Daly—Geology of Pigeon Point, Minnesota. 425 miles (10 kilometers) long; its maximum width is a little more than one mile. Its area is small; yet, on account of a dense growth of moss and trees over much of the interior, a correct delineation of all significant contacts of the rocks must take much time. The scant six days spent on the point did not suf- fice for a complete field review of the geology, but it was possible, with the liberal and highly expert help of Professor Palache, to trace most of the contacts in the eastern third of the peninsula, an area of about 2 square kilometers ; and to run Bigs le fee ss GING 1 al) A LA FEE CO) Ee reel (Our Fic. 1. The eastern part of Pigeon Point. Animikie sediments, left blank ; 7, gabbro; 2, intermediate rock ; 3, red rock. Diabase and gabbro dikes in solid black. Sediments and sill (—3) dip about 15° in a direction somewhat east of south. a few cross-sections in the broader part of the peninsula, to the westward. In general, the maps of the western two-thirds of the penin- sula, shown in Bayley’s Plates XIV and XV, were found not to need any essential change. The eastern third is likewise accurately mapped by Bayley so far as the shore zones are con- cerned. Close study of the less well-exposed interior showed, however, the necessity of significant changes in the mapping of contacts, though no vital improvement can be offered on his remarkably thorough diagnosis of the kind of rocks occurring in the area. The most important alterations of the geological map are noted in the isthmus at Little Portage Bay (“ #” in fig. 1) and 496 R. A. Daly—Geology of Pigeon Point, Minnesota. to the eastward thereof. The red rock of the isthmus has throughout a general east-west contact with the mass of the intermediate rock and gabbro, and not a north-south contact, as indicated on Bayley’s map. The patch of red rock and intermediate rock mapped by him, in a narrow band stretching south toward “ D” from the little bay at “ B,”’ fig. 1, does not exist. Finally, the southern edge of the gabbro locally extends farther south, both to east and west of Little Portage Bay, than as drawn on Bayley’s map. Special care was taken to be sure of these necessary changes and, in general, fig. 1 is believed to be essentially correct, in spite of the fact that the upper part of the gabbro carries some interstitial micropegmatite and is thus not always to be distinguished with ease from the inter- mediate rock. It is seen that the map is distinctly simplified by these changes. The main gabbro is continuous and not cut across by any large body of red rock. The red rock is nearly continuous, lying between the main gabbro on the north and the quartz- ite-metargillite series of the south. The intermediate rock is nearly as continuous; it les between the gabbro and the red rock, again with persistent trend of. outcrop directed nearly east and west. Main Intrusive a Sill... D. Irving and Bayley considered the main intrusive to be a dike. In a letter N. H. Winchell has stated his belief that it is a sill.* The decision between the two views is not obvious, on account of the few exposures of the Animikie sediments along the northern side of the erup- tive mass. So far as the evidence goes, however, the sill hypothesis should be preferred. The matter is so important that some of the detailed field observations are worth stating. Narrow dikes of red rock respectively cut the sediments, the main gabbro, the intermediate rock, and even rock phases which themselves closely approximate red rock in composition ; nevertheless, every one must agree with Bayley and Winchell that the main gabbro, the intermediate rock, and the main body of red rock are transitional into one another, forming a single geological unit. The case is strikingly parallel to the associations of silicic, mediosilicic, and subsilicic rocks in the intrusive sheet at Sudbury, Ontario, in the (Moyie) sills of the Purcell mountains of British Columbia, and in several sills and laccoliths of South Africa. Since the red rock and the gabbro belong to one rock body, the contact of each with the Animikie sediments should be questioned concerning the struc- tural relation of the intrusive. The actual contact of the gabbro with the sediments is exposed at very few places. One of these is described by * See this Journal, vol. xx, page 200, 1905. R.A. Daly— Geology of Pigeon Point, Minnesota. 427 Bayley in the following words (page 23 of his paper): “On the north side of the point, at the mouth of Pigeon river, are slates and quartzites, capped by an overflow of a fine-grained variety of gabbro.”’ Professor Palache and the writer visited that locality and found a dip-section as illustrated in fig. 2. It happens to be situated where two or more dikes of rela- tively fine-grained gabbro merge upward into the main gabbro mass, which itself, at least locally, has a concordant, sill con- tact with the quartzite-metargillite series. In the face of a steep cliff, 300 meters east of the fig. 2 locality and in the line ret 2: 20 40 METERS Fie. 2. Section of floor contact of sill, near mouth of Pigeon river. Gab- bro dikes (dotted) pass upward into the sill gabbro (dotted). @, Animikie sediments. Observed contacts shown by continuous lines ; inferred contacts, by broken lines. of Bayley’s north-south section “ @—6,’ the exact contact was again found, about 20 meters above lake-level. It could be traced for about 25 meters along the strike. Throughout this stretch the contact surface dips gently southward, rigorously parallel to bedding in the sediments except for strictly local- ized and quite minute transgressions, such as are common at most sill contacts. ; At the only other area of sediments exposed in the peninsula, north of the main gabbro (“ A” in fig. 1), the surface of con- tact cannot be located within a stratigraphically vertical interval of about 3 meters. On the small headland the quartzites have their regional, southward dip of 10° to 15°. ‘They are cut by several trap dikes, one of which is composite.* A few meters * The well-exposed composite dike runs north and south. The older com- ponent is a fresh diabase porphyrite, free from both olivine and quartz. Its width measures about 2 meters over all, that is, including the enclosed 498 F. A. Daly—Geology of Pigeon Pomt, Minnesota. south of the most southerly outcrop of quartzite, the main gab- bro appears in a low cliff. Here the gabbro bears feldspar phenocrysts ranging from 2 to 5 centimeters in length. These are commonly parallel, as if arranged by flow in the magma, and lie parallel to the bedding planes of the adjacent strata. The same parallelism of orientated feldspars to the regional dip-plane of the Animikie sediments was observed in the gab- bro at several other localities. Such repetition in the arrange- ment of the gabbro phenocrysts, where that rock is distinetly porphyritic, is mexplicable on the dike hypothesis. As Bay- ley hints (page 23), the phenomenon rather suggests that the gabbro was injected after the manner of a typical sill. Hig. 3, N. re) ee 20 METERS S: Fic. 8. Section of roof contact of sill, made east of ‘‘C,” fig. 1, and similar to other roof sections to the west. @Q, contact-metamorphosed Animikie quartzite with metargillitic interbeds; ?, intermediate rock; 3, red rock. Blocky character of upper contact illustrated diagrammatically. Again, the gabbro, especially near its northern contact, has a principal system of joints or rift, dipping 5° to 20° south- ward, the other joint systems being roughly perpendicular to that rift. These structural details are also consonant with the sill hypothesis. The contact of the red rock with the sedimentary rocks is much better exposed. With a few short interruptions, quartz- ites and interbedded metargillites extend all along the southern younger component. After the consolidation of the porphyrite, this older dike was split, on a nearly-central plane, and a red granite dike was injected along that plane. The granite dike is a little over one meter in width; so that the whole composite, as far as visible, is composed of porphyrite and granite in nearly equal proportions. The porphyrite shows chilled edges. The granite does not, but is coarse-grained throughout and locally even peg- matitic. The granite, almost wholly made up of quartz and alkaline feldspar, has been much kaolinized and otherwise altered, as if hydrothermally. Its original structure was the hypidiomorphic-granular ; no granophyric inter- growth is found in the thin section. The geological relation of this good example of composite dikes to the sill, only a few meters distant, could not be determined. Lt. A. Daly—Geology of Pigeon Point, Minnesota. 429 shore of the peninsula. At many places their contact with the red rock, for distances of from one to ten or more meters, meas- ured down the dip, is seen to be clearly accordant (fig. 3). In general, the surface of contact dips southward, under the quartzites, at angles of from 5° to 20°, averaging about 15°. Parallel to that general direction is an unusually well devel- oped system of rift joints in the red rock (see fig. 8 in Bayley’s memoir). These are also parallel to a distinct, though less con- spicuous, rift system in the adjacent intermediate rock ; and roughly parallel to the main rift system in the gabbro, as already implied. Assuming that these rifts are parallel to the Fig. 4, oO 100 200 METERS ee | Fie. 4. Dip-section of the Pigeon Point sill through ‘‘ A,” fig. 1, where the sill is relatively thin. Q, Animikie quartzite with metargillitic inter- beds ; D, diabase and gabbro dikes; 1, gabbro; 2, intermediate rock ; 3, red rock. original cooling surface of the complex intrusive—a probable interpretation—the conclusion that the intrusive is a sill is again indicated.* Nevertheless, the upper contact of the intrusive is by no means always concordant with the bedding of the roof quartz- ites. In the low cliffs of the south shore the red rock may be seen at some points to cross-cut the sedimentaries at high angles. There the red rock fills blocky re-entrants in the roof. The maximum observed degree of cross-cutting is a couple of meters, measured at right angles to the bedding of the sedi- ments, but it may be locally several times that amount. Thus, in spite of the fact that bedding of the roof rock and the con- tact surface are on the whole roughly parallel, the roof exhibits re-entrants which seem to be small-scale analogues of those depicted in Barrell’s study of the intrusive stock at Marysville, Montana.t As at Marysville such sharp-angled embayments * The wonderfully perfect rift in the red rock is clearly a primary struc- ture and in no way related to relief of load by erosion or to the present topography. This Pigeon Point case strongly suggests that the rift in batho- lithic granite is similarly related, in a genetic way, to the forms of the cor- responding batholithic roofs. For these larger granitic bodies also the explanation of the nearly flat rifting through cooling contraction still seems to be the best. ue fai Professional Paper 57, U. 8. Geol. Survey, 1907, page 72 and plate IT. 430 R.A. Daly—Geology of Prgeon Point, Minnesota. in the country rock are best explained as due to its dislocation along fractures, accompanied by downstoping of the blocks immersed in the magma. At other points along the contact, much more intense shattering of the sedimentary rocks is in evidence, and swarms of quartzitic xenoliths show that mag- matic stoping has, once again, been “ caught in the act.” In summary, the field facts seem to show that the Pigeon Point body is a sill (fig. 4), modified, as to the detailed form of its roof, by heat shattering and moderate overhead stoping. So far as could be observed, the lower contact is more per- fectly concordant with the Animikie beds, the cross-cutting there proved being restricted to rare local offsets of a few centimeters.* Tilting after Injection of the Sill.—Almost all of the many diabase and gabbro dikes cutting the sediments follow master joints, developed at right angles to bedding. The east-west dikes therefore now dip about 75° northward (see tig. 4), since the quartzites dip about 15° in a southerly direction. It is simplest to assume that the joints were formed when the strata lay flat; and that the dikes, some of which merge into the gab- bro at its lower contact and thus seem to represent feeders for the sill, were injected before the Animikie beds were tilted. If so, the great sill was injected before being turned up into its present position. Yet absolute proof of this hypothesis has not been found. (See page 446.) During or after the tilting the sill and sedimentaries have been affected by a few dip-faults of small displacement. One fault has offset a trap dike cutting the main gabbro at the extreme end of the point. Some of the irregularities in the ground-plan of the complex eruptive may, perhaps, be explained by similar faults. Stratiform Structure of the Sill.—Bayley’s view that the principal intrusive of Pigeon Point is a practically vertical dike, does not agree well with his explanation of the red rock as the product of the contact fusion of the sediments by the primary gabbro. Except for a few dikelets, the red rock is resiricted to the southern side of the main gabbro. If the gabbro were injected as a dike, there is no apparent reason why the sediments along the northern contact should not be fused in the same way. * The time spent in the field did not permit of full study of the relations between the small patches of gabbro mapped by Bayley on Plate XV, to the southward of the red rock. The patch on Fisherman’s Point is clearly a thick dike with normal chilled contacts. The other three patches may represent one or more other dikes older than the sill, or they may be residual chill phases of the sill magma, analogous to those observed locally at the roof of the Duluth laccolith. A dip fault may explain the offset of the red-rock band in the area mapped on Plate XIV of Bayley’s memoir. These uncer- tainties do not seriously affect the conclusion that the main eruptive is a sill. Fe A. Daly—Geology of Pigeon Point, Minnesota. 431 On the other hand, the asymmetry of the complex presents no mystery if the whole igneous mass is, in essence, a sill dip- ping gently southward. (See figs. 1 and 4.) Many direct field observations, as well as numerous homologies among the sills of British Columbia, Ontario, South Africa, ete., indicate that the gabbro of the Pigeon Point body forms a thick layer at the sill floor; and that this is, in succession, overlain by much thinner layers of intermediate rock and red rock. The prin- ciple of gravitative differentiation is obviously suggested. Segregation of Red Rock through Gas Tension.— Gravity has not been the sole control im the separation of the red rock magma. Narrow, short dikes of red rock cut the intermediate rock, the gabbro, and the quartzites underlying the sill gabbro. The dikes seen in those quartzites appeared in all cases to peter out within a distance of about 10 meters below the lower con- tact of the sill, and to be, in a sense, apophyses from the sill. The existence and location of these dikes imply that the red rock differentiate was fluid longer than the gabbroid differen- tiate ; and that the red rock magma was injected into the sill floor, into the roof, or into the already solidified gabbroid phase, by virtue of strong gaseous tension in the red-rock magma. That this magma was charged with gas in large amount is shown by the great development of miaroles or drusy cavities in the red rock of sill and dike. Similar miaroles were observed in the intermediate rock, but never in the gabbro. The unusual driving force which must have been resident in the red-rock magina is suggested also by the lengths of numer- ous, exceedingly thin veins of red rock in the sediments of the roof. “ Ribbon Injections.” —A very striking proof of enormous tension in the red-rock magma was found in other minute, though locally numerous bodies, which may be called “ linear injections” or, perhaps better, “ribbon injections.” On the south shore, immediately west of the little bay marked “ C” in fig. 1, the roof of the sill consists of micaceous and feldspathic quartzites enclosing thin, originally argillaceous beds. Into two different layers of the argillite, separated by about 12 deci- meters of quartzite, the red rock has been injected in the form of nearly straight, flattened needles, or thin, narrow rib- bons of relatively great lengths. The exposures in three di- mensions here happen to be almost perfect, so that one can make out the form and relations characterizing this novel kind of intrusive body. The smaller ribbons are one millimeter or less in thickness, 5 to 10 millimeters in width, and of various exposed lengths of 10 centimeters to one meter. The largest ribbon seen is 432 FR. A. Daly—Geology of Pigeon Point, Minnesota. one to two centimeters in thickness, 4 to 8 centimeters wide, and 2°5 meters in ascertained, minimum length. These lath- like bodies all lie in the bedding planes of the argillite. The low side walls of each ribbon are usually sharp and are per- pendicular to the bedding. None of the smaller ribbons ap- pears to have uparched the overlying sedimentary rock, or to Fig. 5. d | e Fie. 5. Cross-sections (a, 0, c) and partial ground-plan (d, e) of ribbon injections of red rock (dotted), cutting thin metargillitic layers (2), between beds of micaceous quartzite (7). Figs. d and e represent the two extremities of an observed ribbon seen in plan. With other ribbons the kind of termi- nation shown ind is the more common. All drawings are to natural scale. Diagram c refers to the largest observed ribbon. have caused the slightest depression in the rock underlying the ribbon. The cross-section is here conspicuously rectangular (fig. 5,a@ and 6). On the other hand the largest ribbon has decidedly arched its roof (fig. 5, c). At their extremities the ribbons either thin out to sharp points (fig. 5, d) or end abruptly R. A. Daly—Geology of Pigeon Point, Minnesota. 483 at master joints in the country rock, witha square termination, like that of a commercial lath (fig. 5, e). The injection mechanism of the largest ribbon is, in appear- ance, much like that of a laccolith with a lateral conduit. The thinner ribbons are not so obviously explained. They seem to follow joints in the argillite, and to have made room for themselves by crowding aside and compacting the argillaceous material on each side. No other feasible interpretation of the rectangular cross-section has yet been imagined. Where erosion has exposed the top of a ribbon as well as the argil- laceous layer on each side, the ribbon looks like the filling of a sun-crack in the sediment. However, the cross-section of the same ribbon always shows normally bedded sediment im- mediately below the red rock ; so that any genetic connection with sun-cracks is improbable. The explanation by lateral crowding presents patent difficulty. Each shale lamina con- cerned may have been prepared for the required sudden con- densation through preliminary contact metamorphism; the main contact of the great sill is not more than about 10 meters below these gently dipping metargillites. Perhaps the heating of water, specially abundant in the shale, prepared the material for sudden condensation under great stress. Some of the probable feeders of the ribbon injections are visible in low cliffs along the shore. The feeders are true dikes or veins, always very thin (one millimeter or less to three millimeters in maximum thickness) and often of capillary dimensions. They are interrupted and some seem to peter out, both above and below, in the cliff sections. None of these dikelets could be traced with certainty into a ribbon. The ribbons regularly cross the planes of the dikelets at high angles. In none of the observed ribbons is the immediate roof or floor diked by red rock. Hence the ribbons cannot be parts of co- terminous dikelets abruptly widened at shaly horizons. Each ribbon chamber appears to have been forced open by magma which entered from one end. The forcing of a nail into wood by very strong, steady pressure is an analogy, though the injected red rock was, of course, not rigid. The viscosity of the ribbon magma must rather have been of an extremely low order. Time felations of the Magmatice Phases.—Because the Pigeon Point gabbro is cut by dikes of the red rock, a few observers have argued that all of the red rock here belongs to a magma distinctly younger than the gabbro magma. Accord- ingly, the intermediate rock has been explained as the product of the fusion of the older gabbro and its peripheral inter- mingling with the red-rock magma. This view has little to commend it. The irregular main layer of red rock averages 434 RB. A. Daly—Geology of Pigeon Point, Minnesota. probably less tham 30 meters in thickness; it is hard to believe that its original content of heat could have sufficed to melt and impregnate the gabbro to the extent demanded by the hypoth- esis. The contact of the two chief phases has not the char- acter of that between two members of the normal composite sill in other parts of the world. ‘This hypothesis also fails to. account for the practical restriction of the red rock, except for the small dikes above noted, to the upper contact of the gabbro. It fails to explain the micropegmatitic (red-rock) material often seen, under the microscope, to fill the interstices of the feldspar-pyroxene fabric in the gabbro, scores of meters stratigraphically below any level which could have been seriously affected by the hypothetical red-rock injection. Finally, this hypothesis fails to account for the local develop- ment of the intermediate rock between the main gabbro and the roof quartzites, with no red rock present in the cross-see- tion at all; the intermediate rock, here as usual, passing gradually into the gabbro. : These objections have great cumulative: weight. One is compelled to assume that at one stage much of the gabbro and much or all of the red rock were simultaneously molten; that the gabbro froze first and the red rock froze last, so as to be capable of diking both gabbro and intermediate rock. In summary, the writer considers the main body of eruptive rock in the peninsula to be a unit, a sill which has been dif- ferentiated under gravity. The intermediate rock represents a shallow layer marking incomplete differentiation between the gabbro below and the red rock above. Because of special concentration of gas in the red rock, its temperature of con- solidation was relatively low, and small tongues of its magma were driven down, into the gabbro, as well as upwards, into the sedimentary roof, because of the abnormally high gas- tension. Differentiation necessarily Postulated—At the close of his paper (p. 118), Bayley stated that the red rock “ is only the final stage in the alteration of the slates and quartzites by the gabbro.”” In other words, he regards the red rock as the crystallized product of a secondary magma. Apparently he nowhere mentions the principle of differentiation in connection with the problem, but considers the red rock as sedimentary- rock material which has been fused 7m sztu. Yet his average analyses of red rock and sedimentaries are systematically con- trasted, as shown in the accompanying Table I, taken from p- 118 of his paper. The red rock is poorer in iron and magnesia, and, as distinctly, richer in soda and potash. Lt. A. Daly— Geology of Pigeon Point, Minnesota. 435 TABLE I. 1 2 3 4 5 Si0, 74:00 F242 70°31 73°64—-74:22 BON} Jig ioe "34 “40 tr. tr: Al,O, 12°04 13°04 12°81 10°61-11°25 18°32 Fe,0, ‘78 68 7°26 6-24— 7-45 8:11 FeO 2°61 2°49 °88 ‘77— 1:04 °85 MnO 05 09 ers none MgO "42 "58 2°03 1°48— 1°57 3°54 CaO *85 "66 "60 "°36— °56 LOS BaO “12 Shy aia’, “ge a SrO tr. tr. ens cA fi Na,O 3°47 3°44 2° 9 1°67— 3°04 |e is K,O 4°33 4°97 1°90 1°08-— 1°65 3°43 Li,O tr.? iF. eae aie H,O 86 1°21 2°29 3°24 EO: 06 20 pants Bates Cl ine tr aphp age 99°93 100°33 100°20 100°18 1.—Analysis of the powder of three specimens of porphyritic red rock. 2.—Analysis of the powder of seven specimens of granular red rock. 3.—Mean of the analyses of three unaltered quartzites and one, slightly altered, metargillite. 4,—Range of oxides in three quartzites free from contact metamorphism. 5.—Analysis of one specimen of unaltered metargillite. The present writer’s field and microscopic study has led to the belief that the systematic differences are typical of red rock and sedimentary rock throughout practically the whole outcrop of the sill roof. Hence the mere fusion of the sediments by the heat of the gabbro magma probably cannot explain the red rock as now constituted. If part of the quartzite-metargillite roof has been simply fused 2m sitw, without being dissolved in the gabbro, the red-rock magma could have originated from the fused product only if the ‘latter had undergone some differ- entiation. Still more clearly, the red rock cannot be assumed to have originated from a solution of sediments zm the gabbroid magma, unless the resulting hybrid magma has been differen- tiated. Competing with both of these hypotheses is a third: that the red rock and gabbro are the two poles of differentia- tion in a primary magma, which has never been affected by assimilation. In any case differentiation has to be reckoned with and it is this feature that, more than anything else, makes the genetic problem difficult. As a preparation for further discussion it is well to note the approximate relative amounts of red rock and gabbro in the sill. Taking 15° as the average dip, the thickness of the exposed part of the sill varies from 125 meters to 200 meters 436 Rk. A. Daly—Geology of Pigeon Point, Minnesota. or perhaps a littlé more, with an average of about 160 meters. Of the exposed mass, nearly one-eighth by weight is red rock ; the remainder is eabbro, associated with an almost negligible proportion of intermediate rock. Nature of the Magma Differentiated.—lf the red-rock material were a differentiate of a purely primary magma, and if the actual weight ratio of red rock to gabbro is given by the surface exposure, the composition of the assumed parent solution (neglecting volatile substances which have escaped from the sill chamber) may be roughly calculated. The result is shown in Column 3 of Table II. Columns 4, 5, and 6 give respectively the average composition of diabase, gabbro, and basalt as world types. TABLE II. 1 2 3 4 9) 6 si0, 49°88 72°42 52°70 50°12 =. 48°24 48°78 Al1,0, 18°55 13°04 17°86 15°68 17°88 15°85 erO: 2°06 68 LTeSGe. 4°55 3°16 5°37 FeO 8°37 2°49 7°64 6°73 5°95 6°34 MgO 5°77 58 o°12 5°85 Yeo 6°03 CaO 9°72 66 8°59 8°80 10:99 8°91 Na,O 2°59 3°44 BoD) 2°95 2°55 3°18 K,O 68 4°97 1°22 1°38 89 1°63 1.—Analysis of mixed powders of five fresh specimens of Pigeon Point gabbro. 2.—Analysis of mixed powders of seven specimens of granular red rock. 5.—Calculated composition of the mean rock of the sill, giving ‘‘1” Bue weight of seven and ‘‘2” the weight of unity. 4.—Mean of analyses of 20 typical diabases. 5.—Mean of analyses of 41 typical gabbros. 6.—Mean of analyses of 161 typical basalts. The calculated mean composition of the sill is seen to be not far from that of the average diabase, gabbro, or basalt of the world. The result would be little affected by considering also the volatile matter that has escaped from the chamber. Hence one must seriously entertain the hypothesis that the red rock has been differentiated from a gabbroid or diabasic magma essentially jike that represented in the narrow, and therefore quickly chilled, dikes of Keweenawan age in Minnesota. However, there are two chief objections to that view. Though red rock is similarly associated with gabbro or diabase in the Logan sills farther north, and in the gigantic Duluth laccolith to the west, many of the thick Keweenawan injections are constituted wholly of gabbro or diabase without any red- rock phase. In other regions, many comparatively thick sills and dikes, of dates varying from the pre-Cambrian to the Terti- ary, contain no acid differentiate which is comparable in purity R. A. Daly—Geology of Pigeon Pot, Minnesota. 437 and relative volume to the red rock at Pigeon Point. One or more special conditions must, therefore, be assumed in order to explain these remarkable Minnesota bodies, if due to pure differentiation. No such condition has yet been dedueed from the field facts ; until some are found, the pure-differentiation idea must remain in the realm of speculation. On the other hand, specially rapid and thorough differentia- tion of an originally basic magma might be thought to have been induced by its absorption of much water from the walls of its conduit. The Pigeon Point magma rose through a great thickness of Animikie strata before reaching the level of the sill chamber. The sediments were then hydrous, probably more so than now. If the original gabbroid magma absorbed some water from them, the temperature of consolidation was lowered and the magmatic life thereby lengthened. Any ten- dency to spontaneous differentiation would become more.effec- tive merely because of the longer time available, and the resur- gent water itself might be an independent cause of magmatic differentiation. If this were true for the sill, similar differentiation might be regularly expected in the thick Keweenawan dikes cutting the upper Animikie sediments; since dikes, crossing the bedding, are favorably situated for the absorption of connate water. As far as known, the thicker Minnesota dikes show no special ten- dency to extreme differentiation.* | Hwidences of Assimilation—The second objection to the pure-differentiation hypothesis is more telling. There is an unmistakable consanguinity between the red rock and the Ani- mikie siliceous sediments, just as there is between the micro- pegmatitic phases of the Moyie sills of British Columbia and the quartzites invaded by them. The consanguinity is both mineralogical and chemical. In neither respect are sediment and red rock identical ; yet qualitatively the two formations are closely parallel, and one cannot rest in the belief that their similarity is accidental. On the contrary, their qualitative likeness suggests that the red rock is the result of syntexis plus differentiation. For a distance of from 5 to 20 meters, measured at right angles to the roof contact of the sill, the quartzites and metar- gillites have been strongly metamorphosed by the magma. Their maximum alteration is like that in the many xenoliths close to the shattered roof. Nearly always the contact between red rock and sedimentary rock is sharp and the bedding planes * Other things being equal, the volume of water absorbed by an ordinary dike would be less than that entering the magma of a conduit to a great sill, so that the postulated differentiating influence of water might be more con- spicuous in the sill than in the dike; yet some effect should, by the hypo- thesis, be illustrated in visible thick dikes. 438 R. A. Daly—Geology of Pigeon Point, Minnesota. of the latter are usually evident. The originally gray or greenish quartzites become of a deeper and deeper red tint, according to the degree of their metamorphism. When most changed the quartzites have a color nearly identical with that of the red rock. More locally, drusy cavities, lined with well-terminated quartz and feldspar crystals, are developed in the roof quartz- ites. These are of habit similar to that of the yet more abun- dant druses in the red rock. Both the reddening and the formation of drusy cavities in the roof rocks might be regarded as effects of water-gas and other emanations from the red rock, in which the same features were developed because of its own content of these gases. By this hypothesis the color change is attributed to the chemical influence of the gas; the existence of the cavities, to its high tension. Thus, the two characteristics mentioned would be considered to have no significance in the problem of the red rock. Yet some mode of origin for the emanating gas itself is demanded. A critic of the foregoing hypothesis may reply that it puts the cart before the horse, holding that the mag- matic gas involved was largely water-gas derived from the sedi- ments, as really implied in Bayley’s theory of the red rock. The consanguinity of red rock and sedimentary rock is more clearly suggested from the fact that each bears indigenous micrographic intergrowths, composed of quartz and feldspar. Some of this material has doubtless been introduced into the roof sediments, through impregnation, from the red-rock magma, but careful microscopic study shows this explanation to be unacceptable for much of the micropegmatite developed in the sediments. The original quartzites and metargillites contain quartz, feldspar, sericite (also paragonite ?), chlorite, biotite, and iron oxides. In an early stage of the contact metamorphism the quartz grains are seen to have been embayed by minute tongues of their alkaline and aluminous cement. These minute, but relatively long and narrow, tongues can only be due to mutual solution of quartz and feldspar, or quartz and general cement, aided by connate fluids. Where the mutual solution has not gone beyond the incipient stage, the product is generally obseure. Ina slightly more advanced stage, most or all of the tongues in an attacked quartz grain are filled with alkaline feldspar, mixed with accessory material. The feldspathic parts of the tongues commonly extinguish simultaneously under the microscope. With further metamorphism this new feldspar is found to be intergrown micrographically with quartz, indicat- ing a quite local, but complete, mutual solution. The micro- R. A. Daly— Geology of Pigeon Point, Minnesota. 439 pegmatite so formed is indistinguishable from that im typical red rock, of which micropegmatite forms an essential part. The thin sections in which this series of changes can be traced show no sign of fracturing and no indication of any channel through which magmatic material has been introduced. In other words, the micropegmatite has apparently been formed of material original in the sediment, without importation. Professor Grout, in a personal letter, has raised the question whether the intergrowths in the metamorphosed sediment have been produced through the agency of juvenile water-gas emanating from the magma. Of this there is no evidence in the thin section. On the other hand, chemical analysis of the sediments collected where they have not undergone any con- tact metamorphism proves that they now bear a relatively high percentage of connate water. In Keweenawan times these beds were less affected by regional (static) metamorphism and were, as above noted, probably even more hydrous. That the connate water must have acted as a solvent in the highly heated sediments is an obvious fact; to attribute the visible solution to the influence of juvenile gas is a speculative assumption, unsupported by facts won from a study of the thin sections. The matter is admittedly delicate, but the present writer pre- fers to dwell on the simpler inference from observed facts. The indigenous character of the micropegmatite observed in the contact aureole is further suggested by its occurrence in the floor quartzites, next the gabbro, in which the granophyric material is almost or quite absent, and where, to judge from the actual mineralogy of the rocks, magmatic gases were never greatly coneentrated. The problem is identical with that studied in connection with the micropegmatite of the Moyie sills of British Colum- bia. There the roof aureoles are sometimes richly charged with typical micropegmatite, distributed through the bedded quartzites. Chemical analysis shows that little or none of this quartz-feldspar intergrowth is to be explained by emanation of feldspathic material from the magma.* According to Pro- fessor Grout the same chemical relations hold at Pigeon Point. While, then, feldspathization in many other contact aureoles is not to be doubted, that principle seems not to account for much of the micropegmatite in the country rocks of the Pigeon Point sill. Nor is it likely that the close field association of micropeg- matite in the red rock with micropegmatite in the metamor- phosed sediments is a pure accident. The material is of a nature too specialized for that. The preferable explanation is probably to be found in the principle of “ ultra-metamorphism,” * Cf. Memoir 38, Geological Survey of Canada, 1912, page 243. Am. JOUR. Dy oa SERIES, Vou. XLIII, No. 258.—Junz, 1917. 440 &. A. Daly— Geology of Pigeon Point, Minnesota. the red-rock material having been chiefly derived from dis- solved quartzite aud metargillite. fted-rock Shells Surrounding Xenoliths.—One set of field observations made by Bayley and repeated by the present writer at first sight appear to corroborate that theory most emphatically. In gabbro, in intermediate: rock, and in red rock alike, are xenoliths of quartzite which are completely enclosed in shells of typical red rock. These small masses of red rock thus, in the field, look like direct contact fusions of the sediment. The shells vary from one centimeter to 50 centimeters or more in thickness. In general their contacts with the enclosed quartzite and with the enclosing igneous rock are fairly sharp. The quartzite of the xenolith often shows clear bedding and always a texture different from that of the surrounding red rock, so that there is seldom much doubt as to the limiting surface of the non-magmatic material. Just east of the little bay marked “ C” in fig. 1, a large group of these shelled xenoliths is exposed along the shore. In the group a series of xenoliths can be found, wherein the relative thickness (and volume) of the red-rock shell increases, the vol- ume of the respective xenoliths decreasing, until thick shells of red rock are seen to enclose mere shreds of quartzite. Finally, at the end of the series, no quartzite is seen, the whole, roughly globular mass being composed of red rock. All of these bodies lie in a general matrix of intermediate rock. The masses of pure red rock, one or two decimeters in diameter, have the appearance of being ‘‘ ghosts” of completely fused xenoliths. That explanation would seem the more probable since xeno- liths of feldspathic gabbro (torn from a dike older than the sill) have no such envelopes of red rock; thus a genetic con- nection between sedimentary rock and red rock is the more readily credited. Lawson found quartzite xenoliths in the diabase of the Logan sills, north of Pigeon Point, and noted the reddening of the diabase in their vicinity.* The importance of the red-rock shells first came to the writer’s attention in the course of a personal discussion with Professor Grout. He seems to have been the first to see clearly that mere fusion of the xenoliths zm sitw'is not the whole explanation of the red-rock shells. He has made a special study of the large rimmed xenolith in the intermediate rock at the locality mentioned on page 110 of Bayley’s memoir. After a detailed microscopic and chemical investigation of this xenolith and its red-rock shell, Professor Grout has concluded that the shell is much richer in alkalies, especially potash, than * A. C. Lawson, Bulletin 8, Geological and Natural History Survey of Minnesota, 1893, page 30. R. A. Daly—Geology of Pigeon Point, Minnesota. 441 the quartzite of the xenolith. The writer has confirmed his view by microscopic study of the same and of other rimmed inclusions in the intermediate rock. As shown by the thin | sections, each shell contrasts with its xenolith in carrying at least twice as much potash and also more soda. Moreover, the shells examined are nearly uniform in composition and therein are equivalent to the main mass of typical red rock of the sill; while the xenoliths vary considerably in their original content of alkalies. As Professor Grout has pointed out, the differ- ences of alkaline content between the shells and xenoliths are of the same order as those between the alkali content in the average analysis of the main red rock and the alkali content of the average Animikie sediment of the region. (See Table L.) (1) Assuming the red-rock shell to have resulted from the fusion of the xenolith, without addition of material from the general magma, a large amount of silica, iron, and water must have been expelled from the secondary melt. It is conceivable that one or more compounds of iron and silica, aided by the water, might rapidly diffuse into the general magma, leaving behind the less volatile feldspar and some free silica. By such . differential diffusion feldspar would be more concentrated in the shell than in the original quartzite. As yet no compelling evidence either for or against this hypothesis has been found ; however, one may well doubt the quantitative value of the imagined process. (2) Is the red-rock shell rich in alkalies because these have been, as it were, sweated-out from the hot interior of the xenolith? The connate water of the inclusion must have at- tained high gas-tension, which might be conceived to have led to the expulsion of a water-feldspar-quartz solution, analogous to an ordinary, low-temperature pegmatitic magma. This second guess as to the cause of the concentration of feldspar in the shell at once meets the difficulty that certain shells have volumes greater than their respective enclosures. (3) On the whole, it seems more probable that, while some fusion of each xenolith has taken place, some of the soda and potash, and perhaps alumina, concentrated in the red-rock shells have been derived from the general magma. How this was accomplished is a residual question, as yet without adequate answer. At least half of the soda and potash in the feldspar of the average shell must, apparently, have been imported, the remainder being the alkalies of the xenolith fused in situ. If the oxides of sodium, potassium, and aluminum, in the presence of water, were highly diffusible in the roof phase of the general magma, those oxides might conceivably have been fixed, both in chemical combination and in position, by the free silica of the partly melted xenolith. The mystery as to the phases 442 R. A. Daly—Geology of Pigeon Point, Minnesota. actually present in a natural magma and as to their capacity for diffusion is so profound that this third speculation cannot be profitably discussed. Difficulty of explanation cannot alter the fact that the red- rock shells appear to be due in part to separation of material from the general magma. These envelopes are, therefore, not direct proofs of an origin of the main red-rock mass in mere contact fusion of the Animikie sediments. Nevertheless, the discovery otf indigenous micropegmatite in the contact- metamorphosed sediments shows that some red-rock material has been generated by contact fusion. Magmatic Stoping.—As Bayley noted, the roof rock of the sill is locally much shattered. Besides the isolated xenoliths, evidence of the disruptive action is given in comparatively thick breccias of roof quartzite, now cemented by subordinate amounts of red rock. Xenolith and breccia block still remain near or at the roof, evidently because the magma was nearly frozen when the enclosure of roof fragments took place. Dur- ing the long preceding magmatic period, other shattering must have occurred, and, because of the smaller magmatie vis- cosity then ruling, the roof blocks sank into the heart of the sill.* As in the case of the normal batholith, the efficieney of fusion and syntexis can here be estimated only after the fate of the sunken blocks is considered. Inall batholiths the deeper levels are never exposed, so that the subsequent history of downstoped blocks can only be inferred. The outcrops of the Pigeon Point intrusive are sufficiently continuous to warrant the statement that xenoliths are very rare at depths greater than 20 meters below the roof. If the initial, hot magma did actually shatter and stope more effectively than the nearly frozen magma, one is compelled to believe the older xenoliths to have been melted and more or less completely dissolved in the primary gabbro magma. If this “abyssal” assimilation strongly supplemented any assimilation at the roof, some dif- ferentiation must be postulated, for the gabbro has been cleansed from nearly all the material which is not present in normal gabbro. The actual stratiform structure of the sill implies, as already noted, gravitative control in the separation, which would presumably progress simultaneously with the stoping and internal solution of blocks.t+ * The lighter shales would float in the initial gabbro. Practically all xenoliths would sink in the acidified magma. Blocks of the dominant quartzites would probably not sink all the way to the sill floor, but would come to rest at intermediate levels. There high temperature was long main- eins evieh the result that the blocks would tend to be completely dissolved or rused. + On the shore, east of the bay marked at ‘‘C”’ in fig. 1, the red rock cuts an older, nearly vertical dike of coarse anorthositic gabbro. For a distance of about 30 centimeters from their contact, the red rock is charged with Re. A. Daly— Geology of Pigeon Point, Minnesota. 443 The probable importance of stoping is indicated by the pro- nounced irregularity of the sill roof. The reader will recall that, at many places, the red rock sharply cross-cuts the roof strata for several meters, the eruptive occupying blocky re- entrants in the roof. One result is the development of local, dike-like contacts, some of which were emphasized by Bayley.* (See fig. 3.) On the other hand, the floor contact, where seen, is that characteristic of a typical sill, with observed cross-cut- ting to the extent of only, at most, a few centimeters, measured at right angles to bedding. Correspondingly, the eruptive at the floor, nearly the normal gabbro in composition, has a con- tinuous chilled contact and shows little or no evidence of havy- ing there shattered or assimilated the sedimentary rock. Since roof and floor originally matched, the roof must have been roughened by some kind of magmatic activity ; that roughen- ing is most readily ascribable to stoping. The reason for the practical restriction of stoping to the upper contact is discussed in a following paragraph. Capacity of the Original Magma for Assimilation and Contact Fusion.—The syntectie theory is very commonly dis- missed from sympathetic consideration by those who emphasize the “enormous” amount of latent and other heat required to fuse or dissolve rock. For the present case, as for many other debated eases, the heat demanded may be regarded as “ enor- mous” in absolute measure but as small when eompared with the initial heat supply. In the Pigeon Point area, trap dikes and sills only a few centimeters thick were observed. Such thinness, combined with great length, demonstrates some superheat in their diabasic or gabbroid magma. The main sill itself, less than 200 meters thick and probably more than 10 kilometers long, could not have been emplaced unless it were somewhat super- heated. The temperature of its magma would be raised by large plagioclase individuals which seem to have crystallized from a hybrid magma, formed by the solution of the gabbro in the invading red-rock magma. This contact phase of the red rock is specially dark in color because of the abundance of a femic mineral, now completely chloritized. Thus the red-rock magma was capable of performing some assimilation, even at a late stage in its history. If that deduction is correct, it tends to fortify the assumption that hydrous, feldspathic quartzite, related to red rock in composition could be wholly dissolved by the initial, hot, gabbroid magma. *The ‘‘gabbro” (really a coarse porphyrite) shown in fig. 3 of Bayley’s paper belongs to a thick dike, which is certainly older than the red rock and probably older than the main gabbro. The cross-cutting quality of the red rock at this point may be explained either by stoping or by local up- faulting of the sill roof along the contact of the massive, strong dike. In any case the section does not invalidate the conclusion that the main erup- tive isa sill. 444. R. A. Daly—Geology of Pigeon Point, Minnesota. friction in the very act of injection, if that magma were initially nearly frozen and therefore highly viscous. Super- - heat to the extent of 100° or perhaps 200° C. can be safely assumed for the melt when the main injection took place. On any hypothesis of origins this original magma was gabbroid in composition. Its temperature at the time of emplacement was probably at 1100° C. or higher. The freezing temperature of the red rock is not known; but, on account of its richness in water, was doubtless well below 1000° C., if not as low as 800°. Taking the masses of the gabbro and red rock as seven to one, and assuming fair values for the latent and specific heats involved, one may roughly calculate the initial superheat required for the assimilation of the sediments in volume sufficient to furnish the observed amount of red rock. The result is to show that the thermal problem is not so portentous as it is thought to be, in such cases, by certain petrologists. Assimilation in the Sill Conduit.—But there is no necessity of assuming all the solution or fusion of the Animikie sedi- ments, corresponding to the volume of the red rock, to have taken place in the visible sill chamber itself. The Pigeon Point eruptive rose through nearly the whole thickness of the great Animikie series. Those rocks are quartzites and metar- gillites like those exposed in the peninsula. Some solution of the sediments during the uprise of the initial, hot magma must be regarded as not impossible. The sill injection may have been a single act, the magma rising, through one or more simple dike-passages, from the earth’s deep interior. Or the magma may have occupied one or more temporary sill chambers before reaching the horizon of the visible sill. The two modes of injection may be briefly described as respectively legato and staccato. It the injection were staccato in quality, some assim- ilation in the temporarily occupied chambers might have been brought about. Material then dissolved. would tend to be mixed through the original gabbroid magma, because of the later movements; the final separation of the solute occurring in the chamber finally occupied. Whether the eruption was legato or staccato, the rising gab- broid magma doubtless received an accession of water, which was thermally expelled from the Animikie sediments, then pos- sibly wetter than now. Such resurgent water, added to that absorbed from xenoliths and main contacts in the visible sill chamber, would become concentrated at the sill roof. Through the consequent depression ‘of its freezing temperature, the magma at the roof was not quickly chilled, as was that at the floor ; but was able to continue stoping and also some marginal solution of country rock. Stoping and vertical currents inci- Lt. A. Daly—Geology of Pigeon Point, Minnesota. 445 dental to differentiation must have stirred the upper layer of magma and prevented rapid chilling at the roof. To this rather bold sketch of the probable events, objection may be urged that the magmatic gas, so clearly forming an original part of the magma of the drusy red rock, may have been purely juvenile and in no degree resurgent. Such a speculation is founded on a dificult thesis, for it denies the efficiency of gas tension to drive connate water into the initial magma from the hot walls of the conduit. Some petrologists will have it that gas can move centrifugally from an intrusive body, but never in the reverse direction, into the intrusive magma. They think of the contact surface of the magma as a one-way gate. This remarkable conception surely needs criti- cal examination. The gas moves because of differential pressure. In contact aureoles developed in relatively imper- vious, hydrous sediments, the steam pressure must, in very many cases, rise enormously. If the pores of the sediments were originally full of water, such pressure might be very much greater than that of the initial magma at the level concerned. To explain dehydration in such contact aureoles, it is quite gratuitous to assume that all the water moves centrifugally. The certainty of pegmatitic emanation in the late magmatic stage proves nothing at all regarding the direction of migration for gases during the long preceding stage of higher tempera- tures and gas pressures. In the present case the conditions, including the existence of many horizons of impervious shales in the Animikie series, seem toimply the necessity of expulsion of much water-gas from the wall-rocks into the gabbroid magma. Further, no good reason is in sight for crediting this particu- lar body of magma with a much higher proportion of juvenile gas than that fairly attributable to many Keweenawan injec- tions in Minnesota. On the other side, no one can doubt that the Animikie sediments, especially the shaly beds, were during Keweenawan time rich in water. So true is this that it would not be utterly absurd to regard the abundance of gas, probably water-gas, in the red rock as another suggestion of consanguin- ity between that formation and the Animikie sediments. Again, the heat producing gas tension in the wall-rocks of the conduit, and also aiding in the preparation of the tempera- ture suitable for solution of their solid material in the magina, was not merely magmatic heat. Before the gabbro eruption, the lower beds of the Animikie series and the unconformably underlying rocks had been heated because of their burial under the thick, upper part of the Animikie series, perhaps already covered by early-Keweenawan lava flows. General earth heat had thus begun the process of superheating the connate water, 446 R.A. Daly— Geology of Pigeon Point, Minnesota. with the final development of steam pressure which tended to expel the water from the wall-rock into the conduit magma. How great this effect was cannot be estimated without a knowledge of the geothermal gradient in the Keweenawan period. That it was steeper than now is indicated by the efficiency of static metamorphism in pre-Cambrian time. Pos- sibly the conduit walls of the Pigeon Point eruptive, where passing through the wet sediments, had a general temperature of 200°C. or higher. Experiments have shown the ample power of water, at temperatures little above 200°, to promote the solution of siliceous materials. Herein, perhaps, is an explanation of the fact that very many of the known gabbroid sills and laccoliths bearing acid, granophyric differentiates, have pre-Ordovician, if not pre- Cambrian, dates of intrusion.* Before leaving this subject, one other speculative point may be noticed. The Pigeon Point eruptive is precisely en axe with the Duluth laccolith, which in its essential petrography, in its gravitative differentiation, and in its general structural rela- tions is a colossal replica of the sill under discussion. The known outcrops of the two bodies are separated by little more than 30 kilometers. Is the sill an apophysis from the 200-kil- ometer laccolith? Was some syntexis accomplished in the vaster chamber before the offshooting sill forced a way east- ward as far as Pigeon Point? In the face of this possibility, is it safe to deny the efficiency of magmatic assimilation simply because of facts observed at the visible contacts of the sill ¢+ Summary on the Origin of the Red fock.—Assimilation of the Animikie or underlying rocks had three possible loci: in the sill conduit, including one or more dikes and perhaps laceco- lithie or sill enlargements; at the main contacts of the visible sili chamber; and in the heart of the sill, where downstoped blocks are concerned. Differentiation of the syntectic material had two possible loci: in the conduit, and in the visible sill chamber. The actual differentiation, implied by the existing rock types in the sill, had two phases: that controlled directly by gravity, and that controlled by gas tension. * Compare list in the writer’s ‘‘ Igneous Rocks and Their Origin,’ New York, 1914, pages 230 and 344 ff. + Elsewhere (Memoir 38, Geological Survey of Canada, 1912, page 200) the writer has described a mechanism by which syntectic magma, formed in one sill chamber, may be injected into a new, higher sill or dike chamber, where differentiation unaccompanied by further assimilation of country rock is possible. The country rocks of the second chamber may be chemically little related to those of the chamber where the assimilation occurred, or to any differentiate in the second chamber. The origin of the red rock in the Duluth laccolith is not here specially considered, but the size of that body warrants a sympathetic reception for the syntectic theory when applied to its red rock. \ R. A. Daly—Geology of Pigeon Point, Minnesota. 447 The observed characters and geological relations of the sill were established after both assimilation and differentiation had been completed. In view of so many complexities, it is not astonishing that the problem of the red rock still awaits definitive solution. The solution must, of course, depend on a patient application of the principle of inference from known facts. Of these the most significant, in the writer’s opinion, is the development of indigenous granophyric intergrowths of quartz and feldspar in the contact-metamorphosed sediments. The typical intergrowth has properties identical with those of the abundant micropeg- matite of the red rock. That the red-rock material was derived from dissolved or fused sediments is further suggested: by the thorough reddening of the Animikie strata where strongly metamorphosed ; by the discovery of isolated druses, like the druses of the red rock, in the same strata; and by the unusual abundance of gas, presumably water-gas, in the red rock. Some genetic, though not easily deduced, relation between red rock and quartzite is implied by the presence of red-rock shells around quartzite xenoliths and their absence at xenoliths of the older gabbro. Supplemented by the principle of differentia- tion, Bayley’s idea of contact fusion seems to give one essential element ina valid explanation of the progressive, and ulti- mately complete, replacement of quartz xenoliths by red rock. The principal difficulty with the assimilation- fusion hypothesis is the higher alkali, especially potash, content of the red rock, as compared with the Animikie sediments already analyzed. Bayley states that individual strata carry 75 per cent of feld- spar, but such beds must be rare, and it is unsafe to assume for the average sediment a percentage of alkalies as large as that characterizing the red rock. The alkalies concentrated in the red rock may, in part, have been derived from the gabbroid magma. Combining with the silica and alumina of the Auni- mikie shales, these juvenile alkalies may have risen to the sill roof dtiring the general differentiation. The analyzed gabbro has, in fact, proportions of soda and potash which are abnor- mally low for gabbro. The same feature appears in other Instances where granophyric and gabbroid rock have differen- tiated from each other.* As the proximate mode of origin for the red rock, differentia- tion must be assumed. Both syntexis and mere fusion én situ, though probably important, have been masked by differentia- tion. The mechanism of magmatic separation represents an unsolved problem. Specifically, future investigation may well be devoted to the question as to how the feldspar molecules *R. A. Daly, Igneous Rocks and Their Origin,’’ New York, 1914, page 320. 448 R.A. Daly— Geology of Pigeon Pot, Minnesota. have been concentrated in the red-rock envelopes surrounding xenoliths and in the red-rock layer, thin or thick, between the main gabbro and the sill roof. Until that mechanism is under- stood, a final decision concerning the origin of the red rock must be delayed.* Nevertheless, the facts now in hand seem to show the origin of the red rock to lie in both assimilation and differentiation, rather than in the differentiation of a wholly primary magma. In any case, be the fate of past and present speculation what it may, Pigeon Point will long engage the profound interest of petrologists, for its mysteries are the mys- teries of igneous rocks in general and at but few other locali- ties are the rocks representing the two commonest magmas— the basaltic and the granitic—better or more suggestively exposed. Harvard University, Cambridge, Mass. * While fully believing in fractional crystallization as a cause for the diversity of igneous rocks (see Journal of Geology, vol. xvi, 1908, pp. 401- 420), the writer finds much difficulty in applying that principle as the sole explanation of the Pigeon Point differentiation. The matter is not here dis- cussed. R. G. Van Name—Temperature Coefficient. 449 Art. XL.—On the Temperature Coefficient of a Hetero- geneous Reaction; by R. G. Van Name. [Contributions from the Kent Chemical Laboratory of Yale Univ.—celxxxviii. | In the course of a series of studies on the rates of solution of metals, the results of which have been published in four pre- vious papers,* the writer has found that certain of these reac- tions are exceptionally well adapted for accurate measurements of reaction velocity, having distinct advantages in this respect over most other types of heterogeneous reactions. Since our knowledge of the effect of temperature on the velocity of heterogeneous reactions is rather limited, it has seemed desirable to utilize the experience gained through the work above mentioned, in. the careful measurement of the temperature coefficient of the velocity of a single reaction of this type over a considerable range of temperature. For this purpose the reaction between cadmium and a water solution of iodine in potassium iodide has been selected, and the velocity determined at seven temperatures covering the range from 0° to 65°. | The choice of this reaction was due not only to the excep- tional convenience and accuracy with which iodine can be titrated, but also to the fact that cadmium can easily be obtained practically free from objectionable impurities, can readily be rolled out into sheets, and when so rolled has a very finely crystalline structure, so that it dissolves in the iodine solution without any appreciable roughening of the surface, thus preserving a constant surface area. : The procedure employed did not differ in any essential respect from that described in an earlier paper,t and therefore calls for no general description. A few points, however, which are only very briefly or incompletely treated in the previous articles, will be taken up in detail here. The following data apply to all the experiments: Reaction vessel, a thin beaker 11™ in internal diameter. Diameter of cadmium disk 38:3", thickness 0°57". Composition of liquid: KI, 0°5 molar; H,SO,, 0°01 molar; I, (at start) about 0:02 molar. Volume of liquid, 600 cm® at start, 20 cm® taken for each titration. Thiosulphate used in titrations, 0°02 normal. Rate of stirring 200 revolutions per minute. Experiments were conducted at 0°, 15°, 25°, 35°, 45°, 55°, and 65°. Only at 25° was the temperature found to be prac- * Van Name and Edgar. this Journal (4), xxix, 237, 1910; Van Name and Bosworth, ibid. (4), xxxii, 207, 1911; Van Name and Hill, ibid. (4), xxxvi, 548, 1913, and (4), xlii, 301, 1916. + The second of the articles cited in the preceding foot-note. 450 R. G. Van Name—Temperature Coefficrent tically the same in the reaction beaker as in the water of the thermostat. At 15°, and at 35° and above, the thermostat was regulated to that temperature which would give the desired temperature in the beaker. Even this was not sufficient in cases where the difference in temperature was large, since the difference was then found to increase with decrease in the sur- face of contact between the beaker and the reaction liquid, that is, with the depth of the liquid in the beaker, which decreased about one-fourth during the course of the experi- ment. In such cases the temperature of the thermostat was gradually varied as the volume of the liquid diminished. Thus, to maintain the temperature of the liquid in the beaker at 65° it was found necessary to set the thermostat to 67°7° at the start and gradually raise its temperature during the experiment to 68°2° at the end. By careful use of this ex- pedient the temperature, even at 65°, where the error was largest, was maintained for the most part within 0°2° of the correct value, greater variations occurring but rarely, and then only for very brief periods. For the experiments at 0° the thermostat was kept at that temperature by the frequent addition of liberal quantities of fine snow. The temperature in the beaker, directly measured, was always between one and two-tenths of a degree higher, averaging + 0°15°. The rate of the reaction follows the equation v C, The reaction velocities were calculated by substituting the observed values of ¢ in this equation, A¢ being the duration of a single reaction period, i.e. the time interval between the removal of two consecutive samples of the solution. These in- tervals were generally ten minutes in length, but somewhat shorter reaction periods were employed at the highest tempera- tures, and longer periods at the lowest. The velocity con- stants so obtained were first corrected, as described below, for variations in the rate of stirring. So corrected they represented the apparent rate of the reaction, but not in general its true velocity, since in many cases a further correction was needed to compensate for the effect of evaporation from the solution. Corrections for Variations in the Rate of Stirring.—The average rate of stirring during each reaction period. was deter- mined and a correction, based on the assumption that the reaction velocity varies as the 4/5 power of the rate of stir- ring,* was afterward calculated and applied to the observed * An empirical relation shown in a previous paper (this Journal (4), xxix, 251, 1910) to be approximately true for an apparatus of this type and dimensions. of a Heterogeneous Leaction. 451 value of the velocity constant A. The method of determining and controlling the rate of stirring was as follows: To the axle of the stirrer was attached a mechanism which rang a bell at every 100th revolution. A stop-watch was started at the first ring of the bell after the beginning of the reaction period and was stopped at the nearest ring to the end. ‘The elapsed time so recorded was that required for an even multiple of 100 revolutions, from which the average rate for the period and the necessary correction to A were readily calculable. Fur- thermore, if the rate were exactly 200 revolutions per minute the rings would evidently coincide with the passage of the second hand of the stop-watch over the even minute and half minute on the dial.* Any variation in the speed was imme- diately shown by the failure of this coincidence, and was promptly corrected by adjusting a rheostat in series with the stirring motor. In this way the average rate was kept, in the great majority of cases, within 0-2 per cent of the correct value, and only in rare cases did the correction to be applied reach 0°5 per cent. In most of the experiments, therefore, the effect of these corrections, sometimes positive and sometimes negative, upon the final result (i.e. the average value of A for the whole experiment), was nearly or wholly negligible. Corrections for Hvaporation.—Except at the lower tempera- tures the effects of evaporation had to be taken into account, since the liquid in each experiment was stirred in an open vessel for a period varying from 40 to 90 minutes. Such evaporation may affect the apparent reaction velocity, as cal- culated from equation (I), in two ways: first by decreasing the volume v, and second by altering the values of the concentra- sions of iodine. The first error drops out if we use the actual value of v in calculating A, which has been done in the present work. Since the rate of evaporation is proportional to the area of the free surface of the liquid, which remains nearly constant,+ the decrease in volume per minute may be regarded as constant. In the experiments of the writer this rate of evaporation was determined either by special blank experiments made under like conditions, or by determining the actual loss during the experiment itself, by carefully measuring the volume remaining at the end. Since the rate of evaporation depends somewhat on the external conditions the latter procedure is prefer- able, and was employed in most cases, especially when the rate of evaporation was large. The total change in volume up to the middle of each reaction period was then calculated, and the * Another watch was of course needed to fix the beginning and end of the reaction period. + In reality the concayity of the surface produced by the stirring increases slightly as the volume of liquid diminishes. 452 R. G. Van Name—Temperature Coefficient value of » so found, i. e. its mean value, 4 (v,+4,), for the given interval, was used in the calculation of A. This method was followed in all cases at 35° and higher temperatures. At 25° a simpler procedure, equivalent in effect, was employed instead. (See p. 453.) Below 25° corrections to v were found to be negligibly small and were therefore not applied. The second error resulting from evaporation, namely, change in the iodine concentration, might conceivably be practically zero if iodine and water evaporated in nearly the same ratio as that in which they were present in the solution. Under the conditions obtaining in the present investigation, however, evaporation of iodine predominated, blank experiments with- out a cadmium disk giving always a positive value of the velocity constant. The problem of correcting for errors so introduced was encountered in one of the previous investiga- tions* and is briefly discussed in the published paper. Al- though the corrections needed in that work were rightly calculated and properly applied, the discussion as printed unfortunately contains a rather obvious error. The following is a corrected and more detailed statement of the case :+ Evaporation of the solute, iodine, (assuming constancy of the liquid surface) obeys the equation — sus = Kc in which m is the mass of iodine and A’ a constant. Simply transformed this becomes de ne edt oa a, (1) Evaporation of the solvent, water, follows the equation = a = K” which is equivalent to de ae eres I = (III) In equation (III) vw is a variable, though a constant in (II). As an approximation, however, equation (III) may safely be integrated upon the assumption that v is constant and equal to 4(v,—v,), the arithmetical mean of its limits, provided that these limits are close together.t In the integrated * This Journal (4), xxxvi, 5438, 1913. +To replace, on page 545 of the article just cited, the second paragraph and the first nine lines of the third paragraph. The error referred to occurs in the 11th line from the bottom of this page, where the word ‘‘ added” should appear in place of ‘‘ subtracted.” tThe error introduced by this approximation when v; and v2 differ by 2 per cent is only 0°19 per cent of the value of K", which is itself only a cor- rection term. In the present series of experiments the effect of this error upon the final result was in all cases wholly negligible. of a Heterogeneous Reaction. 453 equation below, v is understood to have this special value. Integrated for constant v, equations (IT) and (III) yield, respec- tively, A’ = In 7 7 6 7°66 TAO TG EGON S80 ee 774 We6Se 40267 «ai Gosent Cy 7°67 i don teOae - eu: =o tOG 7°70 General av. 7°67* Corr. for evap. — ‘05 Temperature 35° 9°70. 479-58. 29°67.- 9°60 9°69 9°59 956 952 928 9:58 9°58 9-4 59°80)" 9°68) 9°68: 2926 eee General av. 9°68 Corr. forevap. — ‘18 Temperature 45° 12141 OS ae 24 ato es 12°28 WESY, aa ee DS: Sess 11°78 12D 212-07 4. Olt ta: 11°96 19S e825 e993 11°98 Goneral av. 11°99 Corr. for evap. — 18 Temperature 55° 14°63 14:52 14:48 14:70 14°63 14 OS AOA 4 ATA: 14°87+ 14-50) pal4e22 S14 SO 14 Ae 1437+ Generalay. 14°62 Corr. for evap. — 36 Temperature 65° U8 S17 Go SiO 18 A6 17°76 106s L18a2 Sasa -97 18°02 AS Peet elie) eat ydel. 17°44 1 See al 768 269 dita eee 65 17°55 bf O04, EAS OG ier by 700 17:09 General av. 17°97 Corr. for evap. — ‘64 0°87 11°81 14°26 16°93 Reduced to 7°67 by application of the correction + Observed values in these two experiments have been multiplied through- out by the ratio 7°62/7°23, for reasons given on p. 455. of a Heterogeneous Reaction. 455 tively insignificant difference. At the other temperatures this procedure has not been followed, but all experiments not evi- dently faulty have been included in the table. While working at 55°, the last temperature to be studied, the beaker used as reaction vessel was unfortunately broken, after only one experiment (No. 27) had been made. Although anew beaker was procured which had very nearly the same dimensions it was found by a series of careful determinations at 25° that the value of A (corrected) was 7:23 for the new beaker as compared with 7-62 for the old one. Experiments 28 and 29, at 55°, were carried out with the new beaker, and the observed velocity constants have therefore been multiplied Fic. 1. ~ ao REACTION VELOCITY ~ oO on 0° 15° 25° 35° 45° 55° 65° TEMPERATURE throughout by the ratio 7:62/7:-23 to make them comparable with the rest of the table. The value of A for 0°, 3°72, was derived by a:short extrapo- lation from the observed value for +0°15°, (8°74). The observed relation between reaction velocity and temper- ature is shown graphically in fig. 1. Flwidity.—The above results derive much of their interest from their bearing. on the diffusion theory of heterogeneous reactions, and for this reason the change in the fluidity of the solution with the temperature was also measured, on account of its possible influence on the thickness of the diffusion layer. The solution used was made up to approximate to the compo- sition of the reaction liquid near the middle of an average experiment of Table II. It contained, per liter, 83:3 grm. KI, 2°8 orm. L,, 4 grm. Cdl,, and 0:1 mol H,SO,,. Table III gives the values of the density d, the absolute viscosity 7, and the fluidity ¢, for this solution at the different temperatures in question, together with the values of the ratio Am. Jour. Sc1.—FourntTa Series, Vou. XLIII, No. 258.—Junz, 1917. 456 LR. G. Van Name—Temperature Coefficient K/bT, the reaction velocity divided by the product of the fluidity and the absolute temperature. ‘It will be observed that this ratio is very nearly constant, its maximum variation beimg only about three per cent. The possible significance of this fact will be considered later. TABLE ITI. Density, Viscosity and Fluidity. OF 15° 20° 39° 45° 50° 65° d 1:0683 1°0663 1°0688 1:0594 1:0551 1:0499 1:0444 n 701659 -:01095 -00868 ‘00709 :00598 -:00510 -:00438 d 60°3 Ors oe, ae 167°4 196°2 228°3 K orl” 29-6 2.93 99:9 99%) - 99:2 99-9 . aang The fluidity when plotted against the temperature gives a curve of the same general form as that for the reaction velocity (fig. 1), but its rate of increase with the temperature is rela- tively slower. This fact is shown ina different way in Table lV, which gives the temperature coefticients of each expressed as the ratio of increase for every 10° temperature rise. For the interval 0° to 15° the value of this coefficient was calculated for A (and similarly for @) on the assumption that within this interval din /dT = constant. Integrated, this gives log, K,—t0g, HK, = A(T,— Z,) (in which A is a constant whose value is found from the experimental data), and hence EO log Fe = Ae TaBLE IV. Temperature Coefficients of Reaction Velocity and of Fluidity. 0°-15° 15°-25° 25°-39° 30 -40° 45°-55° 5d°-65° A esi — 1.350 1:298 i252 1237 1°208 1°192 vG “ae serfs leo G9) 1'261 1:225 1'186 172 1°163 t Discussion oF RESULTS. The absence of any noticeable irregularity in the graph of the reaction velocity in fig. 1 is an indication of the consistency of the values of A, and their probable freedom from large errors. A similar series of measurements has been made by Meyer Wildermann* of the rate of solution of benzoic acid in water, which was determined at five temperatures covering the range between 1°5° and 60°. * Zeitschr. phys. Chem. Ixvi, 445, 1909. of a Heterogeneous Lreaction. 457 _Wildermann’s velocity constants, K, are given in Table V, Kiss together with the values of a as calculated for the various t intervals by the method described above. TABLE V. Results of Wildermann. 15° si By i A()° 60° K = 1°587 2°851 4524 5°756 9°946 Ars 10° = 1:443 1°408 1°307 1°314 K, A — 2920 3020 2550 2850 This reaction is not as well adapted for accurate measure- ment as the one between cadmium and dissolved iodine, and Wildermann’s duplicate results show a rather poor agreement. The values of A given in the last line of the ee are those calculated from the equation /nK, , the < Tt integrated form of the van’t Hoff equation, din /dT = A/T”, for the effect of temperature on the velocity of a homogeneous reaction. Wildermann, whose article is throughout an attack on the diffusion theory, regards the apparent constancy of A as an argument against that theory. The soundness of this argu- ment, however, is very doubtful, since there is nothing to show that an approximate agreement with an equation of this form is In any way incompatible with the mechanism of the reaction postulated by the diffusion theory. Upon calculating the constant A from the reaction veloci- ties measured by the author, as given in Table II, we obtain the following values, which show a distinct progression and a total variation of over 20 per cent: 0°-15° 15°-5° 290°-39° 30 —-40° 45°-55° 00° —65° AS 2388 2239 2069 2079 1962 1902 It is evident that in this case, at least, the experimental results do not contorm to the van’t Hoff equation. A question of special interest and importance from the standpoint of the diffusion theory, is the nature of the relation between the observed temperature coefficient of the reaction velocity and that of the rate of diffusion of dissolved substances. Unfortunately our knowledge of the effect of temperature on rates of diffusion is very imperfect, since but few direct meas 458 Rk. G. Van Name— Temperature Coefficient urements have been made at temperatures above 25°. Nernst* gives the equation D,= D, (1 + 0°026 (¢ — 18°) ) as approxi- mately expressing the experimental results in the case of salts. For acids and bases the numerical coefficient is 0:024 instead of 0°026. Calculating from this equation the value of sea for , es the different intervals in the range covered by the reaction velocity measurements we obtain the following results: 2:0 —12°5> © Po =2)~ 20-00 30 —45° 45°-55° = §5°-65° D,, 410° ae D, These figures can only be regarded as rough approximations, since the change in Y with the temperature is not, in reality, strictly linear as the equation assumes. Moreover Oholm¢ has shown that the numerical coefficient is not the same for all salts, but varies between the limits 0:023 and 0-027. On comparing these results with corresponding ratios for the reaction velocity, as given in Table IV, we see that with the exception of the lowest value the increase in the reaction veloc- ity with the temperature is throughout slightly more rapid than that of the rate of diffusion. This appears to be at least a partial confirmation of the predictions of the diffusion theory, since the increase in the fluidity with the temperature probably causes some decrease in the thickness of the diffusion layer.t No great significance can be ascribed to this result, however, because of the above mentioned uncertainty in the calculated values of the diffusion coefficients, and because the case of iodine diffusing in an iodide solution is not wholly analogous to that of the diffusion of a simple electrolyte. Although the iodine is largely in the form of tri-iodide ion a probability of complications is introduced by the fact that this ion is itself in dissociation equilibrium with free iodine. It has been proved, for example, that the rate of diffusion of iodine in potassium iodide solutions increases with the concentration of the salt,§ while the reverse would be expected if the case were normal.| * Zeitschr. phys. Chem. ii, 625, 1888. | + Zeitschr. phys. Chem. 1, 309, 1905. + That the effect would be in this direction is practically certain, but noth- ing definite is known about its magnitude. The results of Van Name and Hill (this Journal (4) xxxvi, 548, 1918) would seem to show, however, that the effect of viscosity changes upon the thickness of the diffusion layer is rela- tively small. § Edgar and Diggs, Jour. Amer. Chem. Soc., xxxviii, 253, 1916. This follows from the fact that the velocity of the tri-iodide ion is con- siderably lower than that of the potassium ion, together with the rule (Abegg and Bose, Zeitschr. phys. Chem., xxx, 951, 1899), that the presence of an excess of a common ion tends to impart to a diffusing salt the velocity char- acteristic of its other ion. 1°436 1°282 1°220 1°180 1°153 1°132 of a Heterogeneous Leaction. | 459 ‘ In short, no adequate comparison between reaction veloci- ties and rates of diffusion can be made in the present case with- out a knowledge of the value of the diffusion coefticient of iodine in 0°5-normal potassium iodide at several of the tempera- tures involved. The value of this coefficient at 25° has already been measured by Edgar and Diggs,* but at other tempera- tures its value is unknown. The most that can be said at the present time is that the observed changes in the reaction veloc- ity with the temperature are of very nearly the order of mag- nitude which would be expected from the usual value of the temperature coefficient of diffusion. We have still to consider the significance of the observed proportionality between the reaction velocity and the product of the fluidity by the absolute temperature, which is indicated by the constancy of A/@7. No doubt in many cases the diffu- sion coefficient of a dissolved substance is proportional, over a fairly wide range of temperature, to the product ¢7. This should be at least approximately true in cases in which the diffusion formula of Einstein Dae ] NEO 62rn P DPD = is applicable, since in this formula / and ZV are constants, and the value of /, the molecular radius, can generally be con- sidered as constant for the same kind of molecule at different and not too widely separated temperatures. Hence, ) must be proportional to the absolute temperature divided by the vis- cosity n, or D w $7. The Einstein formula is strictly valid only when the mole- cules of the solute are not dissociated and are large compared with those of the solvent,+ but dissociation, if nearly complete, would not necessarily interfere with the proportionality be- tween D and @7. There is, however, one limitation which may be important. / is the radius, not of the molecule of solute itself, but of the diffusing particle, which in many cases includes an attached or entrained group of molecules of the solvent. Whether P in such a case would vary with the tem- perature or not, would depend upon the nature of this attach- ment, but its variation might often be appreciable, especially over wide temperature intervals. If we assume that the rate of diffusion of iodine, under the conditions of the experiments in Table II, is proportional to the product $77, it follows from the constancy of A/¢7’ that the reaction velocity is closely proportional to the diffusion * Loe. cit. + Hinstein, Ann. der Physik (4), xix, 989, 1906. 460 R. G. Van Name—Temperature Coefficient, ete. coefficient. This, regarded from the standpoint of the dif- fusion theor y, would mean that the change in fluidity between 0° and 65° C. (an increase in the ratio 1: 4) has no appreciable effect upon the thickness of the diffusion layer. On the other hand, it is at least equally probable that in reality D for iodine ‘in potassium iodide solution increases somewhat slower than $7, for the disturbing effect of the thermal dissociation of the tri-lodide ion would act in this direction, as is evident from the fact that iodine diffuses slower in pure water than in aqueous solutions of potassium iodide. If so, the rate of diffusion must increase with the temperature somewhat more slowly than the reaction velocity, in complete agreement with the predictions of the diffusion theory. The truth or falsity of this inference can and must be decided by direct measurement of the diffusion coefficient of iodine at the different temperatures, upon the values of which, as already stated, the settlement of the whole guestion turns. Summary. 1. The velocity of the reaction between metallic cadmium and iodine, dissolved in a 0°5 normal solution of potassium iodide, has been measured at 0°, 15°, 25°, 35, 45°, 55°, and 65°, together with the fluidity of the solution at each of these tem- peratures. 2. The temperature coefficient of the reaction velocity for a 10° rise varies from 1°35 for the lowest, to 1:19 for the high- est temperature interval, and is therefore of about the same order of magnitude as the temperature coefiicient of diffusion. of a binary electrolyte. 3. The reaction velocity is proportional to the product of the fluidity of the solution by the absolute temperature. Berry—A. Sal Fish from the Virginia Miocene. 461 Art. XLI.—A Sail Fish from the Virgina Miocene; by Epwarp W. Brrry. Tue Paleontological collections of the Johns Hopkins University contain two specimens of vertebrate fossils collected many years ago from Tar Bay in Virginia and without other record than “presented by Mr. Dew.” At this locality a few feet of Eocene glauconitic marl of the Aquia formation is over- lain by a considerable thickness of argillaceous beds of the lower part of the Calvert formation. The fossils represent a well-preserved tooth of Physeter vetus (Leidy) and a rostrum of a new species of Istiophorus. The former is a well-known Miocene type and the latter is also considered as coming from the same horizon as the tooth, although this is not conclusive. The genus Istiophorus of Lacépede (numerous authors, as for example Smith Woodward, use Histiophorus) comprises several existing species of large, elongate, compressed, pelagic fishes of warm seas—chiefly Indo-Pacific, but also represented in thie Atlantic. They are popularly known as Sail fishes because of the height of the undivided dorsal fin and have the premaxille vomer and ethmoid united and produced forward to form a rostrum, bony muzzle or sword, which is shorter and rounder than that of the closely related more recent and more special- ized true sword fishes (Xiphias). Istiophorus is a convenient name for the rather numerous fossil rostra of Istiophorus—or Tetrapturus-like forms, of which a number of different types have been recorded from strata ranging in age from the Eocene to the Pliocene. The rostrum, which is more complete and better preserved than is usually the case, may be described as follows : Istiophorus calvertensis, sp. nov. Rostrum subcylindriecal, tapering distad. This tapering is gradual when the rostrum is viewed from the side, and con- fined to the distal one-third from the dorsal or ventral aspect. Bluntly pointed, 31°" in length, in general elliptical in transverse outline. Surface of anastomosing striz. Dorsal surface dis- tinetly less convex than the ventral surface for its anterior two- thirds. The anterior third shows scarcely any trace of the lateral dorsal furrows which about half way to the base are pronounced but not sutticiently emphasized to break the even contour of the surface. From the median region backward these furrows increase in depth and width, flanking a more arched median region, at length becoming toward the base an open lateral gutter, the inner limb of which is flattened and 462 Berry—aA Sail Fish from the Virginia Miocene. with an open obtuse sinus. This sinus becomes more angular as it is traced forward until at one-third of the distance forward it is acutely angular, from which point forward its sides gradually become confluent. The ventral surface is convex, show- ing traces of a lateral suture separating the marginal 4 from the median 4 of the lateral diameter. There is a pronounced median suture more or less distinct for the whole length of the rostrum. This broadens out in the proximal third of the rostrum to form a broad shallow palatine sinus. There are no traces of dentigerous surfaces. Each premaxillary is traversed by a nutrient canal which at the base are large and elliptical with their longer diameters oblique and forming an angle of about 40 degrees. Ventrally they approach within about 2™™ of one another and their maximum and minimum diameters are 12™" and 6°5™™ respectively. Medianly and lying above the premaxillary canals is a central oblanceolate foramina about 7"™ in height and with a maximum transverse diameter of 1™™ dorsad. The component elements of the rostrum are completely coossified and indicate a fish of considerable size. / Vertical Transverse 1/4 distance from base 26™™ OSGi Diameter of rostrum 4 1/2 eto aie 3/4 18-5™™ ggmm The accompanying text figures 1/3 natural size show the dorsal (fig. 1) and ventral (fig. 2) aspects of the rostrum and transverse sections (figs. la, 16, le, 1d) at the points indicated. Oceurrence.—Calvert Formation. ‘Tar Bay, James River,. Prince George County, Virginia. The fragments from the Ashley River marls which Leidy* described as X¢phias robustus and appears to have considered as Eocene in age, and which Hayt referred to Istiophorus and considered Post-phocene in age is similar in general form but smaller and with two dentigerous bands separated by a groove. The age of the South Carolina form is, of course, uncertain since the Ashley marls contain so many mechanically mixed fossils of various ages. As I have remarked elsewhere, this may be true of the Virginia specimen since the Calvert is underlain by the Aquia Eocene. In both cases, however, the Istiophorus rostra are associated witb teeth of the Miocene Physeter vetus (Leidy). L. antiquus (Leidyt) from the Eocene of New Jersey is ee J., in Holmes’ Post-pliocene Fossils of South Carolina, p. 119, pl. 27, figs. 83-5, 1869. + Hay. 708 Pe U. 8. Geol. Survey Bull. 179, p. 402, 1902. { Leidy, J., Proc. Acad. Nat. Sci. Phila., vii, p. 397, 1855. Berry—A Sail Fish from the Virginia Miocene. 463 smaller and more flattened. J. parvulus (Marsh*) from the Upper Cretaceous or Eocene of New Jersey is a much: smaller, more slender, more pointed, and more compressed form. I. homalorhamphus (Copet) is more nearly like the Virginia form, but is considerably smaller, somewhat different in form, being more tapering and more nearly elliptical in transverse Fie. 1. Fic. 2. section. The premaxillary canals are much smaller and the dentigerous area is well developed. This last species is of uncertain age and may be either Upper Cretaceous, Eocene or Miocene. J. eocaenicus (Smith Woodwardt) from the middle Eocene of southern England is much smaller with straight * Marsh, O.C., Proc. Am. Assn. Adv. Sci., p. 227, 1869. + Cope, E. D., Proc. Bost. Soc. Nat. Hist., xii, p. 310, 1869. ¢t Smith Woodward, A., Cat. of Fossil Fishes in the British Museum, Pt. 4, 1901, p. 495, fig. 18, No. 2. - 464. Berry—A Sail Fish from the Virginia Miocene. converging sides and /. votundus (Smith Woodward*) from the Phosphate beds of South Carolina is a very much shorter, wider and more rapidly pointed form. In addition to the foregoing Van Benedent has described a number of rostra under different generic names including a somewhat similar but larger and otherwise different form which he calls Brachyrhynchus teretrirostris (Riitimeyer) and which occurs in the Mio-Plocene of Belgium, France and Italy. LB. solidus (Van Beneden) from the middle Eocene of Belgium is somewhat similar to the Virginia form but more depressed and with much smaller and more widely separated nutrient canals. AXiphiorhynchus elegans (Van Beneden) from the Middle Eocene of Belgium is also somewhat similar, but shows five nutrient canals. These embrace all the forms with which the present fossil has been compared. Johns Hopkins University, Baltimore. Art. X LII.—Haklette, a New Mineral from California; by Esper 8. Larsen. @ Wuitrt making a microscopic study of minerals the author found in the Museum of the University of California a speci- men labelled “ Wollastonite, St. Inez, Calif.,” whose optical properties are different from those of wollastonite, pectolite, and all other known minerals. The mineral proved to be a new species and the name eaklezte (pronounced akeé-el-ite) is proposed for it, after the mineralogist Prof. Arthur 8. Eakle. The largest piece of eakleite in the collection is about three centimeters across; it is free from foreign material and is made up of successive, irregular layers of fibers, some of these layers several millimeters across. It is compact, very tough, and has a hardness of about 63. A specific gravity measure- ment on a large fragment with a balance gave 2°705 and one made by suspending in Thoulet solution gave 2°685. Eakleite fuses at about 2°5 with shght boiling to a glassy, somewhat vesicular globule; it loses its water only at a high heat. It is easily soluble in acid with separation of flaky silica but without gelatinization. : *Idem, fig. 18, No. 8. + Van Beneden, P. J., Recherches sur quelques poissons fossiles de Belgique, Bull. Acad. Roy. Belg., xxxi, pp. 495-518, pl. 1-4, 1871. t Published with permission of the Director of the U. S. Geological Survey. Larsen—Eakleite, a New Mineral from California. 465 The optical properties of eakleite are characteristic. It is pale pink with vitreous to silky luster. Under the microscope tibers give parallel extinction and are seen to be elongated parallel to Z. It is optically positive and the axial angle is very small. The indices of refraction as measured in sodium light by the immersion method are: o =a) Seana OF00T. B = 1°583 + 0:00). y= 1:593 + 0°001. Chemical analyses.—Two chemical analyses of a sample of the mineral, which a microscopic examination showed to be free from impurities and homogeneous, were kindly made for the author by Professor Eakle, and are given in columns 1 and 2 of the table. Column 8 gives the average of the two analyses, column 4 gives the molecular ratios, and column 5 the theoretical composition of a mineral with the formula 5Ca0.58i0,.H,0. Analyses and ratios of Eakleite. 1 2 3 4 5 S10, 50°43 | 49°90 | 50°17 832 5 xX 166 50°16 Fe,O, 0-98" | 11141’ 4-04 6 ~ CaO 45°51 45°39 |. 45°45 Slo 5 xX 168 46°82 MgO tr. tr. tr. Na,O & K,O | none | none | none H,O 3°25 a kL 3°18 il via ROC ee 3°02 100°17 | 99°51 | 100°00 The relation of eakleite to other minerals is not clear, but it may be a calcium pectolite, as some analyses of pectolite show an excess of water over that commonly assigned to the species. Its fibrous structure and general appearance are those of pectolite. 466 O. O. Dunbur—Rensselerina, a New Genus. Art. XLIIT.— Rensselerina, a New Genus of Lower Devonian Brachiopods ; by Cart O. Dunpar.. With Plate II. [Contributions from the Paleontological Laboratory, Peabody Museum, Yale University. ] In the Linden shale of western Tennessee, which is the equivalent formation to the New Scotland of New York, the stout dorsal beaks of an undescribed Aensseleria-like brachi- opod are among the most conspicuous fossils, although the rest of the shell, which is quite thin, is far less often preserved. It is a very characteristic species in the southwestern develop- nent of the New Scotland fauna, but seems not to have reached the Appalachian trough. Its elongate form and partially pli- cated surface, combined with its specialized, septate brachi- dium, preclude its reception in any described genus. It is therefore proposed to give to the new genus which it repre- sents the name /tensselerina, because of its nearer approach to frensseleria than to any other genus of the Centronellide, to which family it belongs. The “genotype is Leensseleerina medvoplicata, n. Sp. It is a pleasure to acknowledge the writer’s obligation to Professor Charles Schuchert of Yale University for advice and criticism in describing the new genus. Order TELOTREMATA Beecher Superfamily TEREBRATULACEA Waagen Family CENTRONELLID# Hall and Clarke RENSSELARINA, new genus Diagnosis.—Shell decidedly elongate oval, often nearly cylindrical, but otherwise having the general form of the later Rensseleerias, though always less broad-shouldered. Shell sub- stance very finely punctate. Of the exterior surface, the pos- terior half and the lateral slopes of the shell are entirely smooth ; but anteriorly the median portion of both valves bears a few simple, rounded, conspicuous plications which become obsolete about half-way to the beaks. The plications are separated by well-defined concave interspaces that are at least as wide as the plications themselves. There is no eardi- nal area and the cardinal and lateral slopes are rounded. The sides of the shell are not incurved as in Geachia, nor steeply angulated as is commonly the case in Rensseleria. The minute beak of the dorsal valve is concealed by that of the ventral valve, which is strongly incurved and closely appressed ; the pedicle foramen is minute and neither resorbed nor abraded through the umbo. 0. O. Dunbar—Rensselerina, a New Genus. 467 The ventral interior differs from that of Stensseleria in having very small vertical dental plates that do not attain the bottom of the valve, while the pedicle or rostral cavity between them is very deep and narrow. The muscle scar is limited to the posterior half of the shell. The diductor scars are elon- gate and slender. Between their more deeply impressed pos- terior portions are embraced the narrow imprints of the adductor muscles. The dorsal interior has a thick and con- spicuous triangular hinge-plate, supported by two thickened crural lamelle. This plate is highly variable. Most commonly it is thickened medially into a shght triangular elevation, the actual cardinal process (see Plate II, figs. 14-18), on which are the scars of the cardinal muscles; but all gradations occur from such an elevation to a decided triangular pit (Plate LI, fies. 14, 15) for the insertion of these muscles, or to the other extreme where a distinct median process arises (Plate II, fig. 18), at the sides of which lie the muscle sears. This tendency toward an elevation of the center of the hinge-plate is in marked contrast with Pensselerza, wherein the plate is primi- tively flat but in later forms always has a depressed cardinal pit. As im most Lower Devonian terebratulids, the hinge- plate is traversed by a well defined visceral foramen, which, beginning dorsad of the anterior edge of the plate, emerges on its surface just below the beak. The adductor muscle scars are narrow, separate, and distinctly impressed. The brachium is distinctive and may be described as follows: The thin, slightly divergent, descending lamelle are broad and stand nearly vertical dorso-ventrally as they leave the wide and conspicuous crural processes; then, diverging more rapidly, the lamelle become narrower and rotate into a horizontal position as the loop attains its greatest width, and by a regular curvature begins to converge (Plate LI, figs. 1, 3). Approach- ing the median line, just in front of the middle of the shell, the loop broadens into a spoon-like plate, drawn out anteriorly into a long acumination which curves gently downward. The edge of the loop, which was ventral when it left the crure, is now anterior and forms the outer margin of this plate, dipping outward at about 40° below the horizontal. From its concave ventral surface arises a vertical median lamella (Plate I, fig. 2) which extends forward beyond this plate, just in front of which it reaches its greatest height. It continues anteriorly to near the front of the shell, where it approaches the ventral valve. Relationships.—This genus is one of the several divergent lines or phyla of the Centronellide. While in general the ex- ternal aspect resembles most Amphigenia and Lensseleria, it presents essential differences. Amphigenia is at once set off 468 C. O. Dunbar— Rensselerina, a New Genus. by its entirely smooth surface and, especially on the interior, by the incomplete loop, the clearly defined double dorsal septa, and the extremely large rostral cavity which apparently has become a true spondylium and is supported by a high median septum. ftensseleria differs in its uniformly striate surface, in its wide, strong, and recumbent dental lamelle whose outer sides rest on the bottom of the valve and which unite medially to make a broad and more or less deep rostral cavity ; but most especially in its simpler brachidinm, which is without the median vertical lamella of the new genus. | Beachia is readily distinguished from ensselwrina by its flat ovate shape, its erect ventral beak which exposes the delti- dial plates, its smooth or finely striated surface, and its inverted margins, but particularly by its simpler ARensselwria—like loop, with a median posterior spine instead of an anterior ver- tical lamella. Lissopleura is for the time being provisionally placed in the Centronellidee, although it is not yet clearly understood. At any rate, it differs widely from the new genus under considera- tion in its rhynchonelloid shape, in its coarse plications which have no interspaces between them, and in its high dorsal septum. One other genus needs comparison. omingerina is based on a single species, /?. gulva, which is a very small, smooth, lenticular shell from the Lower Mississippian of Michigan. The only character which seems to relate this genus to the new one is the presence of a median vertical lamella on the loop. However, this is evidently only a ease of parallel development. The lamella in Romingerina is described by Hall and Clarke as follows:* “The median ridge on the anterior plate of the brachidium is elevated into a conspicuous vertical lamella, ex- tended both anteriorly and posteriorly, being in fact a double plate produced by the abrupt deflection of each lateral branch of the brachidium near the median line; union taking place along the upper edge, which almost reaches the inner surface of the pedicle valve.” It is but a simple step to the development of this type of brachidium out of one of the later Centronellas. Furthermore, the shell of Romingerina has a smooth surface, its beak is erect enough to disclose the deltidial plates, and it frequently shows a faint fold and sinus—all centronelloid characters. The lateral branches of the loop in Centronella are also sharply de- flected before meeting medially, their deflected edges forming a slight ridge. A mere expansion of this character will pro- duce essentially the loop of Leomingerina. * Hall and Clarke, Pal. N. Y., vol. viii, Pt. 2, 1894, p. 271. Winchell’s drawings erroneously show the crura directed dorsally. O. O. Dunbar—Rensselerina, a New Genus. 469 Snch a genus in the early Mississippian, however, can searcely have evolved from one which in the early Devonian has already reached as great a specialization in another direc- tion as is the case in /eensselerina. In the latter genus the vertical median lamella is simple, not double, it has no posterior extension, and it is not formed by a deflection of the descend- ing lamelle of the loop. On the contrary, the halves of the loop meet medially without any deflection or thickening, and - the dorsal side of the median plate is smooth and evenly rounded, while from its ventral side there arises the very thin, simple, vertical lamella. Summary.—-The essential characters of the new genus Rensselerina will be, then, its rensseleeroid shape, its medially plicate surface, and, on its interior, the centronelloid brachidium with its high, vertical, anteriorly directed, median lamella. Rensselerina medioplicuta, new species. Pl. Il, Figs. 1-9, 12-18. Shell with the characters of the genus; length of a mature individual 24™", width 14™", thickness 11™. Six to eight plications on the middle of the ventral valve and one more on the dorsal. About three of these plications in a width of 5™™ on the anterior margin. Hinge-plate extremely variable, being in some specimens excavate, in others raised into a median process. Florizon and locality.—Linden formation; western Ten- nessee. The cotypes are in the Peabody Museum, Yale University. Rensselerina medioplicata var. latior, new variety. Pl. Te iess 10,11. In Professor Schuchert’s collection there is a shell like L?. medioplicata but proportionally much broader and shorter, and with the plications finer and somewhat more restricted to the anterior part of the shell. The length of a fully grown specimen is 20™", breadth 16"", thickness 11™™. There are four plications in about the width of three in 2. medioplicata. Horizon and locality.—Linden formation ; Henry County, Tennessee. Holotype in Professor Schuchert’s collection. 470 C. O. Dunbar—Rensselerina, a New Genus. EXPLANATION OF PLATE II. Rensseleerina medioplicata, n. sp. Fies. 1, 2, 3. Dorsal, side, and ventral views of the loop, restored. In figs. 2 and 3 the visceral foramen which traverses the hinge-plate is in- dicated. Fie. 4, Dorsal view of the posterior portion of a specimen. Fie. 5. Cardinal view of a very obese Speen ea which is almost cylindrical. The beak is abraded off. Fic. 6. Cardinal view of a normal specimen. Fics. 7, 8. Lateral and ventral views of a specimen which shows the normal shape of the species. Fig. 7 is slightly enlarged. The shell is ex- foliated, so the plications do not appear clearly. Fig. 9. Ventral view of another specimen, to show that the posterior portion and lateral slopes are smooth and only the median portion is plicated. Fie. 12. Interior of a ventral valve, to show the deep pedicle cavity, the hinge-teeth supported by small dental lamelle, and the narrow adductor muscle scars embraced in the posterior portion of the diductor muscle scars, x2. Fic. 18. Interior of the posterior portion of a dorsal valve which shows most of the slender adductor muscle scars. x 2. Fies. 14, 15, 16. 17, 18.