<= “8 J a ee Spey ~ it Ytete? “att *t + . 2% Py otelate reevGarars C4 VERE TOTS tad ye HW BEN 4% ' wpe ven ce beet taewa ag! ae pe eee eae envee re weet tat od bared a WC Wa MC eR < 4 soe rave pee Pee ek eee ea VE ae be ay sea we we oe ‘ Parear er reat Wa WO wf TE eee RO OC ee A BO tHe oe Pera e ae terest Pearverrar ar eerie Ye Det er ee P * La forego? Preparer vad QVeypryeryy ces pgp e er Tae Fos fe seve haa f Ada dy tad ORR ot bes re weer o ecgeek ae Re Sol We ee ig ee ee we eC oe 4 % io Wee rear Ra 9h repareuy are Lanna ts eet ee ee ly bee hoe co ec eee ot ovteueed 8 wae vr TR ED RT She a) ’ ee Piers WT Ch eat c fo what Cpe Fg at row fp eae es OW got VS oe ee * eht tee ger yp les Cerne at eee ee sy Rohe cum PAU CAR A eh Ce Pee or er ee ee Vp eee g ) oa Oe Se ae as ’ a tr RTA NOE a eR ROCA LU a8 Lee dine Cr Pee Freeda wpe eer ets Re Tee swat a eer”? ET Ag we, ceded Cop fsa rhe bid hs Gal CAEL SOLS EON Le Parad WC OO NA Pie wh) Bese Sas, Se Led wad Fie p, an ee a: pur teagan FOP ee hoes ? yr eee . son ebal wee 2% oe Ps whew tw? Ee BAT OE bP kt tA) a ’ Pas at en See ig pi rpeaens : : y tT Sit tetera OATS T EMIT att le ty bu WHY Me ray ee er ie ehh mens sperhewpavh yd te wp lp eat + Py A OL Bee FRCL Bt et a te Pr ee ee ee oe he A Le ree, rae vee eur ewapere pou porvueey ANOS OO ORLA Seen sear i 4 evewee ‘ RUN eM i Rk he pe MAA ARH Pekan th Seo ree pe hore vee no eee eA pequy ” * i ee Oe SO Le ACCRA LS Ne ' oe ys Prerariarie Westie Ae a A A ey ey ed Ee oeagi Wks cob: vie gem ti. 0 fia ea « Cee Urge oe EY TS VEG > 6 vebecegrast 4? TOE A AAAs wa ‘ ‘ te Par a Os MD ape ye Cl eee | is bk A PA de * ye ee! TR PPE Dg dos peewee gt RP et HM oe Oe oO he wey ney . ' ) Pye} aust ptm et Pes eye pe Lhe OY Sy R11 OG BA Pe oe Ne Hype er boned vee ove yeu n COye Le OT RES at Peed Ow La et son erreas Oh Ae HA Aa tLe ' vue 3 + } a ae ee ek ROL ea sn ae perme rath ey the rant 1 S edd 8 1 u's ‘ ¥ : rao by ee bt b teteree y bd Rh, b eee ary nya Ce ‘ of Sow yo be ee ath elas 9 a ttatanie taza atheacare hea bee ee on i, bo Pune 4, 0g Ne ED 9 “4 Omit we Cee " Deal ii atta ah ‘ . 8 a ee ee ‘ , peeve 1 o ee alia \ eyoyale ' ‘sh? so atwe yer PEP Rey ree | ‘ ' 14 t+ ’ ina BAAR Ck ee weet 4 ' pee aca hae fell . ' os yh oi dae e | CALA Re ‘ geht iy ‘ yrene Chee ved Pee ee ey OP Ua et nee ene pret epeye ee ge dy rovrsbatyypreyd Arr tipped , ray. R aay We vie ty peruye "y RE aS ROR aru A ieie di Ne E Ba ee Ana LA Ra FP. Yoo ses pe day webby ore ie Ce ee eee Lb ‘ aie bowels i] ‘ 1) 1 Veyy tv yh Ve ye tae) & ’ © 4 “ . : yy . PRT RORIE IC 74 ' OE ee 4 wy tA EADS S| ‘ ‘ ’ eves he \ oF tes pee 4 Y>) ' wee ae ‘ te pen yr yu ‘ a) eee oe he PN ee ee Eo pt ‘ r) edkoeagade ee Pe doa +eey Cvs ey seve t ‘ eae yee Prey ed yt hye eae gs tary _iss eytp eee araty ee RO eee | sree eh e ee ge pa dh Suvrys ' a ce ad ! Palais TENS 114 : Ky Pin eID ~} 3 a ' ‘ gs ony My ‘ ace N Bly yy > eRe) . ’ ab al wy deed i ah une’ oe y thy b OER ahh A Nun ’ y : ea r baw ng da a en i et Daeg ‘heats aN PPe aioe vy Peon een Pree rae ere etsy Ld “ys. Pay) Me aes ener “a ve ; are ae be be ral ‘ “yates Kies duce Bh, By hi yi Ae gate’ # i ky. tis winnie YS 34, ahi Sal's, Pb ay. Fee Fae esi { Pen : hic Teed aed, in ied de af iF aie sinh ue Ee us 4 ath nae ha ee aes 1 Le ai, Ay dares rey 4 Mei ests iat a ee he Perr * ” a deo doe ai Meer ta ; . Rr aR Rates dod ed jedi Bebe d | q dobea kh ed woe COP eg aa PM be ata ons ye been bare) Vea heb Cee ed Fea a gd erro eee eee ee P Ve 23 wei de PPA SA Aerie mere ea Be Le S38 ‘ ee bal ot Wie ya bee a he Harr Fb rere a aie A pa! ie, tarp ORR ORR trails ot Me ot wide th, ti “te ay i Li awe oe UNE Ie POR Bee Fe errant ae ee ere re whe Bt CA I iat SAA tates ow ane I ileal Mal daw eR tok se ed ee er Prana Ue eee eee ae Ge ada bg oh ig hdd ae wae Oe UA dere eee SoG hl bos 8 aw ‘ v-4 isae ‘ tedad bad Saad & ae ae NE 3 ‘ 4% bon : a Pv hal ee ee ee Pe se ad ad Fea des wee 8 , f 4:0. Bie >. tae thas eee Pee debe 8 for gest bhoe dd babe eid ad bd eet ie rated ‘sata oe Pn i ee eves ea ee DY dene 7 be a a DON sata ep iadiada : ne Werth Eee SOS Peon eee oe ay Ratan ” ; ; : tie Ae Me Mare ara aa Pravastatin pie Boh Scant een he aver tee ‘ , auf . Prawns Wrenn eer er ee err ey a ere Hed a UK et We of we wae ‘ ‘ ‘ ' rare adage t on eene ROR aR Pisin ay H se n +t ee mee) we th ae +e LN Ow Se Ba creme en sai rate ec ; Se eC eT gt Wh Xe Le rade mb she ast Bw niubie 4 rar aba Md RC NOL ma tale 2 ua CN : : : ae ve dg eta ls Gdaee wee Liat kr ty ob Lb Teo ‘ dor am 8 . aa Db hae bb bee a? rere ee ree See . aes 4a a ee aev io Fo Uaen ‘ elt eevee i fide eae tbod ‘4 jueecee w 8 ‘ pan ba bene bes eed 54 Dak ot Lead dasee OTe a aca aaa a! es al tana de adhered ae Fat bi blanwaae a a nid ao LA ‘ ied ts a Farina, cer tote is ‘ Sie LP ar acer Ter Ie a ae OC i ‘ iy “ay ne ‘ : 4 ae eee ee ea ph ec bh ota Le ae tee eM wee ree Ie a Be Oe a oe to it ie oe chad oa V4 SoMa ERD ENE a RT ah OR fs Pe ie Lait a yom 4 ec ‘ ‘ ‘ 4: oo ae ~t ry pe tet hd bed tous Oo we EET a eS whan teas OG eedid bode s,—s,, there would be a resultant current away from the electrode in the second case. Applying this theory to the data for the brass electrode, it appears that for 2°2 volts As, and As, must be some 50 per cent larger than s,—s,; that is, there must be a large number of slow-moving 8-particles pro- jected from both surfaces, rays that 2 volts will stop. Such rays are produced by polonium,* radiumt and other radio-active substances. The experiments further show that approximately As,=As, for the same electromotiye force. The assumption made in this theory is that the value of s, does not depend on * J. J. Thomson, Nature, Dec. 15, 1904. Proc. Camb. Phil. Soc., xiii, Pt. 1, p. 39, 1909. + KE. Rutherford, Radio-activity, pp. 151-104. Duane—Emission of Electricity. us that of s,, and vice versa. As the secondary rays depend on the primary, and as there are tertiary rays depending on the secondary, etc., this assumption is probably not strictly correct. : Of course s, and s, may be further analyzed into their con- stituents a and @ rays, secondary and tertiary, ete. 13. We now come to the important question, do the degay curves that can be drawn with the datain Tables I and II coin- cide with curves drawn from ordinary ionization data? It will (oe) “yaLINAO Time in minutes. be seen that they do not. For comparison I measured the cur- rents due to the ionization in air produced by the a rays, when the electrode was made active in exactly the same way as before. The electrode, after being withdrawn from the ema- nation, was placed opposite a hole in a large condenser, and the decay of the ionization current due to the a rays measured. Table III, column 2 contains the valnes of the currents meas- 8 Duane—Lmission of Electricity. ured by the piezo-electric quartz, and column 1 the time in minutes after the electrode had been withdrawn from the ema- nation. In order to compare them, I have drawn this curve and the curve representing the decay of the positive current for the brass surfaces (Table II, column 2) together in figure 2. Curve 1 represents the decay of the a ionization current in air at atmospheric pressure, and curve 2 that of the positive enr- rent in vacuum. The scales of the two curves were so chosen as to make them coincide at ¢= 70. It is easily seen that they do not coincide at other points, the curve representing the charge carried by the rays being much steeper than the other. 14. The explanation of this undoubtedly is that radium B, which under ordinary circumstances produces an almost map- preciable amount of ionization, does emit a very appreciable. number of negatively charged rays. For the sake of compari- son, I have calculated the theoretical curve that one should get under the supposition that radium B emits as much electricity during its change as radium C does during its. This is curve 3 in the figure. It is drawn under the assumption that there is on the electrode, to start with, equilibrium amounts of radium A, Band C, i. e., amounts that are inversely propor- tional to their respective decay constants, and that initially the quantity of radium B present emits as much electricity per second as does the quantity of C present. The scale of the theoretical curve was chosen so as to coincide with the others att = 70, and it appears that the experimental curve 2 is a little steeper even than the theoretical one. In making the calculations I used constants corresponding to decay to half value in 3, 28 and 21 minutes respectively, for radium A, B and C. Itis probable that the last two values, 28 and 21, are too high. Slightly smaller values would bring the theoretical curve closer to the experimental one. All the curves that I have examined representing the posi- tive currents and the 4 algebraic sum of currents, are steeper than the theoretical curve. Curves representing negative cur- rents are usually somewhat less steep than those representing positive currents, but lie much nearer curve 3 than curve 1. It follows that radium B must emit its full share of negative electricity when changing into C. The theoretical curve representing radium O alone coincides so closely with curve 1 that I have not drawn it in the figure. 15. In some interesting experiments H. W. Smith* has shown that radium B produces a slight ionizing effect. He attributed the ionization to easily absorbed rays, probably 6- — rays. The ionization produced by rays from radium B can be shown very easily with the aid of the apparatus used in the * H. W. Schmidt, Physikal. Zeitschrift, vi, 897-903, 1905. Duane—Emission of Hlectricity. 9 above described experiments. In one ease the electrode A was made active exactly as before, placed in position in tube B and the currents measured at atmospheric pressure instead of in vacuum by means of the piezo-electric quartz. Column 3, Table III contains the results reduced to the same scale as those in column 2 for ¢=70. It appears that they by no means coincide except for ¢=70. In fact the data of column 38 repre- sent points that would he much more nearly on the theoretical curve 3 (figure 2) than on curve 1. This can be due only to the ionizing effect of radium B, and indicates that within a $™™ of the surface the rays from radium B produce ionization com- arable with that produced there by the rays from radium C. his large ionization would be explained, if radium B produced a rays having a velocity just greater than the critical velocity required to produce ions. Such an hypothesis, however, is not TaB_eE III. Time in Current in Current in minutes large con. small tube 10 3°O1 4°15 12 2-9 15 2°91 3°84 1b 2°86 20 2°83 3°03 25 2°76 3°30 30) 2°54 3°09 35 2°38 2°89 40) 2°26 nee Oe 45 2°04 50 2°01 | 55 1°83 60 1°70 1°80 65 1°58 | 1°59 70 1°42- 1°42 79d 1°31 absolutely required by the facts, and would contradict some results obtained by Bronson.* | 16. In order to estimate the velocity with which the carriers of electricity are projected from the metallic surfaces the brass electrode and tube were placed between the poles of an electro- magnet and a magnetic field of 996 units produced parallel to their common axis. The currents were measured alternately with and without the field to correct for their decay during the experiments. The measurements indicate that both the positive and negative currents and also their algebraic sums are diminished by the magnetic field. According to the well- * Bronson, Phil, Mag. (6), xi, 806-812, 1906. 10 Duane—Lmission of Hlectricity. known formula for the radius # of the cylinder along the sur- face of which a 8-particle moves in a magnetie field /7, _—m U ne when w is the component velocity normal to the field, m the mass and e the charge of the particle. If any of the particles projected from one surface have velocities so small that /? is less than + the distance to the other surfaces, 0°22™™ they will not reach the other surface. Further, some par- ticles for which / is greater than this will be cut off also, va by the field. Assuming that the ratio < for the carriers of electricity is that of the electrons, namely 1:87 X10’, the velocity for which /7=0-022™ is . u= HR~=4x10° MW It follows that a considerable number of the electrons must have velocities near or less than 4 10°. | 17. The effect of increasing the electromotive force applied to the tube & was also studied, measurements being taken during a series alternately with 1°5 volts and voltages rang- ing up to 80 volts. It appears that both the positive and neg- ative currents are considerably increased by increasing the electric field, and that, up to 40 volts at least, their algebraic sum is decreased. These results can be explained on the assumption that the currents are carried by the #-particles shot off from the surfaces. An electric field between the con- ductors stops some of them. To calculate the velocity that 40 volts would stop we have Lmu'=40 x 10°e or Uu=—3'8X10° It follows, as before, that a considerable number of the elec- trons must be projected with component velocities to the sur- faces in the neighborhood or less than 4X10". 18. Combining the electric and magnetic effects we might calculate values for both < and wu, but owing to the complexity of the rays, and the difficulty of estimating just how many particles have velocities under a given limit, such estimates would not be worth much. The experiments show, however, that the order of magnitude of the effects are such as would be produced, if the currents were carried by §-particles, a considerable number of which had velocities normal to the surfaces in the neighborhood of 4X 10° or less. Duane—ELmission of Electricity. 11 19. Experiments were made to measure the ratio between _ the charge carried by the rays from the induced activity and the charge carried by the maximum number of ions that can be produced in air by the same induced activity when all the a rays are absorbed in the air. As the latter is several thou- sand times the former, it is better to measure the ratio in two steps. In the first step, with the electrode A in the tube JB, the ratio between the currents at atmospheric pressure and in a liquid air vacuum was measured, and in the second, witha smaller amount of activity, the ratio between the currents with the electrode in the tube, and in a larger condenser, was measured, both at atmospheric pressure. The product of the two ratios is the ratio required. ; 20. For the first step, the brass electrode having been made radio-active as before, it was placed in position in tube & and eurrents measured with the piezo-electric quartz at atmospheric pressure. Then the tube was rapidly exhausted by opening a stop-cock communicating with the reservoir containing carbon cooled to the temperature of liquid air, and the currents meas- ured again. ‘The electromotive force each time -was 2°2 volts Gi. e., 50 volts per cm.). The data appear in Table IV. Making the small corrections for the decay of the activity, we find for the ratio of the currents in air to that in vacuum, 61 for ¢= 13, and 69 for ¢= 64. Other experiments gave values TABLE LY. Time — Positive Current 10 52-5 } ty 11 51-4 Atmospheric pressure 13 0°82 14 0-78 I Liquid air vacuum 60 24:0 | Biel Pos! 62 3-1 f Atmospheric pressure 64 7 0322 Liquid air vacuum ranging from 66 to 70 for 60 to 70 minutes after the electrode had been removed from the emanation. Farlier than this the ratio is smaller, owing undoubtedly to the fact that there is then proportionately more radium B en the electrode. 21. To measure the ratio of the current in the tube & at atmospheric pressure and the saturation current when all the a rays produce ions in air, a large cylindrical condenser was constructed 19°8™ long and 17-9°" in diameter. = 3600 X 25 th part of the ionization current the a rays from its radio- 000 activity are capable of producing. Rutherford* has found that each a particle projected from radium produces 86,000 ions. Assuming that each a particle from radium OC produces a number of ions proportional to the length of its path and to the excess of its energy over the critical value, namely, 180,000 ions, it follows that for every a particle projected by radium C there are at least 20, and probably more than 50, electrons expelled from the active surface. 24. Conclusions. (a) A piece of metal made radio-active by immersion in radium emanation emits considerable quanti- ties of negative electricity ; and the rate of discharge decays with the time in such a way as to indicate that radium B when * Rutherford, Radio-activity, p. 454, Duane— Emission of Electricity. 13 changing into C discharges as much electricity as does radium C when changing into D. (b) A magnetic field parallel to the active surface stops part of the emission of electricity ; and an electric field normal to the surface also alters the rate of discharge, the magnitude of the electric and magnetic effects being about what would be expected if the charge was carried by electrons, a considerable portion of which had component velocities normal to the active ee sec (c) The total quantity of negative electricity emitted per second by an active brass surface is at least as large as the ere) part, and probably larger than the oF th part of the 9000 3600 ionization current that can be obtained from the activity, if the a rays from radium C are completely absorbed in the air pro- ducing thelr maximum ionization; 1. e., the number of ions produced by the a rays in air is less than 9,000 times, and probably less than 3,600 times the number of §-particles expelled from the wire during the same time. Assuming that each a particle from radium C can produce 180,000 ions, this means that for every a particle expelled from radium C, at least 20 and proba- | bly more than 50 electrons are emitted from an active brass surface. (qd) The number of ions prcduced by radium B in the air at atmospheric pressure close to (i. e. within 4 a mm. of) the active surface is comparable with the number of ions produced. in the same space by radium C. surface in the neighborhood of, or less than, 4x1 Radium Laboratory, University of Paris, 1908. 14 Prescott—Llvaite fram Shasta Co., California. Art. IL—Zlvaite from Shasta Co., California; by Bast. Prescorr, Stanford Unive ersity. Tue ilvaite described in this article occurs at Potter Creek, Shasta Co., California. The locality is well known to seolo- gists on the Pacific coast, for it was from the caves in the Jarboniferous limestone at this place that Dr. Merriam unearthed the Quaternary vertebrate remains,* and recently it has also attained some economic importance from the exploitation of the magnetite bodies that occur at the contact of this same limestone with an intrusion of diorite. It was during an examination of these ore deposits early in the pres- ent year that the writer noticed the presence of ilvaite. Lindgren+ has cited this mineral as a typical product of contact “metamorphism, and the occurrence and association as seen in Potter Creek are in accord with this view. There were two occurrences noted. On both sides of a six-inch dike cutting through the limestone, a half-inch band of pure mas- sive ilvaite was found, this, in places, sending out rough rec- tangular prisms into the limestone an inch or more in length. A tew feet away, further search was rewarded by a number of ilvaite crystals associated with a coating of eroded quartz crystals on hedenbergite, a more common contact mineral. The crystals are about 7 to 8™ in greatest dimension and are well formed, doubly terminated and symmetrical. The extremes in habit are shown in the figures, but even in the more elon- gated the prism zone is not as well developed as in the erys- tals from Elba. Although the crystals are bright and untarnished, the sig- nals were not distinct, and close measurements were 1mpossi- ble on account of vicinal faces and striations. The forms present are m-(110), s(120), 6(010), o 111), 7 (101), (890) (%) all of which are those more commonly developed in ‘ilvaite with the exception of the doubtful new form (890) found on two crystals, where it repiaces the prism m (110). The following measurements serve to identify the forms, the zone [8, s, m,| taken from one crystal, the zone [o, 7,| from a second, as the two were not found measurable on the same cr ystal. The most striking physical characteristics of the mineral as seen in this occurrence are the submetallic luster and greenish brown streak. The absence ‘of limonite as an alteration prod- uct is noticeable. The cleavage is not prominent and neither the specific gravity nor the hardness would distinguish it from * Sinclair, Cal. Univ. Publ. Am. Arch. and Eth., vol. ii, pp. 1-27, 1904. + Character and genesis of certain contact deposits. Trans. Amer. Inst. M. E., vol. xxxi, p. 227, 1901. Prescott—Ilvaite from Shasta Co., California. 15 Measured Theoretical en, S$). (010 7.120) 36° 40! BO. 52" MAS (l10A 120) 20S OS LO 27. mAmUt (1107110) 66 9°45 Gh ro? Ont, (EEL ROE) 20 7 20 7 oe OU CRE Tee) 40 28 40 15 (890A 120) Ge ss35* 16:6 minerals of similar appearance. Before the blowpipe it fuses readily with slight intumesence to a magnetic globule, yields a small amount of water at high temperature in a closed tube and is readily soluble in hydrochloric acid, giving a gelatinous residue upon evaporation. Moss, Hillebrand,t Hoffman,t Shasta Co., Owyhee Co., Vancouver Theo- Cal. Idaho Island, B. C. retical 3 eee 28°09 29°16 29°81 29°3 PO “32 “52 0°16 a he BO) 7 a 20°80 20°40 18°89 LOG Lie @ eae DOaoS 29°14 32°50 39°2 mG bys 3°24 5°15 2°22 eee. CAO. es 15°89 13°02 13°82 Ras 7 BESO eS ef ire: "15 30 Dae ON e42 ors i ie 08 in aie ven? Be Ore 13 ca soc See 3 UU A aparece 1°62 2°79 «21°62 2°2 BGA (2s ee 100°20 100°41 99°32 100°00 * This agreement is accidental, as the recorded value is the average of many readings on several faces. + Bulletin U. S. G. S. No. 207, p. 45. ¢ Vol. 5, Annual Rep. Geol. Survey of Canada, 1889-90. 16 Prescott—Livaite from Shasta Co., California. The material for the analysis given above was sorted from the massive ilvaite and all por tions that might contain impuri- ties or inclusions were carefully excluded. It was then sub- mitted to H. R. Moss, whose results are tabulated (p. 15) with the two analyses available for America and with the theoreti- eal for HOaFe,™ Fe™ Si,O,, given by Dana in the sixth edi- tion of the Sy stem of Miner alogy. The material was air-dried at 95° C., and the total water obtained by a modification of Pentield’s tube method. The only deviation from the usual course of analysis was in the determination of the manganese by triple precipitation as MnO, with bromine, in the filtrate from a basic acetate separa- tion, and the subsequent precipitation of the chromium with ammonia. Thanks are due Dr. A. F. Rogers, Stanford University, for advice and assistance in the preparation of this note. R. A. Daly—Mechanies of Igneous Intrusion. 17 Art. Wl—TZhe Mechanics of Lgneous Intrusion.* (Third Paper;) by Reeryatp A. Daty, Massachusetts Institute of Technology, Boston. Introduction. Hypothesis of magmatic stoping. Field relations of the typical batholith. Contact-shattering. Relative densities of magma and xenolith. Sinking of the shattered blocks. Problem of the cover. Supply of the necessary heat ; magmatic superheat and its causes. Capacity of superheated, plutonic magma for melting and dissolving xenoliths. Objection founded on rarity of evidences of assimilation at observed wall-rocks. Abyssal assimilation. Existence of basal stocks and batholiths. Differentiation of the syntectic magma. Origin of granite; the petrogenic cycle. Origin of magmatic waters and gases. Conclusion. Introduction.—In the April and August numbers of this Journal in the year 1903, the writer published papers outlin- ing the hypothesis of magmatic stopmg as explanatory of the rise of batholithic magmas in the earth’s crust. The hypothesis had taken form in his mind after some ten years of perplexity as to the mode of intrusion which has actually characterized granite bodies. In Vermont, New Hampshire, British Col- umbia and other regions he had met with this urgent and important field-problem. Everywhere the facts derived from field observations were, in principle, the same; the method of intrusion seemed, for each batholith or stock, to be the same. Since the writing of the two papers the writer has studied in some detail a dozen other large batholiths and as many typ- ical stocks occurring on the southern boundary of British Col- ambia. For all of these also the stoping hypothesis appears to afford the truest explanation of the mode of intrusion. Quite independently Barrell arrived at a similar hypoth- esis, as he attacked, in 1901, the problem of the ‘‘ Marysville bathohth” in Montana. Unfortunately his monograph was delayed im publication until 1907, so that it is only quite recently that geologists have had the benefit of this brilliant and thorough study of intrusive mechanism.t Barlow and. Cole- man have noted their belief in the efficiency of stoping as an intrusive process. At the other side of the world, Andrews has described the great intrusive masses of New South Wales, * Published by permission of the Commissioner for Canada, International Boundary Surveys. +U. S. Geol. Surv., Prof. Paper No. 57, 1907. +A. E. Barlow, Ann. Rep. Geol. Surv. of Canada, xiv, Part H, p. 79, 1904; A. P. Coleman, Jour. of Geol., xv, p. 773, 1907. Am. Jour. Sct.—FourtaH Series, Vout. XXVI, No. 151.—Junry, 1908. 2 18 RR. A. Daly—Mechanics of Igneous Intrusion. and most forcibly shows the value of the stoping hypothesis and of its implied principles in explaining the rocks and field- relations in that state.* Notwithstanding the:support given the hypothesis by the work of these and other observers, the main conception has not met with favor from many working geologists.t A num- ber of objections have been raised, most of which were dis- cussed in the first two papers of this series. Within the last five years an unusually large amount of experimental data has been added to the confessedly meager store of known facts concerning the physies of rocks and rock-melts. These labor- atory results, when fairly interpreted, seem to the writer to dispose of most of the objections. Other objections fall away as soon as they are confronted with the indisputable, long- known facts concerning rocks and igneous magmas. A third class of the objections are more stubborn and still remain among the frank difficulties of the stoping hypothesis. It is, however, the writer’s belief that these difficulties are small when compared to those adhering to the older theories of batholithic intrusion. In this third paper some of the more significant, newer con- tributions of the experimental laboratory to the matter at issue will be noted and discussed. In the ight of the whole body of fact as understood by the writer, he will attempt to make clear the reasons why the various criticisms against the stop- ing hypothesis do not seem fatal to its acceptance. Finally, a new statement of certain important corollaries and tests of the hypothesis will be offered. In their discussion a certain amount of speculation seems not only warranted but necessary. It is obvious that the basis of any theory of the igneous rocks must, in part, consist of speculative assumptions; for every fruitful theory must deal with the earth’s invisible interior. Neither petrology nor geology can afford to leave the problem of the earth’s interior “to the poets.” The advances of mod- ern chemistry have largely been made possible through con- structive speculation as to the nature of molecule and atom; yet molecule and atom are as inaccessible as the core of the earth. In the nature of the case we can never hope to arrive at the final explanation of igneous-rock bodies without building and testing hypotheses of materials and processes in and under the earth’s p ernel. ” Not only petrology but, in marked degree, mining geology is awaiting a stable theory of batho- lithie intrusion, since upon it must. lar gely depend sound pet- - rogenic and minerogenic theory. *E. C. Andrews, Records, Geol. Surv., N. S. Wales, vii, Pt. 4, 1904, and viii, Pt. 1, 1905. + Cf. Science, xxv, p. 620, 1907. R.A. Daly—Mechanies of Igneous Intrusion. no Like the first and second papers, this one does not present a complete discussion of the different topics. On another occasion the writer may publish a fuller statement of the favored solution of the complex problem. Hypothesis of magmatic stopig.—The essential points are the following: 1. Each acid, batholithic magma has reached its present position in the earth’s crust largely through the successive engulfment of suites of blocks broken out of the roof and walls of the batholith. 2. The blocks ( xenoliths) are completely immersed in the magma, partly through the confluence of apophyses which have been injected on joints and other planes of weakness in the country-rock; more often the blocks represent the effect of shattering, due to the obviously unequal heating of the solid rock at magmatic contacts. 3. The sunken blocks must be dissolved in the depths of the original fluid, magmatic body, with the formation of a “ syntectic,”* secondary magma. 4. The visible rock of each granite batholith or stock has resulted from the differentiation of a syntectic magma. In applying the hypothesis to the explanation of actual field-occurrences other general considerations seem necessary. Stoping and abyssal assimilation on the batholithic scale are begun by a primary basaltic magma. This magma carries the heat required for the double action.t The source of the magma is to be found in a general basaltic substratum beneath the earth’s solid crust. The crust is considered as composed of two shells. The lower shell is capable of injection by huge masses from the substratum, which retains open com- munication with the injected bodies. The latter are regarded as then stoping their way up into the overlying shell, in which the resulting derivatives of the syntectic magma are the visible batholithie granites and allied rocks. These subsidiary elements of the problem here to be dis- eussed have been described in the first intrusion paper and, more fully, in a later communication on “ Abyssal Igneous Injection.”{ No one of these additional conceptions is essen- *This very convenient name for a magma rendered compound by assimi- lation or by the mixture of melts, has been proposed by F. Loewinson— Lessing, Comptes Rendus, 7° session, Congres géol. internat. St. Petersburg, 1899, p. 375. + Whether the substratum is actually or only potentially fluid is not a vital question in this connection. T. J. J. See, as a result of his calcula- tions, holds that the earth’s interior may be fluid. He explains the observed rigidity of the planet as due not to its being a true solid but to the direct influence of gravity, which binds the earth-shells so effectively that bodily tides are almost wholly prevented. In any case rigidity and solidity are not synonymous terms. Cf. T. J. J. See, Astron. Nachrichten, v. clxxi, p. 378, 1906. ¢ This Journal, vol. xxii, 1906, p. 195. 20 RR. A. Daly—Mechanics of Igneous Intrusion. tial to the idea of stoping per se. All of them may prove incorrect without invalidating the stoping hypothesis in its main feature. Combining them and the idea of stoping, the writer has constructed a general working hypothesis for the — origin of the igneous rocks. It seems, therefore, expedient in the present paper to discuss the pr oblem in its lar ger aspect. Field Relations of the typical batholith.—& principal faet on which the stoping hypothesis is based has been amply illus- trated in the published descriptions of granite stocks and. batholiths. Most, if not all, of these bodies in their accessi- ble portions have replaced nearly equivalent volumes of the respective country-rocks. They are generally cross-cutting bodies. Their roofs are rough domes or arches, from which large masses of the invaded rocks are sometimes pendant into the crystallized granite. In each of many cases erosion has destroyed much of the roof, and the roof-pendants, still pre- serving: the regional strike of their structure planes, are to-day exposed in section at the erosion-surface. Between the pend- ants and between the main walls of a large batholith, hun- dreds of cubic kilometers of country-rock formations are plainly missing; their place has just as plainly been taken by the granite. A second principal fact is that, so far as granite batholiths and stocks are known, each of these bodies shows a eross- section enlarging with depth.* No one of them has yet exhibited a floor composed of older formations. In relation to visible country-rocks, all of them may be classed as subjacent, rather than as injected, bodies. In relation to the wall-rocks ten or more kilometers below the earth’s surface, each batho- lith may have been truly injected asa kind of gigantic dike, but of this there is no direct proof. The actual observations in the field show unequivocally that the batholithic magmas have worked their way up by replacing and absorbing the country- rocks through the last few kilometers of ascent. Bathe are not laccoliths. A third generally observed fact is worthy of special ater tion. Where erosion has been profound the ground-plan see- tion of the typical stock or batholith is seen to be elliptical and the profile-sections, as already noted, show that the upper- contact surface of the intrusive is dome- shaped. Both in ground-plan and in vertical sections the contact-surface is relatively smooth. Apophysal offshoots do interrupt the wall- rock, but the main-contact lines as mapped on ordinary geolog- ical maps are characteristically flowing lines. Large-scale, *See the numerous sections of stocks and batholiths in Lepsius’ ‘‘ Geo- logie von Deutschland” ; also Barrell’s monograph cited, and the writer’s paper on the Okanagan Composite Batholith, Bull. Geol. Soc. America, xvii, p. 3380, 1906. R. A. Daly— Mechanics of Igneous Intrusion. 21 angular projections of country-rock into a well uncovered batholith are comparatively rare. Such smoothness of main- contact surfaces is that which is to be expected on the stoping hypothesis. A projection of country-rock would suffer spe- cially intense shattering by the magma, which would thus tend to destroy the projection and smoothen the wall of contact. The case is analogous to the familiar exfoliation on sculptured stone in great city fires; architrave, sill, abacus and plinth lose their corners, ornaments in high relief are rifted off, and flutings are effaced. Bowlders of disintegration through weathering furnish other analogies. In detail of form as in the larger field-relations of the typi- eal stock and batholith, therefore, we seem to have cumulative evidence in favor of the theory of replacement and especially in favor of the hypothesis of mechanical replacement. On the other hand, the more intimate becomes our knowledge of these field- relations, the more improbable the “laccolithie theory” becomes. Neither smooth, flowing contact-surfaces against a heterogeneous terrane, nor a general elliptical ground- plan, nor an invariable downward enlargement are expected to characterize a batholith if it is simply a huge laccolith. These summary statements are founded on the writer’s field- experience, and on a tolerably wide study of the geological literature relating to granitic intrusions. ‘The essential idea of replacement rather than displacement is far from new; it has been a lasting merit in the able work of Barrois, Michel Lévy, Lacroix and others, that they have persistently held to this fundamental fact of. field-occurrence. Yet there are to-day many working geologists who just as persistently refuse to recognize the fact of the field. The chief reason for this refusal has undoubtedly been that the replacement of the country-rocks has, until recently, been attributed to their pro- gressive solution on the main. contacts—in other words, to marginal assimilation. The patent difficulties of this one view have prevented many, perhaps most, geologists from subserib- ing to the conclusions of their French colleagues. The proved insufficiency of the marginal-assimilation hy pothesis has thus discouraged belief in that kind of replacement, but it by no means alters the fact of magmatic replacement. On the other hand, this fact will stand, no matter what theories of intrusion may. prevail. So far as recorded, the stoping hypothesis is the only one which recognizes the progressive assimilation of country-rocks as the magma rises in the crust, and, at the same time, explains the common lack of chemical sympathy between granites and their respective wall-rocks. By this hypothesis the preparation of the upper and visible part of the mag- matic chamber is largely a mechanical process, working along 22 Rk. A. Daly—Mechanics of Igneous Intrusion. main contacts; the solution of the engulfed blocks is effected far down in the depths of the magma—by abyssal assimila- tion. The resulting syntectic magma may thus be in strong chemical contrast with the adjacent wallLrock at any one level. Marginal assimilation is not excluded but is considered as an accessory and subordinate phase in the act of replacement. Contact-shattering.—It has been objected. that rocks are good conductors of heat and that, therefore, strong temperature differences with resulting rending strains are not to be expected in the shell of country-rock immediately surrounding a batho- lithic magma. This objection has been recently made by an expert physicist now specially engaged on petrological problems, and evidently needs consideration.* The following table of coefficients of absolute conductivity seems, however, to show, on the contrary, that rock-matter is far from being ranked as a good conductor. The table has peen made by compiling the values noted in the Landolt- Bornstein’s Physikalisch-chemische Tabellen (1905 edition) and in Winkelmann’s Handbuch der Physik. The values for the rocks are of the order expected in view of the familiar proofs of the extremely slow cooling of lava-flows.t k Silver, “aboutmec tive ee sane 1:0000 » Coppers aera as man Omen "9480 Dead 22s Base es ee ae eens SOG (UATE oe ee eee es 0158 Marble Sat tos ay niece ete "00817 Granite messi a 222. (00757 — 008% Gnesi een Ite Perot tase aa °000578—:00817 DPandstone esha he Uh eee "00304 —-00814 Basalt: 2215 ee pale sk en ae gee °00673 DY CULC ie sees he einer ieee "00442 GlgSs) 3. Pelee ona Wee ener ‘00108 -—-00227 Water, abouteee 2. hee 00130 Papert) 222 ee eee "00031 Flannel)? 23a o0e ec enon "00023 Sie ee ee eee "00022 Cork OD a eee ‘00013 Feathers 32.2 2 eee "0000574 * Of. A. L. Day, Science, xxv, p. 620, 1907. +The steepness of the possible temperature gradient in the wall-rock is shown by the fact that, a few days after lava ceases flowing, one can walk on its crust, although the lava just below is at red heat (700°-950° C.) or is yet hotter. For many hours or for several days the gradient at the surface may equal or surpass 500° C. per foot. In the manufacture of calcium-carbide a mixture of limestone and coke is submitted to the action of a powerful electric arc. At the end of a fur- nace-run (about fourteen hours in the plant at Ottawa, Canada) the flow of heat is nearly steady and the temperature gradient in the furnace is about 3000° C. per foot. In this case the diffusivity of the limestone-coke mixture in the interior of the thoroughly heated furnace must ue well below 60 in the Kelvin system of units. R.A. Daly— Mechanics of L[gneous Intrusion. 28 Weber has found that # for gneiss at 0° C. is 0:000578 and at 100° C. 0:000416, showing a very great lowering with increase of temperature.* In fact, through the interval 0°-100° C., & seems to vary about inversely as the absolute temperature.t If this law should hold to 1100° C. the conductivity of average rock at 1100° falls to about 0:001—nearly the value for water, which is famous as a poor conductor. In the present connection the thermal diffusivity («) of rock, rather than its conductivity, is of first importance. If s = specific heat and d = density, we have seas Toe gad For rock at room temperature (20° C.) Kelvin assumed 400 as the value of « when the unit of length is a foot, the unit of time a year, and the unit of temperature one degree Fahrenheit. This value is close to that which represents the average of the determinations made for different rocks at room temperatures, during the years since Kelvin wrote his famous essay.t If « be assumed as 400 at all temperatures up to 1300°C., it is possible to calculate the temperature gradient in the wall- rock of a molten batholith at the end of specified periods of time. For practical purposes the surface of contact may be regarded as infinite; let it further be considered as plane. Under these conditions the following Fourier equation furnishes the datum for calculating the temperature at a point x feet from the contact at the end of ¢ years.§ Inthe eguation =the temperature of the magma; ¢ = the temperature of the wall- rock assumed as initially uniform ; and vw = the required tem- perature. We have :— pa bele 5) = oe” ag — —_ ap. a x : For values of SW which are less than 2°6 the value of the 2 V/ kt integral can be readily found from the table of the probability integral which appears in standard text-books on the Method of * Values taken from Landolt-Bérnstein Phys.-Chemische Tabellen. Forbes and Hall have proved analogous relations for iron and for magnesium oxide ; ef. J. D. Forbes, Trans. Roy. Soc. Edinburgh, xxiv, p. 105, 1867, and E. H. Hall and others, Proc. Amer. Acad. Arts and Sciences, xlii, p. 597, 1907. bee eines Tait, Recent Advances in Physical Science, 2d ed., London, Pp. <(V, 1570. { Trans. Roy. Soc. Edinburgh, 1882. she W. E. Byerly’s Elementary Treatise on Fourier’s Series, Boston, 1893, p. 86. 24 Lf. A. Daly—Mechanies of Igneous Intrusion. : a Least Squares. For higher values of "e the value of the & K integral can, in many cases, be computed by developing it into a series. Kelvin’s value for « is peculiarly favorable for such computation and the corresponding units have been used by the writer in the calculations. Let 6 = 2200° F. (about 1200° C.); ¢c = 400° F. (about 200° C.);¢ = 1, 4, 16, and 100 years; and let « havethe different values shown in the left-hand column of the following table (1) The corresponding temperatures are shown in the other columns. TaBLE I.—Showing values of wu when « = 400 and er SSS SSS Seas tt rages a == teaver. t = 4 years. t = 16 years. t = 100 years - Ope 2007 JE 9200° KF. 2200 EY, 2200° F. OES 703 1947 DOms 20. 1263 1703 1947 4)! 683 1263 1703 80! 408 5 683 1263 100’ ca.400 5a] 1078 ; 1703 NGOs 400 ; 408°5 683 200’ 400 ca. 400 537 1263 320’ 400 400 408.5 400' 400 400 ca.400 683 The table shows that, at the end of the first year, the temper- ature of the rock is but shghtly affected by the magmatic heat at a point 80 feet from the contact, and that the temperature gradient for the 80-foot shell then averages nearly 23° F. per foot. At the end of four years the temperature is but shghtly affected at a point 160 feet from the contact and the tempera- ture-gradient is about 11° I. per foot. But « cannot be nearly so great as 400 in the case before us. We have seen that # decreases rapidly with rise of tempera- ture in rock. The experiments of Weber, Bartoli, Roberts- Austen and Ricker, and Barus show that the specific heat of rock averages about ‘180 at 20° C. and increases regularly with rise of temperature, so that at 1100° C. the specific heat averages about -280.* It follows that thermal diffusivity in rock decreases with rising temperature even faster than the conductivity decreases. At 1100° C., « may, indeed, be only "180 293 1°000 (oss A393 * Sane =) sa | or less than one-seventh, of the diffusivity at 20°C. For rock heated to 1000° or 1200° C. « is, thus, probably not much more than 60 in the Kelvin system of units. * For references see J. H. L. Vogt, Christiania Videnskabs-Selskabets Skrifter, I. math.-naturv. Klasse, No. 1, p. 40, 1904. , R. A. Daly—Mechanies of Igneous Intrusion. 25 It seems safe to assume, first, that the diffusivity of the gradually heated wall-rock may vary from 275 or less to 100 or 150; secondly, that the average diffusivity of an 80-foot shell heated during the first year by adjacent molten magma, will be no greater than 200. If « be regarded as averaging 200 for all periods greater than one year, the four columns pins values of win the table will serve if ¢ is , respectively, 2, 8, 32 and 200 years. As a result of somewhat rigorous calculation, then, it appears certain that the heating of wall-rock by plutonic magma must progress with great slowness and that the result- ing temperature gradient in the shell adjoining the molten magma must be steep for many years after the original estab- lishment of the contact.* Further, Less has proved that rocks have highly variable coefficients of conductivity, some species possessing coefficients twice as high as those of other species.t It is also well known that bedded or schistose rocks conduct heat along and across their structure-planes at quite different rates. Where, there- fore, the wall-rocks about a batholithic mass are heterogeneous, the heat-conduction is variable and expansional stresses must ensue. A rough calculation of the enormous stresses involved in all these processes of differential heating was published in the second paper of this series, where .also an account is given of the practical use which has been made of such stresses in primitive quarrying.t Every great city conflagration leaves manifold evidences of the shattering effects of “the one-sided heating of a rock-mass—in columns, sills, and cornices of oranite or sandstone. There seems, therefore, to be a sheer necessity for believing in contact-shattering through differential heating and expan- sion in the thin shell of a country-rock which encloses a large body of molten magma. The evidence for the shattering is often exceedingly full and clear in the field. The broad or narrow belts of xenoliths so often found just inside the main contacts of batholiths are very hard to explain if those batho- liths are due to laccolithic injection. The blocks are charac- teristically angular; they are generally not arranged with their longer axes parallel, as if they had been pulled off from the * By using the same Fourier equation it is not difficult to show that the loss of thermal energy which a magma suffers by conduction into the country-rock is relatively small, even after the lapse of two or three hundred thousand years. The long duration of the magmatic period in a slightly superheated. plutonic mass of large size becomes easily understood. + Phil. Trans., vol. clxxxiii A, p. 481, 1892. t This Journal, xvi, p. 112, 1903; ef. Ann. Rep. State Geologist of New Jersey, 1906, p. 17. 26 R. A. Daly—Mechanies of Igneous Intrusion. walls by the friction of the moving magma. On the lacco- lithic theory one would expect many of the xenoliths to form elongated smears in the granite rock. This is indeed occa- sionally seen but most exceptionally; as a rule the xenoliths have just that irregularity of form and arrangement which they should have if they had been shattered off by the hot magma just before its final consolidation. Throughout its long, earlier history the magma must, in every case, have had a much more effective shattering power. It may be noted that the shattering of crystals and rock- fragments, when immersed in silicate melts, has often been observed.* The strains are, in such cases, necessarily of a lower order than those developed on the wall of a batholith where, therefore, shattering is even more certainly brought about. | Relative densities of magma and xenolith.—In his first intrusion-paper, the writer published the results of his attempt to calculate the possible specific gravities of the chief types of molten magmas under plutonic conditions. The calculations were based on. Barus’s well-known fusion experiments on diabase. The specimen investigated had a specific gravity of 3°0178; when fused to a glass and cooled to 20° C., a specific gravity of 2°717. He further states that the glasst showed an expansion of 3°9 per cent in “melting” and, as glass, expanded 0:000025 in volume for a temperature rise of 1° C. through the interval 0°-1000° C. and 0-000047 in volume for 1° C. through the interval 1100°-1500°. The “melting” expansion (solidification-contraction) and the varying rate of expansion (or contraction) above and below 1000° ©. seem to show that some crystallization of the melt took place during the experiment. Such crystallization was inevitable under the conditions of the experiment, in which the cooling lasted several hours. , Barus’s curves do not, therefore, show directly the volume changes suffered by pure diabase glass in passing from the molten isotropic state to the rigid isotropic state at room temperature. Excluding the “solidification” contrac- tion, the glass loses but 3°5 per cent of its volume in passing from the molten state at 1400° C. to room temperature; the loss of volume through the same temperature interval was calculated in the first paper as about 8 per cent. Barus found that the net decrease in specific gravity in passing from rock at 20 C. to glass at 20° C. was 10 per cent. For his diabase specimen, therefore, the decrease of specific gravity in passing * Of. C. Doelter and E. Hussak, Neues Jahrb. fiir Min. etc., 1884, p. 18; A. Becker, Zeitschr. d. d. geol. Ges., xxxiii, p. 62, 1881. +‘*Throughout this paper the molten rock solidifies into an obsidian.” C. Barus in Bull. 103, U. S. Geol. Surv., p. 26, 1893. R. A. Daly—Mechanics of Lgneous Intrusion. 27 from 20° C. to molten condition at 1200° C. is possibly only about 13 per cent, instead of about 16 per cent, as noted in the first paper.* Quite recently J. A. Douglas has made a number of very careful measurements of the densities of typical igneous rocks and of their respective glasses, all specific gravities being taken at room temperatures.t . Douglas’s method is reliable and his results accordant. For gabbro he found the decrease of specific gravity, in passing from rock to glass, to be 5:07 per cent. Delesse had found the decrease to be 11°46 per cent, as the average of measurements of two specimens from different localities Barus’s determination, 10 per cent, is intermediate between the two. It seems probable, therefore, that a decrease of 6 per cent in specific gravity (rock to glass at 20° C.) is close to the minimum for the average gabbroid rock, and it is possible that Barus’s 10 per cent decrease is too high for average gabbro. For present purposes it is safer to use the minimum value of 6 per cent. Similar minima for diorite (6 per cent), quartz diorite and tonalite (7 per cent), syenite (8 per cent) and granite (9 per cent) have been estimated from the numerous measurements of Delesse, Cossa and Douglas. Each of these rocks certainly TABLE IT. ; Specific gravity of Specific gravity of same rock Crystalline rock at when molten at (Are a aS TT 20°C. 1000°C. 1300°C. 000°C. 1100°C. 1200°C. 1300°C. E2780" 25737, 2271 2°57 2°56 2°54 2°53 Sappre (| 2-90 2°83" <2°80.;>- 2°66. 2:65. 2-64-.. -2°63 and + 3°00 2:92 2°90 2°75 2°74 2-73 2°72 diorite | 3°10 3°02 3°00 2°84 2°83 2°81 2°80) | 3°20 3°12 3°10 2°94 2°92 2°91 2°91 Quartz-dio- rite and 2°70 2°63 - 2°61 2°46 2°45 2°44 2°43 tonalite 2-80). 2579 2 2561 2°54 2°53 2°51 2°51 ( 260° (2-3 4- 2:52 2°33 2°32 2°31 Trot Syenite {2°70 2°63 2°61 2°42 2°41 2°40 2°40 (2-80 2°73 2°71 2°52 9°51 «950-950 Granite (2°60 9:54 2°52 Bele ers A229. 2:29 and 2°10 2°63. 2°61 2°40 2°39 2°39 2°38 gneiss ZOO G2 fon 2h) 2°49 2°48 2°47 2°47 * Bischof, in 1841, found that basalt expanded 7 per cent in passing to a glass at room temperature, and 10:4 per cent in becoming molten. (Quoted from Zirkel’s Lehrbuch der Petrographie, 2d ed., 1893, vol. i, p. 683.) + Quart. Jour. Geol. Soc., xiii, p. 145, 1907. 28 R.A. Daly—Mechanics of Iyneous Intrusion. expands: in the interval 20°-1300° C. as much as 0°000025 vol- ume per degree Centigrade. (Barus and Reade—see first paper). This average may safely be employed as a means of determin- ing the minimum decrease of density which each rock-type undergoes in passing into the molten condition. On this basis the writer has constructed the preceding table (II), which shows the changes of specific gravity at convenient temperature intervals. | | Table III shows the changes in specific gravity undergone by blocks of stratified and schistose rocks (common country- — rocks about batholiths), as these blocks assume the tempera- ture (1300° C.) of molten magma in which they are immersed. TABLE IIT. Range of sp. gt. Range of sp. gr. at 20°C. at 1300°C. (solid) REACISB Geos F2' de tyh eee 2°60—2°80 2°5 2-29-71 Mica schists | 32 Whe eee Ogos 2°67-3:°00 Sandstone ks yee we aad 2°20-2°75 2°138-2°67 Aroulites 2504. eae e 2°40—2°80 - 2°32-2°71 Pamestone 22% Sac eres 2°65-2'80 2°57-2°71 It appears from these tables that nearly all xenoliths must sink in any molten granite or syenite; most xenoliths must sink in molten quartz-diorite, tonalite or acid gabbro. Many xenoliths might float on basic gabbro but the heavier schists and gneisses must sink in even very dense gabbro magmas at 1300° C. : Giving, then, the highest permissible values to the specific gravities of magmas, it is still true that blocks, such as are. shattered from the wall or roof of a batholith, must sink when immersed in most magmas at atmospheric pressure. As shown in the first intrusion paper, the blocks would likewise sink, though the magma enveloping them hes at depths of ten or fifteen kilometers below the earth’s surface. Sinking of the shattered blocks —\t has been objected to the stoping hypothesis that the viscosity of granitic magmas is too great to allow of the sinking of blocks even much denser than those magmas.* This objection has, however, never been sustained by definite experimental or field proofs. The xeno- liths visible along batholithie contacts have assuredly not sunk far from their former positions in wall or roof and the reason for this must be sought in the high viscosity of the magma. High viscosity is an essential attribute of a nearly frozen magma. The phenomena of fractional crystallization and of magmatic differentiation unquestionably show that each # Of. W. Cross, G. F. Becker, and A. L. Day, Science, xxv, p. 620, 1907, PR. A. Daly—Mechanies of Igneous Intrusion. 29 plutonic magma must pass through a long period of mobility. The most viscous of granitic magmas, the rhyolitic, issues at the earth’s surface with such fluidity that the rhyolite often covers many square miles with a single thin sheet. The absolute viscosity of the Yellowstone Park rhyolites must have been of a low order when many of these persistent flows were erupted.* Even granting that the kinetic viscosity of a plutonic magia is ; thousands of times that of water, 1t seems inevitable that it could not support xenoliths more dense than itself. In a few days or weeks stones will sink through, and corks will rise through, a mass of pitch, the viscosity of which is more than a million of millions of times that of water.t Ladenburg has lately shown that small steel spheres will, in a few minutes, sink through twenty centimeters of Venetian turpentine, a substance 100,000 times as viscous as water.{ Ladenburg’s experiments have verified the generally accepted equation expressing the rate of sinking of a sphere in a strongly viscous fluid : 2gr (d— a’) 9 v were # = the velocity of the sphere when the motion is steady ; g = the acceleration of gravity ; d = the density of the sphere ; d@ =the density of the fluid; 7 =the radius of the sphere; and v = the viscosity of the fluid. § The equation shows that the velocity of sinking varies directly as the square of the radius of the sphere. This fact may be correlated with the observation so often to be made on granite contacts, that large xenoliths are rare. This apparently means that, at the end of the shatter-period, the viscosity is truly so high as to allow of the smaller blocks bemg trapped at high levels in the freezing magma, while the large blocks, with greater velocity, shall have sunk into the depths. C= *See Atlas accompanying Monograph 382 of the U. S. Geol. Survey. King described the great rhyolite flows of Nevada as bearing ‘‘ abundant evidence of true fluidity at the period of ejection.” U.S. Geol. Explor. 40th Parallel. Sys. Geol. 1878, p. 616. Doelter has studied the behavior of a large number of crystalline rocks and minerals during fusion. His results show that the temperature-inter- val between the stage of softening and that of notable fluidity averages, for the basic rocks, about 50° C., and for the acid rocks, about 90°C. (Tscher. Min. u. Petrogr. Mitth. xx, p. 210, 1901.) The interval is not great and it certainly seems unsafe to deny that even the most viscous, because cooled, lavas were fluid in depth. +Jamin et Bouty, Cours de Physique, tome I, 2e fascicule, Paris 1888, p. 135; ef. Daniell’s Text-book of the Principles of Physics, 2d ed., London, 1885, p. 211. +t Annalen der Physik, xxii, p. 287, 1907. § Poynting and Thomson, Text-book of Physics, Properties of Matter. London, p. 222. 1902. 30 R.A. Daly—Mechanics of Igneous Intrusion. Doelter estimates that the pressure of from 7500 to 11,000 meters of rocks increases magmatic viscosity no more than 20 to 80 per cent.* If the increment be anywhere near this value we may be certain that the viscosity of superheated, plutonic magma is relatively low. Becker has calculated that the viscosity, of a Hawaiian basaltic flow, not one of the most fluid, was, at eruption, about fy times that of water. The more fluid rhyolite flows may have viscosity a thousand times greater than that of water. The corresponding viscosities of the same magmas when ten kilometers underground may, then, be possibly no more than from sixty to fifteen hundred times that of water. One must conclude that a xenolith, even very slightly denser than such a plutonic magma, must sink into it. Since such magmas necessarily cool with extreme slowness, there is evidently good ground for believing that an enormous amount of solid rock could be engulfed before practical rigidity is established. The average xenolith must sink in a less dense magma with the viscosity of piteh—yet how much more rapidly in magma possessing the low viscosity which is postulated in any of the ruling theories of plutonic- rock genesis ! Problem of the cover.—The stopmg hypothesis presents an obvious principal difficulty ; it refers to the apparent danger of the foundering of the roofs covering the larger batholiths. Under plutonic conditions (at depths of from three to ten kilometers) the average molten granite would have a specific gravity no higher than 2°40. The average rock of its roof has a specific gravity of about 2°70. If, then, through orogenic Be ee a large mass of the roof- rock became once wholly immersed in the granite, it would not only founder itself but through subsequent buckling the whole roof might collapse and founder in sections. Such a catastrophe has almost cer- tainly not happened in the case of any Paleozoic or later batholithic intrusion. This difficulty has been emphasized by Barrell, who has justly given it a prominent place in his monogr ‘aph. + Lawson speaks of batholiths 100 miles in diam- eter and also finds the necessity of explaming their roof- support as a principal ground of unfavorable criticism.¢ The present writer cannot claim to have solved this problem, but he does not find it to form a fatal objection to the hypo- thesis. In the first place, it seems clear that all the other hypotheses of granitic intrusion are facing the same dilemma. All of them expressly or tacitly postulate some degree of fluidity in each granitic mass as it either replaces or displaces * Physikalisch-chemische Mineralogie, Leipzig, p. 110, 1905. TOR gcitL = WAPNULUE CT z SUDAMAQD “CT aeee ee wood “CG oor e= ehyos “¢ OUR NT \ 64 F. B. Loomis—Rhinocerotide of the Lower Miocene. teeth are not so crowded as in the preceding species. On the premolars the cingulum extends around the front, inner, and rear faces. These premolars are very simple, showing no trace of a crista or crochet. The cingulum on the molars is inter- rupted on the inner face opposite both the protocone and Lili Fie. 15. Aceratherium egrerius C.; the premolar and molar series, one-half nat. size. hvpocone. A crochet is moderately developed, especially on the second and third molars. This and the preceding species show much resemblance to A. occidentale of the Oligocene. Measurements. Total length of the premolar-molar series____-_..-.._..-- 204™™ Length of second molari0? 2022 2 Width of second molara. 222.00 520122 Amherst College, Amherst, Mass. T. D, A. Cockerell—Descriptions of Tertiary Plants. 65 Art. V.—Descriptions of Tertiary Plants; by T. D. A. CocKERELL. 1. A Frog’s-bit from Florissant, Colorado, Limnobium obliteratum sp. nov. Figure la. Blade of leaf nearly circular, with a reniform base; obseurely about 10-nerved, these nerves simple, and exactly as in the living Z. spongia. Length from insertion of petiole to apex about 31™™, from apex of basal lobes to apex 35; breadth 36. Margin perfectly entire, somewhat thickened. The apex is broadly rounded, with no tendency to pointing, such as is seen in the living if spongia. This is the first fossil species of this group from America ; but a similar plant occurs in the Miocene of Europe, and has been named by Heer Hydrocharis orbiculata ; “ distinguished by its circular leaves, which no doubt floated on the surface of the water like the leaves of water-liles.” /Zab.—F lorissant, in the Miocene shales, Station 14 (W. P. Cockerell, 1907). Peabody Museum, Yale, Cat. No. 1001. 2. Two Maples from Florissant. Acer perditum sp. nov. Figure 106. Leaf with the blade deeply trilobed, the sinuses extending about half way to the base; the lobes broad, poimted, the mid- dle one broadest about 9™™" from its base, thence slightly con- tracted basally ; prominent nervures three, but also a smaller one on each side; margin obscurely and rather remotely den- tate. As preserved, the blade is yellowish, while the petiole and principal nerves are dark brown. Length of blade, about 44mm, breadth abont 34; breadth of middle lobe in middle, 1Q"™, Atases Oe length of middle lobe about 244"" ; of lateral lobes 20 or less. There is some resemblance to such fossil species as Aralia notata Ward, but the venation is very different, and agrees with that of Acer. Acer narbonnense Saporta, from the Oligocene, is a somewhat similar leaf, differ- ing however in its margin, and in the much less broadened base. The Chinese A. wilsoni Rehder is similar in general form, but its lobes are far more attenuate, its margins are almost completely entire, and the extra basal (external) veins are absent. There is also resemblance to A. sacchurwm rugelii (Wesm.) Rehd., which, according to Sargent, is a leaf-form sometimes appearing on the upper branches of trees which have on their lower branches the leaves of typical A. saccha- rum. In the southern states, however, the rugeliz form is Am. Jour. Sct.—Fourts Serius. Vou. XXVI, No. 151.—Juty, 1908. 9) 66 ZT. D. A. Cockerell—Descriptions of Tertiary Plants. normal, and frequently the only one present ; so it may per- haps be an open question whether the appearance of rugelic leaves on true séccharum is due to reversion or to hybridiza- tion. In the position of the lobes, the fossil resembles A. pennsylvanicum L. Hab.—F \orissant, in the Miocene shales. Station 14 (8. A. Rohwer, 1907). Peabody Museum, Yale, Cat. No. 1002. Acer florigerum, sp. nov. Figure le. Flower ; pistillate, apetalous, tetramerous. Styles two, sepa- rate to the base , long, exserted, straight or very lightly curved, ie 1. (a) Limnobium obliteratum ; (b) Acer perditum ; (c) Acer florigerum, Sele d} Phaca wilmatte. about 4°" long. No sign of any stamens. Sepals four, lan- ceolate, pointed, only about the basal two-fifths united. In the form of the styles, this resembles A. nigrum Mx., but the calyx segments are nearly as in A. drummondii in form. Only four calyx segments or sepals are preserved, and there were apparently no more. The dicecious tetramerous flowers indicate affinity with such species as A. tetramerum Pax, A. betulifolium Maxim., and A. barbinerve Maxim. FTab.—F lorissant ; Miocene shales (W. P. Cockerell, 1907). It was found at Station 14. I think this flower may safely be T. D. A. Cockerell--Descriptions of Tertiary Plants. 67 assioned to Acer. As it presents some interesting characters, and cannot be assigned with any degree of assurance to any particular species known by leaves or fruits, I give it a sepa- rate name. Peabody Museum, Yale, Cat. No. 1003. 38. A Veitch Pod. Phaca wilmatte sp. nov. Figure 1d. Pod strongly inflated, broad-ovate in outline, 13"" long and 10 broad, apparently thick, as preserved dark red-brown, tipped with a thick strongly curved style about 33" long, and with a short thick stipe about 2™™ long, its union with the base of the pod perfectly abrupt. Calyx very small. In the form ~ of the pod, this is more like P. longifolia (Pursh) Nutt., but the pod being stalked, it is so far related to P. americana (Hook) Rydberg. | Hab.— Florissant, Miocene shales, Station 14 (Wilmatte P. Cockerell). Jam indebted to Miss Alice Eastwood for calling my attention to the affinities of this fossil. Peabody Museum, Yale, Cat. No. 1004. 4. Miocene species of Hydrangea. In 1885 a supposed species of J/arsilea, found in the Upper Miocene beds of the John Day Basin, Oregon, was published by Ward. Lesquereux, in 1888, observing that the plant was certainly not a Jarsilea, but represented a calyx of some sort, referred it to Porana. In 1902, however (Bull. 204, U. S. Geol. Surv., p. 60), Knowlton, following a suggestion from Pollard, referred it to Hydrangea; and this appears to be certainly correct. Hydrangea as been reported to occur, with several species, in the European Tertiaries; but some of the species, at least, are doubtfully of this genus. In addition to H. bendiret (Ward) Knowlton from Oregon, two species have been found in the Miocene shales of Floris- sant. One of these, H/. subincerta, I have published in Bull. Amer. Mus. Nat. Hist., 1908, p. 92; the other, more recently found, is described herewith. Hydrangea florissantia sp. nov. Sterile flower large, the larger diameter about 21™”; sepals nearly round, the larger about 10™™ long, the smaller about 9 (one of the smaller missing in the type); color as preserved light brown, the centre of the flower dark; venation distinct, except peripherally, nearly as in H. bendirez. 68 7. D. A. Cockerell—Descriptions of Tertiary Plants. Evidently close to H. bendirez, but much smaller, more equilateral and with the shorter sepals less truncate. Very different from //. subincerta by the larger size and shape of the sepals. Fic. 2. Hydrangea florissantia, Florissant, Miocene, Station 14 (7. D. A. Cockerell, 1907). Type in Yale University Museum. On the same piece of shale touching the [1ydrangea (as shown in the figure) is what I take to be part of the male inflorescence of a Castanea— presumably that of C. dolichophylla Ckll., which is represented by leaves at Station 14. T. D. A. Cockerell—Descriptions of Tertiary Insects. 69 Art. VI.— Descriptions of Tertiary Insects; by T. D. A. CocKERELL. Part IV. [Continued from p. 312. | (9) Dragonflies from Florissant, Colorado. Melanagrion nigerrimum sp. nov. Wines hyaline basally, to about three cells beyond the quadrangle ; beyond that black to apex (owing to the partial destruction of the membrane, the black is patchy and. irreg- ular; the apical field was perhaps dark brown rather than pure black); first row of costal cells broad, as in Lithagrion ; at the tenth cell before stigma the costal cell is half size of sub- costal, the latter being twice as deep; apex of quadrangle to tip of wing about 245"™ (262 in WZ. umbratum), but the ten costal celis before stigma measure together about 7$"™ (six in M.umbratum). Stiema large, about Bum long, homie five cells below ; costa obtusely bent at stigma, the margin bey ond rapidly descending to apex; eight poststigmatal sectors ; vein Cu,, 2 little before medioanal link ; base of wing agr eeing in general with J/. umbratum. Florissant: one specimen, with reverse ; Station 14 (W. P Cockerell, 1907). Holotype in Peabody Museum, Yale. Lithagrion hyalinum Seudder. Figure 1. A specimen was obtained by Mr. 8. N. Rohwer at Station 17, on a slab with Typha lesquereuar Ckll.,* showing some characters heretofore obscure. There are certainly only two antenodal sectors. The stigma is swollen, and bounds 3% cells below. The total length of the wing is pgm . ; from nodus to stigma 143"; nodus to base PENE 2 louie ila of wing in middle 7", These dimensions are uniformly less than in Scudder’s type, but that was probably an upper wing, while ours is appar- ently a lower; there may also be a difference of sex. (In Enallagma civile, male, I find anterior wing 193"™, posterior 183). There are 17 sectors on costa between odie and stigma (16 in type ZL. hyalinwm), and 14 in the same distance in the subcostal series ; the costal cells beyond the stigma are doubled, which is not at all the case in Melanagrion nigerrimum (in M. umbratum there is a slight tendency to doubling). There are three simple cells,between M, and M, before the doubling begins. These new materials make the genus Melanagrion appear less distinct than when it was proposed, IZ. nigerrimum being in some respects intermediate between Melanagrion and Lith- * On the other side of the slab is Planorbis florissantensis Ckll. (Peabody Museum, Yale). 70 7. D. A. Cockerell— Descriptions of Tertiary Insects. agrion. Characteristic of both is the position of vein M,, originating a long way before the level of the nodus, though not half-way to the arculus;—a condition found to- day in Megaloprepus, one of the Anormostigmatini. The question was raised, whether Lethagrion and Melanagrion could rep- resent the stem which gave rise to the Anormostigmatini; but altogether against this is the position of the long sector between veins M, and M,. Mr. E. B. Williamson writes that he regards the latter character as of considerable significance, and for this and other reasons’ would support Scudder’s reference of the insects to the Podagrion series. He adds, with reference to Fie. 1. Lithagrion hyalinum. M. umbratum: “it is not specialized by reduction, and the nodus is retracted as in Paraphlebia (but not so much as in Anormostigmatini), with which compare the more specialized Argiolestes and Wesolestes, for example.” The following table separates the three Florissant species : Wings hyaline; stigma bounding 33-32 cells below. L ithagrion hyalinum Seudd. Wings strongly infuseated ; stigma bounding 5 cells below. 1. Apex of wings hyaline ; costal cells narrow. Melan agrion umbratum (Seudd. }. Apex of wings dark; costal cells broad. Melanagrion nigerrumum Cxkll. Tinallagma florissantella sp. nov. Figure 2. Wing hyaline, about 23™™ long (base gone) ; nodus to stigma 124°"; nervures and stigma ‘dark sepia brown; subnodus oblique; subquadrangle not ero ossed ; 14 costal sectors between nodus and stigma ; stigma bounding one cell below ; costal cells beyond stigma large; only one “double cell in the series between M, and “the sector M,,, this immediately below stigma, separated therefrom by a single cell; three cells between quadrangle and level of nodus, the third very long, and represented by two cells in the series immediately below ; T. D. A. Cockerell —Descriptions of Tertiary Insects. 71 upper side of quadrangle not or barely longer than inner side ; six cells on lower margin before Cu, begins to zigzag, and ten cells in the zigzag por tion, making 16 cells in all from ‘subquad- rangle and end of Cu, ; three cells between M, and M, before the doubling begins ; costa before nodus scarcely at all arched. Florissant : one specimen ; Station 14 (7. D. A. Cockerell). A poorly preserved leaf of Ficus arenaceetormis Ckll. is on the same slab. The figured specimen is in Peabody Museum, Fie. 2. Enaliagma florissantella. Yale. Some of the characters used to separate fossil Agrio- nines are so variable in recent species as to be of small value. This was pointed out to me by Dr. Calvert, and is very clearly indicated by a series of Enxallagma very kindly given to me by Mr. E. B. Williamson. Thus: (1) Cells between M, and M, before doubling begins. A specimen of £. antennatum (fischeri) has ‘three in ante- rior wing, four in posterior. (2) Cells between quadrangle and level of nodus (supposed to separate the fossil Agrion exsularis Seudd. from A. masce- scens Seudd., the first having three, the second four). Three is the usual number in Enallagma, but E. travi- atum may have four, and in a male £. carunculatum one wing has four, the other three wings three each. (3) Length of upper s side of quadrangle. It is much shorter in anterior than posterior wings of Enallagma civile, E.. antennatum, E. exsulans, and E. traviatum. On the other hand, many undoubtedly distinct recent species are so similar in the wings that it is exceedingly difficult, to say the least, to separate them by these organs alone. In deal- ing with the fossils, therefore, one may well hesitate to assert that species are synonymous, although some of their assigned characters are likely to be of less than specific significance. The Florissant species of this group may be separated as follows: Subquadrangle with a cross-nervure in the middle; sub- nodus almost vertical ; first postnodal cell considerably longer than second ; eleven postnodal sectors. “ Trichocnemis” aliena Scudder. 72 T. D. A. Cockerell—Descriptions of Tertiary Insects. Subquadrangle without a cross-nervure; subnodus obliquepiar ris ruil: 1. Nervures and stigma pale ferruginous ; ten postnodal sec- tors; stigma very oblique, with the inner side as long as the outer. Llesperagrion prevolans Ck. Nervures dark brown or black ; stigma ordinary. . 2. 2. Costa before nodus conspicuously arched ; 11 postnodal cross-veins ; curved basal section of M, very short. Agrion exsularis Seudd. Costa before nodus hardly arched ; curved basal section of M, longer. 3. 3. Postnodal cross-veins 10 to ele upper side of ‘quadrangle longer than inner. Agrion mascescens Seudd. Postnodals 14 ; upper side of quadrangle a little shorter than inner. knallagma florissantella Cll. (probably upper wing). The postnodals in living Hnallagma are from 9 to 11, at least in the species examined. I find the upper side of quad- rangle much longer than inner in /. fischeri and FL, exsulans ; but in an upper wing of £. carunculatum the inner is longer than the upper. The difference between the quadrangles of A. mascescens and £. florissantella cannot be due to their rep- resenting different wings, for the upper wing of mascescens 1s known, and has the upper side of the quadrangle very long. The position of the base of the quadrangle seems to be of some significance : (1) Base of quadrangle conspicuously before level of midmost point between antenodal cross-veins. A. mascescens. (2) Base of quadrangle at or near level of midmost point. EL, florissuntella, A. exsularis, EL. fischerc ; E. signatum, EF. hagent. (3) Base of quadrangle far beyond level of midmost point, not tar from level of first cross-vein. L. carunculatum, EE. civile. Trichocnemis aliena Scudder. Figure 3. A wing was obtained at Station 13 B(W. P. Cockerelt ). The most striking character, the crossed subquadrangle, is unfortunately not visible in our specimen, but there is no reason to doubt that it exists, as represented in Scudder’s figure. Mr. Williamson writes: “ Paraphlebia is the only Agrionine genus known to me with crossed subquadrangle. It is a character that disappears with reduction in the Calopterygine ; e. g., Dephlebia has it rarely crossed. Cyanocharis has quadrangle but not subquadrangle crossed ; Dewndgiia has both quadrangle and subquadrangle T. D. A. Cockerell— Descriptions of Tertiary Insects. 73 crossed; JMicromerus has quadrangle usually crossed and sub- quadrangle with one or two cross-veins. Paraphlebia, men- tioned above, of course is entirely different from Z7richocnemis in all other characters.” With regard to the vertical subnodus, Mr. Williamson writes : « Xunthagri ion erythroneurum has it vertical; Lrythromma NaUS (especially the hind wing) nearly or practically so; 1t 1s nearly vertical in the American Oxyagrion, Argia, yponeur a and Ischnura (at least some of the species). This character, I believe, appears independently many times, and is no crite- rion in itself” of generic relationships. Fic. 38. Trichonemis aliena. Except for the crossed subquadrangle, Mr. Williamson says that he sees no objection to referring 7. aliena to Hespera- grion. “The form of wing is similar; the quadrangle, sub- quadrangle and the relations of their parts to the antenodals are similar; the origin of M, is similar; and the length and direction of the subnodus are ‘not far out of the way. I believe e, in this view, that 7. alzena |Scudder’s type] is a hind wing.” Mr. Williamson further adds: “T have compared Scudder’s figure with wings of Calicne- mis, Hemicnemis (Leptocnemis), Platycnemis and Tatocne- mis,—some of the genera usually associated with Zrzchoenemis. If Scudder’s figure of wing is correct in outline, 7. addenda is not similar to any of the above genera. In arrangement of veins at distal end of quadrangle, Z. a/zena is most similar to Hemicnemis (which genus, of the whole group, has the quad- rangle most dissimilar) and Calicnemis. In arrangement of veins at base of quadrangle it most resembles Platycnenmis. Moreover, in above genera, only in Platycnemis is the sub- nodus nearly or quite vertical. In Zatocnemzs (which is very dissimilar in many characters), the subnodus is very short.” Our specimen, like Scudder’s, has eleven postnodal cross- nervures. The first cross-nervure beyond the stigma is forked below. There are three cells between M, and M, before the doubling begins (four in Seudder’s type). There is a very dis- t+ 7. D. A. Cockerell—Descriptions of Tertiary Insects. tinct brace-vein. The figured specimen is in Peabody Museum, Yale. Hoploneschna separata (Scudder). Two specimens, representing the hind wing, were obtained at Station 14, one by my wife, the other by myself. Seudder’s type, an anterior wing, was referred to Laszeschna; but Needham expressed the opinion that it belonged to Hoplonee- schna (Pr. U.S. Nat. Mus., xxvi, p. 761). The new specimens: are far from perfect, but they show that the hind wing has the following characters: (1) Triangle with a double cell at base, and then four simple cells, varying to a double and three simple cells. (2) Anal triangle of three cells. : (3) M,, after running parallel with M,, separated by a single row of cells, is suddenly bent downwards, and is separ- ated by two and three rows of cells; a character of floploneschna. In my table in Bull. Amer. Mus. Nat. Hist., xxiii, pp. 133, 134, the insect runs to Basiwschna; but the anal triangle agrees with Hoploneschna, and not so well with Basiwschna. On the other hand, as to the stigma //. separata is like Basiceschna, not like Hoploneschna. The region about the triangle agrees almost exactly with Gynacantha. The number of cells in the triangle is variable in _A’schnids. Mr. Williamson, to illustrate this point, has very kindly sent statistics of three very closely allied forms of /schna, based on males. In order to make them clear, [ have constructed cell formule, enumerating the cells in order, beginning from the base. Thus 2, 1, 1, means a double cell am then two sim- ple ones; 1,1, 1, three ‘cells, all simple. rout Wing. 1, . . Indiana species (2 specimens). 1, . . Multicolor (seven). , 1, 1, Multicolor (eleven), Indiana sp. (18), Jalapa sp. (7). 1,1, Multicolor (two), Jalapa sp. (1). Hind Wing. ee al eicolor(one): ily . Multicolor (17), Jalapa sp. (4), Indiana sp. (2). 2, 3 1, ie 1, Multicolor ee Jalapa sp. (4), Indiana sp. (18). The number of cells in the anal triangle Mr. Williamson states is also not of generic importance. He adds: ‘“ The kink in My asa character generally associated with the curving T. D. A. Cockerell— Descriptions of Tertiary Insects. 75 backwards of radial and median supplements (see Coryphe- schna ingens for maximum of both characters, and Vasieschna, e.g., for minimum).” Mr. Williamson concludes that separata cannot go in Lasieschna; and unless a new genus is proposed for it, Hoploneschna seems to be the only genus to receive it. Phenacolestes paratlelus Ckll. This species was described from the apical part of the wing. The base of a wing, probably belonging to P. parallelus, is from Station 14 (Geo. WV. Rohwer). It differs from P. morandus Okll., in having six antenodal cross-nervures, the nervure from subquadranele to lower margin. arising from almost the apex of the former, and cross-nervure in fork of M,,, and M, before level of nodus. The part visible (as far as separ ation of M, from M,) is hyaline. (10) A Longicorn Beetle from Florissant. Saperda (7?) submersa sp. nov. Figure 4. Length about 227"; width of head about 44; width of thorax about 4; of insect in humeral region of ae tra about 745 length of head and thorax about gim antennee rather thick, probably about 16™™ long, but the extreme tip missing. Head dark above, but face and mouth pallid; thorax 4, ‘pallid, darker posteriorly ; elytra black at base (espe- cially on humeri), after which comes a broad ae 44™ long) light area, forming a broad band across both elytra, the remaining portion of the elytra black. Abdomen extending a little beyond tips of elytra. In general build and appearance, this is like Superda. I can- not demonstrate any lateral spimes on thorax and believe there were none, but this part is not very clearly visible. The transverse light area on the elytra recalls Oncideres cingulatus, althongh more basal than in that insect; the antenne are like those of Saperda, not like Oncideres. The rather broad head suggests Mecas rather than Saperda. Type from the Miocene shales of Fiorissant, Colorado, collector unknown. Mr. G. L. Cannon, who kindly placed it in my hands for description, informs me that it has been in the collection at the State Capitol for at least 25 years. Four fossil species of Saperda have been described from Europe. Three are from the Miocene, but one of these (S. valdensis Heer) is not identifiable. University of Colorado, Boulder. 76 = = Wickham—New Fossil Elateride from Florissant. Arr. VII.—WNew Fossil Elateride from Florissant: by H. F. Wickam. Corymbites Latr. C. granulicollis (figure 1).—Body rather short and stout. Head about equal in length and breadth, front apparently transversely rugose; antenne broken, but the remaining por- tion shows them to have been rather slender and only very little serrate, probably not attaining the hind angles of the 1 2 Fie. 1. Corymbites yranulicoliis, n. sp. x 2. Fic. 2. Corymbites primitivus, n. sp. x 2. thorax. Prothorax emarginate and narrower at apex, grow- ing broader with equal lateral curve to about the middle, then arcuately narrowing to a point just anterior to the posterior angles, which are rather markedly divergent and distinctly uni- | carinate; the disk with small closely placed granules, each with a minute central puncture. These granules become much finer and more crowded near the sides, and a median basal area (which may have been canaliculate) is nearly devoid of them. Elytra finely alutaceous, finely and sharply striate but not punctured. Anterior leg (the only one visible) short, second, third and fourth tarsal joints about equal. Wickham—New Fossil Elateride from Florissant. 77 Length (of entire insect) ‘96 1n., of elytron about °60 in., of prothorax about °23 in.; width of prothorax about ‘26 in. I place this insect in Cor ymbites trom the general form ; the shape of the prothorax strongly recalls that of C. carbo Lec., and C. wreipennis Kby. One specimen (Cat. No. 1, Peabody Museum, Yale), Station 14;8. A: Rohwer. 0, primitivus (figure 2).—Form rather stout, head finely and rather closely punctured. Prothorax emar oinate at apex, front angles obtuse, sides broadly arcuate, more sharply in front of the middle, the oreatest” width being at about one- third of the length, hind angles distinctly carinate but rather short and not strongly divergent; disk finely densely subru- gosely punctate, less closely along the middle. Elytra with fine sharp impunctate strie, interstitial spaces finely irregu- larly punctured. Legs and antenne invisible. Length, entire, 87 in., of elytron 50 in., of prothorax along median line 20 in.; width of thorax °23 in., of elytron about Sin. One specimen (Cat. No. 2, Peabody Museum, Yale), Station 13; Geo. N. Rohwer. Here, again, I have placed the species by its general appear- ance, the truly generic characters all being obscured. Melanactes Le Conte. MM. cockerella (figure 3).—Body moderately elongate. Head narrower than thoracic apex, antennee attaining base of thorax, basal joints obscured, the seven distal ones subequal in width and but slightly serrate, each very little longer than wide, front fairly closely but not coarsely punctured. Prothorax slightly broader than long, narrowest at apex, arcuately wider to a point a little behind the middle, thence slightly narrowed to near the hind angles which are somewhat (but not markedly) divergent, disk finely and closely punctate towards the sides but much more sparsely and a trifle more coarsely about the middie; the marginal bead of the pronotum is very distinet, but it is uncertain whether the hind angles are carinate or not. Elytra apparently distinctly alutaceous, striate, the striz fine and marked at their bottoms with rows of moderately deep slightly elongate punctures which are separated by intervals arranging approximately the lengths of the punctures, Legs invisible. Length, entire, ‘94 in., width of prothorax, slightly behind the middle, -26 in., of elytron °37 in. One specimen (obverse and reverse, Cat. No. 38, Peabody Museum, Yale), Station 14; Mrs. W. P. Cockereli. 78 Wickham—New Fossil Elateride from Florissant. Fie. 8, Melanactes cockerelli, n. sp. x 2. In lite, this insect must have been about the size of JW. densus Lec., or Ml. piceus De G., resembling the former very closely in thoracic and elytral sculpture. Named after a good friend and ardent entomologist, Pro- fessor Cockerell, from whose hands the foregoing species were received. Iowa City, lowa G. Edgar—FEstimation of Iron and Vanadium. 79 Art. VIIl—The Estimation of Iron and Vanadium in the Presence of One Another ; by GrawHAM Ener. {Contributions from the Kent Chemical Laboratory of Yale Univ.—clxxvi. ] Tue difficulties in the separation of vanadic acid from iron were early recognized. In 1877 Bettendorf* noted that am- monium hydroxide precipitates from a solution containing these substances a yellow vanadate of iron not entirely decomposed by an excess of the reagent. He therefore added an excess of ammonium phosphate and boiled the solution, whereby the iron was converted into ferric phosphate and precipitated as such on the addition of ammonium hydroxide, while the vanadium remained in solution. Carnott used repeated precipitation by ammonium hydroxide, ammonium acetate or ammonium sul- phide, this process serving to remove iron but not aluminium or chromium. According to Classen, t vanadic acid may be separated from iron in ores by fusion with sodium carbonate and sodium nitrate, and extraction of the melt with water, the vanadium going ‘into solution as sodium vanadate. Accord- ing to Arnold,§ in the analysis of steel, a fusion with sodium per- oxide and extraction with water serves to obtain the vanadic acid in solution, free from iron. Fritcherle,| in the analysis of carnotite, precipitated iron together with uranium and vana- dium with an excess of sodium carbonate, then added sodium hydroxide and boiled for some time, the vanadie acid alone going into solution. Blum, in separating vanadium from rather large amounts of iron, as in pig iron, states that the separation is only complete when, after precipitation of the iron as hydroxide or basic acetate, the precipitate is redissolved, tartaric acid added and the iron precipitated by means of an excess of ammonium sulphide, the solution being allowed to stand for several hours. Campagne** separates the greater part of the iron in ferrovanadium and vanadium steel by repeated extraction of the hydrochloric acid solution with ether, afterward determining the vanadium by repeated evaporation with concentrated hydrochloric acid, removal of this acid by means of sulphuric acid, and titration with potassium, perman- ganate. He states that if it be desired to determine the iron pres- ent, the original solution, without extracting with ether, may * Poggendorff’s Ann. der Phy., clx., 126-181. + Chem. News, lvi, 16 t Amer. Chem. Journal, vii, 349-353. S$ Electrochemist and Metallurgist, March-April 1902. | Chem. News, lzxxii, 258. *| Zeitschr. Anal. Chem., xxxix, 156-157. ** Ber. der Dtsch. Chem. Gesel., xxxvi, 3164. 80 G. EKdgar—Lstimation of Iron and Vanadium. be boiled repeatedly with hydrochlorie acid to reduce the vana- dic acid, the hydrochloric acid removed by means of sulphuric, and the vanadium titrated with permanganate. If then the solution be reduced with hydrogen sulphide, the excess of gas removed by boiling and the solution again titrated with per- manganate, the difference between the two titrations will give the iron present. Glasmann™* effects the separation by adding potassinm iodide to the sulphurie acid solution, removing the free iodine by means of sulphur dioxide, neutralizing with po- tassium hydroxide, and precipitating the iron by means of a mix- ture of potassium iodide and iodate. The liberated iodine is removed by sodium thiosulphate, the ferric hydroxide filtered off and the vanadium estimated in the filtrate. | Among other methods of separation may be mentioned that of Classen, + who precipitates the iron electrolytically from a solution of the double oxalates, and Myers, t who deposits the iron using a mercury cathode, the vanadium in both eases remaining in solution. In view of the dithculties attendant upon the separation of iron and vanadium, and in view of the increasing importance of such substances as ferrovanadium and vanadate of iron, composed almost entirely of these elements, it has been thought advisable to present a method by which vanadium and iron may be estimated in the presence of each other. If a solution containing vanadic acid and iron be reduced by means of sulphur dioxide the reoxidation by potassium perman- ganate proceeds according to the equation 5V,0,+10FeO +4K MnO,=5V,0,+5Fe,O,+ 2K,0+4Mn0 (1) If this solntion, after titration, be passed through a column of amalgamated zinc in the Jones reductor, the receiving flask being charged with a solution of ferric sulphate, the reduction is carried, in the case of vanadie acid, to the condition of V,O,$ and the reoxidation by permanganate proceeds according to the equation 5V,0,+10FeO +8K MnO,=5V,0,+5Fe,0,+4K,0+8Mn0 (2) The difference in the number of cubic centimeters of per- manganate used in the first and second titrations is evidently used in oxidizing the vanadium from the condition of V,O, to V,O,, and multiplied by the factor 00456 (for exactly N/10 solu- tions) gives the amount of vanadic acid present. This being known, the iron present may be calculated from the amount of permanganate used in either titration. * J. russ. Phys. Chem. Ges., xxxvi, 77. + Ber. Dtsch. Chem. Ges., xiv, 2771-85. tJ. Amer. Chem.*Soc., xxvi, 1124. S$ Gooch and Edgar, this Journal, xxv, 238. Ferric Alem ak Il in KMnO, KMn0O, Receiver N/10x 9545 V205 V.0; Erroron Fe,O; Fe.0; Error a = taken found V2.Os5 taken found Fe.03 em3, cm?. erm. erm. grm. grm. grm. grm. 35 31°90 58:02 0:1136 0O711387 +0:0001 0°1487 071436 —0-0001 ag so) 58:04. O1186. O:1138 .+0°0002 0 1437. 0:1435. -—0:0002 35 aioe 9 5S'00.- O°HI386 O1138. + -0002. -0°1437 .0°14383 —0-0004 30 31-90. 58:00 _0O°1136 O-1136 ° +0:0000. 0°1437 0:°1437 +0:'0000 20 75-30 > 38-35. 0:°0568. 0:0568— +0°0000 _0°1487. .0:1423 —0-0004 20 95°29 38°30 0°0568 0°0566 —0:0002 0°1437 0°1483 —0:0004 20 15°98 29:02 0°0568 0°0568 +0°0000 0:0719 O°0721 . +0:0002 50 oso | OO, wOO4. 0-1 1.04% ».— 00009. (O°1437 071442 . + 0°0005 50 38°45 77:60 0°1704 0°1704 +00000 0O°1437 0°1488 +0:°0001 50 aes 77-58 9071704 -0-1703-. —0'0001 0°1437 071439 -+0:0002 35 22°50 48°60 0°11386 0°1136 +0°0000 0°0719 0:0720 +0°0001 35 22°50 48°60 011386 0°1136 +0°0000 0°0719 0°0720 +0°0001 35 92°45 48°58 0711386 0°1187 +0°0001 0°0719 0°0716 —0:0003 20 15°97 29°07 0°0568 0°0570 +0°0002 0°0719 O°0718 —0O:0001 * Inaug. Diss., G. Edgar—Estimation of Iron and Vanadium. 81 In the experiments in Table I three standard solutions were used, viz. :— | 1. A solution of potassium vanadate, slightly acidified with sulphuric acid, and containing 11°36 grams of vanadic acid to the liter. This solution was standardized by the method of Holverscheit.*, 2. A solution of ferric alum, slightly acid, with sulphuric acid and containing 14°37 grams On ferric oxide ‘ the liter. This solution was standardized by the very accurate method of Newton. t+ 3. An approximately tenth normal solution of potassium per- manganate (3°16 grm. per lit.), standardized against a N/10 solution of arsenious oxide. The details of manipulation were as follows :—Measured portions of the ferric sulphate solution were mixed with por- tions of the solution of vanadic acid, and a current of sulphur dioxide was passed through the slightly acid mixture until the color had changed from red into green and finally into a clear blue. A few centimeters of dilute sulphuric acid were then added, and the solution heated to boiling, the current of sul- phur dioxide being replaced by one of air-free carbon dioxide. When the last traces of sulphur dioxide had been removed, the flask was cooled in running water, the atmosphere of carbon dioxide being maintained, and when thoroughly cool, titrated with potassium permanganate until the color had changed from blue into yellowish green. The solution was then heated to Berlin, 1890. + This Journal, xxv, 343. Am. Jour. Sct.—FourtH SERigs, Vout. XXVI, No. 151.—Juty, 1908. 6 82 G. Ldgar—Estimation of Iron and Vanadium. 70°-80° and the titration completed at that temperature. The solution, having now a volume of 100-150°™, was passed through a column of amalgamated zine ina long Jones reductor, being preceded by 150° of hot dilute (23 per cent ) sulphuric acid and followed by 100°™ of dilute acid and finally .200%™* of distilled water. The receiving flask, containing. an excess of ferric sulphate, was kept cool by means of running water, and its contents, after the addition of sirupy phosphorie acid to remove the color of the iron, were titrated with permanganate until the color had changed from bluish green to yellow, and the color of the permanganate began to be persistent and destroyed only by shaking. The flask was then heated to 70° 80° and the titration completed in the hot solution. The results given in the table show that iron and vanadium may be readily estimated in the presence of each other by two oxidations with potassium permanganate, following reduction first with sulphur dioxide and last with amalgamated zine, _ under the conditions described above. In conclusion, the author desires to thank Prof. F. A. Gooch for advice given in the course of the work. Uj Browning and Palmer—Estimation of Cerium. 83 Arr. LX.—On the Estimation of Cerium im the Presence of the other Rare Earths by the action of Potassium Ferricyanide ; by Eee E. Browniye and Howarp E. PALMER. [Contributions from the Kent Chemical Laboratory of Yale Univ.—clxxvii. ] Tue work to be described was undertaken to determine how completely the oxidation of cerium from the cerous to the ceric condition may be effected by potassium ferricyanide in alkaline solution, and how completely the measure of the oxi- dation can be registered in the amount of potassium ferrocy- anide formed, according to the following equation : 2K FeO.N, + Ce,O, + 2KOH = 2K,FeC,N, + H,O + 2CeO,, For this work a solution of pure cerous sulphate was made and standardized by precipitating measured and weighed por- tions with a definite amount of a standard solution of sodium oxalate, filtering, igniting the cerium oxalate, and weighing the ceric oxide obtained. As a check on this method, the excess of the sodium oxalate over the amount used for the precipitate was determined in the filtrate by titration with potassium permanganate, and this amount was subtracted from the whole amount of sodium oxalate used. From this result the cerium present can be readily estimated, the amount of sodium oxalate used in the precipitation being known.* The ferricyanide solution used was made by dissolving 2 grams of carefully selected crystals of potassium feieneetde in 100° of water. About 20™* of this solution were used in each determination. | The procedure was as follows: To measured and weighed portions of the cerous sulphate solution 20° of the ferricy- anide solution were added, and potassium hydroxide in solu-. tion to complete precipitation. The precipitated hydroxide was filtered off and the filtrate and washings, amounting in volume to from 200° to 250°", after being made distinctly acid with dilute sulphuric acid, were titrated with a standard solution of potassium perman- ganate until the presence of the permanganate color showed the oxidation of the ferrocyanide to the ferricyanide,t accord- ing to the equation : 5K,FeC,N, + KMn0O, + 4H,SO, = aie BEC INE Habs Oia MnSO, - “4H, O. By this equation and the pr eceding one the amount of cerium present can be readily calculated. Each day before the ferricyanide was used a portion of 20° of the solution was acidified and titrated with the permanganate to color, and the amount necessary, generally from one to three drops, was subtracted from the amount of the perman- * Browning and Lynch, this Journal [4], viii, 457, 1899. + Sutton’s Volumetric Analysis, [IX edition, page 209. 84 Browning and Palmer—Estimation of Cerium. ganate used in the actual determination. It is of interest to. note that ferricyanide solutions kept in clear glass bottles for a week or more showed variations of only a drop or two in the amount of permanganate taken up. All the various operations in this process were carried on without warming the solution. The filtrations and washings were generally made under gentle pressure, and required on an average not more than fifteen to thirty minutes. In Table I the results obtained with cerium present alone are given. TABLE I. Ce taken, caleu- Ce found, calcu- lated as Ce,O; lated as Ce.O; Error erm, grm. grm. (1) 0°1834 0°1819 —0°0015 (2) 0°1376 0°1380 +0:°0004 (3) 0°1834 0°1829 —0°00085 (4) 0°1834 0°1829 —0°0005 (5) 0°18384 0°1834 +0°0000 (6) 0'1376 0°1885 +0°0009 (7) O36 01371 —0°0005 (8) 0°1376 0°1374 —0°0002 (9) 0°1376 0°1380 +0°0004 (10) 0°1834 0°1824 —0:°0010 (11) 0°1326 0°1335 +0:°0009 (12) 0°1326 0°1328 +0°0002 Solutions of the salts of the other rare earths, containing about 0-1 grm., were treated by the same method, and the failure to obtain evidence, by the permanganate, of the for- mation of potassium ferrocyanide, showed that there is no oxidation of these salts by the ferricyanide. In Table IL the results obtained by estimating cerium according to this method in the presence of the other rare earths are given. TABLE II. . Ce taken, caleu- Ce found, calcu- Other Rare Harths present, lated as Ce.03; lated as Ce.O; Error calculated as oxides grm. erm. erm. erm. (1) 0°1328 0713385 +0°0007 0 IThO, (2) 2 04327 0°1322 —0°0005 Orl (3) 0°0266 0-0275 + 0°0009 Orkneys (4) 0-0267 0°0272 +0°0005 Op ipse (5) 01824 0°1326 + 0°0002 0-1Y,O (6) 0°1326 0°1323 —0°00038 ORE eee (7) 0°0266 0°0264 —(0°0002 Ocal be A ais (8) 0:0264 0°0271 +0:°0005 Ogi wes (9) 0°1376 0°1370 —0°0006 0:15La,O, ae! O, (FO), 0-1 VOr 0°1091 — 0'0010 0°15 «“ (11) 01324 0°1332 +0:°0008 0°03ZrO, This method presents no difficulties in manipulation and 1s especially adapted to the rapid estimation of cerium in rare earth mixtures. Gooch and Weed—LEstimation of Chromium. 85 Art. X.—The Estimation of Chromium as Silver Chro- mate; by F. A. Goocu and L. H. Wexsp. [Contributions from the Kent Chemical Laboratory of Yale Univ.—clxxviii. | Ir has been shown by Autenrieth* that when chromic acid is added to a boiling solution of silver nitrate, or when a soluble chromate or dichromate is added to a solution of silver nitrate previously acidified with nitric acid, or when silver chromate is treated with nitric acid, silver dichromate is formed ; and that, on the other hand, it is silver chromate which is precipitated when silver nitrate in excess is added to a solution of a soluble dichromate, cold or hot, the reaction proceeding according to the equation 4AeNO,+K,Cr,0,+H,O = 2Ae,CrO, + 2KNO, +2HNO,. The characteristics of both silver dichromate and _ silver chromate have recently been summarized and further studied by Margosches,t but so far as we know there is in the literature no account of procedure for the exact quantitative determination of either chromium or silver based upon the characteristics of either of these salts. The solubility of silver dichromate in water and in ordinary solutions is such as to preclude the use of this substance as the final product of a quantitative process depending upon precipitation. The solubility of silver chro- mate in a moderately large volume of water is not inconsider- able, and the solvent action of free acid, even acetic acid in ‘quantity, is marked. We have found, however, that the pre- cipitation of silver chromate is practically complete in a solution only faintly acid with acetic acid and in presence of a large excess of silver nitrate. If such a precipitate is collected in the fitering crucible and washed with a dilute solution of silver nitrate, until no other impurities remain, silver chromate does not dissolve, and the excess of silver nitrate may be re- moved by the cautious use of water without appreciable effect upon the precipitate. The present paper has to do with the determination of chromium as silver chromate. In the experiments of which the details are given in the table, the general treatment just described was put to the practical test. Given amounts of potassium dichromate were weighed out and dissolved in hot water, as in experiments (11) and (12), or given amounts of a solution of potassium dichro- mate of known strength were run from a burette into a beaker and heated to boiling, as in experiments (1) to (10). To the hot solution of the dichromate was added, drop by drop, a * Ber. Dtsch. chem. Ges., xxxv, 2057. + Zeitschr. anorg. Chem., xli, 68; 1, 231. 86 Gooch and Weed—Estimation of Chromium. solution of silver nitrate in considerable excess, and the mix- ture was again brought to the boiling point. Ammonium hydroxide was added until the clear liquid became colorless and turned litmus paper blue; then acetic acid was added cautiously until the reaction of the solution was distinetly acidic to litmus. After standing for at least a half hour the precipitate was filtered off on asbestos in a perforated crucible, washed first with a dilute solution of silver nitrate and then with 20°" to 80° of distilled water applied in small portions, and dried with gentle heating to a constant weight. In exper- iments (1) to (10), the drying was done in an air bath heated to 135°; in experiments (11) and (12), the crucible and pre- cipitate were heated gently over the free flame. In experi- ments (7) and (8) the precipitation was made in presence of 5 grms. of ammonium nitrate, and in (9) and (10) in presence of 5 grms. of sodium nitrate, to test the effect of each of these substances upon the process. In no case did the filtrate, with the washings, show by the lead acetate test the presence of a chromate. The Precipitation of Silver Chromate. AgNO; Used in Volume at AgeCrO, K.Cr20; precipi- precipi- — === ——s . taken tation tation Found Theory Error Crimes erm. cm’. (1) 0°0921 0°4248 100 0°2072 0°2076 —0°0004 (2) 0°0921 0°4248 100 0°2073 0°2076 —0°0003> (3) 0°0921 0°4248 100 0°2075 0°2076 —0°0001 (4) 0°0921 0°4248 100 0°2074 0°2076 —0°0002 (5) 0°0921 0°4248 100 0 2075 0°2076 —0:°0001 (6) 0°0921 0°4248 100 0°20738 0°2076 —0°0008 (7) 0:0921 0°4248 100 0°2073 0°2076 —0:0003* (8) 0:0921 0°4248 100 0°2075 0°2076 —0:0001* (9) 0°0921 0°4248 100 0°2080 0°2076 +0:0004 ¢ (10) 0:0921 0°4248 100 0°2070 0°2076 —0-0006 + (11) 0°5801 3° 150 1°3087 1°3082 +0:°0005 (12) 0°7352 3° 200 1°6573 16574 —0-0001f *The precipitation was made in presence of 0 grm. of NH,NOs. + The precipitation was made in presence of 0 grm. of NaNOs. t An excess of 1°™° of 40% acetic acid was added before filtering. From the results of these experiments it is apparent that accurate determinations of chromium taken as the chromate or dichromate may be secured by precipitating silver chromate in presence of an excess of silver nitrate, making the solution ammoniacal and then faintly acid with acetic acid, transferring the precipitate to the filtering crucible, washing with a dilute solution of silver nitrate, and, after other soluble impurities have been removed, finishing the washing with small amounts ‘of water applied in successive portions. C. Barus—Standardization of the Fog Chamber. 87 Art XI.—Wote on the Standardization of the Fog Cham- - ber by the aid of Thomson's Electron ; by Cart Barus. 1. Advantages.—Of all the methods which I have tried to evaluate the coronas in terms of the number of nuclei which they represent under given conditions of exhaustion, the above method is the most promising and expeditious. A single experiment need take but a few minutes. Incidentally the observer learns whether the negative and positive ions have both been captured ; for on using the tables of coronas which I developed heretofore, the value of ¢ may be computed, and the result must coincide with Thomson’s value. 2. Method.—My first experiments were made with a metal plate in a fog chamber, both the coronas and the current being observed successively, without changing the adjustment. But this was abandoned for a method m which a cylindrical con- denser is employed as follows. being the coefficient of the decay, are (apart from secondary radiation) —n = 6(N*—n*)— L/ev, (1) SSW (2) where (is the capacity of the system, I= 9rin UV e/ (log R/R,) (3) where V/V is constant for very small voltages. Hence there would be a second method of reaching ¢ in “terms of 6, or the reverse, if equation (1), where J is essentially a function of time, were integrable. Nevertheless the equation is available at once if V is large enough to make the current constant. In this case one may write, if VV nuclei are found in the fog chamber without electrical current, while 2 occur in the con- denser with current, e= CV / (300r(R,?— R,)lb(N?—n’)). Here 7 must be negligible compared with VV. Values so obtained (an Exner graduated electroscope suffices) are quite consistent among themselves; but the data come out 20 to 30 too large if 6 = 1:1X10~° is assumed. Installing a plate condenser in the fog chamber, I noticed that for a charge of 100 or 200 volts the coronas between the condenser plates were of about the same character as the coronas without, while the large chamber is filled with ions at the highest voltages. Possibly therefore such ionization as reaches the inside of a condenser by diffusion may account for the excessive currents; or there may be increased production due to secondary radiation. It is interesting to note that potentials of 100 to 200 volts are sufficient to eject dust par- ticles from the condenser, very small and not numerous it is true, but sufficient to make it possible to catch all the ions only after these dust particles have been precipitated in one or more exhaustions.* 5. Conclusion.—The good values of é¢ obtained under widely varied conditions in the present very rough experi- ments, show that the present method is not unworthy of development, with a view to the further measurement of this * The complicated conditions encountered here will be restated elsewhere. 90 ©. Barus—Standardization of the Fog Chamber. important constant. For this purpose I am at work on a redetermination of the nucleation values of the coronas, using as a source of light the virtually monochromatic mercury lamp. This is sufficiently intense and the coronas admit of the more definite optical interpretation.. Elsewhere* I pointed out that for large coronas the greater part of the fog particles evaporate ; thus even at 2 = 200,000 particles per cubic cm., about one-half evaporate and one-half subside. Hence the corona method is here alone available for counting particles. 1 also showed that in the case of coronas the interference phenomenon superposed on the diffraction phenomenon may be treated in a way similar to the lamellar erating, consisting of alternate strips of thin and thicker trans- parent glass; that the given types of coronas must follow each other in the ratio of 5, 4, 8, 2, 1,0 for their particle diameters and an increasing size of coronas ; that the ratio of fog particle diameter and interference plate thickness, dj D, tor, the same color minimum in the interferences is d/D = n/(n-1), where v is the refractive index, or about 7 to 8 in both eases. It must therefore be possible to compute the nucleation cor- responding to a given corona at a given exhaustion and tem- perature, purely from optical considerations of diffraction and | interference, as indicated. I hope to report the results in the near future. Brown University, Providence, R. I. * This Journal, xxv, p. 409, 1908. + Tbid., p. 224, 1908. Chemistry and Physics. eaet SCQEHN PPETC INTELLIGENCE. I. CHEMISTRY AND PHYSICS. 1. Volumetric Method for Chlorates.—The best known volu- metric processes for determining chlorates are the iodometric ‘method and the method depending upon the oxidation of a fer- rous salt. The latter is the more rapid of the two, and it has been extensively employed. To complete this reaction, however, boiling for about ten minutes is necessary. KNecur has recently described a new method for this determination, based upon the use of titanous chloride for the reduction. This reagent acts more rapidly than a ferrous salt, so that no heating is required. The procedure is as follows: 50° of standard titanous chloride solution (of which 1°°=:0015 of iron or thereabouts) are run into 5°° of concentrated hydrochloric acid contained in a conical flask through which a current of carbon dioxide is maintained. Then 10° of the chlorate solution (1£ in 500°°) are added. After a lapse of not less than three minutes, potassium sulphocyanide is added, and the excess of titanous chloride estimated by titra- tion with iron alum solution until a permanent red color is obtained. For the estimation of chlorate in bleaching powder the available chlorine due to hypochlorite is estimated in the usual way by adding potassium iodide and starch and titrating with hydrosulphite in the presence of acetic acid. A second por- tion of the solution is then titrated with titanous chloride, the result giving total chlorine from hypochlorite and chlorate.— Jour. Soc. Chem. Industry, 1908, 434. H. L. W. 2. Atomic Weight of Radium.—Mme. Curie’s first deter- mination of this atomic weight gave the number 225. Subse- quently she obtained witha larger amount of material (about four decigrams of radium chloride) the higher number 226-2. T. E. TuHorpPeE has now repeated the determination of this interesting constant, and has obtained three results, 22.68, 225°7, and 227°7, which agree satisfactorily with those of Mme. Curie, although he used a much smaller amount of radium chloride, only 6 or 8 centigrams, for the determinations. Both investigators have used the same principle, the comparison of the weight of radium chloride with that of the silver chloride produced from it by precipitation. To avoid losses by transferring, Thorpe made use of small glass-stoppered flasks for all of the operations—weigh- ing the radium chloride, dissolving it, precipitating with silver nitrate, washing the silver chloride by decantation, and drying and weighing in a single flask. Spectroscopic evidence is given that the radium chloride was free from all but the minutest traces of barium. It had been very carefully purified by the usual course of fractional crystallization. There are several circumstances which may affect the determination of the atomic ‘92 | Seventific Intelligence. eight of radium. Thorpe states that the chloride gradually increases in weight when exposed to the air, apparently on account of oxidation by ozone, the presence of which can be per- ceived by the odor, and by other tests. Another peculiarity of the radium salt is its action upon the vessels containing the solution. It gradually changes the color of colorless rock crystal vessels to deep purplish black, and these as well as porcelain and glass vessels appear to be sliehtly attacked eden with the formation of silicates.— Chem. News, xevii, 929. H. L. w. 3. The Polyiodides of Potassium, Rubidium, and Cuesium.— Using solubility methods, together with analyses of the undis- solved residues, Foore and CHALKER have determined the poly- iodides of potassium, rubidium, and caesium existing at 25°, and have found positive evidence of the existence of KI,, KL, RbL,, CsI,, and CsI,, while they found no evidence whatever of the existence of RbI,, RbI,, Csl, and CsI, which had been supposed to exist by Abegg and Hambarger, who used somewhat similar physical methods, but did not analyze the residues. It is to be noticed, also, that Abegg and Hamburger did not find KL, the first of these compounds that was discovered. It appears that while Abegg and Hamburger’s work was correct in principle, there must have been some irregularity in their solubility determina- tions, leading to incorrect conclusions.—Amer. Chem. Jour., OOO VION HE Wie 4. A Volumetric Method jor Copper.—A. process based upon the titration of cuprous thiocyanate with potassium iodate solu- tion in the presence of strong hydrochloric acid has been worked out by Jamieson and others of the Sheffield Scientific School. The reaction corresponds to the equation 4CuSCN + 7KIO,+ 14HCI=4CuSO, + 7KCl+71C1+4HCN +5H,0. The titration is carried out in a glass-stoppered bottle with a liquid containing about half of its volume of concentrated hydrochloric acid in the presence of a little chloroform. The disappearance of the iodine color in the chloroform marks the end of the reaction, and it is exceedingly sharp and delicate. The presence of filter paper does not affect the result. This general method of titration is due to L. W. Andrews, but it was not applied by him to thiocy- anates, to which it has now been found to be applicable. The authors give details for applying the method to copper ores and alloys in such a way as to remove interfering substances. Test analyses showed excellent results, and the method appears to be a very rapid and accurate one.—Jour. Amer. Chem. Soe., Xxx, 760. Hi. TS ave 5. Lhermodynamics of Technical Gas- Reactions; by FE. Waser. Translated by Arthur B. Lamb. 8vo, pp. 356. London and New York, 1908 (Longmans, Green & Co.).—This book consists of a _ series of seven lectures which have been considerably enlarged for publication. The mechanical theory of heat is developed, as the author says, from its very foundations. Then a number of reactions, which are important industrially, are treated from a theoret- Chemistry and Physics. 93 ical standpoint in a very thorough manner. The lectures are: (1) The latent heat of chemical reaction and its relation to reac- tion energy; (2) and (3), Entropy and its significance in gas reactions. (+) Examples of reactions which proceed withont a change in the number of molecules. (5) Some examples of reac- tions involving a change in the number of molecules. (6) Deter-. mination of the specific heat of gases. (7) Determination of gaseous equilibrium, with a theoretical and technical discussion. H. W. F. 6. A Search for Fluctuations in the Sun’s Thermal Radia-- tion through their Influence on Terrestrial Temperature ; by Srmon Newcoms. Trans. Amer. Phil. Soc., xxi, pp. 309-387.— The problem of variations in solar heat radiation as affecting. terrestial temperatures is discussed in a thorough, impartial way in this memoir. The conclusions are none the less interesting because essentially negative in character, although the observa- tions made by Langley and later at the Astrophysical Observatory at Washington have seemed to indicate a different result. New-. comb shows that a careful study of the annual departures of tem- perature over many regions in equatorial and middle latitudes, indicate a fluctuation corresponding with the period of solar spots. The maximum fluctuation, however, for tropical regions. is only 0°13° C., or, in other words, the amplitude of the change is 0°26° C., less than one-half degree Fahrenheit. The corre- sponding fluctuation of the sun’s radiation is, hence, concluded. to be 0:2 of 1 per cent on each side of the mean. In addition, ‘there is some inconclusive evidence of changes having a period of about six years, which may be plausibly attributed to changes. in solar radiation. Apart from these changes the evidence at hand indicates that solar radiation is subject to no change pro- ducing a measurable effect upon terrestrial temperature; the magnetic, electric, and radio-active emanations may be left out of account, as their thermal effect is inappreciable. The ordinary terrestrial phenomena of temperature, rainfall and winds are thus uninfluenced by changes in the sun’s radiation. That wide changes of temperature may occur, as those noted in 1903, when the temperature in Russia and Siberia, for example, was more than 20° F. above the normal, is interesting, but it is argued that these fluctuations cannot be attributed to changes in the radia- tion from the sun, because they do not extend to regions (i. e. the equatorial) where such changes would have their greatest effect. I]. Grouoey. 1. Karly Devonic History of New York and Eastern North America; by Joun M. Crarke. N. Y. State Museum Mem., IX, pp. 366, pl. 48+. Albany, 1908.—To the important series of New York State Museum publications is now added the sump- tuous memotr bearing this title, the peer of a notable line of pred-- ecessors and decidedly the crowning achievement on the part. 94 Scientifie Intelligence. of one of James Hall’s most distinguished pupils. Critics are sparing nowadays in bestowing the terms monumental upon a scientific treatise, the adjective so often savors of hyperbole ; and yet in the present case it must be allowed that any less superlative epithet would fail to denote the high character of Dr. Clarke’s magnum opus. Space is here lacking for an ade- quate estimate of its contents, and the reviewer is perforce con- fined to a general appreciation. | The work ‘is first of all a vast repository of information on a singularly complex subject, embracing great wealth of detail. To the practical student also, it commends itself as a digest in which all the essential facts of its theme are collected, classified, analyzed and interpreted with scrupulous exactness. Finally, the net result is systematized with the critical poise and acumen that mark the experienced investigator who brings to his task a broad grasp of cosmic problems, and whose mental attitude has been determined by the successful conquest of a large group of nature’s secrets. The book betokens all these qualities and more; for on the humanistic side one is delighted by the literary skill with which the author handles his material, besides many a brilliant discursus on manners, customs, history, scenery, of an enchanting region. Naturally, as the title indicates, the chief objective aims of the memoir consist in a presentation of the essential features of the geology and paleontology of that time- interval in the Paleozoic of eastern North America with which our author is perhaps the most familiar, and on which he is recognized at home and abroad as an accomplished master. More particularly it deals with the origin and relations of the Lower Devonian rocks of Gaspé peninsula, with an elaborate discussion of their fossil remains, an investigation for which the author’s earlier researches on the Guelph and Naples faunas of New York State served as a fitting prelude. This volume is the fruit of several years of thoughtful study and patient effort, and it may be safely said that a work of this kind will never be superseded. And yet, such are the manifold resources and complexity of his material that the author assures himself and his readers that “the facts here brought together are but suggestions for further study of this fertile field.” In this connection there comes to mind a German saying: “ Wir sind alle Schuldner unserer, Vorginger,” for in speaking of earlier workers Dr. Clarke pays a generous tribute to the memory of Logan, Billings, Dawson and other pioneer heroes of the Canadian Sur- vey. Of Sir William’s Geology of Canada it is said: “To a student of Gaspé geology, this is the compendium and guide.” And with reference to Palaeozoic Fossils we find this: “It is our good fortune to be able to cite this work so frequently that our pages may almost seem its memorial.” Yet we fancy that these twain explorers who handed along the torch are the very ones who would be most surprised at the large increment of knowl- edge and perfection of method that are signalized by the hand- Geology. 95 some volume before us, illustrated by its beautiful lithographs and numerous full-page illustrations, several of which are in poly- chrome and most excellently done. Among _ paleontologists Plates A and £B will command attention from being camera drawings by the author that recall his previous illustration of Dictyospongia i in Memoir 2 of the same series. One hundred pages of letterpress are devoted to an exploita-_ tion of the geology and physiography of Gaspé, this part being a substantial elaboration of the author’s preliminary sketch of the geology of Percé (published in the Report for 1903), and the remainder of the volume consists of a technical discussion of the Gaspé invertebrate faunas. Upwards of 70 new species are described, and the characters of others are redefined. Especial interest attaches to the author’s discussion of the origin, distri- bution and relations of Lower Devonic faunas, and the lines of their dispersal and invasion over different areas, this phase of the subject being treated with great ingenuity, and displaying keen philosophical judgment. It is to be noted, for instance, that the origin of the Gaspé sandstone is explained on the theory that it “‘ was an Old Red lake in the same sense as those of Scot- land and that in which the Oneonta and Catskill sands of New York were laid down.” Due prominence is given to the fact that. late stages of the Oriskany betray a large percentage of incom- ing migrants which form, as it were, the advance guard of the next organic invasion. Dr. Clarke accordingly concludes that the original determination of the age of the Gaspé beds as prac- tically equivalent to the Oriskany of New York is insufiicient, and he is able to trace a passageway by which the Hamilton contin- gent of this fauna entered the region from the Appalachian gulf, moving eastward amid lagoon conditions along the Atlantic border and thence into western Europe. This whole matter is admirably summarized at pages 250-252, and as the work is one that requires to be consulted by all students of the Devonian era, we cannot do better than recommend the reading of this section and others germane to it at first hand. Ose 2. Publications of the United States Geological Survey.— Recent publications of the U. 8. Geological Survey are noted in the following list (continued from voi. xxv, p. 150): ToroerapHic ATLAS.—Sixty-five sheets. The sheets, 38 in number, embracing Connecticut with portions of the adjacent States have been collated and bound in a permanent volume which will be of great value to those interested in the region named. ‘These volumes have been distributed by Congressman Lilley, the plan having originated with R. W. Thompson, private secretary to Senator Hawley. Bulletin No. 117 is bound in with the volume. Foutos.—-No. 154. Winslow Folio, Arkansas——Indian Territory. Description of the Winslow Quadrangle; by A. H. PurpvueE. Pp. 6, with 3 maps. 96 Scientific Intelligence. No. 155. Ann Arbor Folio, Michigan. Description of the Ann Arbor Quadrangle ; by I. C. Russettand Frank Leverett. Pp. 15, with 3 maps. No. 156. Elk Point Folio, South Dakota,—Nebraska—Iowa ; Description of the Elk Point Quadrangle; by J. E. Topp. Pp. 8, with 3 colored maps. No. 158. Rockland Folio, Maine; by Epson S. Bastiy. Pp. 15, with 5 colored maps. Buttetins.—No. 319. Summary of the Controlling Factors of Artesian Flows; by Myron L. Futter. Pp. 44, with 7 plates and 17 figures. No. 325. A Study of Four Hundred Steaming Tests made at the Fuel-Testing Plant, St. Louis, Mo., in 1904, 1905, and 1906; by Lester P. BRECKENRIDGE. Pp. 196, with 76 figures. No. 326. The Arkansas Coal Field ; : by ArTaur J. Coutisr, with Reports on the Paleontology by Davin Wuire and G. H. Girty. Pp. vi, 158, with 6 plates and 29 figures. No. 327. Geologic Reconnaissance in the Matanuska and Talkeetna Basins, Alaska; by StpNey Paige and ADOLPH Knorr. Pp. 71, with 4 plates and 4 figures. No. 328. The Gold Placers of Parts of Seward Peninsula, Alaska, including the Nome, Council, Kougarok, Port Clarence, and Goodhope Precincts; by ArtHuur J. Coniier, Frank L. Hess, Puiip 8. Smits, and Atrrep H. Brooks. Pp. 3438, with 11 plates and 19 figures. No. 329. Organization, Equipment, and Operation of the Structural-Materials Testing Laboratories at St. Louis, Mo. ; by Ricnarp L. Humpsrey, with preface by Joserpn A. Hommes. Pp. vi, 84, with 25 plates and 9 figures. No. 330. The Data of Geochemistry ; by Frank WiGGuEs- WORTH CLARKE. Pp. 716. See vol. xxv, p. 458. No. 831. Portland Cement Mortars and their Constituent Mate- rials. Results of Tests made at the Structural-Materials Testing Laboratories, Forest Park, St. Louis, Mo., 1905-1907 ; by Ricu- arp L. Humpurey and WiiiaM Jonpay, IR. Pp. vii, 130, with 20 plates and 22 figures. No. 382. Report. of the U.S. Fuel: Testing Plant, at St. Louis, Mo. January 1, 1906, to June 30, 1907. JosEPu "A. Horates in charge, ip, 1-299, No. 333. Coal-Mine Accidents; Their Causes and Prevention. A preliminary statistical Report ; by Cuarence Hatyi and WALTER O. SNELLING, with introduction by Josrpu A. Hotmes. Pp. 21. No. 334. The Burning of Coal without Smoke in the Boiler Plants: a Preliminary Report; by D. T. Ranpatit. Pp. 26, with 3 tables. | No. 336. Washing and Coking Tests of Coal and Cupola Tests of Coke, conducted by the U. 8. Fuel-Testing Plant at St. Louis, Mo. January 1, 1905, to June 30, 1907; by RicHarp Mo.tenkE, A. W. Betpren and G. R. DeztamateR, with intro- duction by J. A. Hotmes. Pp. 1-76. Geology. 97 No. 389. The Purchase of Coal under Government and Com- mercial Specifications on the Basis of its Heating Value: with Analyses of Coal delivered under Government Contracts ; by D. WeWANDALL. ~ Pp. 27. No. 343. Binders for Coal Briquets. Investigations made at the Fuel-testing Plant, St. Louis, Mo.; by James E. Mitts. Pp. 56 with | figure. W ATER-SUPPLY AND IreiGATION Paprers.—No. 212. Surface Water Supply of the Great Basin Drainage 1906; by E. C. La Rus, THomas Grieve, Jr., and Henry THurretu. Pp. iv, 98, with 2 plates and 2 figures. No. 213. The Surface Water Supply of. California, 1906. With a Section on Ground Water Levels in Southern California. (Great Basin and Pacific Ocean Drainages in California and Lower Colorado River Drainages) ; by W.B. Crarp. Pp. 219, with 4 plates and 2 figures. No. 214. Surface Water Supply of the North Pacific Coast Drainage, 1906; by J. C. Stevens, Rospert Foiuanspes, and HE. C. La Rue. Pp. vi, 208, with 3 plates and 2 figures. — No. 215. Geology and Water Resources of a Portion of the Missouri River Valley in Northeastern Nebraska ; by G. E. Con- DRA. Pp. 59 with 11 plates. No. 217. Water Resources of Beaver Valley, Utah; by Wits T. Ler. Pp. 57, with 1 plate and 3 figures. 3. Maryland Geological Survey ; Witt1am Burtock CLarRkK, State Geologist. Volume VI, pp. 572, pls. 51, figs. 19, with map. Baltimore, 1906.—Part I of the present report on the physical features of Maryland, by William Bullock Clark and Edwin B. Mathews, is a somewhat elaborate compendium of the geological and geographical features of the state, including the physiography, geology, mineral resources, soils, climate, forestry, etc. ‘The report on the highways of Maryland, by A. N. Jobnson, and state highway construction, by Walter Wilson Crosby, make up Parts III and IV, while Part V is a history of the origin, boundaries, etc., of the counties of Maryland. A new geological and soil map accompanies these descriptions. A large part of the material contained in this volume has been previously published as de- scriptive matter relating to Maryland’s exhibit at the Louisiana Purchase Exposition and in its present form will reach a greater number of readers. IB 1. Ge 4. Iowu Geological Survey, Samun. Carvin, State Geologist, James H. Luxs, Assistant State Geologist. Volume XVII; Annual Report for 1906, with Accompanying Papers. Pp. 588, pls. _ 62, figs. 44. Des Moines, 1907.—The report for 1906 deals largely with the economic resources of the state and includes a study of Portland cement and the geology of quarry products in general. There are included analyses of coals, limestones, chalks, clays, shales, and marls and an account of tests of the Iowa building stone. A new geological map of the state, compiled by T. E. Savage, accompanies this report. H. H.-G. Am. Jour. Sci.—Fourtu SeRies, Vor. XXVI, No, 151.—Juxy, 1908. [eRe 98 _ Seiten tific Intelligence. 5. Wisconsin Geological and Natural History Survey, E. A. Bires, Director.—The Wisconsin Survey has recently issued four road pamphlets, of 24 to 54 pages each, by the Highway Engi- neer, A. R. Hirst. The topics discussed are : Earth roads, Stone and Gravel Roads, the Earth Road drag, and Culverts and Bridges. 6. Geological "Map of Cape of ” Good Hope.—Four new sheets of the Geological Map, by A. W. Rogers and A. L. Du Torr, have been issued. No. 42 covers the region between Kimberley and Hopetown, and No. 52 an area from Mafeking westward to longitude 24° and from the Molopo river southward to latitude 26° 30’. No. 49 is the Kuruman and No. 50 the Vryburg sheet. H. EG 7. Variations Periodiques des Glaciers, XIIme Rapport, 1906 ; par Dr. Ep. Bruckner et E. Murer. Extrait des Annales de Glaciologie i1, March, 1908, pp. 161-198. (Fréres Borntrae- ger Hditeurs). Berlin, 1908. Also Zeitschrift fiir Gletscher- kunde ; Band II, Heft 3, pp. 161-234. Berlin, 1908.—The report on elaciers for 1906 presents facts similar to those of the last tive years. In the eastern Alps, of twenty-six glaciers reported three remained stationary and one was advancing. The Italian glaciers are all in retreat, and in Savoy and in the Pyrenees many small glaciers and even certain névé fields have disappeared. In the Bukhara mountains one glacier in the Pierre le Grand chain has a marked advance. In the glaciers of North America there has been a decided shrinkage, with the exception of the remark- able glacier in Yakutat Bay, described by Tarr. H. E. G. 8. The Ceratopsia ; by Joun B. Hatcurr, based on prelimin- ary studies by Oruniet C. Marsu. Edited and completed by Ricuarp 8. Lutt. Monograph XLIX, U.8. Geological Survey, pp. 198, pls. 51, and 125 figures in the text. Washington, 1898.— This volume constitutes the third of six extensive monographs planned by the late Professor O. C. Marsh on the extinct verte- brates of North America. One, on the Odontornithes, or toothed ‘birds, was published in 1880 ; asecond on huge horned mammals, the Dinocerata, in 1884, while the present volume on the horned dinosaurs has just aa The remaining three are in course of preparation. Under Professor Marsh’s direction, many of the illustrations for these volumes were made both lithogr aphic and on wood, and a series of preliminary notices, largely descriptions of new species, were published from time to time in this Journal. In the present instance, the notices were 16 in number, the lithographic plates 19, while of the woodcuts there were 28. ‘After Professor Marsh’s death, in 1899, Professor H. F. Osborn, who succeeded the former as Vertebrate Paleontologist to the U.S. Geological Survey, assigned the Ceratopsia volume to Mr. J. B. Hatcher, the discoverer and chief collector of this remarkable group of reptiles. Hatcher in turn carried the work forward, add- ing many of the remaining text-figures and plates as well as 157 printed pages of the text. After having completed the morphol- ogy and specific descriptions, Mr. Hatcher died on July 3, 1904, and it became necessary for a third author to pick up the threads Geology. 99 of the task and carry the work to completion. This has been done by Professor Richard 8. Lull, of Yale University, who has edited the whole volume and, in addition, added the final section including the generic and specific summary; the geology and physiography of the Ceratopsia localities and the discussion of the evolution, probable appearance, habits and causes of extinction of the race. The book includes an extensive biographical notice of Mr. Hat- cher by Professor Osborn, who expresses the hope that the vol- ume may prove to be a lasting monument to the rare and noble spirit of John Bell Hatcher. ILL. MisceLLAnrous Screntiric INTELLIGENCE. . |. Harvard College Observatory, Epwarp C. PIcKERING, Director.—Recent publications from the Harvard Observatory are noted in the following list (continued from vol. xxv, p. 460) : Annats.—Vol. XLIX, Pt. II. Peruvian Meteorology; by Soton I. Bartey. Observations made at the Auxiliary Stations 1892- 1895. Pp. 107-232 with 80 tables and 2 figures. Vol. L. Revised Harvard Photometry. A Catalogue of the Positions, Photometric Magnitudes and Spectra of 9110 Stars, painiy of the magnitude 6°50, and brighter. Observed with the 2 and 4 inch meridian photometers. Pp. iv, 252, with 4 tables. Vol. LXI, Pt. I. Researches of the Boyden Department ; by ~WixiiaM H. Pickerine. Pp. vi, 103, with 3 plates. CircuLaR No. 136. Comparison Stars for U Geminorum. Pp. 3. 2. Publications of the Allegheny Observatory of the Western University of Pennsylvania. Vol. I, No. 3. The Orbit of «a Andromede. By Rosperr H. Baker. Pp. 17-24. Vol. I, No. 5. The Orbit of Algol from Observations made in 1906 and 1907. By Frank Scuiesincer and R. H. Curtiss. Pp. 25-33. 3. Carnegie Institution of Washington.—Recent publications of the Carnegie Institution are given in the following list (con- tinued from vol. xxv, p. 163): No. 66. High Steam Pressures in Locomotive Service ; by Wituram F. M. Goss. Pp. 144. No. 82. The Physiology of Stomata; by Francis E. Luoyp. Pp. 142, with 14 plates and 40 figures. No. 85. Index of Economic Material in Documents of the States of the United States. New York 1789-1904. Pre- pared for the Department of Economics and Sociology of the Carnegie Institution of Washington ; by ADELAIDE R. Hasse. Pp. 553. iiss the same for Rhode Island; by ApErz~atwEr R. Hasse. Pp. 95. No. 92. Guide to the Archives of the Government of the United States in Washington; by Craupr H. Van Tyne and Watpo G. LeLranp. Second Edition, revised and enlarged, by 100 ~ WSeventific Intelligence. W.G. Leann. (Revised edition of Publication No. 14.) Pp. Ri e2y. 4. Maryland Weather Service; Ww. B. Crarx, Director. Volume Ii, pp. 515, with 169 figures and 24 plates. Baltimore, 1907.—The first volume of the Maryland Weather Service; issued in 1899 (see vol. ix, 234), contained a general account of the phys- lography and meteorology of the state. In the present work the climatic features of the city of Baltimore are made the subject of detailed study, this being based upon a series of observations extending over a period of nearly a century. The systematic thoroughness and minuteness with which Dr. O. L. Fassig has gone into this subject, and the fulness with which the various topics are illustrated by tables, figures and charts, makes the volume, as a whole, almost unique in meteorological literature, and gives it much more than a local interest. The introduction (pp. 21-26), on the operations of the service, has been prepared by the Director. 5. The Apodous Holothurians. A Monograph of the Synap- tide and Molpadiide ; by Huserr Lyman CrarK. Smith- sonian Contr. to Knowledge, vol. xxxv, 231 pages ; 13 plates.— This is a very much needed and useful work on a group of holo- thurians that is comparatively little known. All the genera and Species are described, the original descriptions and figures being copied when authentic specimens were not available. The anatomy and histology are given pretty fully in many cases. It includes many new species, mostly from the deep seas, and some new genera, as well as new limitations of old ones. There is a full biblography and index. Ay eye 6. American Association for the Advancement of Science.— The summer meeting of the American Association will be held at Hanover, N.H., from June 29 to July 4. The American Physical Society will meet in conjunction with Section B. Various interest- ing excursions, to the White Mountains and elsewhere, have been arranged. . 7. International Catalogue of Periodicals.—Professor KE. Guarini has recently issued a catalogue of 4063 periodicals classified by country and subject. MM. Dunod and Pinat, Paris, are the publishers. OBITUARY. M. Avserr Larparent, the eminent French geologist, died in May last at the age of sixty-seven years. Professor Kart Aucusr Moésius, Director of the Zoological Museum at Berlin, died on April 26 at the age of eighty-three rears. : M. Pierre J. A. Bikcuamp, the veteran French chemist, died on April 15 at the age of ninety-two years. Dr. Roserr Cuatmers, of the Canadian Geological Survey, died on April 9 at the age of seventy-four years. Witiiam A. Anruony, Professor emeritus of physics, electri- cal and mechanical engineering at Cooper Union, New York City, died on May 29 at the age of seventy-three years. Relief Map of the United States We have just prepared a new relief map of the United States, 48 x 32 inches in size, made of a special composition which is hard and durable, and at the saine time light. The map is described in detail in circular No. 77, which will be sent on request. Price, $16.00. WARD’S NATURAL SCIENCE ESTABLISHMENT, 76-104 College Ave., ROCHESEEM ER: Nw Y. Warps Natura Science EstaBuisHMent A Supply-House for Scientific Material. Founded 1862. Incorporated 1890. DEPARTMENTS: Geology, including Phenomenal and Physiographic. Mineralogy, including also Rocks, Meteorites, etc. Palaeontology. Archaeology and Ethnology. Invertebrates, including Biology, Conchology, ete. Zoology, including Osteology and Taxidermy. Human Anatomy, including Craniology, Odontology, etc. Models, Plaster Casts and Wall-Charts in all departments. Circulars in any department free on request; address Wards Natural Science Establishment, 76-104 College Ave., Rochester, New York, U.S. A. CONTENTS. Art. I.—Emission of Electricity from the Induced Activity of Radium ; by W. Duane If.—Ilvaite from Shasta County, Cal.; by B. Prescott... Ili.—Mechanics of Igneous Intrusion (Third Paper) ; by RA. Dary IV.—Rhinocerotide of the Lower Miocene ; by F. B. Loomis V.—Description of Tertiary Plants; by T. D. A. Cocx- eee r- ee ee ewe ww ew He ewe eee VII—New Fossil Elateride from Florissant ; WiIcKHAM alt Wer ion of Iron and Vanadium in the Presence of One Another ; by G. Epear 1X.—Estimation of Cerium in the Presence of the other Rare Earths by the action of Potassium Ferricyanide ; by P. Ee Brownie and GH: Eh, PAIMER 22) (3 oe X.—Estimation of Chromium as Silver Chromate; by F. A. GoocH and 1, EL W emp cus ee eee XI.—Standardization of the Fog Chamber by the aid of Thomson’s Electron ; by C. Barvs SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Volumetric Method for Chlorates, Knecut: Atomic Weight of Radium, THorpsz, 91.—Polyiodides of Potassium, Rubidium, and “Caesium, Foorr and CHAaLKER: Volumetric Method for Copper, JAMIESON : Thermodynamics of Technical Gas-Reactions, ¥. HABER, 92.— Search for Fluctuations in the Sun’s Thermal Radiation through their Influence on Terrestrial Temperature, S. Newcoms, 93. Geology—Early Devonic History of New York and Eastern North America, J. M. Cuarkn, 93.—Publications of the United States Geological Survey, 95. —Maryland Geological Survey: Iowa Geological Survey, 97.--Wis- consin Geological and Natural History Survey: Geological Map of Cape of Good Hope, A. W. Rocmrs and A. L. pu Torr: Variations Périodiques des Glaciers, XIIme Rapport, 1906, Ep. BRucKNER et EH. Murer: Cera- topsia, J. B. HatcuEr, 98. Miscellaneous Scientific Intelligence—Harvard College Observatory : Publica- tions of the Allegheny Observatory of the Western University of Penn- sylvania: Carnegie Institution of Washington, 99.—Maryland Weather Service: Apodous Holothurians, H. L. CLarxk: American Association for the Advancement of Science: International Catalogue of Periodicals, 100. Obituary—M. ALBERT LAPPARENT: Karu A. Mosius: M. Pierre J. A. BECHAMP: ROBERT CHALMERS: WILLIAM A. ANTHONY. Librarian U. S. Nat. Museum. Wor. XXVI. | AUGUST, 1908. Established by BENJAMIN SILLIMAN in 1818. THE AMERICAN JOURNAL OF SCIENCE. Eprron: EDWARD S. DANA. | - ASSOCIATE EDITORS Proressorns GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or CamMBripcE, PROFESSORS ADDISON FE. VERRILL, HORACE L. WELLS, L. V. PIRSSON anp H. E. GREGORY, or New Haven, Proressor GEORGE F. BARKER, or PHILADELPHIA, Proressorn HENRY S. WILLIAMS, or IrHaca, Proresson JOSEPH S. AMES, or BaLtTrMorge, Me. J. S. DILLER, or WASHINGTON. FOURTH SERIES VOL. XXVI-[WHOLE NUMBER, CLXXVI] No. 152—AUGUST, 1908. NEW HAVEN,. CONNECTICUT. 1908 THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. Published monthly, Six dollars per year, in advance. $6.40 to-countries in the Postal Union ; $6.25 to Canada. Remittances should be made eithér Siy ‘monéy-orders, registered letters, or bank checks (preferably on New York banks ; OUR SUMMER BULLETIN We have just issued a 10-page Bulletin covering in brief every department of our stock, the additions to which have been so numerous it was absolutely necessary that we publish a new list, as it is impossible in an advertisement to say much on this subject. Send for this list. As a special inducement we are allowing a discount for sales made during July and August of 20 per cent. on Common Minerals, and 10 per cent. on Fine and Rare Minerals, Polished Specimens, and cut gems. The departments treated in this list > are ; Showy Minerals, Rare Minerals, New Finds of Minerals, Gems, Rough and Cut, Geological Specimens, Ore Collections, Indian Relics, etc. Reconstructed Sapphires (the latest discovery). These stones have the same hardness and specific gravity as a genuine stone and the color and beauty is equal to that of a genuine sapphire. We have them in stock from 4 to 2 k. at $8 per carat. No PETE RELY 81—83 Fulton Street, New York City. THE AMERICAN JOURNALOF SCIENCE [FOURTH SERIES.] ——_____ $9 —__—__— Arr. XII.—The Réle of Water in Tremolite and Certain Other Minerals; by E. T. Auuen and J. K. CLemenrt. Object of the Investrgation.—A study of the composition of tremolite was undertaken with several objectsin view. In the first place, it belongs to the series of calcium and magnesium metasilicates, all the other known members of which have already been studied in this laboratory. Secondly, it is the simplest known amphibole with the exception of kupfferite, and consequently affords an advantageous opportunity for further study of the relations which exist between the amphi- boles and the pyroxenes. For a complete investigation of this kind, pure minerals are necessary, and since natural minerals can rarely be, classed as such, we sought to prepare tremolite synthetically. But, the ordinary methods proving prea to the task, we proceeded to astudy of the natural mineral, order that we might find out what elements were really essen- tial to it, as well as what physical conditions were necessary to its existence. One of the most important questions which presented itself in this connection was whether the mineral must be synthe- sized by wet or dry methods; whether it was hydrous or an- hydrous. Some preliminary work which we had done proved that it was incapable of existence above about 1000°-1100°, a temperature at which it is still in the solid state. Attempts to form it by heating a glass of the same composition below this temperature, or by rapid cooling of the melt, failed, indicating not only that the natural mineral must have been formed from solution, a conclusion in accord with geological evidence, but further, that it could be obtained in no other way. At first, molten salts were tried as solvents, on account of the difficulty of working with water at temperatures much above the Am. Jour. Sci.—FourtH Series, Vou. XXVI, No. 152.—Aveusr, 1908. 8 102. Allen and Clement—Role of Water in Tremolite. ordinary. These efforts also proving unsuccessful, the idea suggested itself that the mineral was perhaps hydrous and therefore must be made by the agency of water, in spite of the general opinion among mineralogists that tremolite is anhydrous. Material for Study.—In order to establish the true compo- sition of tremolite, we made a careful selection of specimens from five different sources. The color of the specimens indi- eated that they contained very little iron and the microscopic examination showed them quite free from inclusions, though some of them required separation from comparatively coarse grains of other minerals intergrown with or adhering to them. This was done with heavy solutions, either potassium mereurie iodide, or mixtures of methylene iodide and benzene. After it had been ascertained by the microscope that each specimen was as pure as it was practicable to get it, it was carefully ana- lyzed. The results are given in Table I. An inspection of the five analyses (I-V) shows that water* is present in all of them and ranges from 1°72-2°50 per cent, averaging 2°17 per cent in the two purest specimens. It may therefore be regarded as TABLE I].—ANALYSES OF NATURAL TREMOLITES. Localities 1 2 3 4 4) 6+ Tt Ham Ossin- Gouver- Rus- Kd- Rich- Island, ing, neur, sell, wards, ville, Lee, Alaska INS Ye: INE OYE New. Nave NONE Mass. S10, -- - -58°59 5:35 56°92 56°36 58°24 54S 57 69 Biss ce Sree ‘07 10 06 "O4 Gi. 14 evi O ees “LO 1s | 1°65 1°88 °60 1°30 1:80 Bes) «co. te alt "36 “61 "43 18 none 1S 6 ee 3 a: none 1:01 none 22 55 mn soes aes “Oil ae 04 1:28 07 trace MgO __.24-78 23°87 23°81 22°97 25°16 24°85 24°12 CAO 82 13°95 14°02 12°51 12°82 10°85 12°89 13°19 RiaiO sk V2 "42 1:22 “94 82 67 "48 = Caen 10 ae “60 “74 19 ‘D4 "22 MeO es 2 2:37 2-27 2-01 1:72 2-30) Zo) 166 , Oh eas none =allat 1:03 1-23 "24 17 37 99°95 99°80 100°21 100°38 100535 THOS) Stee O equiv- alent to | iH 20° ‘00 ‘05 43 9) 10 "32 "15 99°95 99°75 99°78 . 99°85 100°25. . 99°87 = HO0r0g * The water was determined by absorption by calcium chloride. See Hillebrand, Analysis of Silicate and Carbonate Rocks, Bull. U. S. Geol. Surv., 305, p. 62. + Analyses 6 and 7 were made by Penfield and Stanley. Allen and Clement—Roéle of Water in Tremolite. 1038 an essential constituent. While work on this subject was in progress, Penfield and Stanley* published the analyses of a number of amphiboles, including two tremolites, all made on very carefully selected material, in which it was found that all contained notable quantities of water and fluorine. While our results confirm theirs in regard to water, it may be noted in passing that fluorine is entirely absent from one of our speci- mens and occurs in quite insignificant quantities in one or two others. It is not, therefore, to be counted an essential constit- vent. The role of the water.—The next question to present itself is: What part does the water play in the constitution of the mineral? Is it chemically combined or dissolved, i. e., does it escape at one or more temperature points with a sudden change in phys- 1 ical properties, or is it given off grad- ually through a range of temperatures, the physical properties varying with the change in composition 4 Tammann,t in attacking a similar problem with the zeolites, used the method originated Jp by Van Bemmelen. The powdered minerals, in small beakers, were left to stand, at con- stant temperature (19°), in desiccators with sulphuric acid of various concentrations, the vapor pressures of which were known, until the weight became constant. This method had the disadvantage of removing only a frac- tion of the total water, 1-1 per cent to 23 per cent, according to the mineral, leaving one still in doubt about the major part of the water. This difficulty is obviated by the more direct method ot Friedel,t who heated the zeolites, at progressively higher temperatures, in a current of air which was saturated with water at an approximately constant tempera- ture, this temperature remaining nearly the same in all the experiments of any one series. He found thus for several zeolites true equilibria at every temperature. The Apparatus.—Some preliminary experi- ments on tremolite suggested that the vapor pressure, even at higher temperatures, would probably fall practically to zero within a limited time. We therefore adopted Friedel’s method, * This Journal (4), xxiii, p. 23, 1907. + Zeitschr. Phys. Chem., xxvii, 323, 1898. { Bull. Soc. Min., xix, 363, 1896; xxii, 5 and 86, 1899. T SSS aN SSNS S SSS SSSSSSSSSSSSSSSSSSSSSDDArSMN Yy Y G= Air current, T= thermoelement. 104 Allen and Clement—Réle of Water in Tremolite. only heating in a current of dry instead of mozst air.* The air was dried by concentrated sulphuric acid. The apparatus in‘ which the dehydration was earried on is shown in fig. 1. It is a cylinder of Berlin porcelain, closed at one end and glazed on the inside to make it impervious to gases. The upper end is molded so as to form a groove, into which fits the iron cover. There are also two inner covers of porcelain, each 1™ thick, which rest on lugs baked on to the inner wall of the cylinder. These keep the temperature more nearly constant and protect the iron cover from excessive heat. All three covers are perforated near the edge to allow the passage of two porcelain tubes. One of these, which is closed at the bottom, reaches down to the charge and carries the thermo- element. The other is open at both ends, reaches nearly to the bottom of the cylinder, and admits the current of dry air. One of the tubes may be fitted closely to the iron cover by a “ fibre” ring, the other may be left loose enough to allow the escape of the air. If it is desired to use some other atmosphere in the furnace, mercury may be poured into the groove to more effectually prevent any air from leaking in. The crucible whicn holds the tremolite rests on a platinum triangle which is supported by a hollow cylinder of fine white clay. This whole apparatus in an upright position is then slipped into a platinum resistance coil furnace which envelopes it to within 2 of the top. The furnace is heated by a storage battery and the tem- — perature can be regulated very closely indeed for a period of many hours. | Conditions of HExperiment.—There was no difficulty in maintaining the temperature within 5° without much atten- tion, except when the same battery was required intermit- tently for other work. In such cases there were sometimes aberrations of 10° or more. As the time of experiment is not essential, a fall in temperature is of no importance; a half hour or more at a temperature as much as 5° below the point aimed at was not counted. On the other hand, if the temperature ran 5° or more above the point, the work was rejected unless the charge underwent further loss at the same temperature on the following day. The loss of water was so very slow that an aberration of this kind scarcely ever affected the results. Besides, our object was only to determine the form of the curve with sufficient accuracy to settle the ques- tion whether the loss of water was continuous or discontinuous. Regarding the conduct of an experiment only a few words of explanation are needed. The mineral was generally in the form of a rather coarse powder, ground only fine enough to pass a screen of 100-120 meshes to the linear inch, because * The question of equilibrium is raised on p. 117. Allen and Clement—Réle of Water in Tremotite. 105 fine powders are known to absorb moisture from the air.* The crucible containing the powder was introduced into the furnace, rapidly heated to the required temperature and kept there during the rest of the working day. Then it was quickly transferred to a sulphuric acid desiccator, cooled and weighed. The time of cooling was usually about 20 minutes. There is some chance of error here, in that the mineral may, have absorbed moisture or air during the cooling. Friedel} found that a nearly dehydrated chabazite absorbed 2°34 per cent of air when left in a desiccator several hours. Such errors must be small in tremolite, for the time of cooling seemed to make 2 Loss in milligrams. We ee 600 80 Temperature. 200 400 ro Oo Tooo Curve showing loss of water in tremolite from Gouverneur, N. Y. The losses are too small but the form of the curve will be seen to resemble closely those in fig. 3, where the work was more exact. no difference in the apparent weight and the substance did not show a tendency to gain on the balance except after heating at the lower temperatures where the losses were small. The heating at each temperature was continued until prac- tically constant weight, i. e., until the loss in 5-6 hours was not more than 1-3 tenths of a milligram. Curves showing loss of water with temperature.—The first work was done on the tremolite from Gouverneur, N. Y. The results are plotted in fig. 2. It was afterwards found that the work was quite imperfect because, instead of heating to a constant weight, what was considered an ample period of time *Day and Allen, this Journal, xix, 127, 1905. Mauselius, Arsbok Sver- iges Geologiska Undersékning No. 3, 1907. W. F. Hillebrand, private communication. Much more complete treatment in paper about to be pub- lished. + Bull. Soc. Min., xxii, p. 15, 1899 ; see also Hillebrand, loc. cit., p. 50. 106 Allen and Clement—Role of Water in Tremolite. was allowed for complete dehydration. The losses are: there- fore too small, but since the general form of the curve is — similar to that of the others, and the form of the curve is the most important point, the curve is given for the sake of com- pleteness. The curve for the tremolite from Ham Island, Alaska, the purest of all the specimens, was determined with the greatest degree of completeness. On account of the tedious nature of the work,* fewer points were determined on the other curves, and in none of them was the dehydration carried to the end, but only so far as to show beyond doubt that all the curves were of the same general form. An excep-- tion should be made of the specimen from Russell, N. Y. (IV-IV, fig. 3), which contains less water and more iron than any of the rest. There is an interruption in the continuity of the curve between 835° and 865° where no water appeared to be given off. A partial explanation may be found in the fact that oxygen appears to be absorbed here; at any rate the powder, at first slightly greenish, becomes brown on continued heating. This absorption might partly offset the loss of the water. The data from which all the curves are plotted are found in Tables 2, 3 and 4. TaBLe IJ.—Loss or WATER BY Heat. TREMOLITE (Ham Island, Alaska). / 2 grams powdered mineral taken. Tem- Time Reena Weight | Loss Time | per- | Weight | Loss ature Oe nrs: £5 26-685) 200 1/23 hrs.)890°| 26°6753] °0112 1152 “ 500°| 26°6798) :0067 || 2/3 “© 1890°) 26°6751) °0114 AS) oe 500°) 26°6798| 0067 || 3)5 “© 1894°| 26°6744| -0121 15 “| 750°|26°6781| -v084|| 4142 “ |890°| 26-6743) “O122 254 “| 750°| 26-6780) -0085 || 1/2 — |920°| 26°6739| “0126 2 | 6 “| 801°-26-6781| -0084|| 216 “ |918"| 26-6722) -0143 9144 13 805°! 26°6778 "0087 37 “ 1918°| 26 6721] °0144 1/5 66 1I845—850° 296°6774 0091 | 1/6 a 933° 26°6701 "0164 2'6 66 | 845° 296°6776) *0089 2\6 ce 933° 26°6680| °0185 1|4 cc 863° 26°6772| “0093 3/6» a0 933° 26'6667 "0198 916 pee 363° 96-6769 0096 46 “¢ 1930°| 26°6658| °0207 ae SH Bene DReene ‘101 || 2 {9882] 26°6646, -0219 Al | Sele ae 613 ~ |930°| 26:6644] -0221 45 it 860° | 266761} :0104 D0 teas) 862°] 26°6763] -0102 1/44“ 877 | 26°6758| -0107 QiQd 6 | 875°| 26°6756, -0109 * Tf any fluorine escaped during the dehydration of the other tremolites, it must have been very little, for the microscope did not detect any lack of homo- geneity, while, when the Gouverneur tremolite was heated 240 consecutive hours at about 950°, the loss was 0°5 per cent in excess of the water and the mineral showed a very evident change. Allen and Clement—Réle of Water in Tremolite. 107 TABLE IJI.—Totat Loss AT DIFFERENT TEMPERATURES. TREMOLITE (Ham Island, Alaska). : P= Percentage Per cent of aus cae OEE loss total water 2 102 hrs. 500° GeO ha 3S 14°5 LO TSUE oie Brot “49 18°3 gi « S05 a Sere 43 18°8 11 “ 1845-850° Td Sead 45 19°6 26 ss SES 5 a LOZA: 5° D2 22°5 a ss Soe ee NO; 9.3> & *o4 23°4 15 es BOF 2D es ‘61 26°38 15 198-9207) 4a § Bae 31°] 32 9S 0—9G3 ee Doan 6 Vee AT°7 Proportional | loss for 2 gr. | 5 84“ POOc ts = NOs Oia ic 4-4 20) 9°5 gi“ SU Oe tee ESO te 52 "26 11:2 10 eo AN Or Deel Oem a6 6°2 31 13°4 re < SOO | epics Rs 6-9 85 14:9 fos < GO0> 22024)“ 8°2 ‘41 Ieee (eS ee YP as on oa ek S°7 “43 18°8 Dae | BOO es Q Dri ace 9°0 °45 19°4 10 fe DIRS 84-2: 66 13°7 “68 - |: 929-6 about | 100 GEO") shT5:8n < 46°3 2°31 100°0 Interpretation of the curves.—lt will be seen that all the curves (fig. 3) rise very slowly and in nearly a straight line until a point approximating 850° is reached, when they bend strongly upward. The point seems to vary somewhat with the composition of the mineral. The curves appear to be smooth ; still one might suspect that so strong a change in cur- vature indicated some abrupt change in the physical or chemi- eal condition of the mineral. The microscopic evidence shows, however, that the crystal form, with such optical properties as can be quantitatively measured—extinction angle, index of refraction—remain almost unchanged. In the purer specimens from Ham Island and Ossining, and in that from Gouverneur, there is no essential difference between the mineral before and after heating, except in the development of bubbles through- out the mass, which increase in number as more water is lost. In the Edwards specimen, a beautiful parting parallel to the base continued exactly as it was before heating. There is a change in the color of the specimen from pink to dull green- ish, winch i is probably due to the absorption of oxygen by the manganese oxide, 1:28 per cent of which is present. This, however, has nothing to do with the bend in the curve, for the color change was just as noticeable in a portion of the min- 108 Allen and Clement—f6le of Water in Tremolite. eral which had been heated below 850°. The water in all the specimens is lost gradually as the temperature rises without any sudden change of properties or loss of homogeneity.* It cannot, therefore, be chemically combined, even though it is : Cee ee 2 a zo : 2.2 ee ee 3° a 20 : eee eal iS 3° a ; OT ae ie ia Io - bdo ° Loss in milligrams per 2 grams xe) al ° Temperature. 200 400 600 800 1000 I. Tremolite from Ham Island, Alaska. II. ye cS OSSiming. aN avs ITI. gd ‘Edwards, N. Y. IV. pe ‘* Russell) NSW. * Making the above-mentioned reservation regarding the specimen from Russell, N. Y. Allen and Clement—Réle of Water in Tremolite. 109 TABLE [V.—TotTau Losses OF WATER AT DIFFERENT TEMPERATURES. : ‘ emper-.| Loss | Percent | Per cent Tremolite from aule eee mg. loss Beco Ossining, N. Y. 113 hrs.| 500° LOe ie end 24°7 2 er. taken eS 840-5 ° 16°5 °82 33°39 29% © | 870-5° Legis) 89 39°4 Gis © | 904° 2G elon 59-2 95S | 928° 37°6 1°88 85°1 Gouverneur, N. Y. 100° ahs? 2 gr. taken 200° 3°3 | 320° 45 | 608° ee 805° 9°8 880° al gyi 920° 13°7 980° 20 Russell, N. Y. 12 hrs. 500° ee "38 22°4 2 gr. taken 10 | 800° 970 | °45 26°2 8 ae eg 9°4 47 27°3 Ain. BABS 9-4 "47 27°3 ~ Co EOS 2 9°4 "47 Mee | Narre 202 12) yal 41°4 Bawards, N.Y.° |: 11.“ | 500° 9°9 ‘50 19°8 2 gr. taken ple O00) tlie tio "59 23°8 plese Odo ale 14 | AGO ii O48 BS COMA ea 14°4 QE ea 2 BES peers} 905° 21°2 | 1:06 | 42-4 given off so very slowly at a temperature of 900°. That it is mechanically held seems entirely improbable, for there is no indication of a spongy structure to be found by microscopic analysis. If capillary pores exist they are submicroscopic. The phenomenon is in all probability molecular ; the water is therefore te be regarded as dissolved, and the mineral as a solid solution. Mineralogists and chemists are wont to regard the retention of water at a high temperature as proof of chem- ical combination, but all ideas agree that a true hydrous com- pound cannot lose water without becoming inhomogeneous. The behavior of tremolite is comparable with that of ‘the zeo- lites in the two essential points above mentioned. In the latter, however, the quantity of water is very much greater, and presumably for this reason the change in birefringence and volume which they undergo when dehydrated is much more noticeable. In the zeolites, as in tremolite, dehydration is very slow and requires in some cases a temper ature of 500° for its completion. 110 Allen and Clement—Roéle of Water in Tremolite. Friedel’s data for chabazite, analcite and mesolite are plotted in figs. 4 and 5. The similarity in the form of the curve of mesolite to that of tremolite will be seen at once. Friedel did not class the zeolites with solid solutions, but Tammann did so, and as such they are now generally regarded. They have remained as a unique class of minerals. The behavior of trem- olite now shows that the zeolites are not the only ones which 4 2. 18 16 I0o 200 300 ! 400 500 Curves showing loss of water from analcite (1) and chabazite (IT) at differ- ent temperatures. Plotted from Friedel’s data. Bull. Soc. Min., xix, p. 376, 1896 ; xxii, p. 14, 1899. contain dissolved water, and it occurred to us that the class might not be uncommon. Dissolved water in other minerals.—From this standpoint, quartz, wollastonite, garnet, adularia, kupfferite, diopside and beryl were studied. They were coarsely powdered to avoid Allen and Clement—foéle of Water in Tremolite. 111 condensation of moisture from the atmosphere, and then heated. Water-clear quartz from Middleville, Herkimer County, N. Y., ground to 40 mesh and dried thoroughly at 110°, lost only 0°10 per cent of its weight on blasting. Trans- parent crystals of adularia from St. Gothard, Switzerland, treated in the same way, lost 0°12 per cent; wollastonite from Natural Bridge, N. Y., lost 0°27 per cent, and lime garnet from Piedmont, Italy, lost 0-26 per cent. These losses are -¢o0 small to be of much interest in this connection, but the remaining minerals, kupfferite, diopside and beryl, were found to contain much lar ger quantities. Kupfrerite.—The specimen examined was from Edwards, N. Y., where it occurs intergrown with the tremolite, from which it was separated by heavy solutions. It is prismatic and fibrous, straw-colored or white, has an index of refraction y = 1°62, and shows parallel extinction. A portion was ana- lyzed and the results are appended. The amount of this was small and the portion which was experimented on was obtained afterwards and was somewhat purer. Calculated for MgSiO; Analysis which kupfferite approaches =) Seema 59°29 phd ee eee 60 per cent i) 03 Ma ac3 354 A) hg 4 Bee 59 EA eee ees 29 (D2) die ae 06 i aes 271 ee 2) 30°98 io) ae as 1°26 ae one: Si) 4) REE See 19 a "20 ee ee 3°80 99°83 O equivalent to F ‘08 99°75 Heated in dry air, the mineral behaved as follows: TABLE V. Charge, 2 grams Time of Tempera- Total loss Loss per day heating ture in mg. in mg. 6 hours 400 5°9 5*9 34 oe ee 5°2 tt 7 65. 2c 5°8 + °6 112 = =Allen and Clement—Réle of Water in Tremolite. TABLE V (continued). Charge, 2 grams Time of Tempera- Total loss Loss per day heating ture in mg. in mg. 4+ hours 600 6°5 “i 6 ce 6¢ 79 a7 6 ce (5 8:0 °8 64 66 ce 9:0 1:0 54 « i 9°4 4 6 ce (55 9°9 “5 64 (%5 66 10°5 6 65 (13 (59 10°9 “4. 63 (35 (<3 11°3 “4 Ga vy 12°0 ‘a 9) Curve showing loss of water from mesolite at different temperatures. Plotted from Friedel’s data. Bull. Soc. Min., xxii, pl. opposite p. 89, 1899. After about 60 hours’ heating at 600°, the mineral had 0) » : therefore lost a 0-6 per cent of water and was still losing | at about the same rate, viz., about °5™S per day. An estima- tion of the vapor pressure of the water over the mineral was made by measuring the rate of flow of the air through the furnace, taking into aceount the water which escaped during the same time. During a day of six hours, this amounted tu 1570°, while the average loss of water was about ‘5™%, which, under these conditions, would occupy a volume of approxi- mately 1°. The partial pressure of the water vapor would Allen and Clement—fRoéle of Water in Tremolite. 1138 therefore be about 1/1500 of an atmosphere or about 0°5™ of mercury. As it appeared that the loss of water under these conditions goes on indefinitely, a new series of experiments was made in which the mineral was heated in air which had been bubbled through 65 per cent (by weight) sulphuric acid having a vapor pressure at 25°, about the room temperature, of 8™™, The results follow: TaBLE VI. Time of Tempera- Total loss Loss per day heating ture in mg. in mg. 6 hours 600° 12°8 "8 = 9 66 12°5 we G= ce ce 13°3 °S ; ee &< 13°7 “4 It will be seen that the average loss is practically the same as before. At this point, a small portion of material was removed from the crucible and examined microscopically. No essential change had taken place in its optical properties. The remainder was now ground fine enough to pass a screen containing 150 meshes to the linear inch, and a new charge of —1°6428 grams was taken. This was heated at 820° in an atmos- phere saturated with water vapor, which at the temperature of the room should have a partial pressure of about 23™™. The following are the results : TABLE VII. Time of Tempera- Loss Loss per day heating ture in mg. in mg. 54 hours 820° 13°9 13°9 ios a6 17°5 3°6 6 ce ¢ 1 8°4 0:9 es a 19°4 1:0 See & 19°6 0°2 64 44 66 19°7 O'1 At this stage, the mineral having lost altogether 2°05 per cent of water, or 50 per cent of the total quantity, another portion was removed and examined microscopically. It still remained kupfferite, though a secondary change had taken place, due probabiy to the oxidation of the manganese. The color had become dull green, and dark brown patches, partially transparent, were visible on some crystals. This secondary change made further experiments useless. The material was probably absorbing oxygen, in which case the total loss would not represent all the water which escaped. The homogeneity of the substance is preserved then during the earlier stages of dehydration at least, though the experiments did not prove that the loss of water was continuous. If this were the case, however, and the water in kupfferite were dissolved, we can 114 Allen and Clement—fole of Water in Tremolite. readily understand how, by the rapid cooling of a melt of the composition Mesi0,,* there is formed an anhydr ous substance having the properties of an amphibole and very closely resem- bling kupfferite i in particular, though the latter contains nearly 4 per cent of water. It has already been stated that the amphiboles analyzed by Penfield and Stanley all contained water. The following are their results : Actinolite, Greiner in Tyrol, H,O =i ef Russell, N. Y., s = WEGe Krager6, Norway, (6 Se a Pierrepont, N. Y., Bieri Hornblende, Cornwallis, N. Y., == leae i. Renfrew, Ont., OS eee r. Ellenville, N. Y., == one More recent results by Blasdale confirm these figures and include also glaucophane :+ Actinolite, Berkeley, Cal., H,O above 100 = 1°784 = San Pablo, Cal. iy inte 5) 6 Tremolite, ag ue (oO Ses Glaucophane, a ve eee Ja 669. 30 673) 6: Seem MgO OMG ROM SM aaa se ui| 3°48 It appears that the only determination affected by this amount of barium is that. of magnesium oxide. The average of the two determinations of magnesium oxide made without remoy- ing barium is 0°27 per cent higher than in the original material. Assuming that barium is precipitated as phosphate [ Ba,(PQ,),| and weighed with the magnesium pyrophosphate in the deter- mination of inagnesium oxide, the product of the factor for converting magnesium pyrophosphate to magnesium oxide, and the weight of barium phosphate derived from five milli- grams of barium oxide, equal 0:0024 grams. This accounts for the fact that the magnesium oxide figure is only 0:27 per cent too igh when 5 milligrams of barium oxide are present. The presence of barium did not affect the accuracy of the determination of calcium oxide. The method for separating barium gives accurate results, and introduces no errors into the other determinations. The purification of barium sul- phate by solution in concentrated sulphuric acid is necessary. In one of the determinations the weight of barium sulphate plus impurities before purification corresponded to. 09 sper cent instead of 0°5 per cent of barium oxide. I wish to thank Prof. H. W. Foote for his advice. Sheffield Chemical Laboratory, New Haven, Conn., May, 1908. / [ety Mixter— Heat of Combination of Acidic Oxides. 125 Art. XV.—The Heat of Combination of Acidic Oxides ‘ with Sodvum Oxide, and the Heat of Oxidation of Chro- mium ; by W. G. Mixtsr. - Sean from the Sheffield Chemical Laboratory of Yale Univ. ] Iris the purpose of the writer to accumulate data on the heat effect of the union of acidic oxides with sodium oxide, and to determine if the position in the Periodic System and the magnitude of the atomic weight of an element have a marked influence on this heat effect. Much has been accom- plished by Thomsen, Berthelot and others who have derived the heat of formation of salts from the observed heat of neu- tralization in solution—a method not applicable in all cases to salts which hydrolyse largely. The reaction with sodium peroxide avoids errors due to hydrolysis and gives fairly accu- rate results, as shown in a previous paper* in which 2Na,O,,C, = 133500° was the observed heat and 132500° that derived from Thomsen’ s data. Asa test of the method two determinations were made and in each rather more than two grams of rhom- bic sulphur were burned in a bomb with an excess of sodium peroxide. ‘The heat effect for one gram was 5275° and 5267° respectively ; mean 5271° and for 32 grams of sulphur, 168670° The heat effect of Na,O, SO, is derived thus — 3Na, OF Het) == 168700° 3Na, O, 30 — 58200° Na,0,S,30 = 226900 8,30 == LOS 2005 Na@O,,SO. (= 123700° From Thomsen’s data we have 29Na,S,40 = 328590°. 9Na,O = 99760°t S,30 = 103200° Na, O, SO, = La OaOs Thomsen used Bekétof’s result for 2Na, O; deForcrand§ considers it too high and that 91000° is probably more accu- rate. The calculated heat effect of Na,O , SO, will not be changed by using this number. S. W. Parr| mentioned that oxygen is sometimes liberated in combustion with sodium peroxide and the writer has found * This Journal, xxiv, 154 + Thomsen, There ee Untersuchungen, ii, 204. t Ibid., iii, 232. §C. R., exxvii, 1449. IJ. Am. Chem. Soc., xxix, 1606. 126 Mixter—Heat of Combination of Acidic Oxides. that it is necessary to allow for oxygen taken up or set free. The correction of 1°2° for 1 milligram is based on the heat of formation of sodium peroxide from the oxide which according to deForerand* is 19390°. For oxygen liberated there is a loss of heat which is to be added, and “for oxygen absorbed the gain is to be subtracted from the heat observed. The best way to find the change in the oxygen content of the bomb is to connect it after a combustion with a manometer. To avoid excessive detail in the data given of the work only the heat equivalent of the oxygen liberated or taken up is stated. Most of the work was done with a bomb of 500% capacity and in the work on chromium the oxygen correction was large. To obviate the correction or make it insignificant a sterling silver bomb of 100° capacity was made. It proved to be admirably adapted for calorimetric work with sodium peroxide. The water eqnivalent of this bomb and calorimeter can was 109°. The large bomb was used in the experiments in which the water equivalent was over 3200 grams, and the small one in those in which it was less than 3100 grams. When a molten mass is in contact with the cold bomb it solidifies at once and the combustion is not complete. Hence it is better to put the peroxide mixture into a thin silver cup which is in contact with the inner surface of the bomb at only a few points. In order to make ignition certain the bomb was filled with oxygen, as it was found that with air in it the burning iron often failed to start the combustion. The carbon used is the firiely divided form made from acetylene and the heat effect of its reaction with sodium peroxide is taken as 11100°F per gram of carbon. The carbon gives the temperature needed to effect the combustion of other substances and also reduces the peroxide to the sodium oxide required in the reaction with an acidic oxide. The initial temperature of the experiments was between 18° and 19°. Boric Oxide. The heat effect of the combination of boric oxide with sodium oxide has been determined. The oxide used in the experiments, made by fusing boric acid in a platinum dish, was pulverized and weighed in a stoppered bottle. It was exposed to the air as short a time as possible on account of its hygroscopic character. The reaction of boric oxide on an excess of sodium peroxide yields the orthoborate thus : 3Na,0, + B,O, = 2Na,BO, + 30 No perborate results, as shown by the following experiment : *O. R., exxvii, 574. + This Journal, xix, 434. 127 Mixter— Heat of Combination of Acidie Oxides. A mixture of 1:314 gram of boric oxide and 5 grams of sodium peroxide was heated in an ignition tube. The loss in oxygen was 0-950 gram ; calculated 0-901 gram. Were perborate only formed no oxygen would have been liberated, and if metaborate was the product only one-third as much oxygen would have been set free. Experiments. 1. 2 3 PrarterOxide.. 2. 2 2. 1°269 1°858 1°8862 grams Jl) 0-661 1-080 1°0384 °_“ Sodium peroxide. ..-- a 20° 20° ss Water equivalent of 572 Gill ee ees 3435° 3465° 3431" 3 Temperature interval 23019" 4°280° 4°149° Heat observed -.-_---- 9202° 14830° 14235° Heat of oxidation of Sean. ee 8 SS —7337° —11988° —11526° Heat of oxidation of iron for ignition ---- — 80° — 64° — 80° Heat absorbed by oxy- Gem Siven: Off 0.2.2 “+ 96° + 60° +131° 1881° 2838° 2760° For 1 gram of B,0, uniting with sodium Cit Cele eee 1482° 1522° 1463° The three results are respectively 1482°, 1522°, and 1463°. The average is 1489 for 1 gram and 104200° for a gram mole- eule of boric oxide reacting with sodium oxide to form sodium orthoborate. Note on Boron.—The only thermal data on boron are those of Troost and Hautefenille* and Berthelot,+ who determined the heat of formation of boron trichloride and tribromide and the reaction of these halides with water, and from the results they calculated the heat of formation of the trioxide. Since they give no analysis of the boron used, the purity of it is doubtful. Moreover, Moissant has shown that boron prepared by reducing the oxide with sodium or magnesium is not pure. He removed the magnesium which the impure boron contained by fusion with boric oxide, taking precautions to prevent forma- tion of nitride. It may be added that while at work on boron it was learned that another investigator, whose results are not published, considers that none of the methods described in the literature yield pure boron. Some of the observations of the writer in regard to the burning of boron are interesting and * Ann. Ch. Phy. (5), ix, 74. + Ann. Ch. Phy. (5), xv, 215. ¢ Ann. Ch. Phy. (7), vi, 296. 128 Mixter—Heat of Combination of Acidic Oxides. may be briefly stated. When a mixture of impure boron and carbon was burned in oxygen under pressure the boric oxide produced volatilized and condensed as a fine white powder, and considerable boron carbide was formed. Impure amorphous boron and also erystalline boron containing aluminum reacted with explosive violence with sodium peroxide. The heat of oxidation of boron may, therefore, be easily found by the per- oxide method when pure boron is available. Aluminium Oxide. The amorphous oxide used was prepared by igniting a pow- dery form of hydroxide. For crystalline oxide, crystals of corundum were taken. These were pulverized in a steel mor- tar, the powder digested with hydrofluoric acid, then sulphuric acid was added and the mixture heated until fumes of the latter acid escaped. Then the oxide was washed. It was white and was found to be free from lime, iron and silica, and to contain a trace of magnesia. Both pr eparations were floated j in water and only the more finely divided portions retained. In order to deterniine the alumina remaining after a combustion the silver vessel containing the solid product was placed in half a liter or liter of cold water. The fusion dissolved rapidly owing to the presence of sodium peroxide. After solution the silver piece was removed and an excess of nitric acid added and the alumina filtered off. It was washed first with water and then with ammonia to remove any silver chloride present. This residue of alumina was deducted from that taken for an experi- ment. There is nothing in the literature regarding the solu- bility of ignited alumina in alkaline solutions other than the statement that the more intensely the oxide is heated the slower it is taken up by alkalies. In order to learn if the residue of alumina mentioned in the experiment is likely to dissolve so ~ as to cause an error, the following tests were made with finely divided alumina which had been heated in a platinum crucible over a large blast lamp. In one test 4 grams of alumina and 20 grams of sodium peroxide were mixed and about 400° of hot water were slowly poured upon the mixture. The violent reaction between the peroxide and water gave at once a boil- ing concentrated solution of sodium hydroxide. After a few minutes an excess of nitric acid was added and the solution filtered. The alumina found in the filtrate was 2°5 per cent of the quantity taken. In another test about the same quan- tities of the mixed oxides were added to the surface of warm water. In this case no alumina went into solution. Since the solid residues from the combustions were chiefly sodium ear- bonate and aluminate, and contained much less sodium peroxide Mixter— Heat of Combination of Acidic Oxides. 129 and alumina than used in the tests mentioned, it is evident that any error due to solubility of alumina is insignificant. Moreover, varying portions of sodium peroxide do not affect thermal results. Experiments. 1 2 oe oxide (amorphous) =—- 3'357 grams 4°313 grams “Sy. ik Festdues 2 — 026 “ == GO ‘Sao Teackine. a= = Sioa sv ANG 5 0 wl ELD DLs a a ile eon eee loa iS aio ale ae. 66 Sadia peroxide. =.= 222-2 - 21° ee 25° e Water equivalent of system_. 3528: 0a o e Temperature interval .--- ---- 4°245° 4°479° fleatwobserved 22-2 5-22.2 221. 14977° 15941° «of oxidation of carbon_. —13520° —14363° 6é ce 66 66 irov a3 005) HOS — 60° 1397° 1518° For 1 gram of amorphous alumina combining with sodium oxide 419° 365° 3 + turaiorum oxide (crystalline) 4°517 grams 4:038 grams im residue =: Sl LGA os ma ote een as Sy reactine 2 3343 3 =i SHO, 26S 2 ei es i Os809etu, 354 O-814, 6 Sodium peroxide-..- ---- d Iie 3 ve = Water equivalent of sy stem... 3087 ee USO: 4 Wemperature interval... _..- 3°225° 3°355° Blea observed. 25 on Le 9956° 10185° ‘“¢ of oxidation of carbon-- SS 9035° 6¢ &é ce 6¢ TOM... <2 EOC — 64° 945° 1086° For 1 gram of crystalline alumina Doone 310° In experiment 3 the pressure in the bomb was 16™ higher after the combustion than before, and the calculated correc- tion for the oxygen set free was 30°. This is not included in the result above, as changes in pressure were not observed in experiments 1, 2 and 4 with alumina. In experiments 3 and 4 about two-thirds as much carbon was used ag in 1 and 2. In order to find if the ratio of the carbon to the alumina infli- ences the result, a calorimetric test was made in which the amounts of carbon and amorphous alumina were nearly the same as in experiments 3 and 4. The result was the same as in 1 and 2, as shown in the following experiment : 130 = =Mrxter—Heat of Combination of Acidic Oxides. Experiment 5. esi oxide (amorphous)= V2. ites 4°237 grams ge TA ROSIE. fee ee Nee ees —0:195 « rectee es RESCH Me ee eae 4°042 0% Cat oon (2 RR ee eS he 17 ee Ne. Te eee On 78 inti Sodium: peroxide sateasy le. =e eee so) eae 14° Woater, equivalent of system: 22e) 2. ee 2947" Temperature ciiGeryal 2 os eae bes iene 3°484° Heat observers: he. oes ae Nie ae 10267 * .ofroxidation, Of carbon 222 212 72 2eee — 8635° 14 66 Co TEE ONG Wee CaP et Graco er —64¢ 1568° For 1 gram of amorphous alumina.___.._--. Jee In the reaction, 2Na,O,+C=Na,CO,+ Na, 0, 1 part of car- bon produces 5 parts of sodium oxide. In the formation of sodium metaluminate according to the equation, Al,O,+Na,O =%NaAlO,, the ratio of the quantities of the oxides is 1 to 0° 6, ak for the formation of the orthoaluminate three times as much sodium oxide is required. The following table of results, giving the calories evolved for one gram of “alumina and the ratio of the alumina to the sodium oxide, shows that sodium orthoaluminate could not have been formed in experiments 3, 4 and 5, as there was not sufficient sodium oxide produced to form it. It should be noted in this connection that too little oxygen was set free in the combustions to indicate any material difference in the quantity of sodium oxide formed. Na,O Calories for required to form No. of 1 gram Na,.Q 9 ——-—+— — Exp. Al.Os C Al.Q3 ' formed NaAlO, Wa aon 1 419 1-2 3°38 (amor.) Ore 23 a 2 365 IES: 4°2 s 6°5 2°5 7°6 3 283 - 0°8 3°3 (crys.) 4° 2s 6° 4 310 0°8 3°2 5 4° iS 6° 5 388 0°8 4° (amor.) 4° 2°4 72 It is evident that the chief product in the reactions 1s sodi- um metaluminate, but possibly mixed with other aluminates. Assuming that the reaction was essentially the same in all of the combustions, we have for the heat of combination of 1 gram of amorphous alumina with sodium oxide a mean of ex- periments 1, 2,and 5 of 390°, and for a gram molecule 40,000°. For crystalline alumina it is 30, 000°, Hence the transformation of the amorphous alumina into the crystalline form is accom- panied with the heat effect of 10,000°. Mixter— Heat of Combination of Acidic Oxides. 1381 Chromium. The thermal constants of chromium are of considerable im- portance. In Landolt and Bernstein’s Physikalisch-Chemische Tabellen, p. 439, the statement is made: Die Bildungswaerme der Chromverbindungen kann nicht angegeben werden, weil keine Reaction untersucht wurde, an der “metallische Chrom betheiligt ist. Then, too, it is an interesting element to study, forming basic and acidic oxides, both of which yield stable salts. Since chromium and its sesquioxide do not burn in oxygen, it is necessary to resort to indirect methods in deter- mining the thermal constants of chromium compounds. This has been done in solutions by Thomsen, Berthelot, and others. The sodium peroxide method i is a better one, and the reactions are Cr + 3Na,0O, = Na,Cr0, + 2Na,O Cr,O, + 3Na,O, = 2Na,CrO, + Na,O No perchromate is formed as shown by the following result: A mixture of two grams of chromium trioxide and 6 grams of sodium peroxide was placed in an ignition tube closed with a calcium chloride tube to absorb escaping water. On heating gently the mixture glowed. The loss in weight was 0°386 gram ; calculated 0- 32 gram. If perchromate had been for med, less oxygen would have been given off. Moreover, it is im- probable that sodium per chromate can exist in a molten mass containing sodium oxide. To Dr. C. H. Mathewson I am indebted for a fine specimen of crystalline chromium made at the Goldschmidt factory. The metal was pulverized in a steel mortar and the powder was digested with hydrochloric acid to remove the iron. Analysis proved it to be free from aluminium and silicon and to con- tain 0-7 per cent of iron. The last may have been from the mortar. Metallic chromium as a very tine powder will un- doubtedly burn readily with sodium peroxide, but heat was necessary to effect the reaction with that used, and carbon was therefore added to the mixtures. After a combustion the product was dissolved in water, the solution made acid with nitric acid and the metal remaining was separated and weighed. Experiments. 1 2 3 Metal taken i225 Fk 2-000 2°5000 2°500 grams Pig t EEO UTE Lys ee 0:061 0°0043 0:021 e ore DEE. «Yo sche ieee 1°939 2°4957 2°479 sc Maeresisrtlin: 3 he Ss ers 1°925 9°4782 D-AGilow 3 Micertiees ot 6) 2 es Stak gad ae 0°014 0-'O175 OF oa emia 0 A se J en 0°2456 0°3767 0°3800 <“ sodium peroxide. —.--£22 .~ 20° 20° 20°5 é Water equivalent of system 3428: Bd 14: 3487" 5 Temperature interval . ----- 2°601° =3°510° 3°520° 182 Mixter—Teat of Combination of Acidic Ouides. 1 2 3 Meahiobsenved eo: see. as 8916° 12334° 12274¢ “ oxidation of carbon.. —2726° —4181° —4218° 3 a vO aM metal and for ignition — 94° — 90° — 92° “ due tooxygenabsorbed —288° —492° 472° | 5808° 7571° 7492° For1 gram of chromium burn- ing with sodium peroxide-. 3012° 3055° 3044° In the experiments 1, 2, and 8 a bomb of 500° capacity was used. As the correction for oxygen absorbed was large, a determination was made with the 100% bomb. The residue insoluble in nitric acid was collected on a Gooch filter, dried and its weight found. Next the carbon was burned off, and finally the weight ot unburned chromium was obtained. The pressure in the bomb was 29 less after the combustion than before. Experiment 4. Metal taken: as alee ee ee eee 02°0520 grams eee UM OMEN: Oe ae ree en es “OOLS a COS Se TR ETAL 52. SF Goh ict eal 8 aes pee 2-05 055) 68 Chromiuin 206s a ee ee ee 2036) ae SD gra SS Cg 0:0143 34 Carbon stakem io soo ia ee ene 0:36 a0 Unb urMe dl 7k nother es 0:0017 oe 36 |ONE Ue Octo lpedtur Nn ake wragrs Pts Teale Sree 0°3144 Ss Sodium, peroxide: hs 2o oa ee eee elk oe Water equivalent ofthe systems 4 = 2984: ye Temperature mterval 935 22 ee ee 3°282° Heat observed. Veer weir ae eee 9793° OOF OMIAAWOn, OLLear bone res eee —3489° pore se es ce LOM Mae MMe tale wee — 23° ugier ee ci for somitionay — 26° dune to oxygcen/absorbed s3525aa-= — 52° 620385 For 1 gram of chromium geese eee 3046° Evidently the correction made for oxygen taken up in the first three experiments was correct. The results are 3012°, 3055°, 3044°, and 3046°. In the first one the correction for oxygen was not made with the care that it was in the others, and hence it is better not to include it in the final value. The mean of the other figures is.8048° for the reaction of 1 gram of crystalline chromium with sodium peroxide, and for 52°1 grams it is 158800°. Mixter— Heat of Combination of Acidic Oxides. 188 Chromium Sesquioxide. Amorphous chromium sesquioxide was prepared by heating an hydroxide. The crystalline oxide was made according tO.” Ditte’s* method of melting together equal parts of pure potas- sium dichromate and sodium chloride until the evolution of oxygen ceased. After cooling, the soluble portion of the prod- uet was dissolved in water and the crystalline powder obtained,, washed, digested with hydrochloric acid and washed again. The crystals were so small that some passed throngh filter paper. Under the microscope they appeared unmixed with any amorphous substance. Tested by the spectroscope the preparation proved to be free from sodium and potassium. The following results show the heat of the reaction between chromium sesquioxide and sodium peroxide: Experiments. 1 2 3 4 oe Besquioxide. 2... 4°236 3°831 7°436 6576 grams ee unburned 0°062 0°018 0°805 OPO Gers F °F * burned - - 4-174 3°813 6631 GAG << (LiL LL 2 hee eee eee 0°516 0°456 0°4015 OA bie K eee burned ——— 0°007 0-006 0°151 COI2 4S SOMERS. 2 8 ee 0°509 0°450 0°2505 Of sc perm@emeiopieroxide __-..__-..._.- a 15° 21° 20° e Water equivalent of system -.. 2828: 2902- 2936: 3077" eB Temperature interval --_--- eos SO fp see Od 2022" 3°038° Heat observed -___.. .---- Se 8623° 7844¢ (ts 9348° “ of oxidation of carbon. ---- —5650 —4995 —2780 —4584° Beer * of iron forignition —16 —10 —62 —48° ** due to oxygen absorbed or | LTTE ES es eae —69 =F. +14 — 30° / 2888 2769 5751 5686° For one gram of chromium sesqui- oxide, reacting with sodium + 692 726 867 878° 2 es ee ee Crystalline sesquioxide was used in experiments 1 and 2 The results are 692° and 726° and a mean of 709° for the heat of the reaction of 1 gram of crystalline chromium sesquioxide with sodium peroxide. Fora gram molecule it is 108U00°. The experiments 3 and 4 were with amorphous sesquioxide. The results are 867° and 878° and the mean is 872° for the heat of the reaction of 1 gram of amorphous chromium sesqui- oxide with sodium peroxide. For agram molecule it is 132000°. * C. R., exxxiv, 336. Am, Jour. Sct.—FourtH SEriIges, Vot. X XVI, No. 152.—Aveusr, 1908. 10 i384 Mixter—Heat of Combination of Acidic Oxides. In experiment 3 the considerable quantity of unburned sub- stance is due to the fact that the mixture was in contact with the cold bomb and was not in an inner silver cup. Chromium Trioxide. The chromium trioxide was free from sulphuric acid and sufficiently pure for the purpose. It was fused, allowed to cool in a desiccator and then coarsely pulverized. A little ses- quioxide remained after a combustion. The amount of tri- oxide equivalent to it was deducted from the trioxide taken. As the heat of oxidation of the sesquioxide to the trioxide is small the error due to reduction is insignificant. The results following show the heat effect of the combination of chromium trioxide with sodium oxide. Experiments, 1 2 Chromium trioxide taken___.__._-- 4°000 4°782 grams oc ve RECUCed 2 42a ee 0:097 0°176 me ee sd combined 922. -22 3°903 4-606" 2." Car Wome ee eae ea et ae eee 0°515 0°523 66 Sodium perouder ero at - eee 10°6 16° “ Water equivalent of system __.-.... 3408° 3504: e Temperature anterval 25.22.22 422 2°514° 2°595° fleet Observed sae erties serene 8567° 9093° Of oxidation ol canbonm 25 2. — 5676 —5805° ane rf ‘¢ iron for ignition —62 —625* «¢ absorbed by oxygen given off +187 +324° 3016 3590° For 1 gram of chromium trioxide - - 113 Take The mean of the two results is 772° for 1 gram and for a gram molecule of chromium trioxide it is 77000°. The heat of formation of the oxides of chromium is derived from the above data, thus: oma Oo 4+ Cr = Na CrO.) 2Na Oe) eee 158800° aNa Olu: 80. — 3Na, On 4 ot Pies Seen 58200° NOs ier + 30 = NaC rOrer te an eee 217000° NaO +. €0rO, = Na,CrO) 4. ee ee Cr. 4. 30% = CrO: 4 ee ee Oe Mixter— Heat of Combination of Acidic Oxides. 135 3Na,O, + Cr,O,(crys.) = 2Na,CrO, + Na,O + 108000° a Orch sO Na Oona eh Poe Le 58200° 2Na.0 + Cr,0O, + 30 = 2Na,CrO, +. -..---- 166200° men a OOO )e =e 2 NTO) er ee oe le 154000° GrOxXcrystalime) + 307 = 2CrO, +.) 2s.-.-- 12200° MEO Cr Oe 280000° Weel (erys) 4-930 = FOr Oo es. 3s 12200° cere oO) =. Cr O-(erystallime)) 4°. 267800° _ Bees 3: = Cr.O.(amorphous) +. = --...2.--- 243800° Amorphous CO, = crystalline-Cr,O, + ..--..-.-« >. 24000° Thomsen* found for Cr,H,O,, O,, aq = 2CrO,+18913° and Berthelott+ gives Cr,O, precip. +O +eau.=2CrO, crys. +16400°. The changes in the oxidation of the hydroxide are different from those in case of the oxide and hence the above results ean not be compared with that of the writer. They all, how- ever, show that the heat of formation of chromium trioxide from the sesquioxide is small. Berthelott derived from reac- tions in solution the following: OCrO,+K,O=K,CrO,+47800°. This appears to be too low when considered in connection with the 77000° found for Na,O+CrO, = Na,CrO,, since the heat of formation of potassium salts is commonly greater than | that of sodium salts. Tungsten. Metallic tungsten used in the work was prepared by reduc- ing the oxide with dry hydrogen at the highest temperature attainable in a gas combustion furnace. Even after ten hours a little water came off, showing that the reduction was not complete. During the first hours occupied in the reduction a little ammonia was formed from the atmospheric nitrogen contained in the hydrogen, but the reduced metal was free from nitrogen. The test was made by heating a mixture of the metallic powder and soda-lime. No ammonia was given off. The tungsten present in the metallic state was determined by finding the increase in weight when a weighed amount of the metal was oxidized by heating in air and finally in oxygen. The tungsten equivalent to the oxygen taken up was 98°14 per cent. [ron was present to the extent of 0-07 per cent, leaving 1:79 per cent by difference of oxide of tungsten as WQO,. Undoubtedly only the lower oxides were present and there * Thermochemische Untersuchungen, ii, 464. + Thermochemie, ii, 272. ¢ Ann. Ch. Phys. (6), i, 195. 1386 = Mixter—Heat of Combination of Acidic Owides. was iess than 98 per cent of metallic tungsten in the prepara- tion, but the lower oxides give heat when oxidized. Hence we may assume without essential error that the thermal effect is proportional to the amount of tungsten, which is equivalent to the amonnt of oxygen taken up. After each calorimetric experiment the product in the bomb was dissolved in water and the small residue remaining was separated. It dissolved completely in nitric acid, showing that no metallic tungsten remained. The following experiments give the heat of burn- ing tungsten in sodium per oxide; Experiment 1. Punosten C241 x) 00814 te eee 8088 grams W ater equivalent of system ........____- 3596" < Bodiam peroxidegs sey. 2, ee eee One ¢ Mempevature inter vale ss see eee 2°859° Heat, observed 3596 )<-2°8593) = oe = 10281¢" “« of oxidation of iron in tungsten and Used Or someon se 1k te eee —157° 10124° ord cram ok tumor sten): esse 2 ee eee 12528 Hxperiment 2. ihunesten.. 6563) < 10-081 am eee 8377 grams Water equivalent of isy stem 992 === eee 3552: “¢ Sodium peroxide: 4. 5.06 5 i as eee 25° “ Temperature mmtervall: =. sas Sees Gee 3°020° tleat observed 35523" x 02) eases 10727° ‘‘ of oxidation of iron in tungsten and used: for Jomitiona-se 55 = eae —125° 10602° Mor pram. of tumestenca ae) sae ae 1267° The average is 1260° for 1 gram’ and 231200° for 184 grams. For the heat effect of Na,O + WO, we have 3Na,0, +. W = Na, WO, + 2Na,0 - 22> 2eneoue aiNa O30 ==) BINa OF ole aoe PANS Ao 58200° Na OTE eW + 130, ="Nas WON. oun 291000° We Og =! WO) eee eee 196300% Nao Seay = Na WO se) eee eee 94700° * Delépine et Hallopeau, C. R., cxxix, 600. Mixter—Heat of Combination of Acidic Oxides. 187 Summary. eNO i Oe INA Oe SS LL O4200° Na,O + AI,O,(amorphous) = 2NaAlO, + --- ~40000° Na,O + AI,O,(crystalline) = 2NaAlO, + --- 30000° Al,O,(amorphous) = AI,O,(crystalline) + -- 10000° er Or Gy Na OrOrer oil sacee Pasi s ie 77000° Mie ge = Ore ene s 228 TEOOUCS 2Cr + 30 = Cr,O,(amorphous) + ~-.-.----- 243800° ere oO) = Cr O- (crystalline) - 2 -...-2.5 _267800° Gr_O, (crystalline) + 30 = 2CrO, +. ------- 12200° €rO-(amorphous) +: 30 = 2CrO, + °.-:---- 36200° mom NWO. = Nap W Oi) ees ola 94700° The results given have been obtained with substances at hand and it is the intention to complete the work as far as possible on the fourth, fifth, and sixth groups. 138 Phelps and Weed—Acids and Acid Anhydrides. Art. XVI.—Concerning Certain Organic Acids and Acid Anhydrides as Standards in Alkalimetry and Acidimetry ; by I. K. Paetes and L. H. Weep. [Contributions from the Kent Chemical Laboratory of Yale Univ.—clxxix. ] In a former paper* from this laboratory it has been shown that, with cochineal as an indicator, succinic acid may be used as a standard fora decinormal ammonium hydroxide solution quite as accurately as may a decinormal solution of hydro-— chlorie acid, the standard of which is determined gravimetri- cally as the silver chloride. In this paper results are given which show that, in presence of phenolphthalein as an indica- tor, pure sodium hydroxide in solution and also pure barium hydroxide in solution may be determined similarly with suc- cinic acid, succinic anhydride, malonic acid, benzoic acid, phthalic acid and phthalic anhydride, as standards. And, further, it isshown that these organic acids and acid anhydrides react with these alkaline solutions so that each may be used as a standard in acidimetry and alkalimetry with the same exact- ness that is found when these alkaline solutions are titrated in the well established manner with decinormal hydrochloric acid, standardized gravimetrically as silver chloride. For the work given here a solution of hydrochloric acid was made up approximately decinormal by diluting the chem- ically pure acid of commerce in the usual manner. The exact strength of the hydrochloric acid solution was deter- mined by precipitating definite amounts of it in a platinum dish, in some cases, and in a glass beaker, in others, by an excess of silver nitrate, in presence of a few drops of dilute nitric acid. In each case the precipitate of silver chloride was allowed to stand for twenty-four hours before filtering on a weighed asbestos felt in a perforated platinum crucible. The volume in which the silver chloride was precipitated was such that after the precipitation was made it amounted to about 250 cubic centimeters. The sodium hydroxide solution was made up to correspond approximately to the hydrochloric acid solution, by diluting with distilled water, freshly boiled, pure sodium hydroxide, prepared by the action of water vapor on metallic sodium according to the, method of Kister.+ The barium hydroxide was prepared pure by crystallizing twice commercial barium hydroxide out of hot water, washing the crystals after each purification with alcohol. A solution, approximately decinor- mal, was made by dissolving these erystals i in a suitable amount *~ This Journal, xxii, 201, + Zeitschr. anorg. Chem., xli, 474. Phelps and Weed—Acids and Acid Anhydrides. 139 of water and filtering into a closed bottle before diluting with freshly boiled distilled water. Both the sodium hydroxide solution and the barium hydroxide solution were kept in closed bottles, each connected with a three-way-stoppered: burette in the usual manner. ‘These solutions were protected from the action of carbon dioxide in the air by soda-lime tubes. In all the experiments recorded in the tables given below, definite portions of the organic acids and acid anhydrides, in most cases, were treated with distilled water and the solution of sodium hydroxide or barium hydroxide was introduced into these solutions by carefully drawing from the burette until the appearance of color in the solution, due to the presence of phenolphthalein as indicator, showed the reaction to be com- plete. In a few cases, the treatment was special, as is described. Pure succinic acid was obtained by boiling succinic ester, whose purity was established by the fact that it distilled within one-fifth of a degree, on a return condenser for four hours with water containing a few drops of nitric acid. This solu- tion was evaporated to crystallization and the solid product, TABLE I. HCl HCl value of value of Theory in Error Sueciniec Succinic NaOH BaO.H, terms in terms No. acid anhydride used used. of HCl of HCl erm. erm. erm, erm, erm. erm. ie 0°2000 fai 0°1236 ee Ae 0°1235 ‘0001 + II 0:-2000 evi; 0°1238 Saye 0°1235 "00038 + III 0-2000 SEES Onk2 37 Eo Ree 0°1235 "0002 + IV 02000 Sane fs 0°1236 eden. 0°1235 ‘0001 + V_ 0:2000 Bers 0°1236 Pen. 0°1235 ‘0001 + VI 0:2000 ee en 0°1237 lores 0°1235 "0002 + VII 0:2000 ee ete 0°1237 Sees 0°1235 ‘0002 + VIII 0-2000 a ace 071237 Seca 0°1235 "0002 + IX 0:2000 Sess OES 7 Lp Ter 0°1235 0002 + X 0°2000 5 ae 0°1237 sees 0°1235 ‘0002 + XI 0°2000 Cee bgS ier. 0°1238 0°1235 0003 + XII 0-2000 aA ees 0°1237 0°1235 ‘0002+ XIII 60-2000 Eres pee PAO We 071235 ‘0000+ XIV 0:2000 en Edith 0°1236 0°1235 0001+ XV 22 TE OFZ000 0°1458 aoe nid 0°1458 ‘0000+ Bel jt E2000 0°1458 eee 071458 -0000+ XVII zee OcOUD 0°1459 rake 0°1458 ‘0001 + XVIII 2 C2000 0°1458 ae 0°1458 ‘0000+ XIX ta ns eo OOOO Ba Oo 1457 0°1458 ‘0001— BA ae, 622000 uid 3 0°1456 0°1458 ‘0002 — XXI 222 + 02000 fy hiaee 0°1459 0°1458 ‘O001 + XXII eee OF Z2000 ea 0°1458 0°1458 ‘0000+ 140 Phelps and Weed—Acids and Acid Anhydrides. after the removal of the mother liquor by filtering, was recrystallized from distilled water. _ After these crystals had dried in the open air to constant weight, it was found that on standing over sulphuric acid in a desiccator the weight remained unchanged. For the preparation of succinic anhydride, com- Peroni succinic acid was treated with an excess of acetyl chloride and heated on a water bath with a return condenser at 60°, as long as bubbles of gaseous hydrochloric acid were evolved from the hquid. The material, which separated out on cooling, was recrystallized from ethyl acetate. These erys- tals of succinic anhydride were then washed with absolute alcohol and were dried to constant weight over sulphuric acid in a desiccator. The succinic acid used in experiments VI, VII, and VIII of Table I had been dried for more than a year in a desiccator containing sulphuric acid, while that used in experiments IX and X of the same table had been dried for the same length of time over calcium chloride in a desiccator. It is evident from these experiments that succinic acid dried in desiccators over sulphuric acid or calcium chloride for long periods of time is unaffected. Owing to the considerable length of time that is taken by succinic anhydride to dissolve in water even in the presence of some alkali, experiments XX and XXII of Table I were slightly modified. In these the solution was heated until the anhydride completely dissolved before any of the alkaline hydroxide was added. Malonic acid was prepared pure by heating for some hone between 50° and 60° on a return condenser malonie ester, which boiled between limits of two-tenths of a degree, with water in the presence of a few drops of nitric acid. The vol- ume was then concentrated, keeping the temperature of the solution below 60° until crystallization began, the erystals ApNisriioh IB HCl value HCl value Theory Error Malonic of NaOH of BaO.H. in terms in terms No. acid used used of HCl of HCl germ. erm. erm. erm. erm. if 0°2000 0°1404 eee 0°1402 0002+ II 0°2000 0°1403 BY a 0°1402. ‘0001 + Ili 0°2000 0°1402 a. 0°1402 ‘0000+ lV 0°2000 0°1401 apt 0°1402 ‘0001— Vv 0°2000 igs 0°1401 0°1402 ‘0001— VI 0°2000 eee 01400 0°1402 "0002 -- Vil 0°2000 Ne 0°1402 0°1402 "0000+ VIII 0°2000 ieee 0°1400 0°1402 °0002— Phelps and Weed—Acids and Acid Anhydrides. 141 filtered off and recrystallized out of boiling water. The pure malonic acid was then allowed to come to constant weight over sulphuric acid in a desiccator. To obtain pure benzoic acid, benzoic ester was treated with sodium hydroxide in excess, and acidified with hydrochloric acid. The benzoic acid thus precipitated was crystallized twice from water and dried to constant weight ina desiccator over sulphuric acid. TaBLe ITI. HClvalue HCl value Theory Error Benzoic of NaOH of BaO.H. in terms in terms No. acid used used of HEI of HCl germ. eTm. erm. erm. gTm. 1 0°2000 0°0598 ae 0°0597 ‘OOO01 + II 0°2000 0°0599 she 0:0597 0002+ Ill 0°2000 0:0597 nll 0°0597 “0000+ IV 0°2000 0°0598 ss a 0°0597 “0001+ Vv. 0°2000 Bee Mate 0°0598 0°0597 ‘OOO1L + VI 0-2000 aes 0°0597 0°0597 70000 + VII 0°2000 rik & 0°0597 0°0597 ‘O000 + Vill 0°2000 es 0°0597 0°0597 ‘0000 + In all of the experiments in Table IT], alkali in amount nearly sufficient to neutralize the acid was run into the flask, which was then heated. This aided materially in securing the solu- tion of the benzoic acid in the water and did not necessitate raising the solution to the boiling point. TaBLeE IV. 1 HCl Theory Error Phthalic valueof value of in terms in terms Phthalic anhy- NaOH BaO.He of of No. acid dride used used HCl HCl erm. erm. erm. erm. erm. erm. I 0°2000 See 0-0880 pies to 0°0878 0002 + II 0°2000 eet 0-'0880 ie ar 0:0878 "0002 + Il 0°2000 at 1050879 Sie 00878 -0001+ IV 0°2000 te OSES Bers 0°0878 ‘0000+ ee 02000 Clete. Paar 00876 00878 -:0002— VI 0°2000 ees apes 0-O877 0:0878 ‘0001— Vil 0°2000 seeoed Nd seat 0:0878 0'0878 “0000+ Vill 0°2000 bee! nears 00879 0‘0878 “OOO01 + PX. ae 0°2000 0°0986 ee 0'0985 “0001 + xX ee 0°2000 0°0985 wits 0°0985 0000 + XI Tere 0°2000 0°0986 ERY ip 0°0985 0001 + XII L068 0°2000 00987 Paes 0°0985 "0002 + XIII wna 0°2000 eee 0:0986 070985 ‘0001+ XIV wae 0°2000 ise 00985 0°0985 ‘0000+ XV AR, 0°2000 eo 0°0986 0°0985 ‘0001+ a ree 0°2000 2 aa 0'0987 00985 ‘0002 + 142 Phelps and Weed—Acids and Acid Anhydrides. Phthalic acid was prepared by boiling in distilled water some commercial phthalic anhydride. The solution was filtered while still hot; the crystalline product obtained on cooling was separated by filtration, air-dried, and finally dried to con- stant weight in a desiccator over sulphuric acid. The phthalic anhy dride was prepared in a state of purity by distilling 7 vacuo the phthalic anhydride of commerce. The product ob- tained was dried to constant weight in a desiccator containing | sulphurie acid. In Table IV, experiments II and VIII alone were carried on at ordinary temperatures. In the other experiments in this table the titrations were all performed after heating the solu- tion until the phthalic acid or the phthalic anhydride used had entirely dissolved. It is evident from the results recorded in the four tables that succinic acid, suecmic anhydride, malonic acid, benzoic acid, phthalic acid and phthalic anhydride may be used with great exactness as standards for decinormal solutions of sodium hydroxide and of barium hydroxide. As a standard for a solution of barium hydroxide, these organic¢ acids and acid anhydrides are even more accurate in our experience than the determination of the barium hydroxide solution grayi- metrically asthe barium sulphate. In the various tables are given results which show the accuracy with which barium hydroxide may be standardized by the different organic sub- stances when compared with the standard of decinormal hydro- chloric acid established as the silver chloride. This same solu- tion of the barium hydroxide which gave a value of 0-006396 orm. per cubic centimeter in terms of hydrochloric acid when standardized against the organic acids and acid anhydrides, gave avalue of 0:006430 grm. per cubic centimeter in terms of hydrochlorie acid when standardized by precipitating and weighing as the barium sulphate, by the usual procedure for the determination of barium. As standards in alkalimetry and acidimetry, these organic acids and acid anhydrides, in pure state, are equally as accurate as the best previous standard—hydrochloric acid determined — gravimetrically as the silver chloride. The most serviceable of these organic substances tested are those most readily soluble in water—succinic and malonic acids—although they are no more accurate than the other organic acids and acid anhydrides, as is shown by the results given in the tables. Since these substances can be readily prepared in a known state of great purity, their serviceability as most accurate standards 1s evident. a. > a Phelps and Weed—Succinie Acid. 143 Art. XVII.—A Comparison between Succinic Acid, Arsen- ious Oxide, and Silver Chloride as Standards in Lodimetry °Y, Acidimetry, and Alkalimetry; by I. K. Puerps and L. H. WEED. [Contributions from the Kent Chemical Laboratory of Yale Univ.—clxxx.] Tue use of an excess of a mixture of an iodide and an iodate has been suggested by several as a most exact and convenient method for determining quantitatively various acids in alka- limetry and acidimetry. Kyjeldahl* suggested the use of these reagents for determining nitrogen as ammonia in organic substances. Later Furry+ showed that the conditions of con- centration affected the end-point of this reaction. In very dilute solutions there was an after-coloration which be showed to be due to the incomplete action of the mineral acid rather than to the action of carbonic acid in solution. Grédgert showed the applicability in alkalimetry and acidimetry of the lodide-iodate mixture for determining decinormal solutions of acids and of alkaline hydroxides, carbonates, and sulphides. In the ease of the acids, he treated directly with an excess of the neutral iodide-iodate mixture and measured the iodine set free by titration, in presence of starch as an indicator, with sodium thiosulphate, standardized either against a weighed amount of free iodine or against the iodine set free by the action of an excess of hydrochloric acid and of iodate-free potassium iodide upon a weighed amount of pure potassium iodate in solution. In the case of the alkaline hydroxides, carbonates, and sul- phides, the solutions were treated with an excess of standard mineral acid and then this excess was determined in the man- ner described above. The comparison was shown between the results so obtained and those given by direct titration of the solution of the alkalies against decinormal sulphuric acid, in presence of litmus and of phenolphthalein as indicators, and, for the acids, by similar titration with a decinormal potassium hydroxide solution. The agreement was good except in the case of the carbonates where the iodometric results are low, even when the solutions were allowed to stand for thirty minutes before completing the titration. In a later paperS Groéger suggested pure potassium iodate as a standard in 1odi- metry y, acidimetry, and alkalimetry. The results are good but the difficulty in using the process for practical work is not only the necessity of obtaining pure potassium iodate but also the necessity of using potassium iodide free from iodate. * Zeitschr. analyt. Chem., xxii, 366. + Amer. Chem. Jour., vi, 341. ¢ Zeitschr. angw. Chem., 1890, 355. $ Zeitschr, angw. Chem., 1890, 385. 144 Phelps and Weed—Suceinie Acid. The practical difheulty usually found in the use of arsenious oxide as a standard in iodimetry lies, in out experience, in the fact that in most cases the purest resublimed arsenious oxide does not always give a clear solution when treated with an alkaline hy droxide or bicarbonate. To whatever cause this insoluble residue may be due—to a slight action of the alkaline solution upon the glass or to some impurity in the alkaline hydroxide or bicarbonate used—it is in itself sufficient to make desirable, for accurate work, the possibility of checking results obtained with arsenious oxide as a standard. It has been shown in former papers* from this laboratory that succinic acid may be used, in presence of cochineal as an indicator, as a standard for the exact determination of a deei- normal solution of ammonium hydroxide, and, also,. that succinic, malonic, benzoic, and phthalic acids, as well as the anhydrides of succinic and phthalic acids, may also be used, with phenolpthalein as an indicator, as standards for decinor- mal solutions of sodium hydroxide and of barium hydroxide. The results, which are given below, show that succinic acid may also be used ag the standard for work in iodimetry, alka- limetry, and acidimetry, with exactly as much accuracy as the best previous standards for this work—titration against a decinormal solution of pure arsenious oxide or the gravimetric determination of a solution of hydrochloric acid as silver chloride. For this work, solutions of hydrochloric acid and of sul- phuric acid were made up approximately decinormal by diluting in the usual way the chemically pure acids of com- merce. The exact strength of the hydrochloric acid solution was determined by precipitating definite amounts of the solu- tion of hydrochloric acid, in some cases, in a platinum dish and, in other cases, in a olass beaker as silver chloride by an excess of silver nitrate in the presence of a few drops of dilute nitric acid, filtering off on asbestos under pressure in a per- forated platinum crucible the precipitate of silver chloride after allowing the whole to stand in the dark twenty-four hours. The solution of sodium hydroxide was prepared by diluting pure sodium hydroxide, prepared by the method. given by Kiister,+ with distilled water, freshly boiled. The solution of sodium hydroxide was kept in a closed bottle, connected in the usual manner with a three-way- stoppered burette. It was protected from the action of carbon dioxide in the air by soda lime tubes. The exact strength of this approximately decinormal solution was determined by titration against the standard solution of hydrochloric acid, approxi- * This Journal, xxiii, 211; xxvi, 138. + Zeitschr. anorg. Chem. xli, 474. Phelps and Weed—Succinic Acid. 145 mately decinormal, and against pure succinic acid, in the manner described in the. paper to which reference has been made. The exact standard of the decinormal solution of sul- phuric acid was determined by titration against the decinormal sodium hydroxide solution, in presence of phenolphthalein as indicator. . The decinormal solution of arsenious oxide was carefully made up by treating £9500 grams of the purest arsenious oxide of commerce, twice sublimed, in a beaker in fifty cubic centi- meters of distilled water with fifty eubic centimeters of a solution containing about twelve grams of sodium hydroxide, prepared pure according to the method of Kiister.* After the arsenious oxide had been dissolved by gentle warming, the solution was transferred to a standardized liter flask, by using enough distilled water in this transference to make the volume approximately 250 cubic centimeters. This was then satu- rated with purified carbon dioxide and diluted to a liter under proper conditions of temperature. The solution of iodine was made up approximately decinormal by dissolving iodine in an aqueous solution of potassium iodide, the exact strength of the solution being determined by titration against the standard solution of decinormal arsenious oxide, in presence of an excess of sodium bicarbonate, with starch solution made in the usual way as the indicator. The solution of sodium thiosulphate’ was made up approximately decinormal by dissolving in dis- tilled water the pure sodium thiosulphate of commerce. The exact strength of this solution was determined by titrating definite portions of it against the iodine solution, using the starch solution as indicator. In this work, potassium iodide and potassium iodate were first put into solution in such amounts as to be in excess at the end of the reaction. Definite portions of the decinormal hydrochloric acid solution were then run in from a burette, setting free the iodine. The solution of sodium thiosulphate was next added in amount slightly in excess, and this excess was determined either by adding more of the hydrochloric acid solution or by the addition of the iodine solution, until the blue color of the starch indicator showed the reaction to be complete. In all of the experiments recorded in Table I, one gram of potassium iodide was used and fifty cubic centimeters of asolution of potassium iodate, prepared by dissolving 3°3400 grams of the iodate in a liter of distilled water. Presumably on account of the presence of acid potassium iodate in the sample of iodate used, the addition of the potassium iodate solution to the potassium iodide set free a slight amount of iodine. ‘This free iodine was removed either by boiling until * Loe: cit. 146 Phelps and Weed—Succinie Acid. the characteristic color of free 1odine disappeared, or by add- ing portions of a dilute sodium thiosulphate solution until the blue color of the starch solution introduced was bleached. In all of the experiments in this table, also, five cubic centimeters of a potato starch solution were added before the titration was begun. TABLE I. HCl value of - HCl Na2S203 Iodine solution solution solution HCl Error in used used used found HCl No. erm. erm. erm. erm. erm. ( 1) O'1074 0:1075 ad a 0°1075 0°0001 + ( 2) 0°1074 0°1074 deta 0°1074 0:0000 + ( 3) 0°1074 0°1075 Sy ee 0O°'1075 0°O0001 + ( 4) 0°0520 0°0520 ie ee 0:0520 0°0000 + ( 5) 0°1560 0°1562 ee 0°1562 . 0°0002 + ( 6) 0°0645 0-07 12 0°0068 0°0644 0:0001 — ( 7) 0°0968 0°'1068 0°0101 0°0967 0:0001— ( 8) 0°0645 OrO7a2 0°0067 0°0645 0:0000 + ( 9) 0:0968 0°1068 O:0101 0°0967 0:0001 — (10) 0°0484 0°0534 0°0050 0°0484 0:0000 + (ars) 0°0645 0:0748. — 0:0104 0°0644 0:0001— (12) 0:0484 0°0534 0°0050 0:0484 0'0000+ (13) 0°0645 0'0712 0°0066 0'0646 0'0001 + In experiments (1) to (5) inclusive of Table I the excess of the sodium thiosulphate was determined by the addition of decinormal hydrochloric acid to color with starch, while in the other experiments the excess was determined similarly by titration against the decinormal iodine solution. In experi- ments (1) to (7) inclusive, the iodine set free on the addition of the potassium iodate to the potassinm iodide was removed by boilmg as described above. In experiments (8) to (13) inclusive of the same table, the iodine set free was removed by dilute sodium thiosulphate, which was added until the color of the starch solution present was bleached. The solution of hydrochloric acid used to set free the iodine was standardized as silver chloride as well as against a deci- normal sodium hydroxide solution, the exact strength of which was established by titrating with the organic substances as standards, as we have shown ina former paper.* It was shown also in that paper that the standards obtained by silver chloride and the organic substances are in agreement within the limits of experimental error. From an inspection of Table I, in which the action of a decinormal solution of hydrochloric acid, stand- *This Journal, xxvi, 138. Phelps and Weed—Suceinie Acid. | 147 ardized as the silver chloride, as well as the organic substances mentioned above, is brought into comparison with decinormal solutions of sodium thiosulphate and iodine, standardized against a carefully prepared solution of arsenious oxide, it is evident that the two standards, silver chloride and the organic substances, are in agreement ‘with the third standard, a deci- normal arsenite solution made up with the precautions given above. Further, it is evident that the incompleteness of the action of mineral acid on a mixture of iodide and iodate in dilute solution, as shown by Furry,*. does not take place appreciably under the conditions used here with decinormal solutions. This is particularly striking in view of the different procedures used in the experiments recorded. When, however, the action of centinorma] solutions of iodine, thiosulphate, and hydro- ehloriec acid was tested with the iodide-iodate mixture, the phenomena observed by Furry became very apparent, compara- tively wide variations in the results being obtained. That this method of standardization in alkalimetry could also be used in the presence of carbonates is shown in Table Il. In the experiments included in that table, definite portions of the decinormal solution of sodium hydroxide were first drawn from the burette. Purified carbon dioxide was then passed into this solution for different lengths of time, convert- ing the sodium hydroxide to the carbonate or bicarbonate. The solution of sulphuric acid was then added in excess, and the carbon dioxide set free was completely driven out of the solution by boiling in a flask, trapped with the bulb-end of an ordinary calcium chloride tube, to prevent mechanical loss, the boiling being continued until the volume was reduced one- third. After cooling, the excess of acid was estimated in two ways. In experiments (1) to (4) inclusive, the excess of acid was determined by direct titration, in presence of phenolph- thalein as an indicator, with the decinormal solution of sodium hydroxide. The second method of determining the excess of the sulphuric acid was used in experiments (5) to (8) inclusive. To these, after cooling, a solution, containing one gram of potassium iodide and fifty cubic centimeters of the solution of potassium iodate described above, was added, after the free iodine had been removed from this iodide-iodate solution by the addition of a dilute sodium thiosulphate solution in the presence of starch as an indicator. Definite, portions of the decinormal sodium thiosulphate solution were then added in excess, and this excess was determined by titration with the decinormal iodine solution. * Loe. cit, 148 Phelps and Weed—Succinic Acid. TaBLeE II. HCl value of a PRET Pan eat kota y aan Treat- Differ- ment NaOH HeSO. NaOH Na.S.O03; #Iodine ence _ with solution solution solution to solution solution to in terms No. CO, used used coloration used coloration of HCl min, erm. erm, erm. erm. grm. grm. GL) ale O Ou VOW 306". 020293. ae eee eee 0°0001 + (2) 30 O1012) O-1219 0°0205 2 ae Lae ale eae 0'0002 + (8) OM BAO: OPL D245 OO dni? oo coc aed che eae 0°00038 + (4) 30 0°1349 0°1568 O021:6.4-3) a eee ee 0°0008 +- (5) 15 ODOT. OMS O Oe wae serene. 0°0302 0°0013 0°0005 + (OO) a0 OTOH, a OR EKOG se sc = 0°0502 0°0013 0:0005 + C7) sets 0°1349 0°1742 se Ales) es 0.03892 0°0004 0°0005+ (8) 35 OF 3499 OAc see ae 0°0552 = 0°0162 0°0003+ The differences recorded in the last column of Table IT show that the standard of the sulphurie acid is slightly higher than the summation of the sodium hydroxide originally taken, and of the sodium hydroxide or thiosulphate and iodine used to determine the excess of sulphuric acid. From these results © it seems to appear that the differences are presumably to be attributed to two causes—the experimental error and, perhaps, a slight mechanical loss of sulphuric acid during the long boiling. These results show that a non-volatile acid like sulphurie acid may be used with exactness to determine alkaline carbon- ates in either of the two ways described. _ The essential thing in the exact titration with a solution of sodium hydroxide, with phenolphthalein as an indicator, as is well known,* is the absence of carbonate, as is also the case in the use of the potassium iodide- iodate mixture.t . This condition is easily attained by boiling, as was done in these experiments. From these results , it may be seen that the standards of the solutions used in work in iodimetr y, alkalimetry, and acidime- try may be found as exactly by titrating against certain pure organic acids as standards, as against the best known standards usually used in such work—the decinormal solution of arsen- ious oxide, or a decinormal solution of hydrochloric acid, standardized gravimetrically as silver chloride. Succinie acid was used as the organic acid standard in this work because of the ease with which it is prepared in a state of purity and its ready solubility in water, but it is clear from work shown in an earlier paper} from this laboratory that malonic, benzoic, and phthalic acids as well as the anhydrides of succinic and phthalic acids, could be used with equal exactness. * Kuster, Zeitschr. anorg. Chem., xiii, 127. + Groger, Zeitschr. angw. Chem., 1890, 358. t This Journal, xxvi, 138. _ Ford and Tillotsan—Orthoclase Twins. 149 Art. XVII.—On Orthoclase Twins of Unusual Habit; by W. EE. Forp and E. W. Trtvortson, JR. Tue orthoclase twins to be described in the following pages were collected by Prof. L. V. Pirsson during the summer of 1896, while engaged in work for the United States Geological Survey in Montana. They occurred as phenocrysts in an altered tineuaite porphyry which lay as an intruded sheet between black shales near the head of West Armell Creek in the Judith Mountains. The tinguaite sheet is described* as measuring “between ten or twelve feet in thickness with numerous immense feldspar phenoer. ysts, some of them being four by two by one inches across.” The groundmass of the rock is fine- grained with a greenish gray color, the green tone being due to the presence of tine crystals of sevirite, while on the weathered surfaces it is spotted with numerous pits stained yellow with iron oxide. The writers desire to express to Prof. Pirsson their thanks for the opportunity to figure and describe these crystals. The phenocrysts occur as well- developed crystals and, as has been said, are at times of considerable size. They are opaque and are frequently stained on the surface with iron oxide or colored green with a thin coating of egirite. When broken, however, they present a glassy ‘luster and fresh appearance. The crystal faces were too rough to admit of measurements other than those with the contact goniometer, but the forms present were easily identified in this way and by their zonal relations. They were all common forms comprising } (010), c (001), m (110), 2(150), 2 (021) and o (111). v. The conformity in its optical properties to normal orthoclase is rather surprising when the large amount of soda shown to be present by the analysis is considered. The results of the analysis by Tillotson follow: RJA 0 [aCe peo are ager 64°01 I Ore see UNS 20°19 RE): cnn Sogn OS eS 10°48 iat OR Geese tate! ae 5°37 Ao Gash inn erst 100°05 This particular occurrence of the tinguaite rock was not ana- lyzed because of the alteration it had undergone through weath- ering, but a closely similar and fresh rock from Cone Butte was analyzed by Pirsson* and the percentage of the alkalies found by him agree closely with those of the present analysis. In the Cone Butte tinguaite, however, the soda was considered to belong entirely to either albite or nephelite, but in the West * This Journal, ii, 192, 1896. 152 Ford and Tillotson—Orthoclase Twins. Armell.Creek occurrence, supposing that the two magmas were alike, it would all be accounted for by the orthoclase itself. In the present analysis the somewhat disproportionately low percentage of silica is probably due to the small amount of kaolinization which the feldspar showed in the thin sections. Drawing of the Crystals.—The erystals were drawn from a stereographic projection of their forms. ‘There was nothing new in the methods employed, but as concrete examples of the 6 use of the stereographic projection for the solution of such problems are seldom to be met with in the literature, a brief description of the manner of transposition of the poles of the faces from normal to twin position is here given.* According to the Baveno law of twinning the n (021) face becomes the twinning plane and as the angle c~n = 44° 56 1/2’ the angle between c and ¢' (twin position) becomes 89° 53’. For the * For a general discussion of the graphical use of the stereographic projec- tion, see Penfield, this Journal, xi, 1, 1901. Ford and Tillotson—Orthoclase Tivins. 153 purposes of drawing it is quite accurate enough to assume that this angle is exactly 90° and that: accordingly the ¢ face of the twin will occupy a position parallel to that of the 6 face of the normal individual. Figure 6 shows the forms observed of the erystals both in normal and in twin positions, the faces in twin position being indicated by open circles and a prime mark (’) after their respective letters, while the zones in twin position are drawn in dashed lines. Star ting out with the forms in normal position, the first face to transpose is the base c. This form, from the law of the twinning, will be transposed to c’ where ‘it occupies the same position | as 6 of the normal individual, and it necessarily follows that 6 itself in being transposed will come to 6’ at.the point where the normal ¢ is located. In turning therefore the crystal to the leit from normal to twin position, the faces ¢ and 6 travel along the great circle I through an are of 90° until they reach their respective twin positions. We have, in other words, revolved the crystal 90° to the left about an axis which is parallel to the faces of the zone |. The pole of this axis is located on the stereographic projection at 90° from the great circle [and falls on the ~ straight line II, another great circle which intersects zone I at right angles. This pole P is readily located by the stereo- graphic protractor on the great circle II at 90° from c. The problem then is to revolve the poles of the faces from their normal positions about the point P to the vente and through an are of 90° in each case. During the revolution the poles of the 2 faces remain on the oreat cirele [ and as the angle nxn = 90°, the location of their poles when in tw in position is identical with that of the normal position and 7’ falls on top of m. We can now trans- pose the great circle I] from its normal to its twin position, since P remains stationary during the revolution and we have determined the twin position of ¢. The dashed are II’ gives the twin position of the great circle II. The twin position of y must lie on are II’ and can be readily located at y’, the intersection of are [I’ with a small circle about P having the radius Pay. It is now possible to construct the are of the zone II] in its tr ransposed position III’, for we have two of the points, y’ and n’ of the latter, already located. By the aid of the Penfield transparent oreat circle protractor the position of the arc of the great circle on which these two points lie can be determined. On this are, ILI’, 0’ and m’ must also lie. Their positions are most easily determined by drawing ares ot small circles about b’ with the required radii, bao = “63°8', bam = 59° 221/2' and the points at which they intersect are IIl’ locate the position of the poles 0’ and m’. At the same 154. Ford and Tillotson—Orthoclase Twins. time the corresponding points on IV’ may be located, it being noted that IV’ and III are the same are. But one other form remains to be transposed, the prism z. We have already 0’ and mv’ located and it is a simple matter with the aid of the great circle protractor to determine the position of the great circle on which they he. Thenasmall circle about 6’ with the proper radius, bAz = 29° 24’, determines at once by its inter- sections with this arc the position of the poles of the 2 faces. The transposition of the faces from normal to twin position. having been made, it is a simple matter to draw the erystal figures from the projection.* It may be pointed out that if it should be desired to make use of the methods of the gnomonie projection for the drawing of the figures the stereographiec projection, as derived above, can be readily transformed into a gnomonic projection by doubling the angular distance from the center of the projection to each pole by the use of the stereographic protractor. But from whichever projection it is preferred to draw the figures, it is thought that the stereo- graphic projection, with the aid of the Penfield protractors, offers the simplest method for the ready transposition of the poles of the faces from normal into twin positions. Mineralogical Laboratory of the Sheffield Scientific School, Yale University, New Haven, Conn., April, 1908. * See Penfield, this Journal, xxi, 206, 1906. J. V. Lewis—Palisade Diabase of New Jersey. 155 Art. XIX.—The Palisade Diabase of New Jersey; by J. VotnrEy Lewis.* THE intrusive trap that forms the Palisades of the Hudson extends, with outcrops several hundred feet thick, from west of Haverstraw, N. Y., southward to Staten Island and, some- what intermittently, westward across New Jersey to the Del- aware River, an aggregate length of about 100 miles.t It is everywhere a medium- to fine-grained dark gray heavy rock, with dense aphanitic contact facies. The typical coarser rock contains, in the order of abundance, augite, plagioclase feldspars, quartz, orthoclase, magnetite, and apatite. The first two occur in ophitic to equant granular tex- ture, and the next two in graphic intergrowths which some- time constitute one-third of the rock; in the contact facies this micropegmatite disappears and scattering crystals of olivine occur. A highly olivinie ledge, 10 to 20 feet thick and about 50 feet from the base of the sill, is exposed in the outerops north- ward from Jersey City for about twenty miles. The olivine erystals, which constitute 15 to 20 per cent of this rock, occur as poikilitic inclusions in the augite and feldspar. Chemically the diabase ranges from less than 50 to more than 60 per cent of silica, with corresponding variation in alumina, ferric oxide, and the alkalis, while ferrous iron, lime, and magnesia vary inversely. ‘The augite is rich in these latter constituents and poor in alumina, giving a great preponder- ance of the hypersthene and diopside molecules. The feld- spars range from orthoclase and albite to basic labradorite. Doubtless there is some anorthoclase since all feldspar analyses show potash. While there is considerable range in the proportions of the minerals, augite usually comprises about 50 per cent of the rock, the feldspars about 40 per cent, quartz 5 per cent, and the ores 5 per cent, constituting a quartz-diabase, with normal and olivine-diabase facies. Basic concentration at the contacts, followed by differentiation by gravity during crystallization of the body of the sill, especially by the settling of olivine and the ores and the rising of the lighter feldspars in the earlier and more liquid stages of the magma, are hypotheses that seem to account for the facies observed and their present relations. Microscopic characters—In thin sections the texture of the rock is usually diabasic, or ophitic; that is, the augite fills the * Read before the New York Academy of Sciences April 6, 1908. Published by permission of the State Geologist of New Jersey. +J. Volney Lewis, Structure and Correlation of the Newark Trap Rocks of New Jersey, Bull. Geol. Soc. of America, vol. xviii, pp. 195-210; also Origin and Relations of the Newark Rocks, Ann. Rept. State Geologist of N. J. for 1906, pp. 97-129. 156 J. V. Lewis— Palisade Diabase of New Jersey. interstices between the interlacing lath-shaped feldspars, or when greatly in excess 1t forms the groundmass in which the feldspars are imbedded. In the coarser-grained portions of the rock there is often developed a granitoid texture, in which the two chief minerals occur in grains of approximately equal size and of nearly equal dimensions in every direction. Augite, the most abundant constituent, is pale green to colorless and sometimes exhibits distinct pleochroism—pale green to hight yellow.* It occurs in plates up to 3 or 4 mill- meters in diameter, and in irregular grains whose forms are determined by the accompanying feldspars. Crystal outlines are rarely observed. In the denser contact facies augite of two generations appears, the earlier as large plates scattered through the denser groundmass in which the augite of later crystallization forms a fine granular filling between the feld- spars. Two forms of twinning often appear, both separately and in combination. That parallel to the or thopinacoid (100) usually produces paired halves, while the basal twinning (par- allel to 001) is more commonly repeated in thin lamellee, which are sometimes exceedingly minute. ! Plagioclase, the chief feldspar and the second constituent in abundance, occurs in characteristic lathlike forms, ranging up to 2 millimeters in length with a breadth one- fifth to one- third as great. In the coarse textures of the granitoid facies these dimensions become more nearly equal, and diameters of 3 to 4 millimeters are often observed. Often the plagioclase presents complete crystal outlines, but very commonly the terminal planes are lacking, the elongated crystals abutting tee. against each other. They are made up of thin twinning lamellae, chietly according to the albite law, but peri- cline and Carlsbad twinning also occur. Zonal structure is ran) quite commonly developed, and fringing the extreme acid borders a graphic intererowth of quar tz and orthoelase is often found. Maximum extinction angles in sections normal to the albite twinning plane range a little under 30 degrees, corre- sponding to acid labradorite. Analyses of feldspars separated by heavy solution have shown that labradorite containing the soda and lime molecules in about equal proportions is the most abundant plagioclase; but other members of the series, present in considerable amount, range to almost pure albite. Orthoclase and quartz in oraphic intergrowth, as noted above, frequently form a fringe about the plagioclases, and fill many of the triangular and irregular interstices. These *TIn the examination of several hundred sections of Newark diabase from New Jersey and neighboring states only monoclinic pyroxenes have been observed. It seems highly probable that the hypersthene that has been occasionally reported in these rocks is simply pleochroic augite. J. V. Lewis— Palisade Diabase of New Jersey. 157 areas are sometimes as much as 3 or 4 millimeters across, and are then distinctly visible in the hand specimen, as in the western portion of the Pennsylvania railroad tunnels at Home- stead. Frequently individual grains of quartz, and less com- monly of orthoclase, are also” observed, attaining in some instances a diameter of 1 millimeter. Magnetite is always present but in greatly varying amount. Orystals are sometimes observed, but most of it, like the augite, is irregularly clustered between the plagioclases, and some- times partly incloses both the plagioclase and the augite. The frequent presence of magnetite secondary from the alter- ation of augite renders it impossible in many cases to distin- euish with certainty that of primary origin. It is probable that masses molded about the other constituents are largely composed of secondary accretions. Biotite is also often present in small amount, and is usually clustered about the magnetite in relatively large irregular flakes. It is strongly pleochroic—deep reddish brown and light yellow. Some secondary biotite, after augite, occurs, but in most cases this is readily distinguished from the primary mineral. Olivine is absent from the great bulk of the rock. It occurs in small amounts, however, near the contacts with the inclos- ing strata, and is exceptionally abundant in the olivine-diabase ledge of the Palisades, constituting as much as one-fifth of the whole. In the fine-grained border facies of the rock it occurs in scattering por phyritic erystals, which sometimes exhibit resorption phenomena = rounded and embayed outlines. Corrosion mantles or “reaction rims” of radial enstatite occasionally surround the larger crystals, and nest-like aggre- gates of it entirely replace some of the smaller ones. In this part of the rock inass the olivine is usually more or less altered into yellowish or brownish serpentine ; but in the olivine-diabase ledge it occurs in numerous perfectly fresh crystals and irregu- lar grains. Most of it forms poikilitic inclusions in the feld- spars and less abundantly in the augites, and it retains a striking freshness and transparency even in the presence of considerable alteration of the augite. Apatite is always in well-formed prismatic crystals, ranging from very minute up to 1 millimeter in length and 0-06 milli- meter in diameter. Itis always abundant in the feldspars and quartz, sometimes im plagioclase, sometimes in the quartz- orthoclase intergrowth, and is rarely seen in the other consti- tuents. Chemical composition.—From a number of analyses that have been made the following are selected to show the range of composition of typical facies of the rock : 158 J. ‘V. Lewis—Palisade Diabase of New Jersey. Analyses of Palisade Diabase. I I III IV SiO Ae aie 60-05 56°78 51°34 49-02 ASO a One ais 8 14°33 POT 0s oom PeiO: £2 aise - 3-29 5°76 2°65 1°54 HeO ek 10-21 9°27 14°14 10:46 MeO stars 0°85 1°58 3°66 17°25 CaQvse St 4°76 5°26 7-44 8-29 Na Ogesee 5. 4-04 3°43 2°43 1°59 KAO RSP Rel 1°75 1°44 0°40 LO eae i 0°66 010 0-69 0°59 POP ee ne! 0-21 0°33 0-18 0°16 DiGi ee 1:74 1°44 3°47 0°99 [Pes Otero aati 0°52 0°36 0°20 0-11 NnQe irl O88 0°25 0°36 0°16 100°52 100°64 100°71 100-70 J. Pennslyvania R. Rh. tunnel, Homestead, 400 feet from the west end. : II. Old quarry near R. R. station, Rocky Hill, 420 feet from upper surface of the trap. Lil. Pennslyvania R. R. cut 420 feet east of Marion station (Tonnele Ave. ), Jersey City. IV. Englewood Cliffs, on the Palisades, 11 miles north of Jer- sey City. From the olivine diabase ledge. ‘T, LI, and IV by R. B. Gage, chemist of the Geological Survey of New Jersey; II by A. H. Phillips of Princeton University (this Journal, vol. vii, 1899, Peco n): In general, alumina, ferric iron, and the alkalis vary with the silica, while ferrous iron, lime, and magnesia vary inversely. Chemically these rocks overlap the andesite-diorite series on the one hand and the most basic olivine-gabbros on the other, and the extremes are characterized by abundant quartz and olivine, respectively. - Classification. —In the older terminology the prominent facies of the Palisade sill would be known as quartz-diabase, diabase, and olivine-diabase, the prefixes quartz and olivine denoting special richness in these minerals. As indicated above, most of the coarse-grained rock, which constitutes by far the greater bulk of the Palisade sill from the Hudson to the Delaware, is decidedly quartzose, this mineral being quite generally pr esent in graphic intergrowth with orthoclase. In the most acid facies micr opegmatite constitutes about one-third of the bulk of the rock, but the average is probably somewhat less than one-tenth. On the other hand, the most basic facies contain 15 to 20 per cent of olivine, but this is confined to the relatively small mass of the pee diabase ledge. Normal diabase, without quartz or olivine, is much less abundant than. J. V. Lewis—Palisade Diabase of New Jersey. 159 quartz-diabase, and this becomes slightly olivinie near the con- tacts. In the quantitative system of Messrs. Cross, Iddings, Pirs- son, and Washington the analyses given above would be clas- sified as follows: I. Class II. Dosalane Order 4. Quardofelic, Austrare Rang 2. Domalkalic, Dacase Subrang 4, Dosodic, Dacose II. Class II. Dosalane Order 4. Quardofelic, Austrare Rang 3. Alkalicalcic, Tonalase Subrang 4. Dosodic, Zonalose III. Class III. Salfemane Order 5. Perfelic, Gallare Rang 3. Alkalicalcic, Camptonase . . Subrang 4. Dosodic, Camptonose IV. Class 1V. Dofemane Order 1. Perpolic, Hungarare Section 2. Dopyric Rang i. Permirlic Section 2. Domiric Subrang 2. Domagnesic, Palisadose The great bulk of the rock, judging from about a dozen analyses and a large number of thin sections, appears to be of _the type represented by the third of the analyses given above, and might be termed a graphi-ophito- to graphi-grano-camp- tonose (III. 5. 3. 4). Notable amounts of the rock, however, belong to the more salic dacose (LI. 4. 2. 4) and tonalose (II. 4, 3. 4) and to the more calcic auvergnose (III. 5. 4. 4, 5). The highly basic ferromagnesian olivine-diabase, of which the fourth analysis above is typical, is a sharply defined facies that may be designated as a povkili-ophito- to povkili-grano-pali- sadose. From the typical development of this rock along the Palisades northward from Jersey City it is proposed that the subrang into which it falls (LV. 1°. 1°. 2), hitherto unnamed, be designated as palisadose. The relations of these various facies to each other and the problem of their origin are discussed in the report on the petrography of the Newark igneous rocks of New Jersey.* It has already been remarked that the hypothesis of gravitational differentiation seems to account satisfactorily for all the condi- tions observed. Composition of the Augite.—Analyses of the augite from this sill at Rocky Hillt and from the very similar intrusive at West Rock, New Haven, Conn.,t yielded the following results : * Annual Report of the State Geologist of New Jersey for 1907.. + A. H. Phillips, loc. cit. t G. H. Hawes, this Journal, vol.ix, p. 185, 1875. 160 J. V. Lewis—Palisade Diabase of New Jersey. Analyses of Augite from the Palisade Diabase. r i. IIL. IV. SiO, Saas 7°72 48°54 50°71 47°10 AlsO: eaieat 3°44 5°50) 3°55 4°55 Hes@ (ae eee 5°98 COTE) n. d. de FeO. SS. 18°34 Dee? 5 15°30 15°20 MeQ 8 = 12°89 7:67 13°63 18°65 CaO ree. Re A’ = 1 TOr97 13:30 Tare IN 22 O pre ee 0°86 oO Tz bs n. d. TOO aie Ors . NnOss ca ons n.d, 1G: 0:80 35 Ory Aarne fee eo 0-00 0°82 iy 1:38 Fnsolpee ewes ae Ss pea deine 0°34 100°95 100°62 100-00 98-67 1 By difference. I. Rocky Hill, N. J. Quarry near the middle of the mp sheet. A HH. Phillips, analyst. Il. Rocky Hill, N. J. Old quarry near the railroad station, about 420 feet from the upper contact. A. H. Phillips, analyst. III. West Rock, New Haven, Conn. G. W. Hawes, analyst. IV. Aakeroe, Sweden. Partial analysis of diallage by H. von Post. * With the exception of smail deficiencies in silica in I, II, and IV, these analyses correspond to the following pyroxene molecules in the proportions indicated : Acmite, NaFe(Si0,), = ac. Hypersthene, (Me, Fe)SiO, ==. by. Diopside, Ca(Mg,Fe)(Si0, i == di. Aluminous molecule, (Me, Fe) (Al, Ke) SiO. = akaw ac. hy. di. alm. Pee T0822 BAO eT ONG. Kine a9 = Ley See OS een cal cea) ARR Y ieee Dee eb VLR. ee be Yisr. sia amen W270 : 330 : 784 Feemopt == D385 3 6°54 : Wore) = 1 = (pe 1S ey tel Naess, a 2 pprox.) 16 Dara a) 522 : 956 Pear Oe = 18S ss CR Dee FN B86": 1 =< 2 Bors Lo Sa Ass epoxy IV. ?+t : 864 ; 808 5 oh a ae OPLOR oro Ole 1 = 13 12 25 2 2 (ep prone These formulas indicate a quite exceptional composition for augite in the great excess ot ferrous iron and magnesia over lime, alumina, and ferric iron. * Dana’s System of Mineralogy, 6th ed., p. 360. + Ferric iron and alkalis not determined. J. V. Lewis—Palisade Diabase of New Jersey. 161 Composition of the Feldspars.—Feldspars were also separ- ated and analyzed from the Itecky Hill material by Phillips and from the West Rock locality by Hawes. Omitting non-feld- spathic constituents and assigning the potash to orthoclase, and the soda and lime to albite and anorthite, respectively, the analyses show the following constitution : Mineral Constitution of the Feldspars. iS Il. III. IV. VE VE | VI 4 ASE Sp. gr. >2°69 <2°69|>269 <2°69 <260 =2°57 7 >2°69 <2°69 Re 2° Steele <5 ee bare 12°6 ad ae} 3] | 6°0 Bw: Albite A7°1 Diss: f GOs6 £023 68°9 64°2 | 24°4 40°9 maerghie 45°4 = 21-0 |26°2- 17"1 SH AV 69°6> 51-9 ab,an, ab;an.z|absan, absan, abead, ab,;an, absans ab;ans I, if. Rocky Hill. Quarry near the middle of the trap sheet. Ill, 1V, V, VI. Rocky Hill, N. J. Old quarry near upper contact. VII, VOI. New Haven, Conn. West Rock. The plagioclase molecules as calculated range from nearly pure albite to labradorite. It is probable, however, that some of the soda is combined with the potash in orthoclase and anorthoclase. On the other hand. it is also quite possible that small amounts of more basic plagioclases would have been found if the beaviest portions had been further separated before analysis. The frequent occurrence of zonal structure in the plagioclases, however, seems to indicate that these minerals do not occur in individualized grains of uniform composition, but have been built up into composite crystals of successively more acidic and more sodic character. In the instances determined the portion with the highest specific gravity constituted more than half the total feldspar s, so that the labradorite molecule undoubtedly predominates. Metamorphic effects —Contact metamorphism has produced an elaborate series of hornfels, rocks characterized by various combinations of feldspar, biotite, quartz, augite, hornblende, tremolite, garnet, spinel, magnetite, muscovite, cordierite, scapolite, vesuvianite, sillimanite, andalusite, chlorite, calcite, analcite, titanite, tourmaline, zircon, apatite, and possibly leu- cite. The various types within the zone of metamorphism seem to vary with the original composition of the shales, and not to depend on relative distances from the contact nor degree of metamorphism. Metamorphic arkose, both in the inclusions and at, the con- tacts, contains besides the usual plagioclase, orthociase, and some quartz, also augite, biotite, epidote, cordierite, chlorite, calcite, tourmaline, and apatite. 162. oS. *V. Lewis— Palisade Diabase of New Jersey. The contact metamorphie effects of the sill at Hoboken, New Jersey, have been described by Andreae and Osann,* who show that it is of exomorphic pneumatolytic character. To the four types of hornfels which they describe, J. D. Irving+ has added five others, and the present studies, in which no attempt has been made to present a complete series of alteration products, have brought out eight additional types of hornfels and four of metamorphic arkose. These might be extended almost indefi- nitely, since they do not occur as sharply defined types, but present various degrees of gradation from one to another. — Furthermore, they do not form zones or belts in any systematic order with relation to the intrusive rock, but alternate irregu- larly throughout all parts of the zone of metamorphism. It is evident, therefore, that the types observed are not the results of varying degrees of metamorphism, but are dependent only on original variations in the composition of the shales and sand- stones themselves. t * Andreae and Osann, Tiefen contacte an intrusiven Diabasen von New Jersey; Verh. d. Naturh. Med. Ver. zu Heidelberg. N. F. V., Bd. I, 1892. + School of Mines Quarterly, vol. xx, pp. 213-225, 1899. { For descriptions of these rocks, and of other igneous rocks besides the great intrusive sill, see ‘** Petrography of the Newark Igneous Rocks of New Jersey,” Ann. Report of the State Geologist of New Jersey for 1907, pp. | 98-169. : F. B. Loomis—New Horse Srom the Lower Miocene. 1638 Art. XX.—A New Horse from the Lower Miocene; by F. B. Loomis. Wettse the series of fossils, which show us the development and radial adaptations of the horse family, is already a large one, there still remain breaks and places where more material is desired. Such an unfilled gap exists between the rich Ohgo- eene Mesohippus fauna and the upper Miocene Protohippus group, just where the transition from the brachydont unce- mented teeth to the hypsodont cemented ones occurs. The finding by Peterson* of the excellent type of Parahippus nebraskensis in the Upper Harrison beds of Nebraska, closed in a part of this gap, and for two or three seasons fragments Fie. 1.—Crown views of the upper and lower dentition of the type of Parahippus tyleri. One-half natural size. have raised the expectation of a Lower Harrison horse. Dur- ng the season of 1907, under the guidance of Mr. Harold Cook, the Amherst ’96 expedition found a prospect, which proved to be the major part of the skull of this much desired type. While the brain case is wanting, a nearly perfect upper and lower dentition of an adult individual is preserved, show- ing an animal closely related to P. nebraskensis, but about a fourth smaller. The following specific description will point out the affinities and characters of the new species. Parahippus tyleri sp. nov. Type, a skull numbered 1079 in the Amherst College Museum, which while lacking the brain case preserves all the dentition except the upper canine and the first premolar of the * Ann, Carnegie Museum, vol. iv, p. 57, 1906. 164 FB. Loomis—New Horse from the Lower Miocene. upper and lower jaw. The specimen was found in the upper part of the Lower Harrison beds, 8 miles northeast of Agate, Sioux Oo., Nebraska; and is named to honor Prof, J. M. Tyler, the organizer of the ‘Amherst expeditions. While having a rather short facial portion, the skull is moderately high and narrow. ‘The individual being described is an adult, only recently matured, as indicated by the moderate wear on the teeth and the fact that the third upper molar is scarcely worn at all. On the upper incisors the pit is deep, being entirely surrounded on the inner side by the raised cingu- lum. This pit seems to be more developed than in P. nebras- kensis. A canine is indicated by a small alveolus a short distance from the third incisor. The first premolar is wanting in this specimen. The second to the fourth premolars, while slightly larger, grade into the molars having the same charac- teristics. While the parastyle and the mesostyle are well devel- oped they are not as prominent as in P. nebraskensis. A remnant only of the cingulum is present on the inner part of the front border of each tooth. The protocone and protoconule unite to make a strong protoloph, but are separated from each other by a narrow constriction. On the metaloph of premolar four and the succeeding molars a small crochet is developed, which while distinct does not however unite with any part of the protoloph; consequently the prefossette is not en.irely isolated. The ‘hypostyle i is strong and notched in the rear. On the lower jaw the pit in the incisors is not as well devel- oped as in incisors of the upper jaw, appearing more like a groove behind a well-marked cingulum. The lower canine is a simple cone of moderate size. The first premolar is indi- cated by a smal] alveolus and must have been tiny. The remaining premolars and molars each have a well-marked cingulum, starting from the parastylid, continuing around the outer border, and back to a tiny hypostylid. The inner wall of each tooth is relatively straight, the upper part being, in little worn teeth, notched to separ ate off the strong parastylid, and again between the paraconid and hypoconid. “The heel of the third molar is moderate in size and simple in form. While very like, and probably ancestral to P. nebraskensis, this species is distinguished by the less pronounced parastyle and mesostyle, by the relatively narrower teeth, the deeper pit in the incisors, and the smaller size. It is a primitive member of the genus, the crochet not uniting with the proto- loph and cement being entirely absent. Geology. 165 Measurements. Length of the upper incisor series._-....---- 28™™ Distance from incisor 3 to the canine.---_ .--- 11 Length of the upper premolar series_.-__---- 59 Heusen of upper premolar d2_.. --__._._.--- 20 Witdil ef upper premolars .—-_-, =... .--+- 26 Length of the upper molar series ----_-.---- 55 Mele mher Uppee mort 2 9 2 ho 19 Mbttnh of Upper molar 2.2) tk 26 Length of the lower premolar series .--.-._-- 56 Length of the lower molar series _.--..-.--- 59 Mrden- eb lower molar foo fos Sos in 14 Amherst College, Department Biology, May 29, 1908. SCIENTIFIC INTELLIGENCE. I. Grouoey. 1. Indisches Perm. und die permische Hiszeit ; by E. Koxen. N. Jahrb. f. Min., Festband 1907, pp. 446-546, and a large paleo- geographic map.—This is a very important paper to all students of Permian and Triassic stratigraphy and to glaciologists. Noet- ling has in Koken a strong supporter for his Indian Permian stratigraphic correlations. The broader views of equivalency put forth some years ago by Tschernyschew are rejected. Koken ‘holds that the Productus limestone of India passes without break into the Triassic, as it also does in the eastern Alps where the Bellerophon beds pass into the Werfen. Beneath the Productus limestone occur the Permian glacial deposits, the material having come from the south or peninsular India. Sands of glacial origin are also seen in the higher limestone and even in the T’riassic— the regolith of the southern lands carried by the rivers into Tethys. Various theories in regard to the probable causes for the Permian glaciation are discussed at length and Koken rejects the carbon dioxide theory of Arrhenius, and, as well, that of the wan- dering of the pole. He concludes that the probable cause is to be looked for rather in changed relations of the continents to equatorial waters, and therefore in the changed streaming of oceanic currents. During Permian glaciation India stood high and Australia was united to India and Africa, deflecting the southern equatorial waters away from the Indian Ocean and the Antarctic region. The melting of the Permian ice was due to the isolation of Australia from India, as along nearly all of the west coast of the former land are found Permian deposits. This Opening again permitted the southern equatorial currents to a®, Jour. Sct.—Fourtn Seriss, Vou. XXVI, No. 152.—Aveust, 1908. 166 Seventific Intelligence. stream into the Indian Ocean. Koken’s generalizations have added value because of his detailed paleontologic knowledge. C. 8. 2. Geological Survey of Western Australia, Bulletin 29. A Report upon the Geology, together with a Description of the Productive Mines of the Cue and Day Dawn Districts, Murchi- son Goldfield ; by Harry P. Woopwarp, Assistant Government Geologist. Part I, Cue and Cuddingwarra Centres ; pp. 93, with 3 maps, 12 photographs, 2 blocks, and 15 plates of sections. Part Il, Day Dawn Centres ; pp. 48 and appendices, pp. 44-53, with 2 maps, 7 photographs, 1 figure, and 8 plates of sections. Perth, 1907.—The general conditions under which gold deposits occur in Murchison are in lenticular-shaped amphibolite belts, “surrounded or sandwiched with granites, the whole being inter- sected by numerous feldspathic dikes.” The oldest rock of the Cuddingwarra and Day Dawn districts is an amphibolite. Grano-diorite occurs as a magmatic intrusion into the amphibo- lite. A remnant flow of vesicular andesite caps Cue hill. The topography is diversified by the presence of “table tops” made of granite “which owes its weather-resisting character to iron oxide which has been drawn up in solution by capillary attraction through leaching of the rocks below”. Gold occurs chiefly in quartz reefs, the production being—Cue, 212,855 ounces, Cuddingwarra, 35,461 ounces, and Day Dawn, 847,692 ounces. BER AIS CE 3. Lllinois State Geological Survey, H. Foster Bain, Direc- tor. Bulletin No 7. Physical. Geography of the Hvanston- Waukegan Region ; by Wattacre W. Atwoop and James W. GoLpTHWalIT. Pp. 93, pls. 14, figs. 52. Urbana, 1908.—It is becoming recognized that geologists are under obligation to present the main facts of their science in such a manner as to be of direct use to teachers and to the general reader. Following the lead of the Connecticut Survey, the Geological Survey of Illinois has planned a series of bulletins dealing with the physical geography of the state and designed primarily to present material more or less well known to experts in a form directly available for the intelligent reader. ‘The volume in the series listed above is well arranged, well written and illustrated and contains inter- esting material. : H. E. G. 4, Map of Vesuvius.—The Instituto geografico militare of Italy, at Florence, has published a new edition (1908) of an excellent map of Vesuvius in colors, on a scale of 1 : 25,000 (2 francs) indicating all determinable lava flows, with their dates, down to 1906; also a map of the volcano in black, in six sheets, scale 1: 10,000 (4.50 frances for the set) ; also two special maps of the cone of the volcano, 1: 10,000, before and after the eruption of 1906 (each, one franc). Of the general map of Italy, 1 : 100,000, by far the most legible edition is the one known as “ systema Gliamas,” in four colors, now in course of publication (1.50 frances a sheet, 27 sheets published ; edition on thin paper prefer- able). W. M.D —_—: Botany. | 167 5. A Pocket Handbook of Minerals, designed for use in the field or class-room, with little reference to Chemical Tests ; by G. MontacuE Butrer. Pp. ix, 298 with 89 figures and 5 tables. New York, 1908 (J. Wiley & Sons).—Mineralogists and others who feel the need of a small volume suitable for the pocket, giving the important characters of the prominent mineral species, will find this work suited to their needs. It is printed in particularly clear, open form, with the emphasis upon essential characters and the omission of unnecessary detail; there are numerous illustrations. A novel feature is the series of tables at the end presenting the characters of the species in condensed form. BES DOTANY: 1. The Origin of a Land Flora; a Theory based upon the Fucts of Alternation; by F. O. Bower, Regius Professor of Botany in the University of Glasgow. Pp. x1+ 727, with frontis- piece and 361 text-figures. London, 1908 (Macmillan & Co.).— Professor Bower has long been recognized as one of the ablest authorities on the morphology of the Pteridophytes, a group of plants to which the present work is largely devoted. He clearly shows that representatives of this group were the first plants to solve successfully the problems of terrestrial life, and that the Phanerogams, or seed-bearing plants, which are now in the ascendant, were derived from the Pteridophytes by further special- ization. ‘The evidence for these opinions is drawn almost entirely from the sporophyte, or asexual generation, the lines of gameto- phytic development in land plants reaching their culmination in | certain divisions of the Bryophytes. The great gap which exists between the bryophytic sporophyte with its continuous spore- cavity and lack of lateral organs and the pteridophytic sporo- phyte with its distinct sporangia and well developed leaves is still unfilled, but three main factors of advance are indicated, namely: sterilization of originally fertile cells; segregation of sporogenous tissue into distinct masses ; formation of roots and of appendicular organs, such as leaves, on the axis or stem. ‘lhe fact is also emphasized that the primary function of the sporo- phyte is, after all, the production of spores, so that, in the evoiu- tion of the Pteridophytes, the sporophyll was probably the first type of leaf to appear, the true foliage leaf arising from the sporophyll by further sterilization. On the basis of these views the author advances the idea that the sporophyte in the original _ Pteridophytes consisted of an axis attached to the soil by a root- system and bearing a cluster of small sporophylls, each with a single sporangium. The closest approach to this condition is apparently to be seen in such a plant as Lycopodium Selago, where the sporophylls are indefinite in position and essentially like the small foliage leaves in appearance and structure. The large and frequently compound leaves which are characteristic of the Filices and Ophioglossales have apparently been derived from small and simple leaves by longer continuance of growth and increase in complexity. Professor Bower designates the theory 168 Scventifie Intelligence. which he defends as the theory of the “strobilus,” or “ strobi- loid” theory. He admits that the theory is hardly capable of direct proof and claims no originality for certain of the views advanced, but the evidence which he presents is clear and usually _ convincing. The strobiloid theory, in which the axis is the part originally dominant, is in marked contrast to certain theories proposed by earlier writers, in which the leaf either precedes the axis or is simultaneous with it in its appearance. A. W.-E. 2. Linnaeus ; by Dr. VaLcKENIER SURINGAR; pp. 106. S. Gravenhage, 1908 (Martinus Nijhoff)—The author gives an account of the more important works published by Linnaeus with an estimate of their value and indicates the influence which they exerted upon later writers. In the course of the treatise he calls attention to certain passages, some of them now almost forgot- ten, which contain the germs of some of the theories which have played an important part in the subsequent development of bio- logical science. A. W. E. 3. Die Algenflora der Danziger Bucht,; ein Beitrag zur Kenntniss der Ostseeflora ; by Prof. Dr. Laxowrrz, Oberlehrer am koéniglichen Gymnasium in Danzig. Pp. vii+141, with 70 text-figures, 5 double plates, and a map. Danzig, 1907 (pub- lished by the Westpreussischer Botanisch-Zoologischer V erein).— The present work represents an important contribution to our knowledge of the algal flora of the Baltic Sea. In the first see- tion the numerous'species occurring in the Danzig Bay are fully - described and figured and artificial keys are provided for the determination of the genera. In the second section the flora is described from an ecological standpoint, and the distribution of the various species represented is discussed. The five double plates give photographic reproductions of the larger algae in natural size and the text-figures show microscopic details. A. W. E. 4. A Text-Book of Botany ; by Professors STRASBURGER, Noiti, Scupenck and Karsten; third English edition revised with the eighth German edition by Dr. W. H. Lane; pp. x+ 748, with 779 illustrations, in part colored. London, 1908 (Macmillan & Go.).—The Bonn Text-Book of Botany has now reached its ninth German edition, the first one having appeared in 1895. The demand for so many editions within so short a time gives an indication of its great popularity, and it is without doubt the most comprehensive and satisfactory botanical text of the present time. When the third English edition is compared with the second, which was published in 1898, the most important changes are to be found in the section devoted to the Phanerogamia. This portion of the work was originally prepared by Professor A. F. W. Schimper, but upon his death was entirely rewritten ' by Professor Karsten. Schimper’s treatment still appears in the second English edition, but 1s replaced by Karsten’s in the third. The three other sections of the book, devoted respectively to Morphology, Physiology, and the Cryptogams, are brought down to date, and a copious index of literature concludes the volume. A. W. E. Relief Map of the United States We have just prepared a new relief map of the United States, 48 x 32 inches in size, made of a special composition whieh is hard and durable, and at the same time light. The map is described in detail in circular No. 77, which will be sent on request. Price, $16.00. WARD’S NATURAL SCIENCE ESTABLISHMENT, 76-104 College Ave., ROGERS EWR. NY: Warn’s Naturat Science EstaBlisHMENtT A Supply-House for Scientific Material. Founded 1862. Incorporated 1890. DEPARTMENTS: a Geology, including Phenomenal and Physiographice. Mineralogy, including also Rocks, Meteorites, etc. Palaeontology. Archaeology and Ethnology. Invertebrates, including Biology, Conchology, ete. Zoology, including Osteology and Taxidermy. Human Anatomy, including Craniology, Odontology, etc. Models, Plaster Casts and Wall-Charts in all departments. Circulars in any department free on request; address Wards Natural Science Establishment, 76-104 College Ave., Rochester, New York, U.S. A. CONTENTS “Page Art. XII.—Ré6le of Water in Tremolite and Other Minerals; | by E. T. —— and. JK: Chemin 422 3a ee ee 101 XI Guerre Determination of the Radium Emana- tion in the Atmosphere ; by G. C. Asaman ___._-.--.= 119 XIV.—Determination of Small Amounts of Barium in Rocks; by. dh. WoCLANG LEY 20 oe 123 XV.—Heat of Combination of Acidic Oxides with Sodium: Oxide and Heat of Oxidation of Chromium; by W. G. MAX TER oc ie Ba es ed ee 125 XVI.—Concerning Certain Organic Acids and Acid Anhy- drides as Standards in Alkalimetry and Acidimetry ; by I. K. PHEwes and Li: Ee Waeep 22 2c es ee 138 XVII.—Comparison between Succinic Acid, Arsenious Oxide and Silver Chloride as Standards in Iodimetry, Acidi- metry and Alkalimetry ; by I. K. Parzps and L. H. WEED Ales oe ee ee eg ee 143 XVIIT.—Orthoclase Twins of Unusual Habit; by W.. EH. Forp and KW. Timorson, Uric oe eee 149 XIX.—Palisade Diabase of New Jersey; by J. V. Lewis _. 155 XX.—New Horse from the Lower Miocene; by F. B. WsOOMIS* leo eg a ee re SCIENTIFIC INTELLIGENCE. Geology—Indisches Perm. und die permische Wiszeit, E. Koken, 165.—Geo- logical Survey of Western Australia, Bulletin 29: Illinois State Geological Survey: Map of Vesuvius, 166.—Pocket Handbook of Minerals, G. M. BUTLER, 167. Botany—Origin of a Land Flora, F.O. Bowrr, 167.—Linnaeus, V. SURINGAR : Algenfiora der Danziger Bucht; ein Beitrag zur Kenntniss der Ostseeflora, Laxowi1z : Text-Book of Botany, 168. ae gilt lla a ia ti r. Cyrus Adler, Librarian U. S. Nat. Museum. ee, ae ee SEPTEMBER, 1908. Established by BENJAMIN SILLIMAN in 1818. AMERICAN JOURNAL OF SCIENCE. Epirorn: EDWARD S. DANA. ASSOCIATE EDITORS Proressorss GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or CamsBrmce, | / | | | | Proressorss ADDISON E. VERRILL, HORACE L. WELLS, | L. V. PIRSSON anv H. E. GREGORY, or New Haven, © | | | | | Proressor GEORGE F. BARKER, or PHILADELPHI, ; Proressors HENRY S. WILLIAMS, or Irwaca, : Proressor JOSEPH S. AMES, or Battmore, - Me. J. S. DILLER, or Wasurncron. FOURTH SERIES | VOL. XXVI—[WHOLE NUMBER, CLXXYI.] No. 153—SEPTEMBER. 1908. WITH SUPPLEMENT. : NEW HAVEN, CONNECTICUT. 1909 THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE™STREET. / Published monthly. Six dollars per year, in advance. $6.40 to countries_in the Postal Union ; $6.25 to Canada. Remittances should be made either by“monéy/ord e158; registered letters, or bank checks (preferably on New York banks). A NEW FIND OF RARE CINNABAR CRYSTALS. We have just received one of the most remarkable consignments of Cinnabar Crystals ever sent to this or any other country. They were so remarkable in crystallization and color that we immediately contracted for the entire output. They are from Wanshanchang (Hamlet of Ten Thousand Hills), Tungyen Prefecture, Province of Kweichow, China. They occur in ordinary and interpenetrating twins of bright ruby-red color. In size the crystals are from} to2inches. The matrix is a pure white quartz,the crystals always occurring in cavities, with quartz crystals; prices $5, $10, $15, $20, $50 and $100, and the supply is limited. OUR NEW BULLETIN. We have issued a new 10-page bulletin, covering in brief all our depart- ments. The announcement of this in the last number of the Scientific Journals, together with the announcement of an introductory reduction, met with so great a success and pleased so many old and new customers that at their earnest solicitation we have decided to extend it for another month. We will therefore during September only, allow a reduction of Twenty Per Cent on Common Minerals and Ten Per Cent on Fine and Rare Minerals, polished Minerals, and CutGems. The departments treated are as follows :— Showy Minerals, Rare Minerals, New Finds of Minerals, Gems, rough and cut, Mosaic Collec- tions, Geological Specimens, Ore Collections, Indian Relies, etc. Write us to-day for this Bulletin. CUT GEMS. Among the most remarkable and rare of our immense stock of Cut Gems are the following :—Green Garnets, Aquamarines, Emeralds, Tourmalines, Zircons, Sapphires, Sta) Sapphires, Star Rubies, reconstructed Rubies and Sapphires, Chrysoberyl Cats-eye, Opals, Topaz, Spinel, Pink Beryls, Sphene, Andalusite, and other precious and semi-precious stones. We shall receive in a few days a very fine lot of rare minerals from Europe ; they are in the Custom House now, and a list of same will be furnished on application. A. H. PETEREIT, 81—83 Fulton Street, New York City. THE AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES.] 0 e—__—__— Arr. XXI.—On the Retardation Of. Aipha Tedys: by Metal Foils, and its Variation with the Speed of the Alpha Particles ; by T. Smira Taytor. [Contributions from the Sloane Laboratory of Yale College. | Introduction. In 1905, Bragg and Kleeman* observed that the a-ray ionization curve, obtained with a metal sheet over the source of the rays (radium), did not suffer the same drop from the normal curve at all points. The portion of the curve corre- sponding to radium C suffered a greater drop than that due to radium itself. The loss of-range of the a particle of radium C in passing through the sheet of metal was thus greater than the loss of range of the a particle from radium in passing through the same sheet. This seemed to be evidence that the slower a particles from radium were a little less affected by their passage through the sheet than the swifter a particles from radium C. Kucera and Masekt measured the amount by which the range of the a particles from radiotellurium (polonium) was eut down by their passage through a sheet of aluminium and a sheet of platinum. They measured this diminution in range produced by the aluminium and the platinum when the sheet was directly over the radiotellurium and when the sheet was 1°9™ from it. They found the diminution produced in the latter case to be less, by 10 per cent for the aluminium and 12°5 per cent for the platinum, than in the former case. They also describe experiments from which one can conclude that the range lost by an a particle in going through two sheets of dissimilar metals is independent of the order in which it passes through them. This appears to be inconsistent with the previous statement. Experiments made by McClung.t Rutherford$ and Levin| are also inconsistent with Brage’s results as well as with the * Phil. Mag., Sept. 1905. + Phys. Zeitschr., xix, pp. 630-40, 1906. ¢ Phil. Mag., Jan. 1906. § Phil. Mag., Aug. 1906. || Phys. Zeitschr., xv, 519-521, 1906. Am. Jour. Sc1.—FourtH Series, VoL. XXVI, No. 155.—SrpremeBer, 1908. 13 170) = —-T. S. Taylor—Retardation of “ Alpha Rays.” first statement of Kucera and Masek above. Rutherford found that the a particle loses energy at a uniform rate dur- ing its passage through aluminium. McClung; Levin, and Kucera and Masek* obtained results indicating that each suc- cessive layer of aluminium foil which is laid upon a radio- active substance diminishes the range of the a particle by an equal amount. On the other hand, Brage+ observed that the stopping power is not independent of the speed as stated by Rutherford, McClung and Levin; Bragg found that the ionization curve, obtained when a sheet of gold foil was directly above the radium C, suffered a greater drop than when the sheet was at a distance of 1°5°™ from it. The range of the a particle when ~ two metals, Al and Au, were placed over the radium,. was not independent of the order in which they were arranged. The range was diminished by a greater amount when the gold foil was next to the radium and the aluminium over it than when the order was reversed. Similar results were obtained for aluminium and tin when their order was reversed as in the above case. Meyert also observed this same effect. Kucera and Masek,§ and Meitner,| ascribe these latter effects, as they observed them, to a difference in the amount of the scattering of the rays by the two metals. Their argu- ment as presented does not seem to be conclusive; for the scattering, if it exists, is a differential effect. In all the experiments cited, the method of obtaining the air equivalents{] corresponding to different positions of the screen has been to measure the range with the source uncovered and then with the screen in place. The difference of these values gives the air equivalent of the sheet. As, however, the variations in the air equivalent are small, this is not a very accurate or sensitive method; since the air equivalent itself (whose small variations are to be observed) is obtained as the difference between two much larger quantities (the ranges) neither of which can be determined with great accuracy. A method of observing the varzations of the air equivalent was sought which should be free from these dis- advantages. Description of Apparatus. The apparatus, as shown in figures 1 and 2, was similar to that used by Brage. The ionization chamber, which was 5™ in diameter and 2™" in depth, was formed by the wire gauze, * Loc. cit. + Phil. Mag., April, 1907. + Phys. Zeitschr., xiii, 425, July, 1907. § Ibid. vii, 19, 1906. || Ibid. viii, 489, 1907. "| By air- equivalent i is meant the diminution in the range of the a particle produced by the foil when placed over the source of rays, or the amount by which the range of the a particles in air is cut down by their passage through the foil. LAOS DS SE a ee Jur T. 8. Taylor— Retardation of “ Alpha Rays.” Fig. 2. Hig, 1c 172 TL’. S. Taylor—Retardation of “ Alpha Rays.” A,and the brass plate, B. The gauze A was carefully insu- lated from the plate B by means of an ebonite ring G. To prevent any possible leakage from A to B a fine copper wire was run in a groove around the inner edge of this ring, then through the ring and around its outer edge, and, after passing through the large ring of ebonite R, to the earth. A second gauze °C was placed 3™™ below A and, being earthed, formed with it a second ionization chamber which prevented stray ions from entering the main chamber. The plate B was connected to one pair of quadrants of a Dolezalek electrometer as shown in figure 2, the second pair of quadrants being earthed. K is a key by means of which the one pair of quadrants and plate B could be earthed or insulated at pleasure. K rested on a brass plate M, which in turn was insulated from the protecting case 2 by being placed upon a block of sealing wax N. The polonium, which was used as the source of rays, being put upon the plate V, could be moved towards or from the — ionization chamber by means of the screw W, of 1™™ pitch, working in the nut X. Its distance from the chamber could be determined by the scale L and the graduated circular dise 7. The metal sheets were placed upon the brass ring ¢, ¢ and could be moved to different positions by means of the screw H, of 1™™ pitch, working in the nut T. The distance of the metal sheet from the polonium could be read from the scale L by means of the pointer P, which extended below the case I’ from the moveable framework emZ. The polonium kindly prepared for me by Prof. Boltwood was deposited upon the end of a copper plug, which could be fitted into a block of brass and adjusted to a definite distance from the top of the block. The opening in the block above the polonium was 4"" in diameter and 6"™ high. This limited the cone of rays given off by the polonium so that all the rays would fall within the ionization chamber. The ebonite was carefully polished to prevent leakage over its surface. The entire framework of the apparatus together with the tin case F and the protecting case ? for the connec- tion of the plate B to the electrometer was grounded. Care was taken to avoid all outside electrostatic effects upon the electrometer as well. Method of Hxperiments. The block contaming the polonium was put upon the dise V and a metal sheet on the ring ¢, ¢, which was then adjusted by means of the screw H until the sheet just touched the top of the block, thus being 6"" from the polonium. With a con- stant potential of —40 volts on the gauze A and the metal sheet at a distance of 6™" from the polonium, the Bragg ion- ization curve was plotted. T. S. Taylor—Retardution of “ Alpha Rays,” 173 In figure 3, curves I and II represent parts of the top por- tion of two Bragg curves®* as obtained when a sheet of gold foil was kept at a distance of 6™ from the polonium. Curve I was plotted from readings taken at the end of the first minute and curve II from readings taken at the end of the second minute. The curves being plotted with magnified ordinates have much greater slopes than they would have if plotted on the ordinary scale. The polonium, with the gold foil 6™™ above it, was then set at a distance of 3-0 from the ionization chamber. As can be seen from curves I and II, this distance was such that the KS] fe NON eae ee a1 iter ay seneee y ] Y iL 6 Fie. 3. The ordinates of curves I and II are the distances in centimeters of the polonium from the ionization chamber. The ordinates of III and IV are the distances in centimeters of the sheet of gold from the polonium when the polonium was kept at a distance of 3°0°™* from the ionization chamber. The abscissas of I and III are the deflections in centimeters of the electrometer needle at the end of the first minute, and the abscissas of Il and IV are the deflections at the end of the second minute. deflections obtained corresponded to two points in the slightly inclined (top) portions of the Bragg curves. With the polo- nium at this distance of 3:0™ from the ionization chamber, the gold foil was moved away from the polonium towards the ionization chamber, and the ionization was measured when the sheet was at various distances from the polonium. It was found that the ionization increased as the distance of the gold sheet from the polonium increased. In figure 3, curves III and IV were plotted with the deflec- tions of the electrometer needle at the end of the first and sec- ond minutes respectively as abscissas, and with the distances * This is the portion of the ionization curve which, as ordinarily plotted, is nearly horizontal. 174 T. S. Taylor—Retardation of “ Alpha Rays.” of the sheet of gold from the polonium, corresponding to these deflections as ordinates (the polonium being kept at a distance of 3:°0™ from. the ionization chamber). Curves similar to III and IV in figure 3 were also plotted for the gold foil when the polonium was set at distances other than 03°™ from the ionization chamber. The distance of the polonium from the ionization chamber was always so chosen that the chamber eut the Bragg curve somewhere in the slightly inclined por- tion. The Bragg curves I and II were determined each time before making the measurements plotted in curves [III and IV for any given distance between the polonium and the ioniza- tion chamber. In this way, several sets of curves similar to the ones in figure 3 were obtained for each metal sheet given in Table II. ! The diminution in the range of the a particle produced by the sheet when 6"" from the polonium (its air equivalent in this position) was determined by first plotting the ionization curve without the sheet over the polonium and then with the sheet over it. The difference between the ordinates of the two curves corresponding to a given abscissa, which was the deflection of the electrometer needle, was the diminution in range of the a particle produced by the sheet, or what is the same, its air equivalent. This is the same as taking the deflec- tion of the needle without the sheet over the polonium when it is at a certain distance from the chamber, and, after placing the sheet over the polonium, noting how far the polonium and sheet must be moved towards the ionization chamber to get the same deflection of the needle. Determination of the Variation in the Air Equivalents. As previously stated, when the polonium, with a metal sheet 6™™ above it, was set at such a distance from the ionization chamber that the chamber cut some part of the slightly inclined portion of the Bragg curve, it was found that, by moving the sheet farther away from the polonium, the ionization increased as the distance of the sheet from the polonium increased. This signifies that by moving tbe sheet away from the polonium the entire ionization curve was pushed upward, so that the part of the slightly inclined portion, which fell just below the ionization chamber when the sheet was 6"" from the polonium, fell within the chamber when the sheet was at a little greater distance from the polonium. This means that the total range of the a particles is greater with the sheet at the greater distance from the polonium ; and the magnitude of this increase in total range (which is also the diminution in the air equivalent of the sheet) can be obtained directly from the measured ionizations, given in curves, III and IV, and the curves I and II, fig. 3. ee -T. 8. Taylor—Retardation of “ Alpha Rays.” 175 The diminution in the range produced by a given sheet when 0-6 from the polonium, having been determined by the direct method as previously stated, the diminution at any other dis- tance was found from the curves in figure 3, which were plot- ted on finely ruled co-ordinate paper, thus facilitating the determination. The curves given in figure 3 were for a sheet of gold (A in Table II), which cut down the range 0°635™ when it was 0-6 from the polonium. The method of making the determination can be illustrated by finding the diminution in range produced by the sheet when it was at a distance of 1°8™ from the polonium. In curve III the abscissa, corresponding to the ordinate 1°8, which is the distance of the sheet from the polonium, is seen to be 1°5. The abscissa 1°5 in curve I corresponds to the ordinate 2°956 in the same curve. Thus the increase in the deflection of the electrometer needle produced when the sheet is moved from 0-6™ to 18°" from the polonium is the same as would be pro-. duced if the polonium with the sheet 6™™ above it were moved from 3°0™ to 2°956™ from the ionization chamber. The total range of the a particles when the sheet is 1:8 from the polo- nium, is greater by [3°0—2°956] 0°044™ than the total range when the sheet is 0°6™ from the polonium. In a similar man- ner we find from curves IV and II this change is equal to 0-042°". The mean value for the two determinations is then 0-043". The barometer stood at 760°2™™ when the curves were plotted, and hence this change of the range when reduced to barometer 760™" is 760°2 XK 0°043 760 Since there is this difference in the total range of the a parti- cles in the two cases, it is evident that the sheet does not cut down the range of the a particles as much when it is 1°8™ from the polonium as when it is only 0°6™ from the polo- nium, and that the air equivalent of the sheet at 1:8™ distance from the polonium is 0°048™ less than at 0-6. In this manner the air equivalent of the sheet was found when it was at various distances from the polonium. In Table I are given the values of these air equivalents for the A sheet of gold corresponding to the distance of the sheet from the polonium and the range that the a particle still had in air at atmospheric pressure when it entered the sheet. The range of the a particle in air at atmospheric pressure was found to be 3-77. This was the distance from the polonium to the upper side of the gauze A when the ionization, due to the a particles from the polonium, produced the first noticeable deflec- tion of the electrometer needle above that due to the spon- taneous ionization of the air in the chamber.* The spontaneous * If the distance had been measured to the middle of the ionization cham- ber the range would be 3°87", which is the value usually taken. —0-043em 176 TS. Taylor—fRetardation of “ Alpha Rays.” ionization of the air in the chamber produced’ less than 0:04™ deflection. This is so small as to be practicably negligible. But had it been larger it would have made no difference ; because it would have the effect of shifting all the curves in figure 3 to the right by an equal amount, thus not changing their relative positions or forms. Column 1 in Table I contains the distance in centimeters of the sheet from the polonium. Column 2 gives the range in centimeters of the a particle when entering the sheet. In column 8 are given the air equivalents in centimeters of the sheet corresponding to the distances in column 1 and the ranges in column 2 as determined when the polonium was set at 3°1™ from the ionization chamber. Columns 4, 5, 6, 7 and 8 contain the same quantities as column 3, as determined with the polonium at 2-9, 3-0, 3-1, 2°8 and, B-Qem respectively from the chamber. TaBLE I. A Au. Pom pon|oamee cbr tl. ee. ee ae entering 3°1 2°9 30 ori 2°8 3°0 Means. the sheet particle 0°6 Boa 0°635 | 0°635 | 0°635 | 0°635 | 0°635| 0°635!| 0°635 0°8 2°97 0°626| 0°628] 0°629)| 0°626| 0°629| 0°630}, 0°628 1°0 2277. 0°619| 0°620| 0°622| 0°618)| 0°621 | 0°623)| 0-621 r2 Da) 7h 0°613/ 0°614| 0°616| 0°611 | 0°613| 0°616)| O°614 1A Nay 7 0°606 | 0°609| 0°608/ 0°603 | 0°605 | 0°609]| 0:607 1°6 Dealig 0°600| 0°601 | 0°600| 0°596 | 0°595 | 0°601 || 0°599 les 1:97 0°592| 0°595]| 0°592| 0°589} 0°584| 0°592)| 0591 2-0 ony 0°582| 0°581| 0°582) 0°579 | 0:574 | 0°581 || 0°580 LD, Veo 0°572| 0:°567| 0°571/ 0°569/} 0°564| 0°570]) 0:569 2°4 1a 7 0°556| 0°548| 0°555| 0°556| 0°557 | 0°5521|) 0°553 The last column contains the average values of the six pre- ceding columns. The agreement between corresponding quan- TABLE IT. I II III IV Metal sheets | Thickness in cms. | Air equiv. in ems. Ratio A Au 0:000127 0°635 4:99 X 10° B Au 0:000174 0°857 4°91 10° A Pb 0°000284 0°923 3:24 aioe B Pb 0°000411 1231 3 Ocal A’ Sn 0°000386 0°886 2°29 10° B Sn 0:000799 1°800 eS) MN) A Al 0°:000338 0'574 1°74 10° B Al 0°000612 1°04 1°70 X 10° A Ag 0°000228 0°676 3°03 x 10° T. 8S. Taylor— Retardation of “ Alpha Rays.” 177 tities in these columns shows that possible errors in plotting the Bragg curves (e. g., I and II in figure 3) produced no errors of importance in the reduction. In this manner, the air equivalent of the metal sheets given in Table II were determined for the various distances of the sheets from the polonium. The values obtained are found in Table III. TABLE ITI. Range in ems. of | 4 aAu|/BAu|A Pb|B Pb|ASn| BSn| A Al| BAl | A Ag entering particle. 3°17 | 0-635] 0°857| 0-923) 1-231) 0-886] 1°800| 0°574! 1:040| 0°676 2°97 | 0°628) 0°849| 0°914| 1°220| 0°879| 1-787) 0°574| 1029] 6-673 2°77 =| 0°621) 0°839}| 0°903] 1°207| 0-871) 1-768] 0°573) 1°017| 0-669 2°57 | 0°614| 0°828} 0°891) 1°193) 0°862| 1°744| 0-572) 1-005} 0°664 2°37 0°607| 0°816| 0°878| 1-178) 0°852| 1°718] 0°571| 0°990| 0-659 2°17 | 0°599} 0°801} 0°861| 17160} 0-840 0°568| 0°970| 0°652 1°97 =| 0°591) 0°783} 0°843| 1°138] 0°827 0°563} 0°949| 0°644 1-77 | 0°580/-0°761| 0°819 0°810 0°555 0°635 1°57 0°569| 0°735 0-797 0°544 0°625 1°37 0°553 0°532 0°613 The values of the air equivalents for each of the metal sheets recorded in Table III represent the results obtained from a series of determinations similar to that given for A Au in Table I. The agreement in the values, obtained for the different positions of the polonium, was equally as good as that for the A Au. These separate tables for each metal sheet have been omitted for the sake of brevity and only the average results given in Table III. The curves in figure 4 represent the results as recorded in Table III. By noting the slope of these curves, one can obtain some comparison of the rate at which the air equiva- lents of the various sheets are changing. For the thinner sheet of Al, the air equivalent is almost constant for the higher ranges, but, as the range of the entering a particle decreases, the air equivalent decreases slowly and, in the lower ranges the decrease becomes quite marked. The thicker Al, however, shows a marked change in its air equivalent for even the high ranges. The statement of McClung, Levin and Rutherford that equal successive layers of aluminium foil diminish the range by equal amounts seems to hold true for thin sheets of foil when the range is high; but when the metal sheet is thicker, or for thin sheets when the range is low, it does not hold. The slight difference however in the air equivalent of 178 T. S. Taylor—Retardation of “* Alpha Rays.” the thin foil when near the polonium and when farther away from it would scarcely be detected by measuring directly the air equivalents in the two positions. This is probably the explanation of the above statement by McClung, Levin and Rutherford, since their determinations were made in this direct way. Fic. 4. The abscissas are the ranges in air of the a particles when they enter the metal sheets; the ordinates the air-equivalents of the metal sheets. The eurve for BSn has the ordinates from 1°718 to 1°80 as indicated on right-hand side of the figure instead of 1°118 to 1°20 as shown on left-hand side. T. 8. Taylor—Retardation of “ Alpha Rays.” 179 Since the air equivalent does decrease with the range of the a particle entering it, the ratio of the air equivalent to the thickness of a given sheet of metal should be less than the same ratio for a thinner sheet of the same metal. This is shown to be true by the last column of Table IJ. From the curves in figure 4, we see that the change is more pronounced the thicker the metal foil and the heavier the metal. The possibility that the observed variations in the ionization may be due to secondary rays is precluded by the fact that numerous direct determinations of the Bragg ionization curves with the metal sheets near the polonium and again near the. ionization chamber showed no irregularities in the curves, as would be expected were secondary rays present in any appre- ciable amount. Thus far, from a comparison of the data given in Table ITI and represented graphically in figure 4, I have been unable to deduce any definite statement relative to the rate of change in the air equivalents, except that, for the same metal, the thicker the foil the more marked is the change and that, for sheets of different metals of about the same air equivalents, the rate of change is in the order of their atomic weights; G.e., Al, Ag, Sn, Au, Pb). Further experiments are in pro- gress and it is hoped that they will furnish the desired infor- mation, when completed. In conclusion, I wish to express my gratitude to Prof. Bumstead, at whose suggestion these experiments were under- taken, for his valuable suggestions, and interest in the work ; also to Prof. Boltwood, who kindly prepared the polonium for me and gave ine many helpful suggestions. Results. The air equivalents of metal foils decrease with the range (i. €., with the speed) of the a particles entering them. The change is very small for thin foils of the lighter metals when the range of the a particles is high; but, when the range is low for thin sheets, or when the sheets are thicker, the change becomes quite marked. In comparing the change for sheets of different metals of nearly equal air equivalents, the rate of change is seen to be in the order of the atomic weights of the metals. New Haven, Conn., June 15, 1908. 180 Lee—Lower Paleozoic Rocks of Central New Mexico. Arr. XXII.—WVotes on the Lower Paleozoic Rocks of Central New Mexico ;* by Witus T. Luz. Tur lower Paleozoic rocks are exceptionally well exposed in the Caballos Mountains of central New Mexico. Gordon and Gratont+ refer to them in a description of the lower Paleozoic formations in New Mexico, but attention is confined by these writers principally to other parts of the territory, and the occurrence, especially of the Cambrian and the Ordovician ‘formations, in the Caballos Mountains, as well as their relations to each other and to the overlying rocks, are of sufficient inter- est to warrant the more definite statements contained in this paper. : The Caballos Mountains form one of the small mountain groups of central New Mexico and occur west of the town of Engle between the Rio Grande and the Jornado del Muerto. The northern part of the mountains consists of a faulted block tilted to the east and the sedimentary formations outerop in the precipitous western face. The lower part of the slope consists of granite, above which occur the Cambrian, Ordovi- cian and Carboniferous rocks as shown in the following section measured about three miles north of Shandon, a mining camp on the Rio Grande. With the Cambrian and Ordovician of this section are correlated similar rocks observed.in other parts of the Rio Grande Valley. A few species of fossils were found in the Cambrian sedi- ments, and these have been identified by Dr. C. D. Walcott of the Smithsonian Institution. Fossils collected from the Ordo- vician rocks have been examined by E. O. Ulrich of the United States Geological Survey, and the quotations given in the paper are from his written report. The Ordovician fossils are not well preserved and specific identification is difficult. A thickness of 1000 feet or more of the granite is exposed in the cliffs. The rock is massive and coarsely crystalline, although schist and gneiss occur in some places. Its surface was apparently eroded in early Cambrian time to a nearly level plain upon which the sedimentary rocks were deposited. The coarse-grained and, in some places, conglomeratic quartz- ite at the base of the Cambrian grades upward into the green shale, in which the Cambrian fossils occur in great abundance at several horizons, some close to the basal quartzite, others near the top. * Published by permission of the Director of the U. S. Geological Survey. + Gordon, C. H. and Graton, L. C., Lower Paleozoic Formations in New Mexico, this Journal, xxi, pp. 390-395, 1906. Lee—Lower Paleozoic Rocks of Central New Mexico. 181 Section of Rocks exposed in the Caballos Mountain three miles north of Shandon, New Mexico. No. Ft. (1) Limestone, blue, massive, fossiliferous, in thick plates separated by dark colored shales (age, lower Pennsyl- ANNs UY Fake 2 A as Se he eS aie 600 (2) Limestone, cherty in places, white to brown (age not determined) ey Meee LN mere ts 500 (3) Limestone, cherty, massive, cliff- making. The following fossils occur near the top: Rafinesquina cf. hingi, Plectambonites saxea, Plectambonites n. sp., Favosites asper, Zygospira recurvirostris (Richmond mutation), Rhynchotrema capax (age, late Ordovician) -.------ 600 (4) Shale, dark green, containing Odbolus ( Westonia) Ston- eanus Whitfield, Obolus sinoe Walcott? (The fol- lowing fossils were found in talus derived apparently -from this horizon : Plectorthis desmopleura Meek, Obolus sinoe Walcott, Lingulella acutangula Roe- mere (ase, upper. Cantbriait)oi23 22565 2003.02 90 Pee aRL Albee fe ee ht see lee Fe. AO (Gp eGranite. 22.25: Ee ee ae SSE eee ae etary See mag ? Cambrian sediments occur also in Cerro Cuchillo, a hill con- sisting of rocks faulted and tilted steeply to the east, that stands a few miles west of the Rio Grande near the northern end of the Caballos Mountains. The rocks were probably originally continuous with those of the Shandon region, but the two exposures are 20 miles apart and are separated by a zone of profound faults. Here, as in the Caballos Mountains, the Cambrian sediments are about 100 feet thick and consist of a basal quartzite resting on granite and overlain by green- ish shale in which were found fragments of trilobites and the same species of brachiopods found near Shandon. The shale is overlain by chertv limestone similar to the Ordovician of the Caballos Mountains but no fossils were collected from it. The Ordovician rocks of the Caballos Mountains are appar- ently conformable with the Cambrian, but as shown below, there is probably a considerable time-break between them. They consist of massive cherty limestone and form a conspicu- ous cliff 600 feet high. The rocks of undetermined age above are less cherty and not so massive, but in some places can be distinguished from the lower chert only by careful examination, the two forming a single cliff 1100 feet high. The fossils from the lower chert indicate Richmond or late Ordovician age, and in the absence of evidence, the upper chert also may be late Ordovician or it may be Silurian. The chert is overlain in the Caballos Mountains by 600 feet of limestone of Pennsylvanian age. No fossils of intervening 182 Lee—Lower Paleozoic Rocks of Central New Mexico. age were obtained and it is possible that both Mississippian and Devonian rocks may be found here, for they are known to occur at Lake Valley and at Hillsboro, a few miles to the west.* About eight miles south of Shandon near the south end of the Caballos Mountains, the older Paleozoic rocks are again exposed, although considerably folded and faulted. No detailed section of them was made owing to their disturbed condition, but the following generalized section with estimated thicknesses indicates their order of succession : Section of Rocks exposed near Red Cabin, eight miles south of Shandon, New Mexico. . Thickness as estimated: (1) Limestone, blue (age, lower Pennsylvanian) .. _.------ 500 Unconformity of erosion. (2) -Shale;:black- fage,. Devonian (?)-): 2.022.255 22a 1-300 (3)- Limestone, cherty (age, late Ordovician) - 22-22 a25emes 500 Angular unconformity. (4) Limestone, cherty (age, early Ordovician) ..-...------ 200 (5) Quartzite (age/not determined) _=_ 9)22 422. a3 eee 300 (Base not exposed.) A few fragmentary fossils were found in the cherty lime- stone (No. 4) of the Red Cabin section. Although they are very poorly preserved, Ulrich recognizes them as Ophdleta ef, | complanata and Hormotoma ct. artemisia (Billings species) and refers them with some confidence to the lower Ordovician (Beekmantown). Cherty limestone about 500 feet thick, that is probably equivalent to the lower chert (No. 3) of the Shandon section, lies unconformably upon the early Ordovician chert in some places and upon the massive quartzite (No. 5) of the Red Cabin section in other places. Fossils were collected from it at the base, near the middle, and at the top, but as Ulrich includes the three lots in one general fauna, they may be combined as follows, together with his notes concerning them; Rafinesquina, n. sp. (characteristic of Western Richmond). Rhynchotrema capax (most characteristic Richmond fossils). Favosites asper (late Ordovician and Clinton). Petraia, sp. undet. Lindstromia, sp. undet. Dalmanella (2? Schizophoria), sp. undet. * Gordon, C. H. and Graton, L. C., Lower Paleozoic Formations in New Mexico, this Journal, xxi, pp. 394-395, 1906. Lee—Lower Paleozorve Rocks of Central New Mewico. 183 Streptelasma or Petraia, sp. undet. Specimen shows only the exterior. So far as can be seen it recalls Silurian apecies rather than Ordovician, Nematopora, cf. granosa Ulrich, a Trenton species of Min- nesota. Schuchertella sp., closely allied and possibly the same as the Silurian S, subpluna (Conrad). Fragments of a brachiopod recalling Silurian species of Stro- phonella. Small subcircular orthoid, suggesting relations to O. (Lhipido- mella) hybrida. Atrypoid brachiopod, possibly Zygospira but distinct from all American species of that genus known. Brachiopod of undetermined affinities. In its general aspect it recalls Russian species of Porambonites, but the surface of the shell is not porous. Ctenodonata, sp. undet. Lophospira, cf. gyrongonia McCoy and medialis Ulrich. Lophospira, sp. allied to gyrongonia McCoy. Fragment of a gastropod, apparently of the same species as the Clinton form doubtfully referred by Foerste to PEROT OUEES sinuatus Hall. Trochonema, n. sp. EHunema sp. near &. robbinsi Ulrich. Eecyliomphalus, sp. undet. A small, incomplete example of a closely coiled species reminding of £. gotlandicus Lindstrom and Hi. contiguus Ulrich, the former Silurian, the latter middle Ordo- vician. Loxonema? The specimen is a fragment (consisting of the greater part of two whorls) reminding rather strongly of the Z. murrayanum described by Salter from Black River limestone in Canada. In a local uplift about one mile west of Red Cabin, I found the following fossils in chert that is presumably the same as that of No. 3 of the Red Cabin section : Calopoecia canadensis Havosites axper : ? Columnaria alveolata Platystrophia dentuta var. Petraia Lindstromia Dalmanella (? Schizophoria) Lophospira Loxonema. South of Red Cabin, about two miles and five miles respec- tively, the latter loc salty about six miles northwest of Rincon, 184 Lee---Lower Paleozoic Rocks of Central New Mexico. 300 feet of cherty limestone is exposed at the foot of the cliffs. The rocks are continuous with the upper chert (No. 3) of the Red Cabin section and yielded the following fossils : Petraia, sp. undet. (near Streptelasma profundum (Black River) and S. caliculuris (Silurian)). 7 . Lindstromia, sp. nov. | Orthis, near davidsoni.and flabellum. Dalmanella (? Schizophoria), sp. undet. Clorinda? sp. undet. Atrypa sp. nov. (near Zygospira putilla and Atrypa margina- iis). 3 ? Huomphatus sinuatus (Hall) Foerste (Clinton species). Lophospira, sp. undet. Trochonema “ — « Eunema, near robbinsi, a Trenton form. Loxonema Nucleospira ? Ulrich states that all of the fossils from near Red Cabin, with the exception of the two lower Ordovician forms, belong at the same general horizon and are of Richmond or late Ordo- vician age, and suggests that the absence of fossils indicating Ordovician horizons intervening between this and the early Ordovician rocks below the uncomformity “ probably points to conditions similar to those prevailing in the vicinity of El Paso 75 miles to the south, where the Richmond commonly rests directly on Beekmantown.” Since a lower Ordovician chert occurs unconformably below the Richmond in the Red Cabin section, and this in turn rests upon a massive quartzite that is not represented in the Shandon section, it is probable that the apparent conformity of the Cambrian and Ordovician in that section is deceptive and that a time interval of considerable duration is represented between the two formations. According to Ulrich, there are certain elements in the faunas that suggest Silurian rather than Ordovician age, but, except- ing the two early Ordovician species, he is inclined to regard the fauna as representing one of the phases of the Western Richmond. Hestates that “ very few of the species are closely related to ordinary described American forms, but they are very similar and perhaps identical with species occurring in the Borkholm limestone of the Russian Baltic Provinces. In the absence of authentic examples of the Russian species, I hesi- tate as yet to identify these New Mexican fossils with them. The associated coral bed represented west of Red Cabin, how- ever, is widely distributed in America west of the Mississippi and occurs also in the Island of Anticosti. Its position is within the upper part of, or just above, the Maquoketa shale of the Mississippi Valley. The highest horizon represented Lee—Lower Paleozoic Rocks of Central New Mexico, 185 east of Red Cabin and also north of Rincon, may be correlated more or less definitely with the Maquoketa shale and this with the upper Richmond of Ohio and Indiana.”’ No fossils were found in the black shale (No. 2) of the Red Cabin section, and it may be Devonian as stated by Gordon and Graton.* It is apparently this shale that these writers have in mind when they state that the Devonian is represented in the Caballos Mountains, although it does not appear from their descriptions that fossils were found or other evidence of age except stratigraphic position obtained. The overlying Pennsylvanian limestone rests in some places upon this shale, in other places upon the Richmond chert, as shown in the Red Cabin section, and in still other places upon the chert overlying the Richmond, as in the Shandon section. No rocks of Mis- sissippian age were found, but their occurrence a few miles to the west suggests that the Mississippian limestones once extended over this region and that they, together with the greater part of the Devonian, were removed by erosion previ- ous to the deposition of the Pennsylvanian sediments. A small exposure of cherty limestone was found beneath the Pennsylvanian limestone in the northern slope of the Robledo Hills about 15 miles south of Rincon. In this chert I found the following fossils : Lophospira, two small undet. species. Lophospira, ? larger species. Trochus ? sp. undet. Bucania ? sp. undet. Trochonema, sp. undet. Hotomaria, sp. undet. Pentameroid shell agreeing with Steberellay except that it has no fold nor sinus. Ulrich regards these as constituting a part of the Richmond fauna just described, but states that they are c.early distinct from other Richmond faunas so far as known from the west and southwest. His statement that more perfect fossils may prove that the fauna is Silurian finds support in the occur- rence of Silurian rocks in the Franklin Mountains 35 miles to the south.t A thickness of only a few feet of the rocks is exposed in the Robledo Hills and their relations to other for- mations could not be determined. A summary of the foregoing statements regarding the lower Paleozoic rocks of central New Mexico may be given as fol- lows: (1) Rocks of upper Cambrian age about 100 feet thick rest upon an eroded plane of granite. (2) Rocks of early * Ibid: pa 39h + Richardson, G. B., U. S. Geol. Survey, El Paso Folio. Am. Jour. Sc1.—Fourta Series, Vout. X XVI, No. 153.—SEPTEMBER, 1908. 14 , 186 Lee—Lower Paleozoic Rocks of Central New Mexico. Ordovician (Beekmantown) age occur in some places, but are apparently absent in other places. (8) Rocks of Richmond age are well developed and rest in some places upon Beekman- town and in other places apparently upon Cambrian. (4) The Richmond is separated from the Beekmantown by an uncon- formity that apparently represents a long time interval. (5) Certain elements in the fauna which is here described as Rich- mond are suggestive of Silurian rather than Ordovician age, and some of the cherty limestone may belong in the Silurian system. (6) Rocks of possible Devonian age occur in the Caballos Mountains, but no fossils were found in them. (7) No Mississippian rocks were found and Pennsylvanian lime- stones rest in some places upon the Devonian (?) shale and in other places upon the Ordovician chert. Washington— Kaersutite from Linosa and Greenland. 187 Art. XXIIL.—On Haersutite from Linosa and Greenland ; by Henry 8S. Wasuineton; with Optical Studies by Frep. Eucene Wricur. Introductory Note.—While visiting the small island of Linosa, off the coast of Tunis, for the Carnegie Institution in the summer of 1905, I found small erystals of a black amphi- bole, accompanied by others of a glassy white, cleavable min- eral, apparently a feldspar,* among the lapilli of a small, parasitic cone of Monte Rosso.t The presence of amphibole erystals here had previously been noted by Speciale.t Similar crystals were also said to be found near I Faraglioni, but this locality was not visited. Apart from these occurrences, am- phibole is quite unknown in the lavas of Linosa, but the peculiar, triclinic aenigmatite (cossyrite) is met with, though not abun- dantly, and a kaersutite-ike hornblende as well, in the lavas of the near-by island of Pantelleria. Chemical analysis showed that the Linosa hornblende is very high in titanium, and that in this, as well as in other respects, it closely resembles the kaersutite of Greenland, which was partialiy described in 1884 by Lorenzen.§ A comparative investigation of the two minerals was therefore determined on, the chemical work being done by me and the optical determi- nations being very kindly undertaken by Dr. Wright, to whom I am deeply indebted for his valuable and hearty collaboration. Through the kindness of Professor N. V. Ussing, of Copen- hagen, we obtained a piece of one of the best of Lorenzen’s original specimens of kaersutite, and we take this opportunity to express our sincere thanks to him for his courtesy and great liberality, without the aid of which the comparison would have been sadly incomplete. The Linosa Amphibole. Physical Characters.—The Linosa amphibole is monoclinic in crystal system and prismatic in habit. It occurs in roughly developed crystals and fragments from 5 to 20™™ long by 3 to 8™™ thick, and bounded by the faces of the unit prism m (110) and the clinopinacoid 6 (010). Terminal endings are usu- ally absent, but a few of the crystals show the common forms p (101) and r (011), which, however, were too imperfect to admit of accurate measurement with the goniometer. Cleavage parallel to m (110) is highly perfect, viving an angle of 55° 22° * A description of this mineral will be given in a separate paper. + H. S. Washington, Jour. Geol., vol. xvi, p. 10, 1908. tS. Speciale, Boll. Com. Geol. Ital., vol. xv, p. 2, 1884. § J. Lorenzen, Medd. Gronl., vii, p. 27, 1884. 188 Washington—Kaersutite from Linosa and Greenland. (Wright) with observed limits of + 2’, measured on cleavage fragments with a two-circle voniometer with reducing attach- ment. The reflection signals were fairly sharp and satisfactory. No cleavage or parting after 100 or 001 was noted with cer- tainty, and whenever cleavage is not developed the fracture is conchoidal. The hardness is 6 and the mineral is very brittle. Before the blowpipe it fuses readily to a black, slightly mag- netic bead. Fic. 1. Kaersutite from Linosa. Etch piton 110, HF. x 480. The specitic gravity was carefully determined with the pye- nometer on about 2 grams of selected fragments, entirely free from adhering bits “of scoria or feldspar and quite free from inclusions so far as could be seen with a lens. With this mate- rial the density at 13° was found to be 3°336 (Washington), a figure which may be accepted as representing the true value. Etch figures on m (110) were produced by immersing cleav- age fragments in hot commercial hydrofluoric acid (on a steam bath at 100° ) for a period of 30 seconds. Further action was stopped by plunging the platinum basket containing the frag- ments into cold water. Under these conditions of experiment well-formed etch pits resulted, from 0:01 to 0-04" long and about half as wide. Different stages of development are illus- trated in figures 1-3, which are reproductions of photomicro- graphs of the figures in reflected light. In each ease the vertical edge of the photograph is parallel to the prism-axis. A com- parison of these figures with those obtained by Daly,* and later by Wright,t shows that they resemble in certain features the Hane A. Daly, Proc. Amer. Acad. Arts and Sci., vol. xxxiv, pp. 3383-429, 1899. +F, E. Wright, Tschermak’s Min. Petr. Mitth. vol. xix, pp. 308-320, 1899. Washington—Kaersutite from Linosa and Greenland. 189 pits formed on lustrous basaltic hornblende and in other par- ticulars the etch figures on barkevikitic hornblende. The etch figures on the faces of the prism zone prove with certainty that the mineral is monoclinic, and that it belongs in the general group of the dark, highly ferruginous, aluminous amphiboles, which are commonly referred to the hornblendes. Optically, this amphibole is remarkable in several of its properties. The color is an intense jet black, with highly vitre- ous, splendent luster. The streak is light brown. In common with amphiboles of this group, it is brown to pale brown in transmitted light, and is strongly pleochroie: c=dark brown, 2 3 Figs. 2, 3. Kaersutite from Linosa. Etch pits on 110, HF. Fig. 2, x 360; fig. 3, x 240. almost opaque, b=brown, a=pale olive brown or olive green; absorption, c >b >a. The natural color of the mineral is so deep and the pleochroism so strong that the normal inter- ference colors are greatly modified, and the observation of the optic axes and similar optical phenomena is considerably hindered. Owing to the extreme brittleness of the cleavage fragments, it was found difficult to prepare sections parallel to the clino- 190 Washington—Kaersutite from Linosa and Greenland. pinacoid, and the extinction angle was measured only on the cleavage face m (110). By using the etch figures as a ‘basis for orientation, the extinction angle cAc was found to be +1-4° (in the acute angle 8) in white light,* as indicated by the arrow in figure 1. This direction presupposes the standard erystallo- graphic orientation of the amphiboles by Tschermak. This extinction angle is noteworthy because of its positive character, and it is readily discernible, although so small that Gn the absence of corroborative data) it might be considered to be due to observational error alone in the deep-colored flakes. The dispersion of the bisectrices is very slight, and practically the same value was obtained by using sodium light as that for white light. In convergent polarized hght the interference phenomena are only moderately clear and distinct. The optic axes lie in the plane of symmetry (010), and the optical charac- ter is negative. The refractive indices were determined on a very perfect cleavage flake with an Abbé total refractometer, a solution of sulphur in methylene iodide, with refractive index 1°7882, being used. The observations were made in sodium light, and the following values obtained : Average angle Equivalent refrac- observed. tive index. —— G7 aadlsy 1°:760 iS 657° 78: 1°730 == (HO. Bie 1°692 (ip 0'068, Vie B = 0°029 6B —a = 0°039 The optic axial angle calculated directly from these refrac- tive indices is 2V = 79° 54’. The boundary shadow for a was much more distinct than those for 8 or y, and could be determined with greater accu- racy. In the values given for 8 and y an error of + 0-002 is easily possible. The fact that hght waves vibrating along ¢ and b were strongly absorbed undoubtedly exerted an influence on the relative intensity of the phenomena observed, and caused the faintness of the 8 and y curves of total reflection. A somewhat smaller value for 2V was obtained by meas- uring the angle of the optic axes directly in another section cut approximately normal to the acute bisectrix, by the method described by Wright.| The deep color of the mineral impaired the accuracy with which the determinations of the positions of the optic axes could be made, so that the results are but approx- imate at best. In two different portions of the same section thus measured the values 2V = 71° and 72° were obtained. * Average of 10 measurements, with observed limits +0°6°. + F. E. Wright, this Journal, vol. xxiv, p. 317, 1907. Washington—Kaersutite from Linosa and Greenland. 191 While in general the probable error for this method should not exceed 2°, the intense color of the amphibole has evidently affected this limit appreciably, as it is not probable that differ- ences exist in the chemical composition of the material suth- ciently great to cause the optic axial angle to vary 9°. The optic axial angle apparently lies between the two extremes, Gkxand: 19° 54’, but it is thought that the latter more closely approaches the true value. Chemical Composition.—For the chemical analysis several grams of selected crystal fragments were coarsely crushed, washed free from dust, and the material (dried at 110°) care- fully picked out under a lens. The only adherent impurities were feldspar, limonite, and particles of the scoria, and all frag- ments showing traces of these were excluded. Thin sections of fragments showed but very few small inclusions of magnet- tes but these were separated by treatment with an electro- magnet, only a very small amount being thus extracted. It is believed that the material as finally pulverized for analysis was practically free from extraneous matter. Treatment with acid for purification was not resorted to, as the mineral is partially decomposed by acids.* The methods of analysis employed were those advocated by Hillebrandt+ and the writer,{ about one gram being taken in each case for silica, alumina, etc., and for the alkalies; about one-half a gram for ferrous iron; and 0°8 gram for fluorine. The alkalies were determined by Lawrence Smith’s method, titanium colorimetrically (a mean of three closely agreeing determinations), and manganese by precipitation with bromine. Ferrous iron was determined twice by the simple Pratt method, freshly standardized permanganate solution being used. The figure given (3°96) is the mean of 3:99 and 3:94. These results indicate that there was no appreciable oxidation of the FeO in the finely ground powder during the interval of a year which elapsed between the two determinations. * In a recent criticism (Geol. Mag., dec. v, vol. iv, p. 161, 1907) of a pre- vious paper of mine, Mr. T. Crook says that evidence is needed of the absence of inclusions of ilmenite, etc., in this hornblende, and he expresses doubts as to ‘‘ the view that titanium enters vitally and in any serious quantity into the composition of ordinary ferromagnesian minerals.” The amount of TiO, found would imply, if existent only as ilmenite, the presence of 17 or more per cent of this, and it is hoped that the details given here will suffice to show that but minimal amounts, if any, of ilmenite or titaniferous magnet- ite were present. As regards his latter doubt, Mr. Crook seems to be unaware of much recent and highly trustworthy work which has been done in chem- ical mineralogy, and which proves conclusively that titanium does enter vitally and often in considerable amount into the composition of the ferro- magnesian minerals. + W. F. Hillebrand, Bull. No. 305, U.S. Geol. Surv., 1907. { H. S. Washington, Manual of the Chemical Analysis of Rocks, New York, 1904. 192 Washington—Kaersutite from Linosa and Greenland. SiO, Tene: 40°85" 681 HO Peete sa 8°47 "106 LY Oi eg ae none NE Oa eee 9°89 097 IDO) eee 8°85 056 Ke@ aie ee 3 96 "055 MiniQrecee so) One 002 INTO ee ee 0:10 001 MoO C35. 12°47 312 ClO seco ee BE 1G) Tap BAO me sos none ING Ono ah Be 22con 032 Ke Oto: 0°63 “007 PU Ogee eons 0°19 "010 Oe es we EOS ‘007 99°98 Hornblende from Linosa.* Washington, analyst. In its general features, this analysis closely resembles many analyses of basaltic hornblendes, such as those made by Schnei- der.t Alumina is, however, decidedly lower, and a most strik- ing character is the very high percentage of TiO,, the amount of which is nearly twice that reported for most basaltic horn- blendes. The character of the material used for the analysis precludes the possibility that this high figure is due to admix- ture of titaniferous magnetite or ilmenite, as no appreciable amount of such inclusions could have been present, so that the titanium must be regarded as belonging to the hornblende molecule. The analysis will be discussed later, in connection with others, and attention need be called here only to the fig- ures for the iron oxides and the percentage preponderance of ferric over ferrous oxide. The Kaersut Amphibole. Occurrence.—A_ peculiar, highly titaniferous amphibole from Kaersut, Nugsuaks Peninsula, on the shore of Umanak Fjord, Greenland, was described by J. Lorenzen,{ who called it kaersutite. According to Steenstrup (as quoted by Loren- zen), the kaersutite occurs in a vein or dike, 2 to 6 inches wide, which cuts a horizontal sheet of peridotite 120 feet thick. It is accompanied by plagioclase, titaniferous ore, an astrophyllite-like mica, and some pyrite, with zeolites, calcite, and quartz as secondar y minerals. This sheet of peridotite is * This earn. has been published in Rock Minerals, by J. P. ae” New York, 1906, p. 530. Ee: Schneider, Dericciae Kryst., vol. xviii, p. 580, 1890. tJ. Lorenzen, Medd. Groeniand, vol. vil, D. Pile 1884. Washington—Kaersutite from Linosa and Greenland. 193 undoubtedly the same as that mentioned by Phalen* in his description of the rocks of the Nugsuaks Peuinsula, who speaks of it as forming a cliff 200 feet high, calls it picrite, and gives a petrographic description. The kaersutite-bearing vein is not mentioned by him. The specimens sent us by Professor Ussing show a rather coarsely granular mass of feldspars pierced in all directions by prisms of the black hornblende, which run up to 3° in length by 5™™ in thickness. The rock is far from fresh, and is stained brown and yellow with iron, and here and there a pale green. Small grains of magnetite or ilmenite and many apatite crys- tals in water-clear prisms 5™" in length are present, and a few small specks of pyrite were seen, but we could not detect with certainty any of the mica megascopically. In thin section the texture is distinetly that of a granitoid rock rather than that of vein material. The structure is decid- edly miarolitic. The most abundant mineral is a plagioclase, in anhedral development, whose extinction angles indicate the average composition Ab,An,. (Extinction angle on 001 = 13°; y about 1570 and a slightly > 1:56.) With this is some alkali-feldspar, which shows some microperthitic features and is apparently highly sodic, though it is mostly cloudy and con- siderably decomposed. The brown kaersutite prisms are promi- nent, and show the optical properties to be described later. Small stout prisms of fresh augite are not uncommon. For the most part they are colorless in the interior and slightly greenish toward the border, but there are also some small anhe- dra of a highly pleochroic, brilliant grass-green augite which occasionally forms a border about the less colored variety, and is apparently the chromiferous augite mentioned by Ussing.t Indications only of the violet angite described by Ussing were observed by us. Small thin plates of light brown biotite occur. These are intensely pleochroic, the color for rays vibrating parallel to the cleavage cracks being a very deep purplish red, while perpen- dicular to this they are pale yellowish brown. They show no analogy with astrophyllite. Some small grains of opaque ore are present, but these only rarely occur as inclusions in the hornblende and no pyrite was visible in our sections. Deep red goethite was observed as an alteration product of the mag- netite or pyrite. Apatite is very abundant, in long, clear prisms, and is a frequent inclusion in the hornblende. It was also noted by Ussing as an abundant constituent. Patches of greenish chlorite minerals occur and are the cause of the occa- sional green color of the rock. *W. C. Phalen, Smithson. Misc. Coll., vol. xlv, p. 194, 1904. + In Rosenbusch, Mikr. Phys., vol. i, pt. 2, p. 237, 1905. 194 Washington—Kaersutite from Linosa and Greenland. A hornblende which apparently resembles that of Kaersut is deseribed by Phalen* as occurring in a fine-grained quartz- monzonite at Alanekerdlak on the Nugsuaks Peninsula. From our examination of the original specimen, kindly loaned by Dr. G. P. Merrill of the Smithsonian Institution, it appears that, while the rock resembles that of Kaersut, except in its finer grain, this hornblende differs from ours, being of a yellow- brown rather than a red-brown color, and with somewhat different pleochroism. The optical characters were not deter- mined, but it would appear to be titaniferous and related to that which we are describing. Physical Properties—The Kaersut amphibole is mono- clinic and forms prisms which reach a length of 3° and thick- ness of 5™" in our specimens. They are bounded by the prism m (110) and the pinacoid 6 (010). The terminations are usually poor, but some crystals show the presence of steep domes or pyramids, whose symbols could not be determined. These are also to be seen in the thin sections. Ussingy gives (110), (011), (101), and more rarely (121) and (010), as ‘the forms observed by him. ‘Twinning was observed in several of the sections. In one case the twinning plane was probably a steep dome, making an angle of 15° “with the cleavage lines, and showing in several places several narrow lamellae due to repeated twin- ning. The cleavage lines cut this twinning trace uninterrupt- edly. This crystal is shown in fig. 4, from a photograph kindly made for us by Prof. J. Volney Lewis. The extinction is parallel to the right of the twinning line and 17° to the left. The crystals are seen in thin section to be frequently crossed by narrow cracks (seen as dark lines in the figure). Some of these are irregular, but many are straight and parallel, crossing the cleavage lines of 110 at about 32°. For the angle Oe 110 measured on the cleavage surfaces, Lorenzen (p. 39) gives 55° 29’ (the mean of 55° 25’ and 55° 33°), while Ussing obtained 55°35’. On an excellent cleavage piece, giving sharp -reflexion signals, Wright obtained 55° 35), while on less satis- factory fragments the values were 55° 34’ and 55° 23’. We therefore consider that 55° 35’ best represents the value. For other angles Ussing gives the following: 101,011 = 34° 16’, Ol 4 140 = 76° 51’ 1101. O11 = eee ope cleavage is highly perfect, and where it is not developed the fracture is conchoidal. We found the hardness to be about 6; Lorenzen gives it as 5°5. ‘The mineral is very brittle, and fuses readily betore the blowpipe to a dark, magnetic bead. The specific gravity as determined by Washington with the pycnometer on about one gram of very carefully selected frag- *W.C. Phalen, loc. cit., p. 207. +N. V. Ussing, in Rosenbusch, Mikr. Phys., vol. i, pt. 2, p. 286, 1905. Washington— Kaersutite from Linosa and Greenland. 195 ments was 37137 at 25° C., while Ussing obtained 3°237 (tem- perature not stated) and Lorenzen 3-04 at 18°. The extinction angle on m 110 was found to be ¢:¢ =—78° in the obtuse angle 8, measured in sodium light (an average of 14 readings on different sections). In Li light the extine- tion was ¢ A ¢ = —9°3°, which would indicate that there is some slight dispersion of the bisectrices with e A ty, < ¢ A Cy; That it is shght, however, is evident from the comparatively sharp position of extinction in ordinary light. Ussing gives the + extinction angle as ¢:c = about 10° in the obtuse angle 8, but does not state whether this was measured on 010 or 110. He remarks that it is somewhat greater for red than for green, which is in accord with the observations of Wright. The plane of the optic axes is the clinopinacoid. The refractive indices were determined directly by the method employed in the preceding case, and were found to be as follows: a = 1°676, 8 = 1°694, y = 1-708, the probable error being less than + 0:02. This gives for the birefrin- gences: y — a = 0:0382, y— B =0°014, B-—a=0:018. The optic axial angle calculated from these indices is 2V = 82° 6’, while measurement of a section nearly perpendicular to the 196 Washington—Kaersutite from Linosa and Greenland. Fics. 5, 6, Kaersutite from Kaersut, Greenland. Etch pits on 110, HF; x 120. 7 Fic. 7, Kaersutite, Kaersut, Greenland. Etch pits on 110, HF; x 480 Washington—Kaersutite from Linosa and Greenland. 197 acute bisectrix on the universal stage gave 2V= 81°. This value can be considered only fairly accurate because of the deep color of the mineral which tends strongly to veil the optical phenomena, but it agrees satisfactorily with the calcu- lated value, much better, indeed, than in the case of the Linosa hornblende. Chemical Composition.—The chemical analysis was carried out on about 2 grams of carefully selected fragments, which, after crushing and washing free from dust, were treated with an electro-magnet to remove the few particles which contained ore grains. The microscope indicated that but few of these were present as inclusions, and the very small amount thus removed is in harmony with the observations. The only inclu- sions of note are of apatite, the needles of which penetrate the hornblende to a very considerable extent. As the mineral power was somewhat acted on by acid it was thought best not to remove these inclusions by its use, but to correct the analy- sis for their presence by determining P,O,. Apart from these apatite inclusions the material analyzed was extremely pure, as is shown in fig. 4, in which the enhedral apatites are well seen. it was dried at 110° prior to the analysis, which was carried out by the methods adopted in the previous one. Lorenzen (p. 30) seems to have had great difficulty, using Doelter’s method, in decomposing the mineral with sulphuric and hydrofluoric acids to determine FeO and Fe,O,. He reports the value FeO = 6°61 per cent in one case, but pre- fers to consider all the iron as ferrous in the statement of his analysis. Using the simple Pratt method, I had no difficulty in the solution of the finely powdered mineral in six minutes, and the result given here may be accepted as fully as correct as in the case of the Linosa hornblende, though the amount on hand did not permit of a duplicate determination. Special search was made for tin, as 0°26 per cent SnO, was reported by Lorenzen, but with absolutely negative results. The method adopted was essentially that of Baley as outlined by Classen.* The mineral powder was decomposed by evapo- ration to dryness with nitric and hydrofluoric acids, which would not lead to loss of tin by volatilization as tin fluoride is decomposed by heating. The residue was dissolved in hot dilute hydrochloric acid, filtered, and the filtrate treated with pure zinc, which would precipitate any tin. Only a very slight residue remained, which was wholly soluble in nitric acid. Neither this solution nor the previous filtrate gave any precipitate with H,S or other reaction for tin. We therefore consider that our hornblende contains no tin, and that the * A. Classen, Ausgew. Method. Anal. Chem., i, 1901, p. 184, 198 Washington—aersutite from Linosa and Greenland. presence of tin, as reported by Lorenzen, is very doubtful and certainly cannot be considered as characteristie of the mineral, as suggested by him. Lorenzen’s original analysis and the new one are given below, the last two columns showing the figures of the latter as corrected for the 0°77 per cent of apatite present and recal- culated to 100 per cent, and the molecular ratios. I IT III IV SiO, .5:.. 0.5. Skeh 4188. 89-305 90-50 Sees iQ? Nae yee tere 6°75 10°25 10°31 “129 SOs” Rashes 0°26 none none AWOL eerie. Cee 1441 9 116). 11-90 eae Biei@ ve eee eee none om 1°22 008 MeO soak Os eee = 11°28 8°76 8°81 "122 NGOs a ie, Cle 0°06 0°06 001 INT OM reyes SG oe id Cle none none IMG Onsen: eee 1365) 13°24 13°31 oa CAO fo aan 297 MpSZ79 10°93 "195 NiO: ee eee n. d 2°93 2°95 048 AGAQ) ket os Sees eae 1-06 1:07 "Ol Fy OR amas ed ees mn. dd: 0°59 0°59 ~ -083 POR ae ee n. d 0°32 Paes 100°56 100°17 100°00 I. Analysis by J. Lorenzen, Medd. Groenl., vii, p. 30, 1884. II. Analysis by H. 8. Washington. 3 IIf. Analysis II corrected for apatite. IV. Molecular ratios of IIT. There is every reason for the belief that the material anal- yzed by Lorenzen was essentially identical with that investigated by us, as is also indicated by the agreement in the figures for Si0,, total iron as FeO, MgO, and CaO, so that the differences between the analyses cannot be ascribed to varying chemical composition. Lorenzen’s analysis is seriously defective in the assumption that all the iron is present in the ferrous state, as well as in the non-determination of soda, potash, and water. The last is here of comparatively small moment, but the new analysis shows that about four per cent of alkalies are present, and it is well known that all the basaltic hornblendes and others similar to this contain very notable amounts of soda, with often considerable potash, which cannot be ascribed to inclusions. Lorenzen’s figure for TiO, is lower than ours by about 3°5 per cent, while his alumina is higher by about the same amount. As our figure for titanium was determined by the colorimetric method, which is capable of a high degree of accuracy, and is the mean of three closely agreeing determina- Washington —Kaersutite from Linosa and Greenland. 199 tions, we have great confidence in its correctness. It would ‘seein to be highly probable, therefore, that in Lorenzen’s analysis part of the TiO, was reckoned as Al,O,; and this is the more likely as titanium was determined by him by precipi- tation with sodium thiosulphate, a method which is known to be very uncertain and apt to give either too high or too low results, depending on the amount of acid in the solution and other conditions. In its general features this analysis much resembles that of the Linosa hornblende, especially in the amounts of silica, alumina, magnesia, lime, and alkalies, as well as in the very high titanium dioxide. The only ‘prominent difference is found in the oxides of iron. The sum of these is considerably higher in the Linosa mineral and the molecular amounts are equal, while in the Kaersut mineral ferrous oxide is largely in excess of ferric, the amount of which is very low. Interpretation of the Analyses. The interpretation of the analyses of these two hornblendes in terms of the molecular constitution is rendered subject to grave uncertainty through the presence of the very large amounts of titanium. The uncertainty arises from the fact that this element may be present either as Ti,O, or as TiO, or as both oxides together. Potassium permanganate oxidizes Ti LOE tO TiO, just as it does FeO to Fe,O,, so that if the lower oxide of titanium is present it will appear in the ordinary course of analysis as FeO, the apparent amount of which would thus be too high, and that of Fe,O, would be correspondingly low, while the Ti,O, would be determined colorimetrically or gravi- metrically as TiO, If all four oxides are or may be present simultaneously, the analytical problem becomes complex and somewhat difficult. A promising line of attack is being developed by Gooch and Newton,* depending on the selective oxidation of the Ti,O, by cupric salts, bismuth oxide, or ferric sulphate, which have no effect on the ferrous oxide. While the results recorded are excellent and show the possibility of very exact estimation under the conditions observed, yet it is uncertain if the methods are applicable, at least without moditication, to the analysis of silicate rocks and minerals, owing to their insolubility except in hot hydrofluoric acid and the very ready oxidizability of the hot solutions of Ti,O, and FeO so obtained. It may be suggested that the addition of cupric sulphate to the mixture of hydrofluoric and sulphuric acids employed in determining * Gooch and Newton, this Journal, xxiii, 1907, p. 365; H. D. Newton, this Journal, xxv, 1908, pp. 130 and 348. See also G. Gallo, Chem. Zeitung, 1907, p. 399, and A. Cathrein, Zeitschr. Kryst., vi, 1882, p. 248. 200 Washington—Kaersutite from Linosa and Greenland. FeO might solve the problem. If under such conditions the eupric salt would oxidize the Ti,O, without acting on the FeO, titration with permanganate of two portions brought into solution both with and without the addition of CuSO,, together with the determination of total iron as Fe,O, and of total titanium as TiO, by the usual methods, would furnish all the data needed. The discover y of some such method capable of yielding accurate results under the conditions of silicate analysis 1s now one of the most important desiderata. In this connection a peculiarity in the relations of the oxides of iron and titanium may be pointed out. In whole numbers the molecular weights are as follows: Fe,O, = 160, FeO = 72 ke 142), eo, 0 eo 160), Ti = 144. That is, neglecting the refinement of decimals, a molecule of ferric oxide is equal to two of titanium dioxide, and one of titanium sesquioxide is equal to two of ferrous oxide. Therefore, to oxidize either FeO to Fe,O, or Ti,O, to TiO,, one atom of oxygen, equivalent to one-ninth of the lower oxide, will be needed; while conversely, in the case of reduction of Fe,O, to FeO or TiO, to T:,0,, one atom of oxygen, equivalent to one- tenth of the weight of the higher oxide, will be subtracted. Exactly the same amount of potassium permanganate, there- fore, will oxidize the same weights of iron as ferrous oxide or titanium as sesquioxide to the higher form, as is expressed by the two equations: 10FeSO, + K,Mn,O, + 8H,SO,=5Fe,(SO,),+ K,SO,+ 2MnSO,+8H,0, 5Ti,(SO,), + K,Mn,O,+ 8H,SO,=10Ti(SO,), +K,SO,+ 2MnSO,+8H,0. Fe,O, and 2TiO, on the one hand, and 2FeO and Ti,O, on the other, are mutually interchangeable and equivalent as regards titration by permanganate or other such oxidizing aentg. From this it follows that, if Ti,O, is present and the ferrous iron is determined in the usual way, a percentage amount equal to that. of the Ti,O,-must be deducted from the apparent value for FeO, while an equivalent amount must be added to the apparent amount of Fe,O,, and deducted from that of TiO,,. In the case of minerals whose formulas are simple and well established, readjustment may be made with a fair degree of confidence as to probable correctness, even in the absence of determinations of all four oxides. Such readjustments based on the empirical formula would have still greater weight could it be assumed that ferrous oxide is absent or present in only negligible amounts. This consideration applies to the com- position of schorlomite, which we are justified in referring to Washington—Kaersutite from Linosa and Greenland. 201 the garnet group as suggested by Rammelsberg* and discussed in detail by Koenig,+ in which enough of the apparent TiO, is calculated as Ti,O, to conform to the garnet formula. The ease of the hornblendes offers more difficulties, since their molecular constitution is not well understood at present, and is undoubtedly very complex, as is well known. In a recent important paper, Pentfieldt.and Stanley explain 3 the presence of the sesquioxides by “their introduction into the metasilicate molecule in the form of various basic, bivalent radicals,” the mass effect of the very complex amphibole acid exerting a controlling influence on the crystal form and other physical characters. They also suggest the possibility that the molecule of the amphibole acid has a ring form, analogous to that of the benzene compounds. By assuming various bivalent radicals, composed of R,O, with F, HO, Na a, and R’, which combine with SiO, in the ratio of 1: lla "final residue of ‘Mg, Fe)O and CaO is left which conforms to the same metasilicate ratio. As regards tremolite and actinolite their exact ratios show that the molecule Na,A1,Si,O,,, suggested by Tschermak, cannot be present, as this would “deplete the total silica and destroy the 1:1 ratio.” In their calculations of their analyses of the aluminous hornblendes this molecule was also neglected, and very exact metasilicate ratios were obtained without the assumption of its presence. As this paper is the latest and one of the most suggestive and illumi- nating contributions to our knowledge of the constitution of the amphiboles, a study of our two minerals in its hght will be of interest. The two analyses made by me yield the following ratios, MnO and NiO being reckoned in with FeO: Linosa. Kaersut. “eee ie 5 "659 ya De is 106 EST e899 ee yA yon ‘097 } q 110 Medias 056 { 2008 Re We) = 2042. Sy ee | = OES) eens MQ. 52 2-312 ( ihe fy aoe : Oa0= 25 gee ep. 17" 195 195 | 74s Na 0.52 032) 439 ¢ “643 048 ) 059 | 1) oe ‘007 vee ‘O11 | ; BO. ee 010} 447 033 -038 Me * C. Rammelsberg, Min. Chem., 1875. + G. A. Koenig, Proce. Acad. Sci., Phila., 1886, p. 354. ¢ This Journal, vol. xxiii, p. 23, 1907. Aw. Jour. Sci.—FourtH Series, Vou. XX VI, No. 153.—SEPTeMBER, 1908. 15 202 Washington—Kaersutite from Linosa and Greenland. In both cases the ratio of (Si, Ti)O, to (R’,, R”)O is greater than unity, being 1:22 in the Linosa, and 1:06 in the Kaersut hornblende. Such relations differ widely from those presented by the hornblendes analyzed by Penfield and Stanley, in which, where the ratio differs notably from unity, it is always on the side of an excess of RO over SiQ,,. Assuming first that all the titanium is present as the dioxide, if we calculate the composition of the molecules in terms of the bivalent radicals suggested by Penfield and Stanley, we obtain residues of Si0,, (Mg, Fe)O, and CaO, which differ widely from the metasilicate ratio Si0,: RO=1:1, in that the original excess of silica is here greatly accentuated. This is shown in the table below. Linosa. Kaersut. [(Al,-Fe),O(F, OH),] SiO,. -017 +083 [ (Al, Fe), O,RNa, | SiO, - . ‘039 059 GA, ite) § O ‘Ry SiO, baa en OW) 7 "026 (Me, Fe)O a a te iii: io Ta ES 3/0 dneeee Ca O RCM OAE 5 Spe" Lape ee 25 i Pa, "917 451 "195 565 residual, o1Ofs 5. 2a ere 634 600 If, on the other hand, we assume that some of the titanium is present as sesquioxide, by making the necessary calculations and readjustments (the results of which it is needless to give here), we find that this will increase still more the ratio SiO,: RO, in spite of the reduction in the amount of RO,, thr ough the diminution in the amount of FeO and the taking up of MgO to form one of the complex radicals. We-may therefore assume that all the titanium is present as TiO,, as this shows less divergence from metasilicate ratios. It is therefore evident that, for our hornblendes at least, the presence of some other bases or radicals must be assumed, which will take up this seeming excess of silica and at the same time conform to the metasilicate ratio SiO): RO=T ty Sires may be found in the molecule Na(Fe, AD)Si, O,, which exists as the ferric or alumina extreme respectively in riebeckite and elaucophane, and whose presence in many amphiboles was suggested by T'schermak, though not found essential to the interpretation of the hornblendes studied by Penfield and Stan- ley. This, and the analogous (Mg, Fe) (Fe, Al),S81,0,,, con- form to the normal metasilicate ratio, owing to the trivalence of the basic iron and aluminum; and the molecules may be expressed graphically as ring for mulas, quite analogous to those suggested by Penfield, as is shown by that of riebeckite, as follows : Washington— Kaersutite from Linosa and Greenland. 203 Na—O O—Na NSO. SiGe 0% | | No O Ore Xe He-_Ov, | | poke If we assume then the presence of the bases Na,O and (Fe, Al),O,, which will form the metasilicates Na,SiO, and (Fe,- Al),Si,O, respectively, the composition of our amphiboles may be calculated as follows, the presence of some of Penfield’s radicals being needed to account for the F and H,O and the excess of RO, over the alkalies. The distribution can be made mathematically, so that the whole will conform to the meta- silicate formula, by the use of equations analogous to those used in the calculation of the norms of igneous rocks.* Linosa. Kaersut. [(Al, Fe),0,(F, OH),] SiO,. -017 033 [fAl, Fe) O.R] Si0,..-. 088 0.26 [(Al, Fe),0,Na,] SiO, ..--- es O44 Na(Al, Hoyt O. eek alee "039 OLS (Mig, Fe) (Al, Fe),Si,0,,--. 009 es ue LE a enn eee aa aH i aoe A Oe en ees CaO ae Ss ape Pen aces eee wee ue fe eee Brertcar tC) se eo oo Bee i i, AOL) Leah OME, Assuming that the amphiboles are metasilicates, as is held by most authorities, and which view is greatly strenethened by the work of Penfield and Stanley, it is clear that the com-. position of the hornblendes of Linosa and Kaersut may be rationally explained by the assumption of the presence of mole- cules of the general type (R’,, R’”’) R’”’, 8i,0,,.. The presence of these molecules also seems to be quite unavoidable in the case of such amphiboles as riebeckite and glaucophane, and furthermore they cannot be interpreted only in terms of biva- lent radicals such as those suggested by Penfield and Stanley, though some of these may be assumed to be present. The two authors mentioned do not discuss the question of the presence of such riebeckitic molecules, an omission unhesi- tatingly to be ascribed to the preliminary and, most sadly, unfinished character of their paper. Such a discussion would have been inevitable had their investigation been extended to the glaucophanes, riebeckites, and other highly sodic amphi- boles, as one of us knows to have been the late Professor Pen- field’s intention. So far as can be learned from the published * Cross, Iddings, Pirsson, and Washington, Quantitative Classification of Igneous Rocks, Chicago, 1903, pp. 194, 195. 204 Washington—Kaersutite from Linosa and Greenland. . paper their objection to the introduction of the molecule under discussion lay in the fact that, according to Tschermak’s theory, “a definite basic alumo-silicate molecule is regarded as isomor- phous with Ca (Fe, Mg),Si,O,,.”* As has been shown above, however, such a riebeckite-glaucophane molecule may be regar ded as a metasilicate and may be written structurally as a ring formula, exactly analogous to those suggested by the authors named. It is clear, therefore, that the presence of a riebeckitic molecule need not be regarded as a case of isomor- phism of two chemically and structurally unlike molecules, or as inconsistent with the views of Penfield as to the structure of the amphibole acid and the mass effect of complex mineral molecules. On the contrary, they are in complete accord, as the matter reduces itself, in the last analysis, to the simulta- neous replacement of one hydrogen atom by Na and, three by Fe’’, just as two atoms are replaced by Ca”, (Mg, Fe)", or by one of Penfield and Stanley’s bivalent radicals. These authors noted the highly interesting and probably significant fact that the CaO formed “very closely 25 per cent of the various radicals and bases, or in other words replaces one-fourth of the hydrogen atoms of the amphibole acid.” Without giving all the percentage figures, in the Linosa hornblende, as calculated above, CaO forms 33:8 per cent, and in that of Kaersut 26-3 per cent. The latter approx- imates to one-quarter, while the former is about one-third of the radicals and bases. If this last is substantiated by analyses of other hornblendes, and found to be characteristic of certain kinds, it might be held to indicate that the amphibole acid contains a number of hydrogen atoms which is divisible both by 4 and by 3, such as H,,8i,O,,, or a multiple of this. But our data are at present far too insufficient for more than a speculative suggestion. As is well known, the amphiboles which are high in soda and in alumina or ferric oxide, and which there is good reason to believe contain the riebeckite-glaucophane molecule or basie (Al, Fe)”, such as riebeckite, glaucophane, arfvedsonite, crossite, hastingsite, barkevikite, aenigmatite, and those we have been describing, are all intensely pleochroic and show either very distinctive blue colors or very intense reds and browns. Similarly, the pyroxenes which contain the acmite molecule, or basic Na’ and (Al,Fe)’’’, as aegirite, aegirite-aug- ite, babingtonite, ete., are deeply colored and are characteristi- cally pleochroic, in contrast with the common, generally non-pleochroic pyroxenes, which do not contain the acmite molecule. On the other hand, the amphiboles which contain trivalent Al and Fe only in bivalent radicals, on the theory of * Penfield and Stanley, op. cit., p. 49. Washington— Kaersutite from Linosa and Greenland. 205 Penfield and Stanley, as actinolite or common hornblende, are less deeply colored, are never blue, and are markedly less pleochroic ; and the same is true of the ordinary pyroxenes and augites. Tt is commonly supposed*™ that the blue color of these amphiboles is connected with the presence of abundant iron, and Pirssont has recently suggested that the blue color is due to the presence of ferrous-ferric molecules, the analogy of Prussian blue and altered vivianite being cited. Since, how- ever, a similar color is characteristic of glaucophane and gas- taldite, in which the trivalent element is practically entirely aluminum, ferric iron being either absent or present in small amount, it would seem to be necessary to amend this hypothesis by assuming aluminum to replace the ferric iron either wholly or partially. Similarly the deep browns and reds are supposed to be connected with the presence of titanium and, as pointed out by Brégger,t the intensity of the color increases with increasing content in this element, as is shown by the series barkevikite, basaltic hornblende, and kaersutite and aenigmatite. We have seen above that (Al, Fe)” may enter the amphi- bole molecule either in a bivalent radical, such as those sug- gested by Penfield and Stanley, the radical as a whole acting as a base, or it may itself act as a base, replacing three atoms ot hydrogen in the amphibole acid. It would thus occupy different positions and perform very distinct functions in the molecular arrangement. In these different positions, there- fore, the trivalent element may reasonably be supposed to affect differently the optical and other physical properties of the minerals into which it enters, in analogy with the well- established fact in the chemistry of the carbon compounds. Following out this line of thought, it may be suggested that this basic (Al, Fe)’”, and not that which forms part of bivalent radicals, acts as a chromophore, as such color-giving radicals are known in organic chemistry, where they are especially notable among the aromatic compounds; and that, furthermore, the property of pleochroism may be connected with its presence, this either causing a mineral variety to be absolutely pleochroic, when varieties in which the basic triva- lent element is not present are not so, or intensifying the pleochroism of otherwise weakly- pleochroie complex mineral molecules. “That this chromophoric radical does not consist solely of (Al, F)’”, but contains Na as well, probably im the ratio 1: 1, is indicated by the constant presence of much soda in the peculiarly colored and pleochroic amphiboles and pyrox- *Cf. W. C. Brégger, Grorudit-Tinguait Serie, p. 35, 1894. +L. V. Pirsson, this Journal, vol. xxiii, p. 440, 1907. { Brogger, loc. cit. 206 Washington—Kaersutite from Linosa and Greenland. enes which are under discussion, as well as by the fact that these constituents generally show such a constant ratlo, as pointed out by Doelter.* The partial replacement of Na, by (Fe, Mg)’, sometimes observed, would harmonize this idea with that of Pirsson. The chromophoric titaniferous radical would seem to he more intensely active or color-producing than the alumo-ferric- soda one, since amphiboles high in titanium, but otherwise chemically like those low in this element, are red-brown rather than blue, as is shown by the relations of aenigmatite and arfvedsonite. But its nature is at present difficult to suggest. The ability of titanium to assume some seven states of oxidationt complicates the problem immensely, even though this possible number is lessened by considering only the oxides most commonly met with, Ti,0,, Ti0,, Ti0,. It may only be mentioned here that the violet or blue colors of solutions of Ti,O,, and the yellows and deep browns of those containing TiO,,t such as are produced from colorless TiO, solutions on the one hand by reduction with zine or tin, and on the other by the action of H,O,, may be possibly significant of the con- dition of oxidation of the titanium. The constant, characteristic pleochroism of the colored, common amphiboles, which do not contain the supposedly chromophoric Na (Al, Fe)’” radical, as contrasted with the equally characteristic non- pleochroism of the colored, common pyroxenes (free from the acmite molecule), leads also to the speculative suggestion that the difference is possibly connected. with difference in the structure of the molecule. Accepting provisionally the suggestion of Penfield and Stanley that the amphibole acid has a. closed chain or ring structure, it might be advanced as possible or probable, in analogy with the car- bon compounds, that the pyroxene acid is of the pcm chain type. Or the relations might be the reverse.§ In our present state of complete ignorance as to the sone tution and structure of the mineral molecules any such sug- gestion as is here made can but be regarded as a hypothesis of the most speculative character. But such a difference in structure would be a fundamental one between the molecules of the pyroxenes and the amphiboles, and it undoubtedly exerts a profound influence on the physical properties of iso- mers, as the pyroxenes and amphiboles are regarded with reason as being. *C. Doelter, Zeitschr. Kryst., vol. iv, p. 40, 1880. +P.) B. Browning, Introduction to the Rarer Hlements, p. 61, 1903. + A. Classen, Ausgew. Methoden Anal. Chem., vol. i, p. 765, 1901. SIt is, of course, understood that the terms "4 open chain” and “closed chain” are used as they are in organic chemistry, without implying that they actually thus represent the structure of the molecule, or the arrange- ment of the atoms in space. F Washington—Kaersutite Srom Linosa and Greenland. 207 It might explain the inconsistency of the apparently greater ehemical complexity and hence probably greater molecular | weight of the amphiboles, as suggested by Tschermak, and the higher specific gravity of the pyroxenes, which Clarke* urges as evidence of their greater molecular weight. It might also readily explain the characteristic difference in pleochroism between the two groups noted above. In this connection the analogy of the carbon compounds is of interest. The greater part of these are referred, as is well known, to two great groups; the aliphatic compounds, deriv- atives of methane, with an open chain type of formula; and the aromatic compounds, derivatives of benzene, with a closed chain type of formula. These two groups show characteristic differences in general chemical behavior, and also characteristic differences in some physical properties. Thus, the aliphatic compounds are very rarely colored, while colored compounds are quite common among the aromatic bodies. Similarly solutions of members of the first group seldom show absorp- tion bands, while those of the other, when colorless, often do so. The analogy cannot, of course, be pushed very far, but that such a fundamental difference in molecular structure would not be inconsistent with the alteration of amphibole to pyrox- ene, or the converse change of pyroxene to uralite, is indicated by the convertibility of members of the aliphatic series into those of the aromatic series, and vice versa. It may further- more be observed that very many organic compounds are known which contain radicals belonging to both series, such as toluene (methyl-benzene), ©,H,.CH,, and such pyroxenes as aegirite-augite might be regarded as possible analogues. Correlation and Name.—In the annexed table are given analyses of several hornblendes which resemble those of Linosa and Kaersut in one feature or another. In their gen- eral characters, on the whole, these most approach the basaltic hornblendes, or syntagmatites, as Rosenbush has recently pro- posed that these should be called,t+ especially in the figures for silica, iron oxides, magnesia, lime, soda, and potash. The alumina of our minerals is distinctly lower than in these, but the figures for this constituent in analyses of the syntagmatites (in this sense) are somewhat discordant. That shown in III is very high, while the analyses of Schneider run rather uniformly between about 14 and about 15. Also TiO, 1s much higher in I and II than in the syntagmatite analyses. The analysis (III) of a typical “ basaltic hornblende” shows but little TiO,, much less than in the analyses of Schneider (cf. IV and VI), where it varies from 4°26 to 5-40. The correct- *F. W. Clarke, Bull. No. 125 U. S. Geol. Surv., p. 90, 1895. +H. Rosenbusch, Mikr. Phys., vol. i, 2d half, 1905, p. 236. Washington—Kaersutite from Linosa and Greenland. 208 ‘gop d ‘2061 ‘AIXX "Pg ‘leg “qiqer sonen “4sdyeuE uid woyy ‘neueyontg avou ‘z2yVTq ‘opuUguY ‘poe d ‘Tegt ‘A “4st ayos}iez §—“4sATeue 10U48190,] "BLLOT|OJUe ‘(oytyeusiuee) o411Assop ‘gop ‘d ‘0GgT ‘tax “yskry ‘tyosziez “ysAyeute S1eqsiog ‘puvfuecing ‘yIseye{neN “op tyeULsIUEW ‘og ‘d ‘eget ‘a “Sey ur = “ysATeue UsZUdLOT ‘puvlusery ‘yNsIvu[plesuvy ‘ezyTUOspoasly 01g “d ‘96ST “t ‘TeUtmmor sIqT, ‘qgAyeue UOJSULIIV] ‘OlIejUG ‘UOUUSUNG “e}ISSUT{seH ‘CIP F ‘O6RT ‘TAX “askayy “AT 087107 ‘qskjeue yuI,_ ‘“AVMION ‘YlLAoyIeg ‘opTylaAoyleg 9¢.00L 1-001 61-001 08-001 19-66 79-66 ee Ea ees GI-0 G&-0 ie €9-0 6&-0 1g.0 90-1 66:6 VV-1 49-0 6<6-¢ 89.9 Cc1-8 66-6 80-9 EP-ol 10-G 96-1 G9-P 18-6 ¥c-Ol 69-61 98-0 6&-0 18-0 ce. 1 IL-1 ae eee oie Sere eta amen ae ag eS OX) eine 86-1 00-T cP-0 69-0 GL-0 66-11 L8°6& 88-9& 6V-6& 86-16 66-61 69-11 L6-L 18-G 08-€ 69-61 81-9 6-41 96-7 66-6 cV-V 6G- 11 cy-Il 97-6 or LG-h Soe 6¢-1 Seas GV-VG GG. SV 66-LE G3-EP SIL-P& 9F-69 IIx IX xs XI TITA ITA “TIX IL TITA TIA 46-001 VI-T 1-6 ie || VI-sl 6G-P 08-4 Vé-V1 14-¢ GL-OF TA ‘ose “d ‘06ST T1Ax ‘shay “Iyos}iez + ‘yskTeuR dopleuyog ‘“pyeAse -JSe M ‘USSUTT}IBE, “(epustquasoy oryeseq) oyQeuseyusg ‘eg “d ‘TST ‘AT “UNIT “Jed “UTI “HOSE “48Ayeue qplUypoS “eluULeyog “(epue[quaAoyY o1g[eseq) oyeUISE4yUAG “0g¢ d ‘OGST ‘tax “4shry ‘uyos}iez “ysATeue Tap -Teuyog “ViMmeyog “(epuelquaoy oyeseq) eyyeusRyUAG (OLE 78 S80] g1.0 Sepnpouy) ‘or ‘d ‘20@L ‘MExx ‘Teurnor siqy, ‘4ysdyTeue AoTURyS ‘eimeyog ‘UlIg ‘(epueTquaoy oTyTeseq) oy1yvUISeyUAG ‘qskjeue WOISUIYSe A “PULTUSELyH “Wnsievy ‘o}TjNs19v yy ‘qsX[RUB UOJSUTYSY MM “BSOUTT ‘Ossoy oJUOP “oF1gNSIEV yy SF-00T 99-10L 9F-00T 00-001 86-66 ee ee Oe 8 0 ES ee eo CL Or ee Coie = 101. 86-1 0s 3209-0 Sa ee M76. GOL. Ol. SGic™ = 102 ipo Mexee Who = SED tal Sioa Oye ala USS etal aries pts rans ouou OULOG See Pee oe oe ee 00 S10 ee fi <60:0 «© Sia [en OG Chewaae wae Ve Gi ee 2088) gen Gel CG: 9 Gave aan Bo pl> O0-Gm | CG BC.1l Get Ge ay 68:05) 0hS eo 16.0) hee ere, 99-68 GL.68 96.68 69:68 98.0% ~~~" A iar IL I NEE Washington— Kaersutite from Linosa and Greenland. 209 ness of the analyses made by Stanley under the direction of so preéminent an analyst as the late Professor Penfield can- not be doubted, and it is not probable that Schneider’s figures are seriously in error, especially as he states* that special atten- tion was paid to the determination of this constituent, though he does not mention the method employed. In passing, a few remarks may be made in regard to Nos. IV and V. The exact localities of these hornblendes are not given by either author, but they both are said to come from Bohemia, and the extremely close similarity in the figures for Si0,, Al,O,, FeO, MgO, CaO, Na,O, and K,O makes it proba- ble that they were made on identical amphibole from the same locality. In view of this remarkable agreement of the con- stituents mentioned, the discrepancies observed in the figures for TiO, and Fe,O, are noteworthy. In IV TiO, is higher and Fe,O, lower, while in V the reverse is true. ‘The sum of the two in each case is, however, exactly the same, 13°26. The exact agreement is, of course, a coincidence, but ‘taken in con- nection with the close concordance in the other constituents, and having regard to the methods of analysis, it suggests the ‘explanation that in IV the ferric oxide was determined by reducing the iron by H,S, which does not act upon Ti0,, while in V the iron was reduced by zine, which would reduce the TiO, to Ti,O,, and the latter would appear as Fe,O,, after sub- traction of the FeO. On this supposition the lower TiO, of V, in conjunction with the close agreement between the figures for Al,O,, would be explained by the use of the method of boiling an acid solution with SO,, which is as apt to give too low as too high results, or even possibly by the assumption that the residue left on evaporation of the silica with HF is wholly Ti0,, a not uncommon error, especially in former days. The presumption therefore is that IV represents the trne com- position more accurately than V. The matter, after all, is of little importance, but serves to illustrate the need of a critical study of analytical data in the light of the methods of analysis. The blue arfvedsonitic amphiboles (IX) and the triclinic aenigmatite (X and XI) resemble ours in the figures for SiO, and K,O, and especially in the high TiO, of the Greenland aenigmatite. TiO, was not deter mined in the anal ysis of arfved- sonite and the Pantelleria cossyrite, but it is presumably present in both, and almost certainly very abundantly in the cossyrite, as Brogger has shown. Additional reasons for this belief will be given in a forthcoming paper on the rocks of Pantelleria. These hornblendes differ, however, widely from ours in the lower A],O,, the very much higher FeO and Na,O, and in the very low MgO and CaO. The recently discovered triclinic *C. Schneider, Zeitschr. Kryst., vol. xviii, p. 579, 1890. 210 Washington—Kaersutite from Linosa and Greenland. rhénite (XII) is unique in its low silica, but in other respects resembles our amphiboles more than it does the triclinie aenig- matite, though the Al,O, is remarkably high. On the whole, therefore, our amphiboles may be considered to belong to the basaltic hornblendes or syntagmatites, rather than to the arfvedsonite group, and this general relationship is confirmed by the color, the etch figures, and the negative extinction angle of the Kaersut mineral. The very high TiO,, however, places them in a subdivision apart and, with the small positive extinction angle of the Linosa mineral, indicates a transition toward the arfvedsonite group. — On account of their chemical characters, and also because of the position of the negative extinction angles ¢ A c, barkevikite (VII) and hastingsite (VIII) may also be regarded as transi- tional between typical syntagmatite and the arfvedsonites, though in the direction opposite to ours. A similar relation is suggested by Brégger,* who regards kaersutite as an end member of the basaltic hornblendes. The name kaersutite may well be reserved for such highly titaniferous basaltic hornblendes or syntagmatites, and the mineral from Greenland may be regarded as the type. Whether the same name should be applied to the Linosa horn- blende or not is somewhat doubtful. The two are chemically closely similar, but show a marked divergence in the relative amounts of the iron oxides. Also the physical characters are alike in nearly all respects. The only differences of note are the somewhat higher specific gravity and imdices of refraction of the Linosa mineral, and the difference in the extinction angle, which last would seem to be the more important. Indeed, while that of the Kaersut amphibole is negative and lies well within the limits of the extinction angles shown by the ordi- nary syntagmatites, that of the Linosa amphibole is close to the vertical axis, but slightly positive, and indeed occupies a unique position between the extinctions of the riebeckite-arivedsonite group on the one hand and those of the syntagmatites on the other, though it must be remembered that in the former group the bisectrix which lies nearest the vertical axis is a, not c. The physical characters of our two minerals are tabulated below. The Linosa hornblende might be regarded as an end mem- ber of the highly titaniferous syntagmatites, in which case the name kaersutite would apply to it, or its peculiar extinction angle, and the high ferric oxide might justify the separation of it from this group as a distinct subspecies, to which the name linosite may be given. In view of the uncertainty of our * Brégger, Grorudit-Tinguait Serie, p. 35, 1894. We ashington—Kuersutite From Linosa and Greenland. 211 Linosa. Kaersut. Specific gravity... ---- 3°336 3°137 Crystal system __ =~ monoclinic monoclinic TiO rere Pe Dameot Lys aes CRC Onn Oe Soria ss ten D 4! =e 9! Optical character ------ negative negative Glolonmeeevees ee deep brown chestnut brown Ahserphions 22.2522. 22 c>b>a c>b>a Tie ed Siete eer ae 1°692 1°676 eth a ee Yc es Tees 1°694 TLS Se SRS eee 1-760 1-708 “oe dx a 068 032 A ea dee ae ae 79° 54’ 82° 6: PiGHeTSION ose ose. (?) weak (?) weak knowledge of the true chemical composition and relations of the hornblendes, the numerous varieties that are constantly being observed and often named, and the fact that many well- recognized species show greater divergence i in physical proper- ties and chemical composition than do our minerals, it seems to us advisable not to bestow a new name at present on the Linosa hornblende, but to consider it a kaersutite. ; In conclusion attention may be called to the somewhat remarkable coincidence between the finding of kaersutite both in Greenland and on Linosa, and the similar occurrence of aenigmatite in Greenland and the apparently identical cossy- rite on Pantelleria, which lies close to Linosa. In both the Arctic and the Mediterranean localities these minerals occur in comparative abundance and as material which can be easily studied, while elsewhere kaersutite is unknown and aenigma- tite very rare and the crystals small and unsatisfactory. Locust, New Jersey, and Geophysical Laboratory of the Carnegie Institution of Washington, D. C. bo jock bo £. Howe—Geology of the Isthmus of Panama. Art. XXIV.—The Geology of the Isthmus of Panama ; by Ernest Howe. Introduction. Tus first serious effort to investigate the geology of the Isthmus of Panama was made by Robert T. Hill in 1895.* Although his stay on the isthmus was brief and the survey merely a hurried reconnoissance, the principal features of the geology and physiography were recognized. At the time of his visit little or no work was being done by the French com- pany, few records of borings or soundings were available, and many exposures along recent cuts were already in the grasp of the jungle, so that certain details that escaped Hill remained to be worked out by MM. Bertrand and Zitirchert three years later when they made a report to the New Panama Canal Company. Werking under more favorable conditions, these geologists were able to revise certain of Hill’s views, and especially his conclusions concerning the igneous rocks. In the following pages numerous references are made to the reports of Hill and of Bertrand and Ziircher, from which great assistance has been. derived in the course of my own studies of the isthmian geology. The field work upon which the present paper is based was carried on during a part of the summer of 1906, and in the following dry season from January to April, 1907, nearly five months in all. The work was undertaken for the Isthmian Canal Commission and had to do largely with economic matters ; a brief outline of the geology, however, accompanied the report to the Canal Commission{, a résumé ‘of which has been published in “ Economic Geology ”.§ The purpose of the present paper is to record in more detail certain observations on the geology, and more especially on the stratigraphy, that affect the conclusions reached by Hill and the French geologists. There were unusual opportunities for studying the section across the isthmus because of fresh exposures made by recent excavation along the canal line, and * Robert T. Hill, The Geological History of the Isthmus of Panama and Portions of Costa Rica. Bull. Museum. Comp. Zool., vol. xxviii, No. 5, pp. 149-285. Cambridge, 1898. + M. Bertrand and P. Zurcher, Etude Géologique sur l’Isthme de Panama. Rapport de la Commission. Compagnie Nouvelle du Canal de Panama, etc. Annexe I, pp. 85-120. Paris, 1899. The page references in the present paper are to a reprint of this report. +t Ernest Howe, Report on the Geology of the Canal Zone, Annual Rep. Isthmian Canal Commission, 1907. Appendix EH, pp. 108- 138. § Isthmian Geology and the Panama ‘Canal, Economic Geology, vol. ii, pp. 639-658, 1907. EF. Howe— Geology of the Isthmus of Panama. 218 at many places the information to be obtained from surface outcrops was augmented by a great number of boring records and samples. Special thanks are due Mr. John F’. Stevens, under tae as chief engineer the work was done, for many facilities placed at my disposal. !am also greatly indebted to Dr. William H. Dall for his kindness in looking over ee collections and for many valuable suggestions. Preliminary Statement. The Isthmus of Panama, where it is to be crossed by the canal, consists of sediments and pyroclastics of Tertiary age that rest on an eroded surface of andesitic breccias and associ- ated lava flows, all of which have been intrnded at numerous places, probably in Miocene time, by dikes and large cross- cutting masses of andesite or basalt. During the period of intrusion, or immediately after it, the region was uplifted, and the cycle of erosion thus inaugurated continued to late matu- rity or old age. Before the completion of the cycle, however, another upward movement occurred accompanied by warping or a gentle medial doming. Asa result of the continued ero- sion the basal igneous mass was exposed in the interior and a sub-mature topography developed in regions of harder rocks, while near both coasts less resistant sedimentary beds favored ‘more active erosion and conditions of greater maturity pre- vailed. A third uplift cansed the streams to entrench them- selves in their old or mature valleys, but before this last cycle had advanced beyond its youth, a gradual subsidence began that continued until the young valleys were agegraded and ‘the mature valleys of the second cycle drowned for short distances back from the coasts. A final slight upward movement has elevated the estuarine deposits, formed during the period of depression, a few feet above sea level. Obispo Breccias. The igneous complex that appears to underlie all of the other formations of the isthmian region consists largely of andesitic tuffs and breccias; it has been given the name Obispo on account of its characteristic exposures at Bas Obispo, the northern end of the Culebra cut. From Bas Obispo south- ward to a point between Empire and Culebra the breccias have been well exposed by excavation, but at the time of exam- ination the zone of surface alteration had not been passed, so that although the rocks show their pyroclastic nature the petro- graphical character of the fragments and the matrix is not readily determinable. Fresher “material was collected on the 214 =F. Howe—Geology of the Lsthmus of Panama. Pacitic side of the isthmus in the vicinity of Las Sabanas, and this has been compared with specimens from the other locali- ties which seem to be equivalent. The formation as a whole consists of moderately coarse breccias of andesite of a variety of textures but nearly uniform composition. Fine-grained tuffs have been observed, mostly on the northern side of the isthmus, while the Sabanas occur- rences present an extremely coarse facies. As a rule the breccias of the interior are composed of angular fragments varying from a quarter of an inch to two inches in diameter. Ata quarry near the Sabanas road one block more than six feet in diameter and many twelve or eighteen inches in diam- eter were observed in a cement as coarse as the average breccia of the interior. In a region where outcrops are few and con- tinuous exposures unknown it is not possible to make genera]- izations, but from the observed occurrences it may be said that the usual appearance of the Obispo breccia les between the extremes of coarseness and fineness that have been mentioned. So far as observations have been carried, the rock fragments composing the breccias consist largely if not entirely of pyroxene-andesite. Some are highly vesicular and glassy, others markedly porphyritic, while the majority are dense, even-textured rocks in which imdividual crystals can be made out only with the aid of a lens. In one specimen there were doubtful indications of hornblende. The prevailing greenish . gray color of the rock suggests what is found to be the case on microscopical examination, that the pyroxene, usually augite, has undergone extensive alteration; on the other hand, the plagioclase, which generally has a composition of about Ab,An,, is comparatively fresh, a fact noted in the cases of most of the rocks of the isthmus where the final stage in the decomposition to the surface red clay has not set in. Distribution of the Obispo breccia.—The most northerly occurrences of these breccias are shown by borings at the San Pablo dam site; from that point southward they appear occa- sionally at the surface, and are indicated in all the borings made at intervals of one kilometer along the center line of the canal as far as Empire. Decomposed but otherwise character- istic exposures are to be seen between Mamei and Gorgona along the Panama Railroad, and from Gorgona to Empire, where they disappear beneath a cover of later sediments, the Obispo breccias are the prevailing rocks. Composing the hills north, west, and east of Corozal, and particularly well exposed northeast of Panama in the vicinity of Las Sabanas, are rocks of the same petrographical character that are believed to be a part of the central mass. : FE. Howe— Geology of the Isthmus of Panama. 215 Age of the Obispo breccias.—The only definite statement that can be made in regard to the age of the Obispo breccias is that they are older than the oldest sedimentar y rocks of the region, which, as will be shown presently, carry a fauna consid- ered by Prof. “William H. Dall to correspond to the Claiborne Eocene. Although it is not improbable that the eruption of these andesitic breccias marked the beginning of Tertiary his- tory in the isthmian region, there is no direct evidence that the rocks are not still older. MM. Bertrand and Ziircher* included with the breccias that have been described as the Obispo, certain other fragmental igneous rocks exposed near Bohio, calling the whole ‘ Roche de Gamboa.” From fossil evidence found at Bohio they corre- lated the Gamboa rock with the Oligocene of southern Europe. Reasons are given in a later paragraph for believing that the breccias at Bohio are intimately associated with the oldest sedi- mentary rocks and separated from the Obispo breccias by an unconformity. Bohio Formation. The name Bohio is proposed for the oldest sedimentary for- mation recognized on the isthmus and includes cer tain beds occurring at Bohio, Vamos Vamos, and in the vicinity of Gatun. This name is suggested | in order to avoid confusion in referring to the same rocks described by Hill and Bertrand, but not ree- ognized by them as being parts of one formation. What are believed to be the lowest beds of this formation are conglomerates exposed in the lock site partly excavated by the French near Bohio, and breccias in the quarries at the Bohio railway station. A mile and a half to the west, ata locality known as Vamos Vamos, rocks of the same age are exposed on the south bank of the French canal, and similar rocks, although no fossils have been found among them, occur along the Panama Railroad near Tiger Hill. On the west bank of the canal nearly opposite the mouth of the Gatuncillo the formation is again exposed and the beds carry abundant fossils. At intermediate points borings have shown the pres- ence of rocks lithologically the same as those at Vamos Vamos or near Gatun, but their few fossils have been indeterminable. To the southeast of Bohio borings in the vicinity of Buena Vista and San Pablo have indicated the presence of beds like those at Bohio, while a few isolated patches of conglomerate that rest on the Obispo breccias in the neighborhood of Mamei, Gorgona, and Matachin, are in all probability a part of the Bohio formation. * Etude Géologique sur L’Isthme de Panama, p. 5. 216 F. Howe—Geology of the Isthmus of Panama. —Lithologic character of the Bohio rocks.—ULack of contin- uous outerops and sudden changes in the character of the beds within short distances have prevented an altogether satisfactory study of the Bohio formation. Boring records have assisted at a number of places, but even with their help the relations — at Bohio are not easy to make out. Directly south of the village of Bohio on the opposite side of the river is a hill about seventy-five feet in elevation, through the middle of which a cut was made by the French for the construction of a lock. The sides of the cut are now more or less covered by vegetation but still present excellent exposures of the rocks composing the hill. The section shown is about fifty feet in thickness and consists of beds of coarse conglom- erates, gravels, and sands that strike N. 25° E. and dip about 14° to the northwest. Crossbedding is common and the beds of finer textured rocks are frequently lenticular. The con- glomerates contain many bowlders a foot or more in diameter associated with coarse gravel or cobbles and held in a gritty matrix that is of the same character as the beds of finer sand- stone. The bowlders are of a number of kinds of eruptive rock, the commonest being hornblende-andesites with less abundant hornblende augite-andesite, augite-andesite and latite porphyry ; nearly all are coarse-grained and strongly porphy- ritic. It is noteworthy that of the bowlders examined few resembled rocks observed in other parts of the isthmus. The matrix in which these bowlders lie and the associated sandstones are composed of finer debris of the same rocks; quartz is so rarely present as to be negligible. It is worth noting in this connection that the bowlders of the conglomerate are of com- paratively fresh rock while the matrix and intercalated sand- stones are invariably decomposed, the alteration being to a complex aggregation of epidote, serpentine, kaolinite and seri- cite that causes the rocks to have a distinctly soapy feel. This is a feature common to most of the fragmental rocks not of direct voleanic origin, wherever alteration has not progressed so far as to result in the formation of the surface red clays. Not more than a quarter of a mile north of the conglom- erates exposed at the Bohio lock site, outcrops of a peculiar greenish brown rock are to be seen on the right bank of the Chagres beneath the village of Bohio. The hills to the north and east are composed of the same rock, and just back of the village, quarries are locateu from which building stone has been obtained for the Panama Railroad bridge piers and other purposes. The quarrying operations have exposed a vertical section of the rock about fifty feet thick. The rock is com- posed of fragmental materials of voleanic origin; the imbedded fragments are largely of pyroxene-andesite of a number of dif- EF. Howe— Geology of the Isthmus of Panama. 217 ferent textures, often fluidal and glassy, and varying from a tenth of an inch to one or two inches in diameter. Although many of the fragments are comparatively fresh, the matrix or cement is in an advanced state of decomposition, the second- ary minerals being zoisite, epidote, serpentine, and some doubt- ful kaolinite. As a whole, however, the rock is so extremely altered that it is difficult to determine whether it is a simple voleanic breccia ike the Obispo or a fine conglomerate of vol- eanie debris like that at the near-by lock site. So far as it is possible to make out, I am inclined to believe that it is a vol- eanic breccia, in the sense that it is more or less of eruptive origin, largely on account of the uniform character of the rock fragments composing it as contrasted with the varied pet- rographie character of the bowlders in the conglomerate at the lock site. It is not impossible that the breccia represents a “volcanic mud-flow” as suggested by Bertrand.* The particular interest attaching to this breccia is its inti- mate association with the conglomerates exposed at the lock site. This relation is not shown on the surface, but is brought out very clearly by borings made in exploring possible dam sites in the vicinity. One series of these borings, along what is known in the surveys as the “}’” line, extended from the quarries at Bohio in a southerly direction across the river to the hill through which the French excavated the lock site, a distance of about one-half mile. At the northern and southern ends of this line respectively characteristic specimens of the breccia and of the conglomerate were obtained by the drill, while at intermediate points transitional facies were shown in nearly all the borings. Among other features is the associ- ation of the typical breccia of the quarry with water-laid sand- stones or tuffs, many of which contain carbonaceous matter, while others, coarser grained, are composed of distinctly water- worn material and are held in a caleareous cement. Passing southward the breccias become less and less abundant while the water-worn material increases in coarseness and conglomer- atic beds are to be noted in many of the holes. Layers of fine material interbedded with both the typical breccia and with the coarse conglomerates, as at the French lock site, are shown in practically all the borings. The lowest point reached by the drills was about 194 feet below sea level near the middle of the line of borings. After passing through alluvium filling the;Pleistocene valley of the Chagres, the deepest boring entered about thirty feet of coarse conglomerate; the matrix holding the bowlders was ground up by the drill and none is shown in the samples collected, but * Op. cit., p. 5. Am. Journ. Sct.—FourtTH Series, VoL. XXVI, No. 153.—SzupremsBer, 1908. 16 218 LL. Howe—Geology of the Isthmus of Panama. the presence of certain porphyritic rocks closely resembling those contained in the conglomerate of the lock site makes me believe that this is to be regarded as a conglomerate rather than as a breccia like that of the quarry where little or no vari- ation in the eharacter of the fragments of pyroxene-andesite was found. One hundred feet south of this point another boring passed through fine sandstones or tuffs before striking beds similar to those in the previous hole and at a correspond- ing elevation. About the same distance to the north from the first hole and at the same elevation, fine-grained beds like those of the second hole are also shown by the core, while still nearer the quarries and at an interval again of about one hundred feet fine material mixed with angular fragments of pyroxene and andesite is shown. In the opposite direction a corresponding change to the conglomerate facies is to be observed. Precisely the same transition is shown by a line of borings made from the hill in which the French lock site was excavated in a north- easterly direction across the valley of the Chagres to the hills east of Bohio. : This evidence would seem to indicate that the breccias, best shown at the quarries, were deposited contemporaneously with the conglomerates and sandstones of the lock site. The breccias were in large part, if not entirely, laid down in water, and prob- ably running water, but it is impossible to decide whether the material composing the breccia was transported for some dis- tance by streams before being deposited or if it fell into the water directly from the air. The angular character of the fragments, so far as it may be determined in the extremely decomposed rock, would seem to favor the direct deposition of the breccia in water as a result of voleanic eruption. Eruptions of fragmental material of uniform composition undoubtedly took place during this first period sedimentation, and evidence for this has been found at other points, especially in the vicin- ity of Culebra, as stated in a later paragraph. Southeast of Bohio the character and distribution of the conglomerates and breccias are imperfectly known. Near San Pablo rocks similar to the Bohio breccias are indicated by bor- ings beneath acid tuffs, while it is more than probable that the Bohio beds are represented in the neighborhood of Gorgona by conglomerates exposed along the line of the Panama Rail- road. In most instances it is all but impossible to distinguish between the conglomerates and the volcanic breccias upon which they rest because of the extreme decomposition that both have suffered. The occurrences at Gorgona are fresher, however, and their conglomeratic character unmistakable. The andesitic breccias beneath these conglomerates belong to the Obispo formation and have been traced by borings practi- FE. Howe—Geology of the Isthmus of Panama. 219 eally all the way from near San Pablo to the type locality at Bas Obispo. Hill noted the presence of these conglomerates at the places named and correlated them with the beds at the Bohio lock site.* Vamos Vamos and Gatun beds——-About two and a half miles west of Bohio, on the south bank of the French canal, at a locality known as Vamos Vamos, are exposures of sedimentar y rock that have been described by Hillt and Ber- trand.{ The rocks are impure calcareous shales or marls of a dirty brown color, rich in fossils, and at certain horizons con- tain numerous lar; ge fossiliferous ‘calcareous concretions that at first suggest bowlders in a conglomerate. Following the line of the French canal northwest for nearly six miles, no outcrops of any sort are to be seen until within a short distance of Gatun, where, on the left bank of the canal, are exposures of extremely fossiliferous marls and calcareous sandstones. Similar beds are exposed along the cuttings of the Panama Railroad through Tiger Hill between Bohio and Gatun. No other exposures in this region are known, but a great number of borings made at Gatun in exploration of dam and lock sites furnish much valuable imformation in regard to the character of the rocks. The samples of borings from the lowest sedimentary rocks at Gatun are all of extremely fine- grained, even-textured sandstones which in some places merge into sandy shales. The cementing material is usually calca- reous, but considerable earthy impurity or clay is often present. The coarser components of the rock at Gatun and of that occur- ring along the canal between Gatun and Bohio consist of igneous material; in many specimens grains of nearly fresh feldspar and ferro-magnesian silicates may be recognized, but as a rule the material is in such a finely divided state that it is impossible to say whether it is of direct voleanic origin or has resulted from the degradation of older igneous rocks. At many points fossils are abundant in the core specimens and carbonaceous matter is very generally present. Age of the Bohio formation.—From fossils collected by Hill and others and by myself and examined by Dr. William H. Dall, it appears that the beds described are of Eocene age and contain a number of species typical of the Claiborne and some common to the Upper Tejon of California.§ In the material that I collected at Vamos Vamos Dall has identified the following species : * Op. cit., 188. { Op. cit., p. 179. t Op. cit., p. 6. § Hill, op. cit., Part VI, Report by Dr. William H. Dall upon the Paleon- tology of the collections, pp. 271-275. 220 #. Howe—Geology of the [sthmus of Panama. Lupia perovata Cony, Cytherea Glyptostyla panamensis Dall. Mactra Turritella gatunensis Conr. Corbula Marginella sp. Tellina Natica (cf. eminula Conr. ) Leda sp. Pleurostoma sp. Pyramidella sp. Dentalium sp. In addition to these there are fragments of Ostrea sp., Liocar- dium, and Pecten, and a small ribbed Cardium was noted. From the locality on the left bank of the canal near Gatun practically the same fossils were collected as were found at: Vamos Vamos with the addition of a Cadulus. There can be no doubt that the rocks at Vamos Vamos, near Gatun, and at intermediate points as shown by borings, belong to the same formation, and they have been so regarded by Hill and Bertrand. The relation of these rocks, however, to the peculiar breccias and conglomerates at Bohio is not so clearly shown. Bertrand considers that the breccias at Bohio belong to the same series as those occurring in the central part of the isthmus which I have described as the Obispo formation, and he correlates them with the European Tongrian, that is, the base of the Oligocene in the Paris Basin.section. This determination is .based on fossils obtained from a boring at Kilometer 24°36. Here, according to Bertrand,* small num- mulites of the Oligocene type are associated with Orbztordes that appear to be the same as those of the Pefia Blanca marls (to be described later). The breccia, a part of the ‘“‘ Gamboa rock,” is regarded, therefore, as certainly Oligocene and prob- ably Aquitanian. In the geological profile accompanying Bertrand’s report a special tint is given to the breccias, in order to accentuate the difference between these beds and the higher _ series, and they are referred tentatively to the Tongrian. Hill appears to regard the breccia and perhaps the conglom- erates at the French lock site as parts of an early igneous forma- tion older than the Vamos Vamos beds. My own observations make me believe that the conglomerates at the lock site repre- sent the lowest portion of the formation of which the Vamos Vamos beds and those in the vicinity of Gatun are the upper parts. This opinion is in accord with the local structure, the prevailing dip of the conglomerates, about 15 degrees to the northwest, being sufficient to carry them below the beds at Vamos Vamos. From a boring made by the Canal Commission at the Bohio lock site (Hole 24-b, K 24), fossils were obtained at elevations of from 20 to 40 feet below sea level that are believed by Doctor Dall to represent an Eocene horizon. They consist of fragments of Lucona, Lima, Pecten, Cardium, SOD iCih., ero. £. Howe—Geology of the Isthmus of Panama. 221 Protocardia, and Ostrea, a small Fusus (2), a specimen of Or- bulina, and a Melanian (2) associated with abundant fragmentary plant remains. The abundance of carbonaceous material and the presence of the fresh-water Melanian within a few feet vertically of the salt or brackish water species suggest delta deposits, and with this the physical character of the beds agrees pertectly, the fossils being preserved in calcareous sandstones or shales between beds of conglomerate precisely as are those exposed at the cutting for the locksite. Before reaching these fossiliferous layers the drill passed through a fine breccia like that shown at the quarries at Bohio, interbedded with which are fine-grained sandy layers containing plant remains. Con- glomerates like those of the lock site were found below this breccia, then at about twelve feet below sea level a fine car- bonaceous sandstone layer was encountered, and below this the fossiliferous sandstones. ; On the strength of this evidence I am inclined to regard the conglomerates and breccias at Bohio as members, probably | occurring near the base, of the formation that I have called the Bohio, and that they were laid down as delta deposits at the mouth of a large river, while the Vamos Vamos beds and those near Gatun were deposited contemporaneously in deeper water. It is entirely possible, of course, that the conglomerates may belong to an earlier epoch than the Claiborne, but the evidence for or against this is so meagre that in the absence of anything to the contrary it seems entirely reasonable to place the conglomerates at the base of the Claibornian Bohio forma- tion. No one can doubt the intimate relation between the conglomerates and the breccias nor that they are contempora- neous. It has been shown also that the conglomerate occur- ring in patches near Mamei and Gorgona rests unconformably upon the breccias of the Obispo formation. For this reason and as is also shown in the case of the Culebra beds, I believe that the Bohio rocks with their breccias are younger than the breccias of the Obispo formation instead of being a part of them as suggested by Bertrand. Thickness of the Bohio Formation.—Of the thickness of the Bohio formation very little can be said. Borings on the “TF” line at the Bohio dam site have brought up samples of the conglomerate, sandstone and breccia from nearly 200 feet below sea level, while at least 75 feet more may be added to this as exposed above sea level in the near-by hills. From Bohio to Gatun, where the Bohio formation is covered by younger beds, the distance is about seven miles in a straight line; assuming a uniform dip of only two degrees to the north- west, this would give at Gatun a thickness of nearly 1300 feet. At least 300 feet of the section has been shown in borings at Gatun. 222 . Howe—Geology of the Isthmus of Panama. Culebra Beds. From near Empire to Pedro Miguel, a distance of about five miles along the line of the canal through the Culebra cut, are the best exposures of sedimentary rocks in the Canal Zone. The Obispo formation has been uncovered by excavation from Bas Obispo to a point about midway between Empire and Culebra, where it disappears beneath the sediments described by Hill and others as the Culebra beds. A considerable thick- ness of these beds has been exposed in the deepest. part of the cut, beneath which borings have shown that they extend to at least 40 feet below sea level. A boring at Kilometer 55 passed through 207 feet of Culebra shales and sandstones without reaching the base of the formation. Nearly 175 feet should be added to this section as representing the part already removed from the canal prism at the point where the boring was made. This.would give an observed thickness of nearly 400 feet. Hill estimated the thickness as at least 500 feet* and the total is probably greater rather than less. So far as they have been exposed or explored by borings the rocks are found to be largely soft shales with abundant sandy, conglomeratic, and calcareous layers. At many horizons lentic- ular bodies of limestone occur. Although there are some thick beds of homogeneous pure clay shales, the majority of the rocks, whether sandstones, shales or conglomerates, are richly carbonaceous; lens-like seams of lignite have been found at many places in the cut and remains of trees and plants are abundant. The sandstones are impregnated with lime and in the coarser varieties of the rock films of the cementing calcite are readily visible. The conglomerates are composed of sub- angular rock fragments less than a quarter of an inch in diameter held in an impure calcareous matrix. The material of which all of the sedimentary Culebra beds are composed was derived from an older igneous land mass of which presum- ably the Obispo breccia was a part; quartz is notably absent and the colors range from bluish and greenish gray to nearly black in the richly carbonaceous beds. In the upper part of the section at a number of places along the canal cut between Empire and Paraiso are extensive occur- rences of andesitic breccias. Petrographically the fragments composing them closely resemble those of the Obispo breccia, but they are smaller and more uniform in size than those of the Obispo breccias of the same region. In the central area near Culebra summit these breccias are seen to rest upon the shales or fine sandstones of the Culebra beds. At the same locality there are numerous intrusive masses of basalt, and * Op. cit., p. 198. EF. Howe— Geology of the Isthmus of Panama. 228 faulting and considerable folding have taken place so that the relations of the various rocks are not perfectly clear. There ean be no question, however, that the breccias are a part of the upper Culebra beds, for in several places a simple sedimen- tary contact between the two kinds of rock was observed, and it was even possible to collect hand specimens showing the transition from fine shales to the breccia. The breccias are at many places cemented by calcite of secondary origin. Speci- mens collected at Gold Hill (Cerro Culebra) and Paraiso, although considerably decomposed, may be recognized as made up of fragments of pyroxene-andesite of different textures, some of which have an abundant glassy groundmass. A boring near Paraiso (Kilometer 58) showed similar rocks at several horizons, while interstratified with them were other beds of sediments exactly like those in the lower or main part of the Culebra section. It would appear, therefore, that toward the close of Culebra sedimentation volcanic eruptions took place and that the ejectamenta were laid down conformably on the sediments; a number of such eruptions evidently occurred with intervals of quiescence, during which the deposition of the sediments continued. The association of these breccias with the Culebra sedimentary beds corresponds closely with that of the Bohio breccias and conglomerates with the exception that the Culebra breccias appear in the upper part of the sec- tion while those at Bohio seem to be near the base. Unfortu- nately the state of decomposition of the Bohio rocks prevents satisfactory petrographical comparison between the two brec- cias. In neither case does it seem possible to separate the breccias from the sediments as distinct igneous formations. The relation of the Culebra sediments to the Obispo forma- tion is more definitely shown than the Bohio rocks to the Obispo on the Caribbean side of the isthmus. At the surface the relations are not always clear, but it is possible by means of borings made at frequent intervals to trace the Obispo breccias from Bas Obispo soutb to a point near Empire, where they suddenly pitch in a southeast direction beneath the rapidly thickening cover of the Culebra beds. There is a moderate amount of local folding and faulting in the region and the Culebra beds have a prevailing dip to the southeast, but their inclination is much less than that of the southeasterly pitching surface of the Obispo on which they rest. Near Corozal the Obispo again appears at the surface, while between Corozal and La Boca borings have shown that there are present, beneath the alluvium, sandy shales precisely like many found in the Culebra section. As mentioned later in discussing the age of the Culebra beds, certain limestones that are thought to belong to the upper part of the shale series rest on the 224 LE. Howe —Geology of the Isthmus of Panama. Obispo breccias near Empire. The evidence between Culebra and Empire, at Corozal, and in the vicinity of La Boca seems to indicate that an unconformity exists at the top of the Obispo separating it from the Culebra beds. As has been shown, - similar unconformable relations appear to exist between the Obispo and Bohio formations on the Atlantic slope. Limpire lumestone.—Hill noted an occurrence of massive, semi crystalline limestone near Empire and from its relations in the field referred it to the Culebra beds.* I am not certain that [ identified Hill’s exact locality, but I did succeed in find- ing limestone of precisely the same character in the vicinity of the railway station at Empire on both the east and west sides of the track. Fifty feet east of the new station there are exposures, about ten feet thick, of a very massive cream-col- ored limestone in which I was unable to discover fossils; the outcrop is a small one and its relation to other rocks in the vicinity is not shown. About 150 yards west of this locality on the opposite side of the railway a greenish, impure sandy limestone, hard and compact, occurs beneath thin-bedded ‘eal- careous shales and sandstones, the whole exposure being less than ten feet thick. F ragments of a FPecten were found in this limestone but no determinable fossils. There are no exposures between these outcrops and the canal cut about one quarter of a mile to the northeast, at which point the Obispo formation is found close to where it disappears beneath the Culebra beds. Elsewhere in the Culebra section lenticular bodies of lime- stone, less thick than the massive rock at Empire but other- wise of the same character, are found associated with the eal- careous shales and sandstones, and it seems reasonable, as Hill has suggested, to regard the Empire occurrences as of this nature and belonging to the Culebra beds. The field rela- tions, such as they are, appear to indicate that rocks of the Obispo igneous formation must lie within a few feet of the surface in the neighborhood of Empire, and it is not impossible that the limestone may rest directly upon them. About midway between Empire and Las Cascadas on the Panama Railroad other limestones occur. They are less mas- sive than those at Empire, in places gnarly and crumbling, and being practically at the surface where exposed in the railway cut are considerably weathered. Their color is buff or yellow- ish pink. The few fossils from this locality are poorly pre- served and indeterminable, according to Dr. Dall, but he recognized a nullipore and two species of Pecten and suggests that this may be a reef deposit. =Op7 cit... p: 19a: EF. Howe— Geology of the Isthmus of Panama. — 225 Nearer Las Cascadas bowlder-like concretionary masses of hard crystalline limestone, one to four feet in length, are pre- served in red clays of decomposition near the railway and eanal. Fragments of a Pecten and a tube-like XYylotrya were collected from one of these limestones but their age could not be determined. Age of the Culebra beds.—Hill made a careful study of the Culebra beds, but only succeeded in finding fossils in the Empire limestone, mostly foraminifera that Bagge considered . representative of the Eocene, and as quoted in Hill’s report, he correlated this limestone with the Pefia Blanca marl.* In material that I collected in the vicinity of Las Cascadas from rocks that I believed to belong to the same calcareous horizons as the Empire limestone, Dr. Dall failed to find any species from which the age ot these beds could be determined. Ber- trand and Zircher, however, collected from the same localities and their material was determined by Douvillé as representing the Burdigalian Miocene, or the equivalent of the beds exposed at Kilometer 10 near Gatun.t The French geologists regarded the Empire limestone as belonging to the upper part of the Culebra beds and equivalent to the Vamos Vamos beds of the Caribbean slope. In the vicinity of Pedro Miguel Bertrand and Zurcher found fossils that were correlated with those from Las Cascadas. Fossils from my collection at the same locality were determined by Dr. Dall as probably representing an Oligocene reef deposit, while beneath them a typical Claiborne fauna was found in a compact impure limestone. _ It is probable that the Claiborne horizon at Pedro Miguel is the one that Bertrand and Zircher correlate with the beds at Las Cascadas, since the fauna of both is regarded by them as identical with that preserved at Kilometer 10 near Gatun. It is in collections from the Gatun locality made by Hill and myself that Dall finds his most typical Claiborne fossils. At both Pedro Miguel and Las Cascadas the position of the fos- siliferous beds appears to be at or very near the top of the Culebra section, and in the absence of determinable fossils from lower horizons it seems reasonable to refer the Culebra beds as a whole to the same period as that in which the Bohio forma- tion was deposited. The abundance of carbonaceous shales, seams of lignite and plant remains, together with fragments of a melanian found in the canal eut near Culebra, suggest that most of the beds below the limestone horizons were deposited in fresh water. Limestones of the Upper Chagres.—Following the Chagres headward in a northeasterly direction from Matachin, where the river. turns sharply to the northwest and flows toward the = Op..cit... po ries aba. + Op. cit., pp. 6-9. 226 #. Howe—Geology of the Isthmus of Panama. Caribbean, calcareous rocks similar to those near Las Cascadas and Empire are exposed at many places between Cruces and Dos Bocas, a distance in a straight line of about twelve miles. Between Oruces and Palo Grande compact limestone similar to that at Empire rests on carbonaceous shales like many in the Culebra beds. The rock is partly crystalline and contains in places fragments of shells. About one mile above the mouth of the Chilibre is the lower end of a winding gorge through -which the Chagres flows for nearly twelve miles, the entrance to the gorge being a mile below Dos Bocas. The river, entrenching itself in an old valley, encountered between the points mentioned limestones and calcareous sandstones which offered greater resistance to erosion than the rocks to the north- east and southwest, so that its former meandering course has been preserved very perfectly, and at the outside curves of the meanders are nearly vertical cliff exposures of massive cal- careous sandstone and limestone, in places more than one hun- dred and fifty feet high. Although the length of the gorge following the river is approximately twelve miles, the belt of hard rocks traversed is only about four miles wide. The rocks are all of a light cream or buff color and range from partly crystalline limestones to coarsely granular rocks composed of broken shells, sands, specks of magnetite and occasional pebbles of igneous rocks held in a calcareous cement. Certain layers are well bedded, others massive, while cross- bedding is not uncommon. Fossils, Ostrea and Pecten, are abundant, but I was unable to find any determinable species ; at a locality about a mile below. Alhajuela a bed composed of broken corals and fragments of molluscan shells was found. The similarity of some of these rocks and their fossils with those observed near Las Cascadas is striking, and Bertrand and Zurcher regarded the two occurrences as of the same age.* I agree with this opinion in so far as the Chagres and Las Cas- cadas beds are correlated with the fossiliferous rocks exposed by the canal at Kilometer 10 near Gatun, whose fauna, as has been said, Dall considers Claiborne Eocene ; the French geolo- gists have compared the Chagres rocks with beds near Mar- seilles that grade from Upper Oligocene to Miocene. Bertrand states, on the authority of M. Boutan,t that the limestone near Dos Bocas resembles in part certain marls occurring at Pefia Blanca, and that it contains foraminifera. I am not pre- pared to say that this is not the case, but the structure of these beds as noted both by Bertrand and myself would carry them beneath the limestones of the gorge, and they should corre- spond closely in position to the “limestones observed near Cruces and therefore be comparable to the Empire limestone. *~ Op» cit:, pao: + Op. cit,/p, 10: E.. Howe— Geology of the Isthmus of Panama. — 227 Martius oF PeNa BLAnca. (Lower Oligocene.) Less than a quarter of a mile west of the railway station and quarries at Bohio at the base of a low hill are outcrops of a hard, light yellowish marl. A quarter of a mile still farther west the same rock has been exposed in excavations for a diversion channel made by the French. The rock as a whole may be described as a thick-bedded marl, buff to cream-colored, and containing foraminifera in greater or less abundance, the most typical being Orbitordes fortist. Sparsely disseminated through the rock are minute specks of a dark silicate and frag- ments of feldspar. So far as is known, the only occurrences of this rock are at the localities mentioned between Bohio and Pefia Blanea. Nowhere in the vicinity of Bohio was I able to discover the relation of the Pefia Blanca rock to the Bohio conglomerate or breccia, nor is any direct evidence found in the borings, but from the general field relations it would seem that the marls are younger than the breccias, since less than half a mile east of the marls borings have indicated the presence of the brec- cias and conglomerates many feet below the observed elevation of the marls, while as shown at the lock site exposures, the Bohio rocks dip about fifteen degrees in the the direction of the marls. Unless a fault of considerable magnitude occurs — between the two localities, of which there is no evidence, the marls cannot be older than the Bohio beds, nor is there any indication that they were contemporaneous deposits, borings between Pefia Blanca and Vamos Vamos having passed through rocks of the Vamos Vamos facies of the Bohio alone. Hill states* that he found the foraminiferal marls in uncom- formable contact with the conglomerates at a locality known as Pefia Negra, one mile below Bohio. I was unable to identify this point, but Hill’s observation is in entire accord with my opinion that the foraminiferal beds of Pefia Blanca rest uncom- formably on the rocks of the Bohio formation. Hill regarded the Pefia Blanca beds as older than the Vamos Vamos, basing his opinion on the northwest dip of the Vamos Vamos beds and the position of the foraminiferal marls south- east of them. I have shown, however, that in all probability the Vamos Vamos beds and the Bohio conglomerates belong to the same formation, and the discordance between the Pefia Blanca and Vamos Vamos beds is to be considered as another manifestation of the same unconformity observed by Hill at Pefia Negra. This view agrees with Dr. Dall’s opinion that the Pefia Blanca marls, on account of the presence of the charac- * Op. cit., pp. 178-179. 228 = EE. Howe—Geology of the Isthmus of Panama. teristic species Orbetordes fortis, are of Lower Oligocene age corresponding to the Vicksburg. | Monkey Hitt Formation. (Upper Oligocene.) Between Gatun and the coast at Colon is a more or less hilly region in which numerous exposures may be found along the lines of the Panama Railroad and the French canal. Fossils collected from the rocks at these points indicate that all of the beds are of the same age. ‘The rocks are well stratified, often thin bedded, calcareous sandstones, argillaceous sandstones, marls and shales, usually fine grained and even-textured, and when fresh of a neutral bluish or greenish gray color. They belong to what Hill called the Monkey Hill beds,* the best exposures at the time of his visit being near the Panama Rail- road where it passes through the hills. At the time of my examination by far the best exposures were at Gatun, where extensive excavation had been made for the locks. At this point also it is believed that the base of the formation is shown. In order to avoid ambiguity these beds will be referred to as the Monkey Hill formation, as they were so described by Hill, although in my report to the Canal Commission they were spoken of as the Gatun beds, since they made up a large part of the rocks through which the locks at Gatun were to be con- structed. | From fossils collected both by Hill and myself at Monkey Hill and Gatun, Dr. Dall considers that the Monkey Hill for- mation is equivalent to the Chipola Oligocene, that is, younger than the Vicksburg or the Orbitocdes marls of Pena Blanea. Among the species recognized are: Cardium sp. Turritella Liocardium serratum Linn. Oliva sp. Psammobia Cadulus Cytherea Cerithiopsis Abra Agriopoma Tellina Cyclinella. Chione Directly below and within a foot or two of the point where these fossils were collected at Gatun is a coarse conglomerate also containing numerous fossils that are pronounced by Dall as of the same age as those found near Gatun in the Bohio beds. The conglomerate appears to grade upward into the fine caleareous sandstone or marl that contains the younger fossils, and were it not for the definite indication of greater *Op. cit., 176-177, 208. E. Howe— Geology of the Isthmus of Panama. 229 age from the fossil evidence, I should be inclined to consider this conglomerate a basal member of the Monkey Hill forma- tion ; it is, of course, not impossible that this may be the case and that the fossils were derived from the erosion of the older Bohio terrane. Less than half a mile in a straight line from this locality is the point where the Bohio (Claibornian) fossils were collected at the edge of the French canal near Gatun (Kilometer 10), and the general dip of the beds at this locality is in the direction of the conglomerates exposed at Gatun. In any event the base of the Monkey Hill formation must be very near the conglomerate, whether actually at the top. or including the conglomerate. Unfortunately there are no other exposures in the vicinity to throw light on this matter and little can be learned from the drill records. The conglom- erate is clearly shown by a number of borings, below which are beds of a white pumiceous tuff associated with other con- glomerates similar to the uppermost one, and calcareous sand- stones and marls. Beneath these beds are fine calcareous or argillaceous sandstones of uniform composition that undoubt- edly belong to the Bohio formation. At Gatun about eighty feet of the Monkey Hill formation is shown, but, as in the case of the other sediments, it is only possible to give an approximate estimate of the thickness of the whole formation. [rom rough calculations, assuming low dips of from one to five degrees from Gatun northward, there should be a thickness of at least five hundred feet in the vicin- _ ity of the Monkey Hills near Colon, while perhaps from two to three times this thickness may exist. To the west, where the Chagres leaves its main valley and passes through a younger valley to the sea, a thickness of from one hundred and fifty to two hundred feet is actually shown in the hills, while as much more must exist beneath the present floodplain of the river if the dips observed in the region are at all regular. No rocks corresponding in age to those of the Monkey Hill formation have been observed on the Pacific side of the isth- mus. The lithologic character of these rocks is essentially the same as that of the finer-grained Bohio rocks. They are com- posed almost entirely of the debris of igneous rocks, in some cases the particles being exceedingly fresh, but commonly decomposition and fineness of texture make it impossible to say whether the material was derived directly from volcanic eruption or from the degradation of an older land surface. I am in favor of the latter hypothesis, inasmuch as no evidence has been found elsewhere of contemporaneous eruptions. 230 EF. Howe—Geology of the Isthmus of Panama. Ianrous Rocks. Acid Porphyries and Tuff. The oldest igneous rocks, the breccias of the Obispo forma- tion, have already been described, and it has been shown that eruptions of much the same character continued intermittently for some time during the period of early Tertiary sedimenta- tion. After an interval of quiet, eruptions of an entirely dif- ferent sort of rock took place either at the close of the Bohio epoch or at the beginning of the Monkey Hill. These later eruptives are largely fragmental and are found on both the Atlantic and Pacific sides of the isthmus. The massive rocks of this period are best shown at Ancon Hill, which is com- posed entirely of rhyolite porphyry. Almost completely sur- rounding Ancon Hill, the continuity being broken only on the west by a later intrusion of pyroxene-andesite, are tuffs and fine breccias of the same composition as the rhyolite porphyry of the hill. The tuffs are well bedded and in some places quite massive, individual beds varying from six inches to five or six feet in thickness; they are fine-grained and, where exposed near the surface, usually altered to a white clay. Beds similar to those in the vicinity of Panama are exposed between San Pablo and Tabernilla on the Atlantic slope, the best outcrops being along the Chagres River near the point where the Panama Railroad crosses northwest of San Pablo. These comparatively siliceous rocks are the ones that Hill placed in his Panama formation; he referred to them as rhyo- litic tuffs,* while Bertrand considered them as trachytic.t Specimens that I collected from various points showed greater or less decomposition in most cases and none was sufiiciently fresh for chemical analysis. A microscopical study shows that they are all closely related and that in addition to the occurrences mentioned, certain intrusives in the Culebra beds probably belong to the same series. It is impossible to group all these rocks under one descriptive head such as rhyolite tuff or trachyte tuff; certain facies are distinctly rhyolitie, others trachytic, while forms near quartz-bearing latite are not uncommon. The rock of which Ancon Hill is composed is a creamy- white porphyry with phenocrysts of feldspar in a fine felsitic groundmass; occasional specks of an altered dark mineral are present. Microscopically the rock is seen to be a porphyry with phenocrysts (mentioned in order of relative abundance) of albite, quartz, and orthoclase in a groundmass containing abundant orthoclase, some quartz and a little albite. Slight kaolinization of the feldspars has taken place, and the ferro- * Hill, op. cit., 199-202. + Bertrand, op. cit., 9, 28. E.. Howe— Geology of the Isthmus of Panama. 281 magnesian mineral, altered beyond recognition, has stained the rock slightly with iron oxide; there is no indication that the dark silicate was an important constituent. Another specimen of the same rock from a near-by locality showed little or no porphyritic texture,; a few fragments of orthoclase and laths of oligoclase are pr esent, with many large irregular patches of quartz; there is some magnetite and limonite but apparently no ferro- magnesian miner al. The islands of Naos and Culebra in Panama Bay, between three and four miles from Ancon Hill, are composed of a similar porphyry containing a moderate number of phenocrysts of plagioclase ranging from andesine to a labradorite (Ab,An,), -and hornblende. The groundmass is holocrystalline and con- sists of orthoclase and plagioclase in about equal amount with traces of hornblende, usually altered to chlorite, some magnet- ite, and quartz. Closely related to the rock of Naos Island is an intrusive sill found in the beds at Culebra north of Gold Hill (Cerro Culebra). This rock is a fine-grained porphyry with a fiuidal base, largely glassy, but containing a few microlites of plagio- clase and minute grains of orthoclase. The phenocrysts are mostly andesine although labradorite is present. A dark min- eral, perhaps hornblende, was observed in several specimens, but in all cases it has been considerably altered. In one speci- men it was impossible to decide whether the altered mineral had been hornblende or biotite, the ragged form suggesting the latter. Near Las Cascadas, and associated with the Obispo breccia, a dense rock with a marked fluidal texture is exposed at several places. Under the microscope it is found to consist of a partly glassy groundmass, laths of plagioclase, a little orthoclase and an abundance of fine grains of magnetite. The few plagioclase phenocrysts are considerably decomposed and ferro-magnesian minerals are notably absent. All of these rocks, with the exception of the Ancon porphyry, appear to be intermediate between true rhyolites and latites, with perhaps a stronger tendency toward the latitic form; the impossibility, however, of determining the exact character of the groundmass in most cases pr events a definite classification by microscopical methods, and none of the specimens, as has been said, was sufficiently fresh to warrant chemical analysis. The tuff exposed by the Chagres River in the vicinity of San Pablo consists of small fragments of fibrous or vesicular glass or pumice of low specitic “gravity ; there are occasional ~ par- ticles of plagioclase and orthoclase crystals but no quartz. A greater textural variation is found in the tuffs in the vicinity of Panama; they are less pumiceous and consist of fragments of rock essentially the same as that of Ancon Hill. 232 EF. Howe—Geology of the Lsthmus of Panama. The relation of the tuffs and fine breccias in the vicinity of Panama to Ancon Hill suggests that the pyroclastics were derived from a volcanic center now marked by the massive porphyry of Ancon. The tuffs, as has been said, are well stratified and surround Ancon Hill except in the vicinity of Sosa; they dip away from Ancon Hill at all points and in a few instances have been found to be cut by dikes that appear to radiate from the hill as a center. No local source or center of eruption for the tuffs on the Atlantic slope has been found, and it is possible that they may have been derived from the Ancon eruptions, the distance between the points being only twenty miles. The extremely hght pumiceous character of the San Pablo deposits would favor this view. The age of the acid eruptives is fairly well shown at several places. Near Panama evidence from borings quoted by Ber- trand* indicates that the acid tuffs are younger than the Culebra beds and rest upon them. I was unable to find bor- ing records at the localities mentioned by Bertrand, but the field relations fully justify Bertrand’s view. The Culebra beds disappear beneath the surface of the lower Rio Grande Valley near Pedro Miguel, and at Miraflores the acid tuffs are well exposed at a number of points. Borings made in explor- ing for dam sites between Sosa Hill and Corazal and also across the mouth of the Rio Grande indicate the presence of sediments similar to those of the Culebra beds at approxi- mately seventy or eighty feet below sea level, and one boring near La Boca showed a dike of rock similar to that of Ancon Hull cutting these sediments. On the Atlantic side between San Pablo and Tabernilla there is no surface evidence bearing on the age of the acid tuffs. Borings for a dam site near San Pablo indicate rocks similar to those.of the Bohio breccias, beneath what I believe to be tuffs of the acid series, but the advanced decomposition of the rocks makes it difficult to dis- tinguish one from another, and the supposed tuffs may be sandstones of the Bohio formation. It was mentioned, in describing the Gatun occurrence of the Monkey Hill beds, that beneath the uppermost conglomerate borings revealed the presence of one or more beds of extremely fine white pumi- ceous tuff, and recent excavation for a road from the old Gatun railway station to the encampment at the top of the hill has exposed the tuffs. They are clearly interstratified with the sediments, but as previously stated, it is difficult to determine the base of the Monkey Hill beds, so that the tuffs may be regarded as lying either at the top of the Bohio or at the base of the Monkey Hill formation. In either case there can be no doubt that they are younver than the Eocene sediments and not Cretaceous, as supposed by Hull. * Op. cit., 8-9. FE. Howe— Geology of the Isthmus of Panama. 238 Basie Intrusives. The last phase of eruptive activity of which there is evidence in the Canal Zone is represented by numerous basic intrusives that occur in great abundance in the central and southern _ parts of the isthmus. To the north, on the Atlantic coast, they are exposed in the vicinity of Porto Bello, and about eighteen miles east of Colon. Between Gorgona and Pedro Miguel most of the hills are composed of the basic rocks. In all observed occurrences the rocks are intrusive in the older formations and most commonly occur as large stock-like masses; dikes are numerous but no surface flows have been found. The rocks that have been collectively referred to as basic intrusives are pyroxene-andesites or basalts. All are notable for the calcic character of the feldspar, which is fre- quently bytownite or anorthite. Augite and magnetite are abundant and many of the rocks contain considerable ortho- rhombic pyroxene. The majority of the rocks are pyroxene- andesites, but basalts are common and in one or two instances were found to contain much olivine; practically all have more or less glass in the groundmass and some are vesicular. Although undoubtedly related to the andesites from which the Obispo breccias were derived, none of the Obispo rocks shows the extremely basic character of many of the later intrusives. _At many places in the interior the basic intrusives may be seen cutting the older sedimentary rocks, while the relations near Panama seem to indicate that the andesite of Sosa Hill has been intruded into the rhyolitic tuffs of Ancon. So far no basic rocks have been observed to cut the Monkey Hill for- mation, but the large masses of pyroxene-andesite at Porto Bello are probably intrusive in these sediments althongh no contacts have been observed. The very uniform lithologic character of the Monkey Hill beds, and the lack of any observed unconformity indicating decided crustal movements such as must have accompanied the intrusion of the andésites, would seem to show, in the absence of any more direct evi- dence, that the period of basic intrusions followed the Monkey Hill epoch and may have been the immediate cause of the termination of that long period of quiet sedimentation. Distribution of the Rocks. Unfortunately little is known of the areal distribution of the various formations that have been described. The section exposed along the canal line is fairly complete and in places it has been possible to trace certain of the formations a few miles beyond the boundaries of the Canal Zone, but even if greater areas had been examined it is doubtful if any more Am. JouR Sci.—FourtH Series, Vout. XXVI, No. 153.—SEPTEMBER, 1908. Aly 234 L. Howe—Geology of the Isthmus of Panama. definite information in regard to the limits of the different formations could have been obtained. To a moderate degree the character of the topography may be taken as a guide and the known distribution of the different sorts of rocks thus slightly extended. This applies more especially to areas of sedimentary rocks as contrasted with those of igneous forma- tions. Obispo brecccas.—In the vicinity of Matachin Obispo bree- cias form the hills north of the Chagres, but the surface extension of the formation in this direction can not be great, for, not more than four miles northwest of Matachin, younger sediments are exposed in the neighborhood of Tabernilla and Frijoles. From Matachin southeasterly the breccias occur on both sides of the canal, in places covered by the Culebra beds, beneath which they eventually disappear between Empire and Culebra; they again make their appearance in the vicinity of Corozal and they have been noted at several places in the hills traversed by the old Cruces trail northeast of Miraflores and Pedro Miguel. The breccias are particularly well shown in the rolling country northeast of the city of Panama near Las Sabanas. From borings, the Obispo is known to occur as far north as San Pablo, but beyond that point there is no evidence of its existence. Basic intrusives.—Although the basic intrusives are not limited to any particular part of the isthmus, they are most abundant in the areas characterized by the Obispo breccias. In the Culebra region they are especially numerous, and are believed to make up a large part of the mountainous region southwest of the city of Panama along the Pacific coast; near the shore are many exposures of columnar basalt that have been observed beyond the mouth of the Chorrera River, 18 miles southwest of La Boca. No basic intrusives have been found northwest of San Pablo in the Canal Zone, but at Porto Bello, about 18 miles northeast of Colon, pyroxene-andesites. are exposed at the water’s edge and occur in many of the near-by hills. Acid eruptives.—The rhyohtic rocks and others related to them occur in two general areas. One of these is at Panama, where the well-bedded tuffs surround the central mass of rhyolite porphyry of Ancon Hill. The fragmental rocks underlie the city and extend as far north as Miraflores. Sim- ilar rocks are also found on the southwest side of the Rio Grande opposite La Boca, while massive rock like that of Ancon Hill composes the islands of Naos and Culebra in Panama Bay. Hershey has reported similar rocks more than one hundred miles to the southwest near Santiago.* The * The Geology of the Central Portion of the Isthmus of Panama, Oscar H. Hershey, Bull. Dept. of Geol., Univ. of Cal., vol. ii, p. 244, 1901. E. Howe— Geology of the Isthmus of Panama. 285 second area, in the vicinity of San Pablo and Tabernilla, is probably more restricted than that of the Pacific side. The rocks are best exposed along the Chagres River near San Pablo, but the hills extending northward from Barbacoas, where the Panama Railroad crosses the Chagres, are composed entirely of these acid pyroclastics. Their exact northwestern limit is unknown but probably lies at some point between Buena Vista and Tabernilla. Similar rocks are found in the vicinity of Gatun interbedded with the sediments at that locality. Sedimentary rocks.—A little more is known in regard to the distribution of the sedimentary formations. So far as my observations go, the Culebra beds are restricted to the occur- rences in the Culebra district with a probable extension as far south as La Boca suggested by borings. On the northern side the sediments from the Bohio formation up to the Monkey Hill have a much wider distribution. The sediments extend from Limon Bay along the line of the canal to Bohio at least, and remnants of the Bohio formation have been found as far south as Matachin. The subdued topography characteristic of areas of these young sedimentary rocks extends in a north- easterly direction from Colon almost to Porto Bello and throughout this distance I believe that sedimentary rocks are the only ones represented at the surface. Between this coast belt and the interior, where the limestones of the upper Chagres are exposed, isa region of some elevation concerning which there is no information whatever except quite near the canal line, at Tabernilla and Frijoles, where poorly exposed sediments ‘containing lignite are known to occur. Southwest of the canal line and along the Atlantic slope sediments have been traced for nearly twenty miles up the valley of the Trinidad River. The rocks of this region are of the same character as the Bohio or Monkey Hill formations ; they are exposed at a few places along the water’s edge, but the rock at these outcrops is in an extremely decomposed condition and no fossils were discovered. On the Pacific coast southwest of the city of Panama and beyond the range of hills composed of basic intrusives is the broad low valley of the Chorrera, that, on account of its very subdued topography, looks as if it had been eroded in sedimentary rocks. The Chorrera River rises near the head waters of the Trinidad and it is not improb- able that the sedimentary rocks extend across the divide and down the Chorrera to the Pacific. According to Lull,* coal, uly a lignite, occurs in the drainage of the Rio Indio, * Reports of Explorations and Surveys for the location of Interoceanic Ship Canals through the Isthmus of Panama, E. P. Lull, U. S. N., Washington, 1879. Pp. 30-82. 236 8 £. Howe—Geology of the Isthmus of Panama. and the deposits may be reached from the Pacific side by ascending the Chorrera River. The locality is indefinite, but eannot be far from the head waters of the Trinidad, where similar lignites have been reported. Structure. The broader features of the geological structure of the isthmus are simple. In the Canal Zone the sedimentary rocks of the Atlantic side dip toward the coast at moderate angles while on the southern side of the isthmus the inclination of the Culebra beds is in the direction of the Pacific. It has not been possible to discover any marked characteristic structure in the Obispo breccias ; in the central region the few contem- poraneous flows appear to be nearly horizontal, while no evidence of stratification can be made out in the occurrences near Panama and Las Sabanas. When examined ‘in greater detail it is found that the ineli- nation of the older sediments on the Atlantic side of the isthmus is greater than that of the Monkey Hill beds. The strikes also are different; that of the Bohio formation, best shown at Bohio and Vamos Vamos, being about northeast- southwest, while the Monkey Hill beds at Gatun strike more nearly east and west. This agrees with the evidence of uncon- formity between the Bohio formation and the Orbitozdes beds of Pefia Blanca. No indication of marked faulting has been discovered in the region north of Empire, although local frac- turing and minor dislocation have occurred at places where the folding appears to have been pronounced as at Vamos Vamos, according to Hill, and to a minor degree at Gatun. The acid tuffs at San Pablo are inclined to the northwest at angles of from five to ten degrees; with the exception of isolated patches of Bohio conglomerate near Matachin these are the last beds encountered, following the canal southward, in which any structure can be made out until the Culebra beds are reached near Empire. South of Empire greater structural complexity exists. The Culebra beds are locally folded, but have a general south- easterly dip of from ten to fifteen degrees. At several points the Culebra cut is crossed by small faults, the downthrows of from ten to fifty feet being to the southeast. At the deepest part of the cut, where the canal passes between Gold Hill and Contractor’s Hill, the Culebra shales with the breccias in the upper part of the section have been intruded by basalt and since intrusion have suffered faulting. The mass of Gold Hill has dropped as a block or wedge between the beds on either side, being bounded on the north and south by faults; during FE. Howe— Geology of the Isthmus of Panama. — 237 the process of faulting, the sediments, previously deformed by the intrusions of basalt, were still further twisted and now dip at angles of sixty deer ees or more to the northeast mto Gold Hill. Between Culebra and the Pacific the structure is simple. The Culebra beds, frequently intruded by broad dikes of basalt, are gradually carried by their prevailing southeast dip beneath the level of the Rio Grande valley not far from Pedro Miguel. From Miraflores to the outskirts of the city of Panama the acid tuffs have been locally folded, but in the neighborhood of Ancon Hill and under the town their structure, as already mentioned, appears to be due to initial dips of the beds depos- ited on the fianks of the old Ancon voleano. That is, in tracing the beds from north of Ancon around the hill toa point on the shore close to Sosa Hill the strike is found to swing through an angie of nearly one hundred and eighty degrees ; the dips range from five to fifteen degrees, the steeper inclina- tion being found near the base of Ancon Hill. Reviewing these facts, the structure of the isthmus appears to be characterized by a dominant arch or broad anticline with its axial trend between east-west and northeast-southwest, the crest of the arch being in the south-central part of the isthmus near Bas Obispo. The northern limb of the anticline is of moderate inclination, while the southern limb, near the crest of the arch at least, is steeper. The limestones and calcareous sandstones of the upper Chagres are nearly at the crest of the fold and their prevailing southwest to west dip suggests that the anticline pitches to the southwest. Whether this isa broad structure or comparatively local is not known, no evidence having been found in the country southwest of the Canal Zone. It is probable that the westward pitch is more than a mere local cross fold and due to the same uplift that formed the San ' Blas Range to theeast. Unfortunately there is little or no trustworthy information as to the geology of these mountains other than that they are believed to have a core of granular rock* intruded in late Tertiary time. It must be borne in mind, however, that there is evidence of a decided orogenic movement in this region in late Eocene time that caused the uncontormity at the base of the Pefia Blanca Orbitozdes beds it is not impossible that the supposed southwest pitch of the anticline may be due to the earlier deformation. Washington, D. C. * Hill, op. cit., pp. 211-213. 238 Scientific Intelligence. SCIENTIFIC INTELLIGENCE. I. GEOLOGY. 1. Geology of the Adirondack Magnetic Iron Ores; by D. H. Newtanp, with a Report on the Mineville-Port Henry Group ; by J. FE. Kemp. N. Y. State Mus. Bull. 119, 8°, pp: 182, pls. and maps. Albany, 1908.—This work gives a detailed description of the geology, petrography, and occurrence of the iron-ore bodies in the eastern and northern Adirondack region. It contains also much material relating to the history of the min- erals and mining of the area. It will no doubt prove of great service to those locally concerned in these deposits, and contains much of interest to those engaged in economic geology and in the study of ore deposits. From the descriptions given it would appear that the unusable titaniferous ores are magmatic segrega- tions of the gabbro-anorthosite masses (p. 149); while the purer and worked magnetites, which have a different method of occurrence, are probably, in part at least, due to pneumatolytic processes as suggested by the presence of fluorite in the Palmer hill and other mines (pp. 31-33 and 100). Levees 5. Geologische Prinzipienfragen ; von KH. Rryrer. Pp. 202, 254 figs. Leipzig, 1907 (Wm. Engelmann).—There was a time, in the recollection. of older geologists, when Reyer’s name was associated with the active publication of works on various geo- logical subjects and on the geology of particular regions, marked by a highly theoretical and, at times, imaginative treatment. In the present work the author states that finding his views, espe- cially those relating to the origin of mountain ranges, not gener- ally accepted, he engaged in lines of work other than scientitic, but now, convinced that eventually his experiments and views must prove of service, he feels it incumbent upon him to publish - them in a general statement.. While it would be entirely out of place in this brief notice to enter into a critical discussion of the author’s views on fundamental geological problems, it may be men- tioned that some of the more important subjects treated relate to the origin and manner of igneous intrusions and extrusions, to the part played by volcanic islands, to the igneous phenomena as displayed in the Alps, to the origin of mountain ranges, to ele- vation and depression of the crust, etc. The author says frankly at the outset that he expects opposition to his views, which, in many cases at least, depart widely from those generally held to-day, and in some instances represent theories which have been discarded in the evolution of geological science. While this is true, the work is at least suggestive, and even if the reader does not accept the presentation of the particular theses discussed, he may find a strong sidelight cast on some special problem in which he is interested. ‘ada 0 IVES 2. Die Entstehung der Kontinente, der Vulkane und Gebirge ; von P.O. Kéuter. 8°, 58 pp., Leipzig, 1908 (Wm. Engelmann).— Geology. 239 These are momentous questions to be considered and answered in 58 pages, and therefore the author does not waste time in pre- liminaries, or in: consideration of details. He believes that the view, often advanced, that the features of the earth mentioned in — the title are caused by the contraction of a cold crust settling down to fit a still hot, but cooling and contracting, nucleus, is essentially wrong. He essays to prove that the crust is losing heat faster than the interior, and that this interior, as compared with the outer shell, is relatively stable. The latter, instead of being under contractional stresses, is in a state of tension. The relative movements of the outer surface, which give rise to earth features, are ascribed to the action of water, which penetrating downward to the heated zone below, returns in a hot condition, warming the superincumbent masses and thus causing them to expand. By this mechanism, in various ways, he endeavors to show that the continents, volcanoes, and mountain ranges are formed. While it is not probable that this brochure will be taken very seriously by those who still hold by the nebular hypothesis of the earth’s origin, and will seem to followers of Chamberlin’s planetesimal hypothesis much lke a charge upon windmills, it may be still said that it is well and clearly written, and in places contains suggestive ideas. Bi Wee 3. Geological Survey of Canada, A. P. Low, Director.— The following publications have been recently issued : Annual Report, New Series, Volume XVI, for 19045 this contains Reports A, B, C, CC, G, H and 8 and is accompanied by a series of fourteen maps. Ottawa, 1906. It is stated that this volume is the last to be published in this form, the plan being to present in future each report as a separate publication. Summary Report for the calendar year 1906 ; pp. 206. Ottawa, 1906, The Falls of Niagara; by J. W.W. Spencer, 1905-6. Pp. xxv, 490, with 43 plates and 30 figures. Ottawa, 1907. This volume was noticed in an earlier number (vol. xxv, p. 455). Report on Gold Values in the Klondike high level Gravels ; by R. G. McConnety. Pp. 34, with one plate and a geological map. The Telkwa River and Vicinity, B. C.; by W. W. Leacu. Pp. 27, with a geological map. Ottawa, 1908. Report on a portion of Northwestern Ontario, traversed by the National Transcontinental Railway, between Lake Nipigon and Sturgeon Lake; by W. H. Couuins. Pp. 23, with a geological map. Ottawa, 1908. 4. Geography and Geology of a Portion of Southwestern Wyoming, with special Reference to Coal and Oil; by A: C. Veatcn. Prof. Papers, 56, U. 8S. Geol. Surv., pp. 178, 1907 (=1908).—This is a very important paper for statigraphers and paleontologists, and, especially, for those interested in the discus- sion as to whether the Upper Laramie is to be referred to the Mesozoic or the Tertiary. Theauthor shows that the Cretaceous 240 Scientific Intelligence. of this region ‘has the enormous thickness of over 20,000 feet,” and is a conformable series, beginning with the Bear River forma- tion, and closing with the Adaville formation=Lower Laramie. Then followed “a long period of folding, faulting and erosion.” Angularly unconformable with the Cretaceous series are the Evanston, Almy and Fowkes formations, followed by another period of “folding and erosion of great magnitude though of much less importance” than the earlier one. The Almy and Fowkes formations “have without exception been considered Kocene,” and, as they are conformable with the Evanston, all are regarded by Veatch as best placed in the Kocene. ‘The uncom- formity at the base of this series [Evanston, Almy, and Fowkes] amounts to over 20,000 feet ; that at the top amounts to perhaps 5,000 feet, but this is of much less relative significance than the figures indicate, because the movements of the second disturb- ance were along lines of weakness produced by the first. The physical break between this group and the known Cretaceous beds is thus greater than the break between it and the known Kocene, and, on purely physical grounds, this group would seem to belong rather to the Eocene than to the Cretaceous ” (p. 75—76). Cais: 5. Hinfiithrung in die Paldontologie ; von Gustav STEINMANN. Second edition, pp. 542, with 902 text-figures. Leipzig, 1907 (Wilhelm Engelmann).—This well-known introduction to paleon- tology has been enlarged and brought up to date. It treats of plants (pp. 13-74), invertebrates (75-388), and vertebrates (389— ~ 514). All of the more important groups of forms found fossil are defined and illustrated, so that any beginner in paleontology may obtain a good knowledge of the hard parts of extinct organ- isms. No detailed classifications appear, nor is there any extended discussion of lines of descent. The book presents what is known of the leading forms in each group of organisms in short synoptic form. The illustrations are wood-cuts and line drawings, made especially for this book, and are abundant and adequate. c. Ss. 6. Niagara Stromatoporoids; by W. A. Parxs. Univ. Toronto Studies, Geo]. Ser., 1908, pp. 175-240, pls. 7-14.—Pro- fessor Parks here continues his detailed studies on American Silurian stromatoporoids. At least 34 forms are described, many being new. Chalazodes is the only new genus. C. 8. 7. On an Occurrence of Hybocystis in Ontario; by W. A. Parxs. Ottawa Nat., XXI, 1908, pp. 232-236, pl. 2.—Very excellent material of this obscure echinoderm has been found near Elton, Ontario. The material is described and figured in detail, indicating, the reviewer thinks, cystid rather than blastid char- acters. Cc. S. Il. MiscetuAnrous Screntiric INTELLIGENCE. 1. Publications of the Japanese Harthquake Investigation Committee.—Nos. 22 A and 22 C of this issue have recently been received, also Vol. II, No. 1, of the Bulletin of the Commit- tee. This last contains 8 articles by Prof. F. Omori, all of inter- Miscellaneous Intelligence. 241 est to those concerned with seismology. One of these is an interesting discussion on microtremors and another gives a list of prominent Japanese earthquakes between 1902 and 1907. In No. 22 C an account is given by T. Wakimizu of a new vol- canic island in the Iwdjima group, remarkable because of its ephemeral character. On February 1, 1905, this island was 3 miles in circumference and 480 feet in height, but on the 16th of June it was reduced to alow reef only 1,500 feet long and less than 10 feet high. In June of the present year it is stated that the new island was entirely buried by the sea. Interesting accounts are also given of the other islands of the group, accompanied by a series of excellent plates. Volume xxiv of the Journal of the College of Science, Imperial University of Tokyo, is devoted to an investigation of the sec- ondary undulations of ocean tides, carried out by the order of the Earthquake Investigation Committee during 1903-1906. It is accompanied by 95 plates, maps and charts. The authors are K. Honda, T. Terada, Y. Yoshida and D. Isitani. 2. The Physical Basis of Civilization ; by T. W. HEINEMAN. Pp. 241. Chicago, 1908 (Forbes & Co.).—The author advances the theory in this book that the physical, mental, moral, and social conditions of the human race are due to two compara- tively slight structural modifications of the ape-like ancestors of man: first, the lengthening of the foot by the modification of the great toe, which gave man a position on his feet of far greater firmness than that of any existing apes; and second, the position of the skull with reference to the spinal column, on account of which the erect position of the body is more readily maintained. As a result of the specialization of the feet for sup- porting the body firmly, the hands were left free for grasping and handling and the sense of touch became more highly devel- oped so that a more efficient means of acquiring knowledge of — surrounding objects was opened to man ; likewise, the elevation of the special sense organs high above the ground by the erect attitude widened the range of their usefulness and thus led to the increase of the intelligence of man. The upright position, and consequent exposure of the vital organs to attack, and the comparative defenselessness of the human species, rendered his survival in the struggle for existence dependent upon conduct to a greater extent than in any other known animals, and the extreme disability of the pregnant female made the devotion of the male to his mate and the establishment of the family relation- ship necessary conditions for the survival of the race. B. w. K. 3. General Physics. An elementary Text-book for Colleges, by Henry Crew. Pp. xi, 522 with 40 figures. New York, 1908 (The Macmillan Co.). —This text-book is in its treatment dis- tinctly above the plane of the high school type, and yet, without sacrifice of accuracy, maintains throughout a simple, lucid exposition of the fundamental principles of the subject. The scope of the book is obvious from the title and number of pages stated above. The presentation of the subject matter is excel- lent. There is in it little to suggest the older purely descriptive 242 Screntifie Intelligence. texts and happily less of the more recent tendency toward a presentation so severe as to engender in most beginners a hatred, rather than a love of the subject. The author has adopted, in general, if not uniformly, a happy method of leading up to prin- ciples through a simple discussion of the knowledge already in the student’s possession. The mathematical expressions are developed in a natural and consistent manner and are made to serve effectively both in summarizing the respective topics and in revealing to the student ‘the essential unity of the subject.” Practically all of the illustrations are diagrammatic and clear. Problems are found in ample number and variety. The text lends itself readily to abridgement or amplification, and taken as a whole this appears to be one of the best recent text-books. Di wAQGKe 4. Die Insektenfamilie der Phasmiden ; bearbeitet von K. Brunner v. Watrenwyt und Jos. Reprensacuer. II Liefe- rung, pp. 181-340; IIL Schluss-Lieferung, pp 341-589, with plates xvi-xxvil. Leipzig, 1908 (Wm. Engelmann).—The first part of this monumental work on the Phasmids appeared in the spring of 1907 (see vol. xxii, 398). ‘The two parts now issued | complete the work. Of these the second is devoted to the Clitumnini, Lonchodini and Bacunculini; it has been prepared by the senior editor. ‘The third and concluding part embraces the Phibalosomini, Acrophyllini, and Necrosciini and has been worked up by Prof. Redtenbacher. The whole work is admir- ably thorough and is based, not only upon the very extensive private collections of the authors, but they have also taken advantage of the material in various public museums, especially those in Europe, which have most freely placed their collections at their disposal. The publication of the whole work in the liberal form presented, with its numerous plates, has been made possible through the support of the Imperial Academy at Vienna, the funds being furnished by the Treitl Foundation. 5. Les Dépéts Marins; Lion W. Cotter. Pp. 325, with 35 figures. Paris, 1908 (O. Doin).—This volume forms one of the issues of the “ Encyclopédie scientifique” which is being pub- lished under the direction of Dr. Toulouse. If carried through on the very liberal scale planned, the Encyclopedia will include 40 sections, aggregating about 1000 volumes. The work in hand is one of seven volumes, to be devoted to physical oceanography, which are in charge of Dr. J. Richard. The author, who has studied with Sir John Murray at Edinburgh, presents here an excellent summary of the whole subject of marine depositions, giving the results contained in the well-known work by Murray and Renard on Deep Sea Deposits, and also bringing the various branches of the subject down to date. The concise, systematic treatment of the whole makes it a very con- venient résumé of a subject of more than usual interest. OBITUARY. JAMES Duncan Hague, long prominent as a mining engineer, died on August 5 at the age of seventy-two years. He was one of the geologists of the Survey of the 40th Parallel in 1867-70. SUPPLEMENT. Arr. XXV.—On the Esterification of Malonic Acid; by I. K. Puetes and E. W. Tiruotson, JR. [Contributions from the Kent Chemical Laboratory of Yale Univ.—clxxxi. | Frinxetstetn*™ has shown that diethyl malonic ester may be prepared by dissolving malonic acid in the least possible amount of absolute alcohol before saturating with hydrochloric acid gas. The excess of alcohol is then distilled off, the resi- due poured into water, neutralized with sodium carbonate, and extracted with ether. On evaporation of the ether, diethyl malonic ester, boiling for the most part at 195°, is obtained. He gives no quantitative results. Conrad+ has followed a pre- cisely similar procedure, using the calcium salt of malonie acid instead of the acid itself. The yield given is 70 per cent of that theoretically possible. Quite recently Bogojawlensky¢, by boiling a solution of malonic acid in alcohol under a return condenser for six to seven hours, in the presence of anhydrous copper sulphate or potassium pyrosulphate as dehydrating agents, has obtained yields of malonic ester 68 per cent of that theoretically possible. In former papers§ from this laboratory studies of conditions giving high yields of the ethyl esters of succinic and benzoic acids have been made. And, further, the effect upon the quantity of ester produced, caused by vary- ing three factors in the reaction, was shown. These factors were, first, the quantity of alcohol, second, certain catalyzers, and third, the period of the time of action. In this paper a similar study of the esterification of malonic acid with ethyl alcohol is recorded. for the preparation of pure malonic acid, the diethyl ester was purified by repeated fractional distilla- tions under atmospheric pressure. Portions of malonic ester, * Ann., cxxxiii, 338. tIbid., eciv, 126. t Berichte, xxxviii, 3344. § This Journal, xxii, 368; xxiv, 194; xxv, 39. || In distilling the ester, the following simple modification of the Hempel bead column was found advantageous. To the lower end of the column, which was 15™™ in diameter, was fused a glass tube 7™™ inside diameter and d™ long, the lower end of which was ground off at an angle.. To prevent the beads from falling through this tube, two devices were made use of. Hither A, shown by itself, or B, shown in position in such a column in the 244 Phelps and Tillotson, Jr.—Malonic Acid. boiling within two-tenths of a degree, were hydrolyzed by heating at a temperature of about 50°, a mixture of ester and water in nearly equal amounts with a few dr ops of nitric acid, for some time after the mixture became homogeneous. The solution was then transferred to a porcelain dish and evapo- rated to the point of saturation at a temperature not exceed- ing* 60°, filtered while hot, and stirred while cooling. After recrystallizing from water, ‘the acid was dried, first ‘in the air, and then to constant weight in a desiccator over sulphuric acid. The malonie acid prepared in this manner was proved to be pure by titrating it against standard sodium hydroxidet and barium hydroxide solutions. eut. A is a glass tube 1°™ in diameter, drawn together at either end, and held in place in the column by three ‘‘ tears” fused to it. Bisa smal] glass rod bent into a U shape wide enough to slip easily into the column. To the bottom of the U are fused one or more crosspieces of the same small glass rod, thus forming a grating which allows a ready escape for the vapors and. the return of condensed liquid. The number of crosspieces necessary varies, obviously, with the internal diameter of the column and the size of the beads used. This device is also useful in connection with constricted side- necked flasks, since very little constriction is sufficient to hold it in place. The device can be varied for introducing a bead column in the side neck of a Claisen flask by lengthening the upright rods sufficiently for the device to be held in the angle of the side-neck tube. This modification obviates the necessity of constricting the neck in this case. Even in distilling high boil- ing point liquids like diethyl malonic ester, when the amount of liquid flow- ing back is large, liquid does not collect insuch a column. In practice, B seems preferable to A on account of the more ready back flow of liquid. * F, Lamouroux, Compt. Rend., cxxviii, 998. + This Journal, xxvi, 138. Phelps and Tillotson, Jr.—Matonie Acid. 245 The alcohol used was the alcohol of commerce made as free as possible from water by repeated distillations over fresh cal- cium oxide. Pure zine chloride of commerce was freshly fused and treated with a current of dry hydrochloric acid gas until a clear melt was obtained; then this mass was heated for a short time to expel any hydrochloric acid gas before cooling and granulating. In every experiment recorded in Table I, weighed portions of malonic acid were treated with definite amounts of absolute alcohol, alone, or charged with a known amount of dry hydro- chlorie acid gas with or without a definite weight of zinc chloride, in the special arrangement of flasks described in a former paper for use in esterifying succinic® acid. In all experiments, except those in which the treatment was special, definite amounts of malonic acid with 40°"° of absolute alcohol, alone or charged with dry hydrochlorie acid gas, or in presence of whatever other catalyzer was employed, were heated in a 500™* round-bottomed flask, while the remainder of the alco- hol used containing hydrochloric acid or not, as shown in the table, was boiled in a second 500°* round-bottomed flask, and passed in vapor form to the bottom of the malonic acid solu- tion in the first flask, which was kept at a temperature of 100° to 110° by heating in a bath of sulphuric acid and potassium sulphate. The temperature in the esterification flask was regis- tered by a thermometer dipping into the alcoholic solution and held in place by a three-bored rubber stopper which carried the inlet tube, and also a Hempel bead column, arranged as described in the paper to which reference has been made, to provide an outlet for the vapors liberated in the flask. The product obtained was transferred, with the aid of a small amount of ether, to a separating funnel containing chipped ice, and treated with an excess of an aqueous solution of sodium carbonate. The ethereal solution was washed with . a solution containmg sodium chloride. To recover traces of ester in the wash waters, the carbonate and chloride solutions were shaken out twice, successively, with fresh portions of ether, the ether extracts combined in a 250° side-necked flask fitted for vacuum distillation, with a capillary tube and receiver consisting of a 100° side-necked flask connected through a manometer to an aspirating pump. The low-boiling products, consisting chiefly of ether, alcohol and water, were removed by heating the flask containing the ester solution in a water bath, finally at a temperature of 60° for fifteen minutes after the manometer showed a pressure of 15™™. The water bath was then replaced by an acid potassium sulphate bath, heated to about 140°, and the diethyl malonic ester distilled * This Journal, xxiv, 194. 246 Phelps and Tillotson, Jr.—Matonie Acid. and collected in the 100° side-necked receiver, which was kept cool by allowing a stream of cold water to flow over it constantly. In every experiment in Table I the flow of hot alcohol vapor was started into the maloniec acid solution before the tempera- No. (Ye ise) — (22) Malonic acid ZnCl, grm. grm. 50 Sui 50 re 50 dee 50 Ey 50 pS 50 05 50 te) 50 1°0 50 1:0 50 1°0 50 1:0 50 1:0 50 120) 50 1:0 50 10°0 50 10°0 50 10°0 50 10°0 50 10°0 HON AOLOr, 50 = 10°0 AO O70 TABLE I. Alcohol Reaction with HCl time (FS |) em®, percent hr, min. 200 Lee 1 50 200 132653 epee 200 1°25 1 45 300 1°25 20) 100 1°25 Berth ae 100 1°25 Ae ey 200 1:25 Dts oe 40 1°25 160 oe Bie nays) 200 as J) AO 40 1°25 160 Ee BAS 40: 1°25 160 ae “4545 40 Ho's) 160 ae: 1 40 200 1°25 tis 300 1°25 er O 40 10 | 260 jes Leo 40 10 260 ee 2 30 40 1°25 160 tae ce Mes Ob 40 25 160 ae te G0) 200 WAS DAS 200 10 55 200 10 1 30 400 125 ea) 200 1°25 A) 200 129 books 200 1°25 Ae GRAD 200 1 heaaes) Eee 595) Malonie ester Theory erm, 76°92 76°92 76°92 76°92 16°92 76°92 76°92 76°92 76°92 76°92 76°92 76°92 76°92 76°92 76°92 76°92 76°92 76°92 76°92 76°92 GoO2 76°92 Ae 7 ea per cent 70°2 84°5 89°1 94°0 93°7 84°0 47°0 . 9073 82°6 87:0 90°0 92°71 93°2 ture reached 105°, which temperature was attained within fif- teen minutes after beginning the experiment, and a steady current was maintained afterwards, holding the temperature between 100° and 110° till all the alcohol had been distilled Phelps and Tillotson, Jr.—Malonie Acid. 247 over, except in experiments (5), (9), (10), (21) and (22), where the treatment was special. In experiment (5) the malonie acid, dissolved in the first portion of alcohol shown in the table, and contained in the 500°" round-bottomed flask, was heated on a water bath at 50° for three hours, then connected with a vacuum pump and heated with the water bath at the same temperature until 40° of liquid, consisting, presumably, of water and alcohol, had distilled over. The second -portion of aleohol was then added and the heating on the water bath at 50° continued for four hours, after which about 80° of liquid was removed under diminished pressure as before. Finally, 40° of the third portion of alcohol shown in the table was added and the process of esterification completed in the same manner as in the otherexperiments. In experiments (9) and (10) the current of alcohol vapor at first was slow; in the case of experiment (9), for fifteen minutes; in the case of experiment (10), for an hour, and more rapid for the remain- der of the time. In experiment (21), after treatment in the usual manner with the first portion of alcohol, the low-boiling products were removed under diminished pressure by heating the esterification flask on a water bath at 60° till the manome- ter showed a pressure of 15™™ for fifteen minutes, and the proc- . ess of esterification was repeated with the second portion of alcohol, shown in the table. In experiment (22) the malonic acid, zine chloride and 200™* of alcohol were boiled in the 500° round-bottomed flask under a return condenser for forty-five minutes before continuing the esterification with the second portion of alcohol in the usual manner. To learn whether any ester had distilled with the alcohol during the esterification, several of the alcoholic distillates were cooled with ice diluted with three or four times its volume of water and shaken out separately three times with fresh portions of ether, washing the collected portion with a solution of sodium carbonate, and, finally, with pure water. The combined ether solutions were then fractioned in vacuo as described above. In no case was malonic ester found in the distillates tested. The loss inherent in the process employed for recovering the pure ester from the crude material, produced - in the esterification flask, was determined by treating 75 grm. of pure malonic ester by the same procedure, described above, for the crude material. The loss amounted to 1:25 erm. of malonic ester. From experiments (1), (2), (8) and (4) of Table [it is plain that an increase in either the amount of alcoho] containing hydrochloric acid, or the time of- action, gives an increase in the amount of ester produced. Comparing experiments (1) and (7), absolute alcohol alone, in the absence of zinc chloride, 248 Phelps and Tillotson, Jr.—Matlonic Acid. acting for a longer time, produces a far greater amount of ester than is obtained by the use of zine chloride for a shorter time. However, if, as in experiment (8), a small amount of hydro- chloric acid is present with the zine chloride, the yield is increased 43 per cent over experiment (7) when zine chloride was present but no hydrochloric acid, and 5°5 per cent over experiment (2), when hydrochloric acid was used with no zine ehloride. In experiment (11) the use of 1 grm. of zine chlo- ride with hydrochloric acid gives an increase of 5°5 per cent over experiment (2), where zine chloride was not used. Com- paring experiments (6), (8) and (15), it is apparent that under similar conditions, 1 grm. of zine chloride gives better yields than 0-5 grm. or 10 grm. for the shorter time, while for a longer time, as shown in experiments (10) and (16), 10 grm. of zine chloride seem to give the better yield. Increasing the amount of alcohol driven over in the same time, as shown in experi- ments (11) and (12), gives a decided increase in the amount of ester obtained, while increasing the amount of hydrochloric acid to 10 per cent tends to reduce the amount of ester formed, as is evident In experiments (18) and (19). In experiments (1) to (4) and in eertain of the others, ethyl acetate was . detected by its odor in the first few drops of the distillates, but when the esterification was partially completed on a water bath, as in experiments (5) and (22), before continuing in the regular apparatus, no ethyl acetate could be detected by its odor in the presence of the large mass of alcohol, and, in the case of experiment (22) a better yield of ester was obtained. It is possible that in experiment (5) some ester was lost during the treatments under diminished pressure, as described above. For all the experiments in Table II, sulphuric acid of com- merce sp. gr. 1°84, and commercial alcohol, made anhydrous by the method described, were used. In the experiments in series A, the procedure was precisely as in the experiments of Table . I. Inseries B, malonic acid was treated with the first portion of alcohol and the small amounts of sulphuric acid, shown in Table Il. This solution was boiled under a return condenser for one hour, then the second amount of sulphuric acid was ‘added and the process of esterification completed with the second portion of alcoho] in the usual manner. In series O, the procedure was the same as that described for experiment (5) of Table I. It was found in experiment (1) of series A that when malonic acid was treated in the esterification apparatus in the presence of sulphuric acid, there was a distinct odor of ethyl acetate in the distillate, indicating decomposition of the malonic acid, or the acid ester, but when the alcoholic solution was boiled under a return condenser for an hour, as in experi- ment (4) of series B, the odor of ethyl acetate was not Phelps and Tillotson, Jr—Malonie Acid. 249 observed. There was also some decomposition, as evidenced by the odor of ethyl acetate in the distillate of experiment (2) of series B, when the solution was heated immediately to 105° in the esterification apparatus, but if the temperature was kept at 90° for fifteen minutes, as in experiment (8) of series B, the ethyl. acetate was not detected by its odor in the distil- late in the presence of the large amount of alcohol. For the purpose of examining the distillates more closely for the pres- ence of ethyl acetate, blank tests were made by mixing in a side-necked flask, connected: with a condenser, a solution of aleohol and ethyl] acetate with an equal volume of concen- trated sulphuric acid, and heating the liquid to a temperature of 80°, at the same time passing a current of air through the liquid. About 3° of the distillate was collected in a gradu- ated pipette, containing an aqueous solution of sodium chloride, and thoroughly shaken. When 50° of alcohol and 1:8 of ethyl acetate, which was calculated as the least amount to be found in any of the distillates, if all of the malonic acid not found as ester had been decomposed, was treated in this man- ner, 0°4°™ of liquid separated out, easily recognized by its characteristic odor as ethyl acetate. In treating by the process outlined above the first 50°™* of the alcoholic distillates of experiments (5) and (6) of series B and (1) and (2) of series C, no ethyl acetate could be detected. In order to learn, if possible, whether the loss of malonie acid in the process of esterification was due entirely to decom- position, or whether part of the malonic acid was incompletely esterified, experiments were made to discover if malonie acid could be recovered from water solution. Shaking out a water solution containing sodium chloride and 2 grm. of malonic acid three times with ether gave 0-05 srm. of malonic acid, but when 5 grm. of malonic acid, in a similar solution, were treated eighteen hours in an ether extractor, and the ethereal solution evaporated under diminished pressure at 60° until the manom- eter registered 15™", 4°97 orm. were recovered. A _ blank test on 5 grm. of pure malonic acid showed no loss in drying from an aqueous ether solution at 60° in vacuo. The sodium carbonate wash waters were acidified with hydrochloric acid and treated for eighteen hours in an extractor similar to that described by Van Rijn*, except that an ordinary test tube of about 100° capacity was used instead of the constricted and perforated tube employed by him. The amounts of malonic acid so recovered are shown in Table II. The amount of malonic ester taken up by the sodium carbonate wash waters was determined by ablank test on 75 grm. of pure malonic ester, which was treated with a sodium carbonate solution, as * Berichte, xxviii, 2387. Am. Jour. Scit.—FourtTH Series, Vout. XXVI, No. 153.—SEPTEMBER, 1908. 250 Phelps and Tillotson, Jr.—Malonic Acid. described for the recovery of malonic ester. Upon acidify- ing this sodium carbonate solution with hydrochlorie acid and treating in the extractor as outlined above, 0°30 orm. of malonic acid equivalent to 0°39 grm. of ester was obtained. In calculating the figures given in the last column that stand for the percent of malonic acid, or ester, unaccounted TaBLeE IT. Malonic acid — es -— Malonic ester FYoundin Not Malonic ——_—_*——_— wash accounted No. acid Alcohol H.SO, Time Theory Found Percent water for orm, acm. | orm, hr min orm. soma. . grm. per cent A 40 2:00 ; 1 oD 160" 222) A= 15 8 7G592 70296) 92730 ne is ae B OO; 0-15 = 0 150.5 200. -_- 8 1 — 00. 7692. 367-625 87-90) aslo 7°8 60 O15 1-00 | F250. 140) A855 1 00n) 76:92... 70:00)» Ol 00pm awe 5:8 100 2p O21 5 1 00 32° 250) 200) 1°85 90 —45., 16:92 70-84) 02209. eos A'S 100) 0-1 00 A. 50°. 200, 1°85 1 = 00 71692" Fi50 92-95) aaa is OOS IPSae 9) 50° 200 1°85 2—00 76°92 73°12 95°06 0°36 3°2 MOOT 0S, ie— 00 6 50 +) 300). 11:85 = 235)" 76:92 72°60" 94:38) sare 3°9 100-015, >1—00 7 50 200),- 4:85" 4 — 00 76:92" (70°93) 92:24 1°80 3°2 C 100% 015 to — 00 LOO 22 4 — 00 SOV ALS Oy se ey ] 50 2007 eee v2 00M iG 22 ai 3 es OS ai 0°39 2°4 100 2°00 4-00 100 baa ss 0 OO wero s hee 2) = OO 200. +... 2-00. 76°92 73°94 96:11 0°40 ropes | for, it was considered fair to add to the amount of malonic ester obtained in a single experiment 1°25 grm., the amount found by the blank test to be the loss inherent in the process of recovery of the pure ester. To this sum was added the amount of ester equivalent to the difference between the malonic acid obtained from the sodium carbonate wash waters and 0°39 grm., which as has been shown is the fraction of the Pheips and Tillotson, Jr.—Malonic Acid. 251 1:25 orm. loss in the process of recovery, taken up by the sodium carbonate water, and recovered as malonie acid by the ether extraction. ‘Taking, for example, experiment (2) of series C in Table II, of the 0-40 grm. of malonic acid recovered from the wash water, 0°30 grm. was presumably due to the action of the sodium carbonate on the ester during the process of shaking out with ether.. We have then, 73:94 grm. of ester found, 1°25 erm. of ester lost during the process of recovering the ester and acid, corresponding to 0°18 grm. of ester, remaining unesterified. The total, therefore, of ester accounted for is 75°32 grm. or 97-9 per cent, leaving 2°1 per cent lost, probably through decom- position. In this particular experiment, as well as in experi- ment (1) of series C, a part of the loss is due to vaporization of ester during the interpolated treatments under diminished pressure. | From the results shown in Table II it is evident that under conditions which give the best yield, not more than 0°10 germ. of malonic acid remains unesterified, and, further, that about 1 grm. or two per cent of acid is lost through decomposition. Under conditions less favorable for theoretical yields, a larger amount of acid remains unesterified, and at the same time the amount lost by decomposition is gr eater. Comparing experi- ments (2) and (3) of series B, it is evident that increasing the amount of alcohol with which the malonic acid is treated on the water bath, from 60°™* to 100°", increases the yield by one per cent, while if malonic acid is treated with several portions of alcohol as described above, and shown in experiments (1) and (2) of series C, or if the time of driving over the second portion of alcohol is increased as in experiment (5) of series 6, a much larger yield of ester is obtained. Merely increas- ing the amount of alcohol distilled over, as in experiment (6) of series B, gives no better yield than under otherwise similar conditions in experiment (5). In experiments (1), (4) and (7) it is plain that with small amounts of sulphuric acid, esterifi- cation is not complete, while with larger amounts the yield of ester is apparently not quite as good as that produced by two grams. by varying the time in experiments (3), (4) and (5) of series B, a larger yield of ester is obtained, with increase in the time taken to distil the second portion of aleohol, while if the alcoholic solution of malonic acid is heated for a longer time on the water bath, as in experiments (1) and (2) of series C, the yield is materially increased. The ester obtained in all the experiments of Tables I and II was found to be in a high state of purity, since, on redistilla- tion, it showed no considerable variation in boiling point. From the work here described, it may be seen that in gen- eral, increasing within limits the amount of alcohol used, and 252 Phelps and Tillotson, Jr.—Malonic Acid. the time during which it is allowed to act, produces a more complete esterification of malonic acid. Larger amounts of zine chloride up to 10 grm. appear to increase the yield as in experiments (1), (4) and (17) of Table I, while the use of 5 grm. of sulphuric acid seems to possess no advantage over 2 grm., as is shown in experiments (4) and (7) of series B in Table II. Under similar conditions, however, 2 grm. of sulphuric acid with 200% of absolute alcohol running for one hour and fifteen minutes in experiment (1) of series A in Table II gives a yield of malonic ester nearly equal to that produced by 1 grm. of zine chloride and 300° of alcohol, charged with 1°25 per cent hydrochloric acid and running for 50 minutes, in experiment (12) of Table I, or to that produced in experiment (17) of Table I by 10 grm. of zine chloride and 200°™* of alcohol charged with 1:25 per cent hydrochloric acid. Finally it has been shown that by allowing malonic acid and absolute alcohol to react in the form of apparatus described, a yield of malonic ester, equal to the best described in the litera- ture, may be obtained, while by treating for a long time with alcohol char ged with hydr ochloric acid increases the yield over twenty per cent. The presence of sulphuric acid or of zine chloride and alcohol charged with hydrochloric acid permits the action to proceed in a shorter time. The best yields of malonic ester were obtained by causing the esterification to proceed as far as possible at a temperature below that at which the malonic acid decomposes. This was accomplished, as described above, by heating an alcoholic solution of malonic acid with sulphuric acid on a water bath at 50° for eight hours, and treating the residue obtained in this manner with a fresh portion of alcohol for a period of two hours. By following this procedure, a yield of 96-1 percent of malonic ester was obtained, with only 2-1 per cent lost, either through decom- position of the acid, or acid ester, or by volatilization of the ester. ; } Phelps and Eddy—Purification of Esters. 253 Art. XXVI.—Concerning the Purification of Esters; by J. K. and M. A. Portrs and E. A. Eppy. _ [Contributions from the Kent Chemical Laboratory of Yale Univ. —elxxxii. ] QUANTITATIVE studies of the ester reaction were made under the direction of one of us and published earlier in this Jour- nal.* In that work the exactness was shown with which in the preparation of the ethyl esters of succinic, malonic and benzoic acids, the crude product of esterification obtained in the special form of apparatus used, may be treated to isolate the pure ester. In brief this treatment consisted in shaking out the crude ester with ether, in the presence of an excess of sodium carbonate in solution, distilling off under diminished pressure the low boiling point products, and, finally, the ester. which was collected and weighed. This method of purifying esters is a modified form of the procedure in common use for isolating a pure ester from the crude product obtained during esterification. Otherst have separated by distillation under diminished pressure the water taken up by the ether during the shaking out of the crude ester from the sodium carbonate solution. This method of removing the water at as low a temperature as possible, so that hydrolysis may not take place, is particularly adapted for use in a quantitative study of the ester reaction. The organic acid unesterified and the mineral acid used as a catalyzer is commonly removed by treatment with sodium or potassium carbonate. Where the esters are soluble in water Fischer and Spiert{ varied the procedure by treating the erude product from esterification with an excess of pulverized potassium carbonate, and removed, after long shaking, the potassium salt by treatment with ether. The filtrate was freed from ether on a water bath and fractionated under diminished pressure. A-study is given here of the exactness with which dry potassium carbonate may be used in isolating succinic, malonic and benzoic ethyl esters impure with alcohol, water, unesteri- fied organic acid and small amounts of mineral acid. For this purpose mixtures of the purified esters were made with the substances, as shown in the table, and the ester separated and weighed. Definite portions, 75 grm. each, of carefully purified diethyl succinate, diethyl malonate, or ethyl benzoate were placed in a Claisen flask of 250°™° capacity, chilled in ice, together with 2 grm. of the corresponding acid, 2° of concentrated hydro- * This Journal, xxiii, 368; xxiv, 294; xxv, 39; xxvi, 148. +J. Am. Chem. Soc., xxiii, 1105, 1896. ¢ Berichte, xxviii, 3252. 254 Phelps and Kddy— Purification of Esters. chloric acid of commerce or 1°” of concentrated sulphuric acid of commerce, 10°™* of alcohol nearly absolute and an excess— 10 grm.—of either pure potassium carbonate of commerce or the same freshly fused and cooled before introducing into the ester mixed with impurities. These impurities are pres- ent here in larger proportion than would be expected if the ester were produced according to the procedure referred to earlier in this paper. The Claisen flask was connected for dis- tillation under diminished pressure with a second Claisen flask of 100° capacity used as a receiver. To secure a current of air through the apparatus during the entire operation an open lass tube was used in place of the usual capillary tube. The 100° Claisen was connected either directly or with a glass tube held through a rubber stopper to the larger Claisen, the side neck of which was very short, in the same manner. as has Kster meat eas ester H.SO, C2H;0H H.O cold hot Theory Found Per No. grm. grm. mols. mols. hr. hr. germ. grm. cent | A Cry 50 10 4°5 1 0 3 67°61 52°00 qiGw (2) 50 2 3) 1 + 2 67°61 63°72 94°2 (3) 50 2 9) iP 11 9) 67°61 63°30 93°6 (4) 50 aa, 3) 2 + 2 67:61. - 6Y°85 91°5 (5) 50 2 10 Z 5) 2 67°61 55°84 82°5 (6) 50 2 10 2 10 2 67°61 59°64 88°2 (7) 50 4 3) 0 4 0 67°61 64°96 96°0 (8) 50 4 3) 0 4 2 67°61 64°79 95°8 (9) 50 + 3) 1 4 2 67°61 65°61 97°0 (10) 50 4 10 2 8 2 67°61 57°22 84°6 B CL 250 + 5 1) + 0 70°80 68°21. 96°3 (26 50 + 3) 0 4 2 70°80 66°92 94°5 (3) 50 4 9) 0 0 2 70°80 64°82 91°5 wash waters were shaken out successively with two fresh por- The combined ethereal solutions of the ester tions of ether. were separated from low boiling impurities, distilled under diminished pressure in the usual manner, and the ester weighed. The malonic ester obtained was found to be pure, since it showed no considerable variation in boiling point. In experiments (8) of series A and (3) of series B of Table I, (8) of series A and (2) of series B of Table II, the filtration and process of esterification were omitted. . Enough water was 260 Phelps and Tillotson, Jr.—Cyanacetic Ester. added to dissolve the ammonium chloride and the ester was recovered directly in the manner described above. In experi- ments (4) of series A, (1) of series B of Table I, (7) of series A and (1) of series B of Table II, after saturation with hydro- chloric acid in the cold, the solution was treated directly in the esterification apparatus, the ammonium chloride bein precipitated when the temperature had reached 60°-70° dur- ing the process of ésterification. In experiment (5) of series A of Table I, after saturating in the cold with hydrochloric acid, the excess of hydrochloric acid, alcohol, water and all low-boiling products were removed by heating the flask in a water bath at 50° under diminished pressure, and the remainder was esterified in the usual manner. In this experi- ment, no considerable amount of ammonium chloride was pre- cipitated, and there remained in the distillation flask a viscous fluid which did not distil with the malonic ester and which began to decompose on being heated to a higher temperature. In experiment (1) of series A of Table II, cyanacetic ester was boiled with alcohol sp. gr. 0°825 and sulphuric acid sp. gr. 1°84 under a return condenser for two hours. and as no ammonium salt was precipitated, gaseous hydrochloric acid was passed in for three hours, the ammonium salt filtered off, and the pro- cess continued in the usual manner. The theory for malonic ester given in series A of Tables I and II was calculated on the basis of the analyses of the first sample of ester, given. above, and with the assumption that the substance was a mixture of cyanacetic and malonie esters. This assumption was apparently justified, first, by the fact that, on treatment, it yielded pure malonic ester, and second, on saponification, a mixture of cyanacetic and malonic acids was obtained. It was further assumed that all fractions pre- pared in the same way and boiling at the same temperature, 205°6°-206° (corrected), had the same composition. Experi- ments (7) and (8) of series A of Table II were made using portions of the samples analyzed. The theory given in series B of both tables was calculated on the basis of pure ethyl cyanacetate. This is presumed to be justified by the analyses given above as well as by the yields of malonic ester obtained from it. The loss sustained by saturating an alcoholic solution of malonie ester with hydrochloric acid was determined by a blank test. When 75 grm. of pure ethyl malonate, with 125°™° of absolute aleohol was saturated with hydrochloric acid for the same time and under the same conditions as in the several experiments, 72°02 grm. were recovered, with a loss of 2°98 grm. or 4:2 per cent. This loss must be added to the results given in the table in order to show the extent ‘of the conver- Phelps and Tillotson, Jr.—Cyanacetic Ester. 261 sion by this procedure. From the results given in Table [I it is plain that with small amounts of zine chloride as catalyzer, in the presence of one or two molecules of water, malonic ester may be obtained to the extent of 90 per cent of that theoreti- cally possible, as shown in experiments (1) and (2) of series A. Taking into consideration the loss inherent in the process, we have 94 per cent accounted for. The loss of 6 per cent is doubtless due to decomposition of malonic ester or acid ester in the presence of water and the large amount of hydrochloric acid. When no water was used, as in the remaining experi- ments of the table, a much larger yield was obtained. Under these circumstances the water necessary for the conversion may have been formed in the secondary reaction between the alcohol and hydrochloric acid in the presence of the catalyzer, water and ethyl chloride bemg formed. The presence of ethyl chloride was detected in the alcoholic distillates from these experiments. It is also possible that the reaction took a differ- — ent course, the imido ester hydrochloride first formed reacting with alcohol to form ammonium chloride and the ortho ester which, in the presence of the large amount of alcohol and the eatalyzers, was easily decomposed with the formation of the normal ester and ethyl ether. This same sort of a reaction has been described by Claisen* in the case of acetals, which he finds are easily broken up by catalyzers with the formation of the aldehydes and ether. It is evident from the results obtained in experiments (1) and (2) of series B that the con- version of cyanacetic ester to malonic ester has been nearly quantitative, either because sufficient water has been formed in the secondary reaction mentioned above, or a reaction simi- lar to that pointed out by Claisen has taken place, with the formation of ethyl ether, or both of these actions have gone on simultaneously. That is to say, the water theoretically demanded by the equation for the conversion of cyanacetic ester to malonic ester need not be introduced as such when the conversion is effected by hydrochloric acid im alcoholic solu- tion. Since this is true and since malonic ester is so easily hydrolyzed, the better yields obtained without the use of any water are easily understood. . Experiment (4) of series A, which was treated in the esteri- fication apparatus in order to complete the esterification of any acid ester that might have been formed, if too much water were present for any reason, shows a slight increase in yield over experiment (3) of series A, which was shaken out directly after saturation with hydrochloric acid and treatment on the return condenser. In experiment (5) of series 338°8 _ 296°24 87-4 (2) 200 165 90°-95° 338°8 292°65 86°3 carbonate and extracted the malonic ester with ether. The ether was distilled off, and the ester fractioned under diminished pressure. By recovering and esterifying the acid ester from the sodium carbonate solution, he obtained a weight of malonie ester equal to that of the chloracetic acid employed, or about 60 per cent of that theoretically possible. In an earlier paper* in this Journal, the best conditions for the esterification of malonic acid were studied. In a later papert a study of the conversion of cyanacetic ester to malonic ester was reported. In preparing material for use in those papers the synthesis from chloracetic acid was employed. It seemed worth while to study these syntheses further and apply the results of the two earlier papers to the practical prepara- tion of malonic ester. This paper reports the results of such a study. 7 For the work described in this paper, impure commercial monochloracetic acid was twice fractionally distilled at atmos- pheric pressure, using portions boiling within limits of one to one and one-half degrees. The acid obtained in this manner was found to be about 95 per cent pure as shown by the fol- lowing analyses, made according to Carius: I. 0:1808 grm. of acid gave 0°2608 grm. of silver chloride. Cl=35°64 per cent. II. 0:0939 grm. of acid gave 0:1347 grm. of silver chloride. Cl=35°47 per cent. Calculated for C,H,O,Cl, Cl=37'53 per cent. * This Journal, xxvi, 143. + Ibid, xxvi, 257, Phelps and Tillotson, Jr.—Malonic Acid or its Ester. 269 Potassium cyanide of 96-98 per cent purity was used for the experiments in series A of the table. For those of series B, pure commercial cyanide containing ammonia, and for those of series C, pure commercial potassium cyanide free from ammonia and cyanate was employed. Ananalysis of the sample used in the experiments in series C gave the following results. I. 02372 grm. of potassium cyanide gave ‘4579 grm. of silver eyanide. CN=37'50 per cent. II. 0°2682 grm. of potassium cyanide gave ‘5198 grm. of silver cyanide. CN =87-65 per cent. Calculated for KCN, CN=39°95 per cent. The aleohol made use of in the conversion to malonic ester and that employed in the process of esterification was alcohol of commerce, made as free from water as possible by repeated distillations over calcium oxide. In all the experiments in the table, except those specially treated, 200 grm. of monochloracetic acid, contained in a liter flask, was treated with about 300 germ. of pure hydrous sodium carbonate of commerce, and 50°* of water added to start the reaction. Good results were obtained, in experiments not recorded, using the anhydrous sodium carbonate dissolved in 250°" of water. The advantage in the use of the hydrous salt les in the fact that the solution of sodium chloracetate is kept cold during the process of neutralization, thus keeping the hydrolysis of the chloracetate at a minimum. In fact, under the above conditions, a temperature low enough: to freeze the mass is nearly always obtained. It is usually convenient to hurry the reaction by standing the flask in water at room tem- perature. In experiment (1) of series B, potassium carbonate was used instead of sodium carbonate, but there appeared to be no advantage in its use, since the potassium sulphate formed on acidifying with sulphuric acid was not less soluble than sodium sulphate under the conditions of experimentation. When the chloracetic acid was entirely neutralized, the solu- tion was poured into a solution of 165 grm. of potassium eyanide in 250™* of water heated to 70—-80°, and after the vigorous action had taken place, the solution was boiled for about five minutes to complete the reaction. The solution was then cooled and acidified with sulphuric acid, using the action of a drop of the liquid with logwood paper as a test of acidity. About 100™* of concentrated sulphuric acid were required for each experiment. After cooling the precipitated salt was filtered off and the water solution evaporated to dry- ness. In experiment (1) of series A, this was done at the temperature of the steam bath. In the remaining experiments 270 Phelps and Tillotson, Jr.—Malonic Acid or its Ester. of the table, it was accomplished under diminished pressure ~ by heating the flask in a water bath at 70-80°, collecting the distillate in a side-necked receiver which was kept cool by a stream of water flowing over it continuously. The salt which had been filtered off from the water solution was shaken up in a flask with 200° of 95 per cent alcohol, filtered, and the salt washed on the filter with 100™* of alcohol of the same strength. These aleoholic solutions were then added to the residue, obtained by evaporation of the water solution, warmed in a water bath and shaken till homogeneous. The insoluble salt was filtered off, shaken up again with 100—200°™ of hot 95 per cent alcohol, and again filtered. The combined alco- holie solutions contained in a liter flask were then freed from alcohol, water, and low-boiling products by heating under diminished pressure with a water bath at 60°, collecting the distillate as before. The residue, consisting chiefly of cyan- acetic acid, cyanacetic ester, and some sodium or potassium salt, was treated by the method described in a former paper* for the conversion of cyanacetic ester to malonic ester. For these experiments, 600°° of absolute alcohol with 5™ of sulphuric acid, sp. gr. 1°84, were placed with the product obtained as described above, in a two liter flask fitted with a reflux condenser.and kept cool in a mixture of ice and salt, and the solution saturated with gaseous hydrochloric acid, dried by bubbling through concentrated sulphuric acid. At the end of about twelve hours saturation was usually complete. The ice mixture was then replaced by a water bath and the solution boiled under the return condenser for two hours, passing the eurrent of hydrochloric acid through the mixture continuously. The precipitated ammonium salt was then filtered off, shaken up with 100°™ of absolute alcohol, filtered again and washed with absolute alcohol. The alcoholic solutions were then collected in a liter flask fitted for esterification, as described in a former papert from this laboratory, and 700%™* of abso- lute alcohol distilled during a period of three to four hours through the solution of ester, which was kept at a temperature of 100-110°. The resulting product was purified by treating with ice and a solution of sodium carbonate and shaking out with ether. The ether and all low-boiling products were then removed and the pure malonic ester distilled under diminished pressure, as described in the paper to which reference has been made. The ester obtained in this manner boiled within reasonable limits and was nearly pure ethyl malonate. Further evidence of its purity is furnished by the fact that, on * This Journal, xxvi, 145. + This Journal, xxiv, 194. Phelps and Tillotson, Jr.—Matonie Acid or its Ester. 271 hydrolysis of the fractionated material, pure* malonic acid was obtained. In experiments (8), (4) and (5) im series A, (1) and (2) in series B, and (2) in series O of the table, the reaction between the sodium chloracetate and potassium cyanide was made to take place at a temperature of 90°—95°, by allowing the cold solution of chloracetate to run slowly from a separatory funnel into the hot cyanide solution, an operation which took from fifteen to twenty minutes. In the other experiments the two solutions were mixed and the reaction allowed to take place vigorously, the temperature under these conditions reaching 110°. It was noticed in this reaction that if the chloracetate solution was distinctly alkaline a nearly colorless solution re- sulted, while if acid considerable color developed. The amount of color seemed to increase with the length of time of the reaction. Thus when the solutions were mixed and a vigorous action took place, less color resulted than when the reaction went at 90°—95° for fifteen minutes, but even then the color- ation was small in comparison to that formed in experiments not recorded in the table, in which the reaction was made to take place at the temperature of ice water during a period of twelve hours. The color is possibly due to dark-colored poly- meric products formed from hydrocyaniec acid. Hence it is advantageous to allow the reaction between potassium cyanide and sodium chloracetate to proceed in alkaline solution at a temperature of about 100°. But these conditions are favorable for the conversion of the cyanacetate to a malonate, which is to some extent decomposed in the hot alkaline solution as shown by Van’t Hoff,t who explained the decomposition as taking place according to the following equation: CH,(COOK), + KOH = CH,COOK + K,CO,. For this reason prolonged boiling after the reaction is over is not desirable. On acidi- fying, the slight excess of sulphuric acid produces a small amount of hydrochloric acid, which, in the process of evapo- ration of the aqueous solution of cyanacetic acid, may also cause the formation and decomposition of some malonic acid. This is shown in experiment (1) of series A, in which the acid solution was evaporated on a steam bath. When the volume of liquid became small, decomposition was apparent, through the escape of bubbles of carbon dioxide. This decomposition was minimized in all the remaining experiments, except exper- iment (4) of series A, by evaporating at a lower temperature, under diminished pressure. In order, if possible, to entirely prevent this decomposition through the hydrolyzing effect of the strong acid, the excess of sulphuric acid was removed, in * This Journal, Phelpsand Weed, xxvi, 138. + Berichte, vii, 1382, 272 Phelps and Tillotson, Jr.—Malonic Acid or its Ester. the case of experiment (4) of series A, by adding a solution of sodium acetate in excess, testing the disappearance of acidity of the solution with logwood paper. The slightly lower yield would indicate that sodium cyanacetate had been formed from the sodium acetate and the free cyanacetic acid present. Some loss always occurred during the distillation of the aleohol from the alcoholic solution of cyanacetic acid. The alcoholic distillates, averaging 500™*, from five experiments not recorded in the table but made by the procedure described, were separately distilled through a Hempel column to small volume and the remainder heated at 60°, under diminished pressure, till the manometer registered 15"™. The weighed residues were combined and esterified. From the weight of eyanacetic ester so obtained it was found that when 500% of aleohol, containing some water, were distilled at 60° under diminished pressure, from a solution containing about 180 grm. of cyanacetic acid, there also distilled 7-1 grm. of cyanacetic ester, which is equivalent to 10 grm. of malonic ester, or three per cent for each experiment. It has been shown in a former paper* that the loss of malonic ester, inherent in the process employed for the conversion of cyanacetic ester to malonic ester, was about three grams for seventy grams of malonic ester, or 4°3 per cent. If these constant losses be considered, it is evident that the results shown in the table are approxi- mately 7°3 per cent lower than the amount of malonie ester which would have been obtained had there been no loss during the process. | | From the experiments in series A, in which potassium cyanide of 96-98 per cent purity was used in excess, the results are as good as those in series B, in which pure cyanide of commerce containing ammonia was employed, and even as good as those in series C, in which cyanide free from ammo- nia and cyanate was made use of, other conditions being the same. Thus it would seem that if the potassium cyanide is in excess, impurities in the cyanide to the extent of three or four per cent do not materially affect the reaction with sodium chloracetate. In experiment (2) of series C there was much evolution of carbon dioxide when sodium carbonate was added in the pro- cess of recovery of the pure ester. This was presumably due to the excess of sulphuric acid which was added previously to lib- erate the cyanacetic acid. Too large an excess is not desirable, as it retards esterification, either by holding back water or by forming water and ethyl ether in reacting with alcohol at the temperature 100°-110°, at which esterification took place. The slight variations in yield among otherwise similar experiments may be due to variations in the excess of sulphuric acid present. * This Journal, xxvi, 260. Phelps and Tillotson, Jr.—Malonic Acid or its Ester. 273 From the work recorded here it is evident that the reac- tion between potassium cyanide and sodium chloracetate to form sodium cyanacetate proceeds best in alkaline solution. This reaction may take place vigorously at 110° or slowly at 90°-95° without materially affecting the yield of malonic ester. For the best results the alkaline solution of sodium cyan- acetate should not be evaporated to dryness at a high tem- perature or even boiled for a long time, since these conditions are favorable for the formation and decomposition of the sodium malonate formed. The aqueous solution of cyanacetic acid is best evaporated to dryness under diminished pres- sure at a temperature of 70°-80°, and the alcoholic solution at about 60°. In removing the alcohol under diminished pres- sure there is a constant loss equivalent to three per cent of malonic ester, and in the process employed for converting the eyanacetic ester to malonic ester and recovering the latter in pure form, there is an additional loss of 4°3 per cent. If these losses be considered we have, in the case of the sev- eral higher results shown in the table, from 94°3 to 95-1 per cent of the theoretical amount of malonic ester accounted for. Further, if we take into account that the chloracetic acid employed was but 95 per cent pure, as shown by anal- ysis, it seems that the results given are from 99:3-100 per cent of the theoretical amount for the chlorine content of the chloracetic acid used. That is to say, potassium cyanide and sodium chloracetate, in alkaline solution at about 100°, react to give a theoretical amount of sodium cyanacetate and potassium chloride. Or, again, the theory for malonic ester obtained from 200 grm. of monochloracetic acid of 95 per cent purity is 321-8 grm. and the yields actually obtained, leaving out of account experiments (1) and (4) of series A, average 295°18 grm. or 91°7 per cent instead of 87-1 per cent in the table, in which the theory is calculated on the basis of pure monochloracetic acid. Thus malonic ester in large quantities may be made in nearly theoretical amount from monochloracetic acid if the procedure as outlined above be followed. The essential points are, an excess of potassium cyanide, reacting in alkaline solution ; the addition of sulphuric acid to the solution of sodium cyan- acetate in only slight excess; the evaporation of the aqueous and alcoholic solutions of\cyanacetic acid at low temperature, which is most readily done under diminished pressure; the conversion to malonic ester and the purification of the same by — the procedure described above. Under these conditions, if the chloracetic acid be only 95 per cent pure, about 87 per cent of the theoretical amount of malonic ester may be obtained, or 274 Phelps and Tillotson, Jr—Matonic Acid or its Ester. figuring it on the chlorine content of the chloracetic acid, about 92 per cent, which is an improvement of about thirty per cent on the best previous methods known to us. Pure malonic acid was obtained from ester, prepared in this way, for use in an earlier paper* on the esterification of mal- onicacid. The ester was fractionally distilled under atmospherie pressure with a Hempel column. Portions boiling within two tenths of a degree were converted into malonic acid by heating equal parts of malonie ester and water with a few drops of nitric acid at a temperature of about 60° for some time after the mixture had become homogeneous. The solution was then evaporated in an open dish at a temperature of about 50° to the point of saturation, and the malonic acid which separated on cooling was recrystallized from hot water. In this manner pure malonic acid may be obtained with only slight loss of material. ; * This Journal, xxvi, 248. Phelps and Tillotson, Jr.— Cyanacetic Acid. 275 Art. XX X.—On the Preparation of Cyanacetic Acid and us Ester from Monochloracetic Acid; by I. K. Puetps and E. W. Trittotson, JR. [Contributions from the Kent Chemical Laboratory of Yale Univ.—clxxxiii.] | Ko.ze,* in synthetizing malonic acid, states that if the solu- tion obtained by acting on chloracetic acid with potassium cyanide be acidified with sulphuric acid and extracted with ether, cyanacetic acid is obtained, which on boiling with potas- sium hydroxide is converted to potassium malonate. Miullert obtained cyanacetic ester “boiling above 200°” by treating chloracetic ester in alcoholic solution with potassium cyanide. Finkelstein{ used chloracetic ester and an aqueous solution of potassium cyanide, boiling till all the ester had gone into solu- tion, evaporating to dryness, acidifying the residue with hydro- ehloric acid and extracting with ether. On evaporation of the ether cyanacetic acid was obtained. Meeves§ also acted on chloracetic ester with potassium cyanide in a water solution, evaporated, acidified with sulphuric acid, and extracted with ether. The acid obtained on evaporation of the ether was purified by treating with lead carbonate, filtering off the excess of carbonate and insoluble lead malonate and decomposing the solution of lead cyanacetate with hydrogen sulphide ; but as in the case of the other investigators mentioned above, no results were given by him showing the amount of cyanacetic acid obtainable from a given weight of chloracetic acid. Grim- aux and Tcherneak| caused sodium chloracetate and potassium eyanide to react, extracted the cyanacetic acid with ether according to Meeves’ procedure and obtained about 75 per cent of the theoretical amount of cyanacetic acid. Van’t Hoff] states that. he obtained a nearly theoretical yield from chloracetic acid, but gave no definite information as to how it was obtained. Fiquet** followed in general the same procedure as Grimaux and Tcherneak, acidifying with hydrochloric acid instead of sulphuric acid, and obtained 70 per cent of that theo- retically possible. Noyes, ++ after boiling chloracetic ester and potassium cyanide with methyl alcohol, obtained 50 per cent of the theoretical yield of eyanacetie ester, boiling within limits of ten degrees. In a former paper{t{ it has been shown from the amount of malonic ester obtained that sodium chloracetate and potassium * Ann., cxxxi, 348. +Ann., cxxxi, 350. tAnn., cxxxiii, 338. § Ann., cxliii, 201. || Bull. Soc. Chim., xxxi, 338. “| Berichte, vii, 1382. ** Jour. Am. Chem. Soc., xxvi, 1545. 4 Ann, Chim. [6], xxix, 439. tt This Journal, xxvi, 267. 276 Phelps and Tillotson, Jr.—Cyanacetic Acid. Cyanacetic ester —_ __ Chlorace- Potassium Temperature = —— tic acid cyanide of reaction Theory Found He No. grm. grm. erm. erm. Per cent. A | (1) 200 165 90°-95° 239°4 219°78 91°4 (2) 200 Gop =< 1 LOe 239°4, 224°15 93°6 (3) 200 165 110° 239°4 222-00 92°7 (4) 200 165 110° 239°4 220°00 91°9 B (1) 200 165 0°-5° 239°4 213°30 89°1 (2) 200 165 90°-95° 239°4 2NSe7 2 91°4 (3) 200 165 90°-95° 239°4 220°04 91°9 4) 9200 165 110° 239'4 207°68 86°8 (5) 200 165 110° 239°4 206°63 85°1 eyanide react in alkaline solution at 100° to form the theoreti- cal amount of sodium cyanacetate. The work recorded in this paper is a study of the formation of ethyl cyanacetate from monochloracetic acid, making use of the best conditions for the esterification of cyanacetic acid as described in an earlier paper® in this Journal. | For the work described in this paper the impure monochlor- acetic acid of commerce was fractionally distilled twice at atmospheric pressure, using portions boiling within limits of one and one-half degrees. The acid so obtained was found, by analysis given in a former paper,ft to be about 95 per cent pure. Commercial potassium cyanide of 96-98 per cent purity was used in all the experiments. The alcohol employed in esteri- fication of the cyanacetic acid was made as free as possible from water by repeated distillations from calcium oxide. In all the experiments except those in which the treatment was special, 200 grm. of monochloracetic acid of the purity described above, was treated in a liter flask with about 300 grm. of hydrous sodium carbonate and 50° of water to start the reaction. The temperature of the mass usually became so low that the mixture was frozen. The action may be hastened by immersing the flask in water at room temperature. Good results were also obtained by using about 110 grm. of anhy- drous sodium carbonate and 250™* of water, but since the reaction with the hydrous carbonate takes place with cooling, there is less tendency towards hydrolysis of the chloracetate. The alkaline solution of sodium chloracetate was then poured into a hot solution of 165 grm. of potassium cyanide in 250° of water and after the action had taken place the solution was boiled for five minutes to complete the reaction. To the cooled solution was then added sulphuric acid in slight excess, * This Journal, xxvi, 264. + This Journal, xxvi, 268. Phelps and Tillotson, Jr.— Cyanacetic Acid. 277 using a drop of the solution with logwood paper as a test of acidity. The precipitated salt was then filtered off and the water distilled from the cyanacetic acid and other non-volatile material under diminished pressure by heating the flask in a water bath at 70°-80° and collecting the distillate in a side- necked flask as a receiver, which was kept cool by allowing a stream of water to flow over it continuously. The salt which remained on the filter was washed with 300™* of 96 per cent alcohol, and the alcoholic solution added to the cyanacetic acid residue remaining after the aqueous solution had been distilled under diminished pressure. The mixture was well shaken, filtered, and the salt shaken up with 100°™* of alcohol of 96 per cent strength, filtered and washed with alcohol. The com- bined alcoholic solutions were then evaporated under dimin- ished pressure, heating the flask in a water bath at 50°-60° and collecting the distillate as before, continuing distillation till the manometer showed 15™™ pressure, to insure removal of all the water. To the residue, which consisted chiefly of cyan- acetic acid and its ester, was added 100°™ of absolute alcohol and 5° of sulphuric acid, sp. gr. 1°84. Then the mass was esterified at 100°-110° for 2°5 to 3 hours with 500°™* of absolute alcohol, using the special arrangement of flasks for esterifica- tion described in a former paper* from this laboratory. The crude product was then purified by treating with ice and a solution of sodium carbonate and shaking out with ether. The ethereal solution. was then freed from low-boiling pro- ducts and distilled under diminished pressure in the usual manner. In experiment (1) of series B, the potassium cyanide and sodium chloracetate were made to react in the cold, the flask standing in ice water for twelve hours. In experiments (1) of series A, and (2) and (3) of series B, the cold alkaline solution of chloracetate, contained in a separating funnel, was slowly run into the hot solution of potassium cyanide. The time employed was fifteen to twenty minutes and the temperature was kept at 90°-95° by the heat of the reaction. In the other experiments of the table the two solutions were mixed immedi- ately and the action allowed to take place vigorously, the temperature being usually about 110°. It has been shown in the former paper to which reference has been made, that this reaction proceeds more smoothly in an alkaline solution at a temperature of about 110°, but these conditions are unfavorable for the production of pure cyanacetic acid or ester, since cyanacetates are converted to malonates in hot alkaline solution. For this reason prolonged boiling after the reaction is over is not desirable. On acidifying, the slight excess of sulphuric * This Journal, xxiv, 194, 278 Phelps and Tillotson, Jr.—Oyanacetie Acid. acid acting on the chlorides present produces a small amount of hydrochloric acid, which, in the process of evaporation, also tends to the formation of malonic acid. In series A there was a slight excess of sulphuric acid in each experiment. The ester obtained, after fractioning under diminished pressure, appeared to contain malonic ester as shown by the following analyses: I. 0°5060 grm. of ester gave 51°5°™* of moist nitrogen at 21° and 762°6"™, ( N = 11°39! per'cent | | II. 0°5070 grm. of ester gave 51:2°™* of moist nitrogen at 22° and M65 2B OS Nas 152 percent. | Calculated for C,H,O,N, N = 12°40 per cent. In the experiments of series B, after adding a slight excess of sulphurie acid, as indicated by logwood paper, enough of a saturated solution of sodium acetate was added to destroy the excess of sulphuric acid, logwood paper again being the indi- cator. ‘The ester obtained in this manner, after being fractioned under diminished pressure, gave the following analyses: I. 0°5020 grm. of ester gave 55:2°™ of moist nitrogen at 23° and 762-30m oN 2747 per cent. II. 0°5004 grm. of ester gave 55°0°™ of moist nitrogen at 20° and 150; 2 aN) = 2no 4 per cent.) Calculated for C,H,O,N, N = 12°40 per cent. It is plain therefore that if pure cyanacetic ester is to be obtained, the water solution of cyanacetic acid must not be evaporated in the presence of mineral acids, but in the presence of acetic acid no considerable decomposition takes place. The operation is most conveniently performed under diminished pressure, since the low temperature necessary to remove the water under diminished pressure is unfavorable for the forma- tion of malonic acid. Some loss always occurred during the distillation of the alcoholic solution of cyanacetic acid. This loss has been shown in the former paper,* on the preparation of malonic ester, to be 71 grm. of cyanacetic ester or about three per cent. It has also been shown in a former paper,ft that the loss herent in the process used for recovery of pure ester from the crude product obtained in the esterification flask was 1:67 grm. for 70 grm. of pure ester or 2:4 per cent. Assuming that this percentage error holds for large amounts, we have approxi- mately 5-4 per cent of cyanacetic ester lost during the pro- cedure. In an earlier papert it was shown from the yield of malonic ester obtained that if the purity of the chloracetic acid used and the loss of malonic ester inherent in the process be taken into *This Journal, xxvi, 272. + This Journal, xxvi, 265. + This Journal, xxvi, 273. Phelps and Tillotson, Jr.—Cyanacetic Acid. 279 account, the reaction between sodium chloracetate and potas- sium cyanide was practically quantitative. It is therefore possible to calculate from the amount of malonic ester obtained from the chloracetic acid used, the amount of cyanacetic ester which might be expected from 200 grm. of the same quality of ehloracetic acid. A series of seven experiments similarly carried out gave an average of 871 per cent of malonic ester, with a loss inherent in the process of converting the cyanacetic to the malonic ester of 4:2 per cent. This gives 309-32 grm. of malonic ester, which is equivalent to 218°57 grm. of cyanacetie ester from 200 grm. of chloracetic acid. The loss sustained by the process employed for the recovery of the ester from the crude product obtained in the esterification flask, has been shown in a former paper®* to be 1°67 grm. for 70 grm. of cyanacetic ester, and is presumably 5 grams for the 218°57 grm. under consideration in these experiments. Thus if the ester formed is pure ethyl cyanacetate, we should not expect to obtain more than 213°57 grm., and higher yields may be explained by the conversion of some of the cyanacetic ester into malonic ester. This explanation is supported by the analyses, given above, of the ester obtained in experiment (2) of series A, and appears justified since in the experiments of series A there was an excess of sulphuric acid while the aqueous solutions of cyan- acetic acid were being evaporated to dryness. The higher yields obtained in experiments (2) and (8) of series B are ascribed to the formation of sodium malonate from sodium eyanacetate in the hot alkaline solution. In these experiments the reaction took place slowly for a period of fifteen to twenty minutes, which allowed more hydrolysis to take place than in experiments (4) and (5) of series B, in which the action was over in five minutes or less. The lower results in experiments (4) and (5) of series B are explained on the assumption that the cyanacetic acid, reacting with the excess of sodium acetate, forms an equilibrium mixture and, since the sodium cyanacetate is not dissolved by the alcohol, a loss occurs. The results of the work here described are in agreement with those of a former paper, which show that potassium cvanide and sodium chloracetate react quantitatively in alkaline solution at 110°. They also show that cyanacetic ester of high purity may be prepared in large amounts, and in good yield, from mono- chloracetic acid if precautions be taken to minimize the trans- formation to malonic acid. To this end the reaction between sodium chloracetate and potassium cyanide should take place quickly, the alialine solution of sodium cyanacetate should not be boiled for a long time, and the solution of cyanacetic acid, containing an excess of mineral acid, must not be evaporated * This Journal, xxvi, 264. 230 Phelps and Tillotson, Jr.— Cyanacetic Acid. to dryness at a high temperature. However, it may be quickly and conveniently evaporated by heating with a water bath at 70°-80° under diminished pressure, in which case the cyan- acetic ester formed is of 90-95 per cent purity. Experiments (2), (8) and (4) of series A were performed under such condi- tions. If pure cyanacetic ester is to be obtained, the excess of mineral acid must be removed by adding sodium acetate. In this case the yield is not so good, but the ester, after fractioning under diminished pressure, is pure ethyl cyanacetate. This ‘procedure was followed in experiments (4) and (5) of series B. Pure cyanacetic acid was obtained from the ester prepared in this way for use in an earlier paper* on the esterification of cyanacetic acid. Two parts of water with one of ester and a few drops of nitric acid were heated at a temperature of about 60° for some time after the mixture had become homogeneous. The solution was then evaporated in an open dish, at a tem- perature of 50°—60°, to the point of saturation. The cyanacetic acid which separated on cooling was recrystallized from a mix- ture of ether and chloroform. In this manner pure cyanacetic acid, melting at 66°1°-66°4° (corrected), was obtained in quantity and in good yield. * This Journal, xxvi, 264. Phelps and ELddy—fydrobromie Acid. 281 Art. XXXI.— Researches on the Influence of Catalytic Agents in Ester Formation. Hydrobromic Acid and Zine Bromide in the Formation of Ethyl Benzoate ; by I. K. and M. A. Puetrs and EK. A. Eppy. [Contributions from the Kent Chemical Laboratory of Yale Univ.—clxxxvii. | Gotpscamipt® has measured the rate of esterification of ben- zoic acid with ethyl alcohol, using hydrochloric and hydrobro- mic acids as catalyzers. The results in the use of these two catalytic agents for a temperature of 25° are so nearly the same that he concludes that hydrochloric and hydrobromic acids under the conditions of his experiments have the same efficiency as catalytic agents in ester formation. Goldschmidtt and many others, from similar physico-chemical measurements given in recent literature, have concluded that the efficiency of a given catalyzer depends upon the concentration of the cata- lyzer and upon the degree of dissociation of the catalyzer in alcoholic solution. The amount of ester formed from benzoic acid in presence of sulphuric acid, or hydrochloric acid, or hydrochloric acid and zine chloride as catalyzers, with different amounts of ethyl alcohol acting for different lengths of time, has been shown in a former papert{ in this Journal. The esterification of the ben- zoic acid was made in flasks specially arranged as illustrated in an earlier paper§ in this Journal, where the action of zinc chloride and hydrochloric acid, as catalyzers, was shown in the formation of ethyl succinate. ‘The results given in the paper concerning the esterification of benzoic acid show that under the conditions of the experiments made, increasing the concen- tration of the catalyzers up to a certain limit increases the amount of ester produced with a given amount of alcohol acting in a given time on a given amount of benzoic acid. Further increase in. the amount of catalyzers beyond this hmiting con- centration caused a decided falling off in the amount of ester produced. And, finally, the results referred to show that the yields of ester produced bear no relationship to the degree of lonization of the catalyzer. For example, hydrochloric acid dissociates to a greater extent than sulphuric acid, but with hydrochloric acid as a catalytic agent yields beyond 90-4 per cent were not obtained even when a mass of 50 grm. of ben- zoic acid was treated with 400° of alcohol containing 25 per eent of hydrochloric acid during eight and a half hours, inter- polating a fractionation under diminished pressure to remove low-boiling products, especially whatever water formed during * Berichte, xxviii, 3218. allot 0 le.o.. 40. ear Ip t This Journal, xxv, 39. § This Journal, xxiv, 194. Am. Jour. Sci.—FourtH Series, Vout. XXVI, No, 153.—SrpremsBer, 1908. 282 Phelps and Hddy—Hydrobromic Acid. esterification that had not been removed under the conditions of the experiment, when half of the alcoholic mixture had acted upon the acid. It is, however, a striking fact that two grams of sulphuric acid with half the amount of alcohol acting about one-third of the same time with the same amount of ben- zoic acid gives theoretical yields of ester. In the work recorded here the catalytic action of zine bro- mide and hydrobromic acid at different temperatures on ben- zoic acid with ethyl alcohol is brought into comparison with the similar action of zine chloride and hydrochloric acid recorded in the paper to which reference has been made, as well as in this aper. , ? : For this work ethyl alcohol was prepared as free from water as possible by repeated distillations with calcium oxide. The benzoic acid used was the pure benzoic acid of commerce, For the preparation of pure, dry hydrobromie acid the pure bro- mine of commerce was freed from chlorine by long standing, with frequent shaking, in contact with an aqueous solution of potassium bromide,* before distilling off the bromine. The hydrobromic acid gas was prepared by allowing the purified bromine to act on a mixture of red phosphorus and water, the gas set free being purified by passing first through layers of moist red phosphorus and glass wool contained in one leg of a U-tube, the outer leg of which, to dry the gas completely, con- tained layers of phosphorus pentoxide and glass wool. The hydrobromie acid thus prepared was dissolved in chilled alcohol in the concentrations given in the tables. These concentra- tions were chosen of such values that the hydrobromie acid in these experiments was in molecular proportion to the hydro- chloric acid in the experiments to which reference has been made. Purezine bromine was prepared for use in two different ways. The pure zine bromide of commerce was fused in an atmosphere of pure hydrobromic acid before granulating the melted zine bromide. Zine bromide was also prepared by heat- ing in a flask connected to a return condenser the pure zinc of commerce at a temperature above the melting point of zine with bromide purified, as described above, for the preparation of hydrobromic acid, and then dried with sulphuric acid before distillmg. The pure zine bromide made in this way was melted, as in the case of the commercial sample, in an atmosphere of dry hydrobromic acid, heated to expel any excess of hydrobro- mic acid, and then granulated. The amounts of zine bromide used corresponded molecularly to the zine chloride used in the esterification of benzoic acid in the work to which reference has been made. In all the experiments given in this paper the procedure was * Richards and Wells, Proc. Amer. Acad., xli, 440. Phelps and EKddy—Hydrobromic Acid. 283 similar to that given in the former paper in this Journal on the esterification of benzoic acid to which reference has been made. In brief the treatment consisted in heating at a definite tem- perature a given weight of benzoic acid in the presence of a small amount of alcohol with a certain amount of hydrobromic or hydrochloric acid, or with an alcoholic solution of hydro- bromie or hydrochloric acid, and zine bromide or chloride in definite amount. Into this mixture a known amount of alcohol and hydrobromic or hydrochloric acid in definite amount was driven at a uniform rate in vapor condition. The vapor issuing from the flask in which the esterification took place was frac- tioned by passing through a Hempel column attached to an ordinary condenser. It has been shown in the earlier paper that benzoic ester is retained completely in the flask in which the benzoic acid is esterified. The ester was isolated in pure condition by shaking out the mixture of the crude product with ether, treating with sodium carbonate, and distilling under diminished pressure in the manner described in the earlier work on benzoic ester. It appears from an inspection of the results recorded in Table I that hydrobromic acid as a catalyzer varies in its efficiency according to its concentration, and according to the temper- ature at which esterification takes place. This is seen in com- paring experiments (1) with (10) and (1) with (5), or (10) with (12) and (13). It is also clear that in comparing (8) with (9) that the rate of flow of a given amount of alcoholic mixture is of influence. A certain amount flowing rapidly is able to esterify 10 per cent less benzoic acid than half that amount of alcohol even in shorter time, as is seen by comparing (4) with TABLE I, Ben- Alcohol Time of Benzoic ester Zoic with HBr Tempera- action — Se Se = No. acid ——*~—-—, ture ——-*~—— Theory Found grm. -cm®* per ct. hrs. min. grm. grm. per ct. ees 200 2772 785 90" 2° 10>. Gl-48 41-18 66-98 (2), 50 ZOO TT 90° 1 55 61°48 39°45 64°17 ap. .90- 7200" 62°77, 100° 2 35 61°48 2930 47°66 fe) 30 oS 200-277 100 110" 1 AS 61°48. 27-00. 43°92 (5) ~"50 200g OO 0c ooo 61:48 © 26-40 42:94 faye 00) 200 2-717 2125 “130, 5 A 55° 61°48) 17°70. 28-79 (7) 50 200. 227 0e 125 T4053. 29 5 61°48 15°58 25°34 (8) 50 400 2°77 100° 2 5 61°48 20°31 33°03 eo). 50 400 2°77 100° 4 25 61°48 35°96 58°49 Pees 0 20027-74285. 908 2. 6148 56:05 9117 (11) 50 200 27°74 90°-100° 3 50 61°48 50°96 82°89 (12) 50 200 27°74 SO N0Omee 2. Gl-48- 575 8ae17 (13) 50 200 27°74 100° See O48, AI-05 2 O67, 284 Phelps and EKddy—Hydrobromice Acid. (8). In comparing (8) and (9), lengthening the time of action of the alcoholic mixtures, other conditions remaining the same, increases the amount of ethyl benzoate formed. In the results in Table (11), where zine bromide is used as. eatalyzer with the hydrobromic acid, no difference was found in the amount of benzoic acid esterified by the two samples of zine bromide. Itis seen from a comparison of experiments (6) with (10) and (12) with (20) that under definite conditions the greater the concentration of hydrobromic acid with a given amount of zinc bromide present, the greater the amount of benzoic acid esterified. When these experiments are compared with (1) and (11) of Table I it at once becomes evident that zine bromide as well as hydrobromic acid has here a catalytic effect. The-greater the amount of either catalyzer present . the greater the yield under conditions otherwise closely sim- ilar, as is seen in comparing (4), (14) and (24) together, also (2) with (12). The largest amounts of zinc bromide used seem to retard esterification, as appears when (2) and (12) are com- TaBLeE IT. Ben- Aleohol Time of Benzoic ester ZOIC with HBr Tempera- action ——-H—— —_——— No. acid ZnBrz ——--——\ ture —-—— Theory Found grm. grm. cm? per ct. hrs. min. grm. STM) sper Cie ( 1) 50 1:7 200 2:77 °90°-100° 1 85 61:48 (84-73) Siam ¢ 2) HO Wen BOO Berry 90° 1 50 61°48 36°34 59°19 (3). 50 1°7 * 200 2°77 100°-150" 2°10 61:48) 25a aimee ( 4) 50. 1:77 900° 2°77 125°-150> 9°15) 61-48. 2G eee ( 5) ES Olas Teen AO DASH Tf 100° 2 25 61°48 4245 69°04 ( 6) DOM me 20Ors Dei 5 90° 3 380 61°48 44:06 71°67 (7) HO Ne SOO Bey 100° ‘ 3 © 61°48 49°76) 76093 Sin OO Wet SOO. Peart 100°-110° 38 50 61:48 45:05 7aa7 DON Aa A005 22h 7 100° 2° 8 6148 48 ioeanoaaS (10) 50 1st 200 27-74 85° — 90-35. 6 UAe bbe onsen Ga) RO AO Bey LOO 1 30 61°48 46°36 75°40 2) 4 DOG ZOO nm 2rd 100° 2. 10 61:48) AT AG aaa cB AO mae 300: 2°77. -90°=100". 4 2 2 61:48. 57-06 acm (14) 50 17° 400 2°77 125°-1380° 1 50.6148 41:46 67-44 (15) 50 17°, A400 2:77) 125°-1357 2 30 16148.) 41¢0a aioe (16) DORs. NOON 2 aa, 100° 2 35. 61:48) 5 705g Gai ao0) lie ADO B20 4 100° 4 10 3:61:48, -59: 74 oan (18) Omeleie 400 2°77 125°-130° 4 45 61:48 | 45-100 7aeans (19) 50,17: 200 27°74 90°-100° 1 50 61:48 59°08).9609 Ai” BO 7 200 27°74 100° 2 5 61°48 . 58° 3G794.06 ek 5 Omalaie 900 27:74 °85°—-902 3.7.2) 16148) 158°8a aca (22) 50. 42°5 200 2°77. -85°— 902.2". 15 {61:48 SI Geet (23), 5044 a | 200L M2547 100°-110"7 2° (15 61:48 41 9Saaioeeze (24) 50.42°5 200 2°77 120°-150272, 22" 61:48) 240 ine = = j a = Phelps and Eddy— Hydrobromic Acid. 285 pared with (23), but even with the largest amount, as in (23) of Table II, the yield of ester is greater than when no zine bromide was used under conditions otherwise similar as in (5) and (6) of Tablel. The temperature exerts a decided influence here, again, in the presence of different amounts ef zine bro- mide, as is seen in experiments (2), (3) and (5) when compared with each other, and the same is seen when experiment (15) is compared with (16), also when (17)-1s compared with (18). The same effect of temperature is not seen when zine bromide, as second catalyzer, is present in such large amounts as found in experiments (19), (20) and (21) when compared with each other, also in (22), (23) and (24) when they are compared with each other. The rate of flow of the aleohol seems to have a similar influence in presence of the second catalyzer that it had in the presence of hydrobromie acid alone. Experiments (2) and (6) show that, as do also (16) and (17). In Table III are given results which show the action of hydrochloric acid either alone as catalyzer, or with zine chlor- ide, as a second catalyzer in esterifying at different tempera- tures benzoic acid. It is clear from an inspection of results. that raising the temperature retards the esterification of benzoic acid by means of hydrochloric acid and zine chloride, as it does the esterification of benzoic acid by means of hydrobromic acid and zinc bromide. In the work for the former paper the influ- ence of temperatures as high as 125° was studied. No marked effects at this temperature appeared. Attention was called there to the difference in esterification produced by varying the amount of the eatalyzer, the amount of alcohol, and the time of action. TABLE III. Ben- Aliecohol Time of Benzoie ester zoic with HCl Tempera- action —_—s —_—_— No. acid ZnCl, ——+~—— ture —-—— Theory Found grm. grm. cm? perct. hrs. min. grm. grm. per ct, (1) 50 __ 200 1°25 100°-110° 1 40 61°48 87°86 61°74 (2) ieee 251-25 19s 150° 25 30 61-48 24°45 40-09 (3) 50 40 200 1-25 100°-110° 2 10 61-48 58°33 94°88 (4) 50 i0 200 1:25 125°-150° 2 __ 61°48 50°56 82°24 In all the experiments of Table III, in (4) and (24) of Table Ii and in (1) of Table I, the alcoholic distillate was collected in four portions and the amount of mineral acid was determined in each portion by titrating the diluted distillates with stand- ardized sodium hydroxide solution in the presence of phe- nolphthalein as an indicator. The amount of mineral acid left in the flask from which the alcohol, charged with hydrochloric or hydrobromic acid, was distilled, was similarly 286 Phelps and Hddy— Hydrobromic Acid. estimated. The amount of mineral acid left in the esterifi- eation flask and neutralized with sodium carbonate in the process of recovery of the crude ester in: the experiments (1) and (2) of Table III and (1) of Table I was also estimated. This was done by acidifying the sodium carbonate wash water with nitric acid, the precipitated benzoic acid was filtered off and washed with cold water and the halogen acid in the filtrate was determined gravimetrically as the silver salt. In experi- iment (1) of Table I the total amount of hydrobromie acid in the residues was found to be 1°34 grm., of which 0:07 orm. was in the alcoholic distillate. In this experiment 4:4 orm, of hydrobromic acid were used, leaving 3°06 grm. of hydro-_ bromic acid, which pr esumably formed ethyl bromide. In experiment (4) of Table II the amount of hydrobromie acid found in the alcoholic distillate was 0°0102 grm. In experi- ment (24) of Table II the amount of hydrobromic acid found in the alcoholic distillate was 0°0048 grm. In diluting the first portion of the alcoholic distillate ethyl: bromide was found present in such amount as to make the resulting liquid turbid. These experiments show first that at a temperature 85°-90° about seventy-five per cent of the hydrobromie acid has reacted to form ethyl bromide; second, that the hydrobromic acid which remains as such accumulates in the esterification flask ; and, third, that zinc bromide has a eatalytie effect upon the action of ethyl aleohol and hydrobromic acid, as might have been anticipated. Similarly, in experiment (1) of Table III the total residues of hydrochloric acid were found and amounted to 1°84 grm., of which 0°110 grm. were found in the alcoholic distillate and 1-455 erm. in the esterification flask, leaving 0°185 erm. from a total of 2-025 grm. of hydrochloric acid taken with the aleo- hol, which reacted presumably to form ethyl chloride. In exper- iment (2) of Table III the total hydrochloric acid found in all residues was 1°485 grm., of which 0°615 grm. was in the alco- holic distillate and 0-595 germ. was in the esterification flask, leaving 0°54 grm. from a total of 2-025 orm. taken in the aleo- holic mixture. In experiment (8) of Table III the hydro- chloric acid found in the alcoholic distillate was 0°0019 grm., and in (4) of the same table 0-0017 grm. was found in the alcoholic distillate. | The evidence from these experiments proves that, at a tem- perature of 100°-110°, about ten per cent of the hydrochloric acid present has reacted to form ethyl chloride. At the higher temperature of 125°-150°, a larger amount, about twenty-five per cent, is used in this way. With hydrobromic acid of simi- - lar concentration, this sort of action, at the lower temperature of 85°-90°, goes on to a much larger extent. Thus, in experi- Phelps and Eddy— Hydrobromic Acid. 287 ment (1) of Table I seventy-five per cent of the hydrobromic acid has reacted in this way. Further, the efficiency in ester formation of hydrobromie acid with or without zinc bromide was shown in Tables [and II to be dependent in large measure upon the temperature. The difference in the amount of ethyl chloride formed at the different temperatures, the large amount of ethyl bromide formed under similar conditions of tempera- ture with hydrobromic acid, and the variability in the effi- ciency of zine bromide at different temperatures, would indicate that the failure of zine bromide to act as an efficient catalyzer at higher temperature is due to the almost complete action of hydrobromie acid on alcohol at that temperature to form ethyl bromide and water. The same difference in the effect of temperature is evident with hydrochloric acid, as is shown by experiments (1) and (2) of Table III, but in these cases the action to form ethy] chloride and water makes itself markedly evident only at temperatures that are most unsuited for esteri- fication. And it seems fair to assume that it is this fact that makes hydrochloric acid more advantageous than hydrobromic for use as a catalyzer in esterification. Presumably it is this difference in catalytic action at differ- ent temperatures of hydrobromie acid that explains the differ- ence in the results given by Goldschmidt and those recorded here. Goldschmidt, measuring the rate of esterification at 25°, found hydrochloric and hydrobromie acids equally efficient as catalyzers. In our experiments at higher temperatures they are never equally efficient. In dilute solutions the hydro- chloric acid as catalyzer is much more efticient, while with the highest concentrations the hydrobromie acid is more efficient if the esterification is carried on at a temperature of 85°—90°. At 100°-110° the hydrochloric acid is mueh more efficient, and markedly more so than the hydrobromic acid at a temper- ature of 125°-150°. While the presence of zine chloride or zinc bromine as the second catalyzer increases the amount of ester formed, it yet remains true, as shown by the results given, that raising the temperature above the point where alcohol will just distil from the esterification flask will decrease the amount of benzoic ester produced in the ease of either zine chloride or zine bromide, except where the zine bromide is present in amount almost equal to the weight of benzoic acid taken. Zine chloride in presence of hydrochloric acid is, however, a more efficient catalytic agent, both the hydrobromic acid and the zinc bromide being especially sensitive to any rise in tem- perature. The amount of ethyl benzoate produced in experiments (1) and (3) of Table III is in the opposite ratio to the quantity of the ethyl chloride formed. Hence, the statement found in the 288 Phelps and Kddy—Hydrobromic Acid. literature that the efficiency of hydrochloric acid as a catalyzer is dependent upon the formation of nascent ethyl chloride would not seem to be borne out by these experiments. It is to be noted, on the other hand, that zine chloride, known to be one of the best catalyzers* in the formation of ethyl chloride from alcohol and hydrochloric acid, helps the esterification of benzoic acid. The extent to which hydro- chloric acid has reacted to form ethyl chloride, either with or without zine chloride, is seen on comparing (1) and (8) of Table ILI. The acid found in the esterification flasks in (1) and (2) of Table III shows a tendency to the accumulation of mineral acid in that, flask. This accumulation is greater, naturally, at the lower temperatures. A similar tendency was observed at the temperature of 85°—90° in experiment (1) of Table I with hydrobromic acid. Since in (1) and (2) of Table III the same amount of alcoholic hydrochloric acid was employed, the temperature being the only marked difference, the results are directly comparable. It would seem possible that the accumulation of hydrochloric acid at the lower tem- perature might explain the greater yield of ester obtained. That is to say, the catalytic effect of hydrochloric acid to give ideal yields of ester depends upon a certain concentration of acid or of positive hydrogen ions, other conditions remaining the same. It would seem from these experiments that neither the concentration of the hydrochloric acid nor the concentra- tion of the positive hydrogen ions can be the determining factor in esterification in the presence of zine chloride, since it has been shown that aleoholic hydrochloric acid gives a much higher yield of ester in the presence of zine chloride than when acting under precisely the same conditions without zinc chloride. The well known catalytic effect of zine chloride on ethyl alcohol and hydrochloric acid to form ethyl chloride must have taken place and thus have diminished the concen- tration of the hydrochloric acid. Not only this, but in this action of aleohol and hydrochloric acid water has been pro- duced, which would also be a hindrance to esterification. Yet the fact remains that in the presence of the lower concentra- tion of hydrochloric acid, even with: the water formed. by alcohol and hydrochloric acid in the presence of zine chloride, the zine chloride is a most efficient catalytic agent in esterifica- tion. Another proof that the amount of ester produced is not pro- portional to the concentration of the positive hydrogen ions present has been mentioned earlier in this paper, when it was noted that, under otherwise similar conditions, a small amount of sulphuric acid, which ionizes to a smaller extent than hydro- * Groves, Ann., clxxiv, 372. Phelps and Eddy—Hydrobromie Acid. 289 chloric acid, which might be taken even in large amount, is a more efficient catalyzer in the production of ethyl benzoate. It has been noted in previous papers on esterification and in this paper also, that an increase up to a certain limit in the amount of catalyzer present under given conditions has been helpful in every case studied. The increase in the amount of eatalyzer beyond that limit is decidedly harmful, judging from the amount of ester formed. This would appear to be due to the fact that the catalyzers used, hydrochlorie acid, zine — chloride, hydrobromic acid, zinc bromide, and sulphuric acid, have a strong aftinity for water. It is also a fact, that each of the catalytic agents used reacts with alcohol of themselves to produce water, ethyl chloride being formed with hydrochloric acid, ethyl bromide with ‘hydrobromic acid, and, ethyl ether with sulphuric acid. Consequently when they are present in large amounts together with a high-boiling point ester, abso- Inte alcohol is not able to effect dehydration as completely as is necessary for complete esterification. _ Hence it would seem that to esterify a given organic acid under most advantageous conditions for complete esterification it is necessary to determine experimentally the proper propor- tions of alcohol, the time of action, the most efficient catalyzers, and the most suitable temper ature at which reagents and cata- lyzers interact. It has been shown that concentration of mineral acids as catalyzers does not control the esterification at the temperatures studied, and, further, that the best conditions of temperature for the formation either of ethyl chloride or of ethyl bromide are not ideal ones for esterification. When further experimental data upon this question are at hand it is hoped that the real function of catalyzers in esterification will be made known. 290 Phelps, Palmer and Smillie—Ester Formation. Arr. XXXII.—Researches on the Influence of Catalytic. Agentsin Ester Formation. The Lffect of Certain Sulphates on Benzoice and Succinic Acids; by I. K. Puetps, H. E. Patmer, and R. Sinus. [Contributions from the Kent Chemical Laboratory of Yale Univ.—clxxxviii.] In a former work* published in this Journal it has been shown that almost theoretical yields of succinic ethyl ester may be obtained by passing alcoholic vapor charged with dry hydro- chloric acid into the flask containing the succinic acid. Ina somewhat later papert in this series of researches on esters, under the direction of one of us, it has been shown that theoret- ical yields of benzoic ethyl ester may be obtained from benzoic acid, using sulphuric acid as a catalyzer, while 1f no catalyzer be present, only a trace of the ester is produced under con- ditions otherwise precisely similar. Bogojawlensky? has stud- ied the effect of various morganic sulphates on the esterification of a number of organic acids, using as the indication of esteri- fication the amount of ester produced by heating the mixture of acid, alcohol, and catalyzer on a return condenser. In the case of benzoic acid he obtained a yield of 92 per cent with a mixture of 80 grams of copper sulphate and 1 gram of sul- phuric acid, while with sulphuric acid alone he obtained a yield of only 65 per cent, and with copper sulphate alone no ester whatever was formed. In the work to be described the effect of various acid sul- phates on the esterification of succinic and benzoic acids has been studied. Pure succinic acid was prepared by heating on a return condenser succinic ester, which boiled within 0°2°, with an equal volume of water and a few drops of nitri¢ acid, and recrystallizing from water the pure succinic acid formed, as has been described in a previons paper§ in this Journal. In the case of the benzoic acid, the pure acid of commerce was ‘used. The apparatus was the same as that illustrated and de- seribed in an earlier paper| in this Journal on the esterification of succinic acid. In all of the work absolute alcohol which had been made as free from water as possible by repeated dis- tillations over lime was used. In all cases 40°™* of the abso- lute aleohol were put into the second flask together with the benzoic or succinic acid and catalyzer, and the remainder of the aleohol as vapor was run into the second flask from the first dur- ing the intervals of time indicated in the tables, the temper- ature of the mixture in the second flask being kept by means of an acid sulphate bath at 100°-110° during the action. * This Journal, xxiii, 368. + This Journal, xxv, 39. t Berichte, xxxviii, 3344. § This Journal, xxiii, 211. | This Journal, xxiv, 194. Phelps, Palmer and Smallie—Ester Formation. 291 The catalyzers studied were the acid sulphates of potassium, - ammonium, sodium, pyridine, and aniline. These were made, in the case of the inorganic salts, by heating the anhydrous neutral sulphates with the proper proportions of sulphuric acid in a porcelain crucible until the mass fused together, then grinding in an agate mortar. The organic sulphates were made by adding sulphuric acid in the proper proportions to the pyridine or aniline in the esterification flask in the case of the benzoic acid; in the case of the succinic acid, however, they were mixed before being put into the flask. The sodium sul- phate, the potassium sulphate, the pyridine, and the aniline were the pure anhydrous material of commerce. Pure ammo- nium sulphate was prepared by treating in water solution with an excess of sodium hydroxide the ammonium salt precipitated by the action of hydrochloric acid on cyanacetic ester under the conditions shown in an earlier paper* in this Journal and catching the ammonia evolved on distillation in dilute sul- phuric acid. The salt obtained by evaporating the solution, neutral to litmus, was recrystallized and dried. The proportions of catalyzers used were chosen such that the amount of the sulphuric acid used to form the acid sulphates here was the same amount, or a multiple or submultiple of the amount, used as catalytic agent in the former papert on the esterification of benzoic acid; that is to say, the concentrations of the hydrogen ions present during the esterification in the two researches were in molecular ratio. Of the catalyzers studied, the acid pyridine sulphate was the only one which seemed to go entirely into solution during the esterification. Of the others, the acid ammonium sulphate was perhaps the most soluble, but it did not go entirely into solu- tion, even in the case of the smallest amount which was used. The succinic ester in experiments (1) to (7) inclusive of Table I was recovered, according to previous work,t{ by treating the crude ester in the esterification flask with an excess of solid potassium: carbonate, and heating the flask fitted up for a yacuum distillation with a 100° Claisen flask as receiver, under a pressure. of 15™™, to 100°-110° on an acid potassium sulphate bath until no more carbon dioxide was evolved. The ester was then distilled under the same pressure—15"™—, allowing a stream of cold water to strike the receiver continu- ously during the distillation. The distilled product was then redistilled, the lower boiling impurities being first removed by raising the temperature of the flask to 60° under a pressure of 15™™", before the succinic ester was distilled and weighed. The succinic ester in the remaining experiments of Table I and the * This Journal, xxvi, 258. + This Journal, xxv, 39. t This Journal, xxvi, 253. 292 Phelps, Palmer and Smillie—EKster Formation. benzoic ester in all of the experiments of Table II, except where aniline acid sulphate was the catalyzer, were recovered - by shaking out the ethereal solution of the impure ester from the esterification flask with a solution of sodium carbonate in a separating funnel containing a few pieces of ice and washing with a saturated solution of sodium chloride. The sodium ear- bonate and the sodium chloride wash waters were each extracted separately twice with fresh portions of ether to recover any portions of the ester that may have been carried along with them, and these ethereal solutions were added to the main mass of ester. The ether was distilled off on a water bath, after which the flask containing the ester was fitted for a vacuum dis- tiation. The lower boiling impurities were removed by rais- ing the temperature to 60° under a pressure of 15™™; the ester was then distilled by heating to 140°-150° on an acid potassium sulphate bath under the same pressure—15™"—, the receiver being cooled by allowing a stream of cold water to strike it continuously during the distillation. In the experiments where aniline acid sulphate was used, since the aniline formed on neutralization with sodium carbonate would distil over along with the benzoic ester, the aniline sulphate in experiments (27), (80), (381), and (82) of Table IL was removed by frst shaking up the ethereal solution of the crude ester with water in a separating funnel, and the recovery was then carried out in the usual manner. In experiments (28) and (29) of Table II all the material which would distil over below 150° under 15™™ TABLE I, Suc- Abso- Reaction Succinic ester cinic lute time — -A~- —~ acid Catalyzer alcohol —-+-— Theory Found Per No. grm. grm. em?’ hr. min. grt. | erm ewcend 1. 30 HSs0, 0°5 2001 25. 73% (Oia aes Zo 40s oe 0°5 200 <1. 30° 73°T = 7000S oce 3. 90 Bs 0°5 A400 1° 15° °438°7 — 7245 Bere AS 2150 és 0°5 400 2 Or T3°T 20 eae 5. 50 bf 1:0 200.5 450. F380 G2 oa0 3 1:0 200 1 0. 73897 10-72 Sone ik iho) KHSO, 0°694 200 1 10 738°7 33°40 45°3 8. 50 Fe 9-777 200.1 30. 72:7 56702 ee 9, 50 ce 5°5p4. 200 = .55. 73°7 3654s 10. | 50 (NH,)HSO, 0°587.° 200. 1. 10° 73:7. 39 arene Lin a0 a 2:348 200 1° 15 “732i b9:SoRee lee 1 eAaie 9 @) rs 4°696. 200. 1 45 73°27 sGo285mmoure lon 50 NaHSO, 4:899 900% I 15” 737 69 2aoeae 14. 50 C,H,N.H,SO, 0°907° 200. — 50. 73:7 > 44:35) 36052 15,0250 ee 0'907 200 1 On 187 428s wear 1 O13 1 ea O eae ae 16. 50 CHINH,.H,S0, 0°975. 200 Phelps, Palmer and Smillie—Ester Formation. 293 pressure was collected first and afterward purified in the usual manner by treatment with ether. From an inspection of the results of Table Tit is evident that theoretical yields of succinic ester are obtained with sul- phuric acid as a catalyzer, taking into consideration the loss of ester inherent in the process of recovery. In the case of acid potassium sulphate a greater yield is obtained by an increase, within limits, in the amount of catalyzer. This is shown by comparing (7) with (8). But if the catalyzer is present in larger amount, there tends to be a falling off in the yield, as is shown by comparing (8) with (9). Acid ammonium sulphate accelerates the esterification much more than acid potassinm sulphate. An increase in the amount of acid ammonium sul- hate increases the yield of ester. As is shown in experiment (13), acid sodium sulphate under the conditions of this experi- ment gives almost as good results as sulphuric acid alone. In Table IJ are given the resuits with benzoic acid. It will be seen that acid potassium sulphate does not accelerate the esterification to any great extent, but that an increase in the quantity of the acid sulphate present up to a certain extent in- creases the yield, as appears in experiments (8) and (4); how- ever, the presence of a still larger amount of the acid sulphate, as in experiments (5), (6), (7), and (8), seems to hinder the esterification; the ester ification indeed seems to depend on conditions not yet completely understood. In the case of acid ammonium sulphate, it is evident that the yields increase both with the time of reaction and with the concentration of the catalyzer. From a comparison of (10), (11), and (14), in which the time of reaction was approximately the same, it is seen that the yields increase with the concentration of the catalyzer ; and from a comparison of (9) with (10), (11) with (12), and (13) with (14), in cach of which the amount of the catalyzer present was the same, it is evident that the yields of ester increase with the time of reaction. With acid sodium sul- phate larger yields are obtained than with corresponding quan- tities of acid ammonium sulphate, but similarly the yields are increased as the reaction time is greater, as shown by comparing (15) with (16), (17) with (18), and (19) with (20), and also as the amount of catalyzer present is greater, as shown by comparing (15), (17) and (19), and further (16), (18) and (20) with each other. It is evident that neither the pyridine nor the aniline acid sulphates accelerate the esterification of benzoic acid to any great extent. In comparing these results with the results which were obtained with sulphuric acid alone as a catalyzer in the former paper* in this Journal, it is evident that none of the acid sul- * This Journal, xxv, 39. 294. Phelps, Palmer and Smillie—Ester Formation. TABLE II. Ben- Abso- Reaction Benzoic ester Zoic lute time — —~- — acid Catalyzer alcohol —-+-—, Theory Found Per No. grm. orm. 4 em? shes mini (erin. grm. cent Lo 0 KHSO, 388g e200 a O 61°48 2. 4°4 2. o0 re 1388 200 2 45 61°48 9°64 Tosa somo on 2777 §200°°2 15) “6148 We AG 4, 50 S 2777 200: 3 152) 61438 VO ais 5. 950 ES 5°5654 200 1 40 61°48 11°05 18:0 6. 50 ay 51554 200 1 45 61°48 aie 9°4 (AP LIOY es 5°554 200 2 20 61°48 762) hoes 8. 50 fe 5004 »200...2) 380 61:48 107027 iG Sho) 210) (NH,)HSO, Mel j4 200° 2 10361248 0°48 9°0 10. 50 ee VATs 200.3, , 152 261e48 8°69 14°] ll. 50 oom 2°349 200 2 50 61°48 36°43 59°3 ee WO) os 2°349 200 3750 61:°48> 4130336722 oy a0) “ 4°696. 200°..1°55 61°48 43°83 aiies 14. 50 oy 4°696 200 3 0 61°48 48°56 79:0 15. 50 NaHso, 1°225° 200 <1 °°30° 61:48 —20:995 34a TOs SOW ie 22 de 20 Ogre, O 61°48 40°43 65°8 lifio OO) eg 2°450 200 1 30 61°48 46°00 74°8 18.750 eb 2450 200 3 O 61°48 58:69. 95:5 Legh = a0) rs 4°899 200 1 15 61°48 56°44 91°8 20. 50 tp 4°899 200 3 0 61°48 60°58 98°5 21. 50 CHL.N.H{SO, .0:907 200 3 10) 6148 ““Saleaiiae 22 a 0) eS 0-907 200 4 O 61°48 5°00 8°] 23. 50 ce 1814 200 1 35 61°48 1°46 2°4 24. 50 ef 1°814 200 2 0 61°48 Paee 7) 4°5 25. 50 oe 3°628 200 2 20 61°48 3°86 6°3 26. 50 oc 3°628 200 3 10 «61°48 3°63 5°9 Qe 50 C,H,NH,.H,SO, 0975 200 1 50 61°48 1°39 2°3 28. 90 0:'975 200 2 40 61°48 ‘56 0°9 One) ef 1:950 200 1 15 61°48 3 0°6 30. 50 - 1:950. 200 .3 .20. 61:48 {2:33 imaes 31. 50 $f 3290052200 el love oles 2°60 4°2 Sy AEG) iy 3°900 200 3 10 61°48 4 36 rl phates studied, when present in amounts of equal concentration of acid hydrogen, are as efficient in catalytic effect as sulphuric acid. . Acid sodium sulphate is nearly catalytically equivalent ; acid ammonium sulphate gives less effect ; acid potassium sul- phate considerably less; while acid pyridine and aniline sul- phates are very poor catalyzers. The acid pyridine sulphate gives with succinic acid, which esterifies readily, distinctly more effect than acid aniline sulphate, but with benzoic acid, where esterification is more difficult, they give about the same effects. Aside from the possibility of the acid sulphate in solution being the active catalytic agent, two explanations would seem obvious for the facts. First, that the catalytic effect is due in Phelps, Palmer and Smillic—Ester Formation. 295 these experiments to a dissociation of the sodium and potas- sium acid sulphates into neutral sulphates and sulphuric acid, and of the nitrogen-containing sulphates into free base and sulphuric acid, this sulphuric acid so formed being the active catalytic agent. This explanation, however, fails in the case of the weakly basic aniline and pyridine salts. Second, since in most cases the salts used as catalyzers did not go into solu- tion, and an increase in the amount of the salt used produced noticeable effects, it seems possible that the salts not in solution are active as contact agents. All that may be done, however, is to record our results until future experimentation, to be recorded in this series of papers on catalysis, will make clear the correct explanation of the facts under consideration. 296 Phelps and Hddy—Ester Formation. Arr. XXXIII.—fesearches on the Influence of Catulytic Agents in Ester Formation. The Esterification of Benzoic Acid with Certain Chlorides ; by I. K. and M. A. Pretps and E. A. Eppy. | [Contributions from the Kent Chemical Laboratory of Yale Univ.—clxxxix. | In this Journal* in earlier work under the direction of one of us, the efficiency of zine chloride with hydrochloric acid as a catalytic agent in the formation of the ethyl esters of benzoic, succinic, malonic, and cyanacetic acids has been shown. The esterification was carried on at a temperature of 100°-110° in an apparatus arranged for the purpose. In the work given in this paper, in the specially arranged flasks illustrated in the work on succinic ester referred to, the catalytic action of certain chlorides in presence of hydrochloric acid in small amount in the esterification of benzoic acid with ethyl alcohol is brought into comparison with the similar action of zine chloride with hydrochloric or of hydrochlorie acid alone as catalytic agent. The esterification of this acid with- out a catalytic agent is shown to be almost none at all. TABLE I. Ben- Abs. Ale. Benzoic ester ZOiC with HCl Time — re No. acid Catalyzer ——+— —— Theory Found erm. grm. cm? \per ct... hr.min. grms-) 2rme apemecns (1). a arenes 300) 222.0538 o. "6148 o ena ae ( 2) 50 i eiaae 200°°1°25 9 (2 2 6148 AA ees ( 3) 50 eee 200° 1:25 -3 30 - 61:48 “OOS 8e asees ( 4) 50 By aN 200. 1:25 4 LL 61°48 ol oes (.5)°50 ZnCl, 0°50 200 1°25" 2... 6148" a8 26 eams ( 6) 50 vi 1°00 - 200. 1:25 2 2. 61:48" S286 8 sso80 (27) 00 ES WOO BOO PAB 74 -. 61:48 59:23 soitre ( 8) 50 2 100) * 3007 4:25:13 .-° 61°48, 6072795 Sako ( 9) 50 i NOLO O00 eel oan _. 61°48: 60°79 49389 (10) 50 “10°00. ‘400. 1:25 4°. 22 6148) Gia Sao (11) Ome Nel 07861022005 aloo ad .. €1°48. 38°14) 62:0 (12) 50 ae 8°60 200 Nea a2 -. 61°48 42°00 68°3 (13) 50. ICI 1°09. 900 41°25 (2 lo 6148 ae es (14) 50 2 OSes OOo Shai 2 2 61:48 37-235 606 (15)'50 - InCl 0°62. 200° 1°25 ° 2 °° 20. 61:48 - 42 A eee (16) 50 ef 62207 200] Ae 2antreZ 226 1:48 (Se O aioli (17) 50 IEE C079" 200) 125.82 .. 61°48 36°00 58°6 (18) 50 ae LS 0 200721 22 _. 61°48 32:95 -53°6 (19). 50. Cu@l, © 1:00) 1200: 1:25) 422) 20 61-46) soo gam 87°7 (20) 50 or 9:87" 200 1°25 °-35> 20°" 61:48) 59:62am one (CuCl. : BS. (21) 505.05 2 58 200 125 2 .. 61:48 59°05 961 * This Journal, xxiv, 194; xxv, 39; xxvi, 143; xxvi, 264. Phelps and Kddy—Ester Formation. 297 The pure benzoic acid of commerce was used in all the work recorded here. The aleohol was made as anhydrous as possi- ble by repeated distillations over caleium oxide. A known amount of hydrochloric acid gas, dried by passing through concentrated sulphuric acid, was passed into a given: weight of alcohol in the cold before diluting to definite concentration. The pure zine chloride of commerce was made anhydrous by heating it to the melting point while a current of dry hydro- chloric acid passed, then expelling the excess of hydrochloric acid by further heating before granulating. Potassium chlor- ide was made pure by fusing potassium chlorate purified by recrystallization from hot water. Pure ammonium chloride was obtained by reerystallizing the product formed on treating with sodium hydroxide a water solution of ammonium salt precipitated in the conversion of cyanacetic ester to malonic ester in the procedure described in a former paper® in this Jour- nal, and collecting in pure hydrochloric acid the ammonia gas . TABLE II. Ben- Abs. Alc. Benzoic ester ZOIC Catalyzer with HCl NG ee =a No. acid HA —~—w Theory Found grm. grm. em? per ct. hr. min. grm., grm. per cent fee) CaCl 1-00 200 1.25 °3 15-61-48 50°30 81-8 (2) 50 et 14200 5s Oe. 61-48) 14564: 23:9 eaeoO ee stCly 116 200 9125 22 2105. 6148. 32°63. 5371 ( 4) 50 peel 60570 Oma loin 6229 2 6148 632-3457 59:6 [e205 Bal tos, 200 125 .2. 15 61-48 38°70 63:0 ( 6) 50 Soe 5302200) 41:25 2.3.10 61-48 43:26. 70:4 Meese oO) 1-005 2008 1225.38.15. 61-48 41:11. 67-9 ( 8) 50 ee OS 00 leo 2 OA eae ol 7 0n 84 ul ( 9) 50 ep OsS0 e200 nl 25s 2 he GLAS 60-65, 98:7 ies to Cl 2346, 200 1:25 2. 15 61°48 43°65.° 71-0 Gib). 50 oe ee tod 200) Neto 2 61-48 46°39 55 Ch) (12) 50 at Ie OO ned De eG le 30:57. 497 2 (13) 50 PO 2005 25 0 Fe GIA 2190". 34-5 (14) 50) to f14-60 200 #0502 61-48 40:00. 65°! 2 eg s0) onl, 709677200) 1-25 13. 20: 61748... 54:91 . 89:3 (16) 50 OO OOO eon Oe 61-48) 58-86) 95-7 (17) 50 Se CO 200m alan Be a G48 60:00 - 97-8 GS) 50 EbCl, 2:03) 200.125 92), £10. ..61:48. 38:70 62:9 (19) 50 “203i 200K: Ob eee 6148. 36°58 59'S (20) 50. SbCl, 110 200 31:95 3° 15-61-48 51:98 84-6 (21) 50 — SALOU e200 ieee) 61748 58°80. -95°6 (72)50 Bill 154 200 1:25 1 26-61-48 36°93 60°'1 (23) 50 eo 00 yee ee G48 51-35 83°5 (24) 50 1540 200-125. 2 20 61°48 59:48 96°8 * This Journal, xxvi, 148. Am. Jour. Scr.—Fourts Serius, Vou. XXVI, No. 153.—Srpremper, 1908. 298 Phelps and Hddy—Ester' Formation. set free. ‘The commercially pure chlorides of copper, barium, and strontium were made anhydrous by drying: in an air- bath at 100°. Hydrous aluminium* chloride was prepared by precipitating the commercial salt in water solution with hydro- | chloric acid gas and drying the product in a desiccator. Lead chloride was made by reerystallizing and drying in a desiccator the product obtained by precipitating pure lead nitrate with hydrochloric acid. In case of all the other chlorides, the preparation of which is not given, the commercially pure material was used. The various chlorides were used in such molecular ratio that the chlorine should be present in the same amount as in the experiments with zine chloride, which have been published previously. The procedure is the same as has been given in the earlier work to which reference has been made. In the experiments alcohol, with hydrochloric acid in the concentrations indicated in the tables, was driven over as vapor from a 500™* round- bottomed flask into the mixture of benzoic acid and alcoholic hydrochloric acid containmg the additional catalyzer in a second 500°™ round-bottomed flask carrying a modified Hem- pel column through which vapors passed to a condenser. The temperature of the second flask was kept between 100° and 110°. In most cases the crude ester from the esterification was recovered by extraction with ether in the manner outlned in earlier work on catalysis in ester formation. Where on account of a large amount of such catalyzers as bismuth, anti- mony, or tin chloride, an ether extraction would be impracti- cable, the mass of ester with the low- boiling products was distilled from the esterification flask to a 100@* Claisen flask before neutralizing with potassium carbonate and recovering by the proceduret+ published by us earlier in this Journal. It is evident from an inspection of the results given in the tables that, as has been seen in all the previous work on esterification, the amount of ester formed, other conditions remaining the same, varies with the kind and amount of the catalyzers present, with the time of Cen, and with the quan- tity of alcohol. Certain of the catalyzers appeared under the conditions of experimentation to dissolve completely in the alcoholic mixture. This was observed in the use of the chlor- ides of zine, lithium and tin. In the cases of the chorides of copper, calcium and mer cury in the higher condition the smaller amounts only seemed to go “into solution completely. Alumin- jum chloride in the smaller amount used went into solution at first but later in the experiment was precipitated out. The * Gooch and Havens, this Journal, ii, 416. + This Journal, xxvi, 208. Phelps and Eddy—Ester Formation. 299 bismuth and antimony chlorides with the hydrochlorie acid in the concentration taken did not give a clear solution in any case, as is also true of all the other chlorides employed as cata- lyzers. In interpreting the results given in the tables some slight account should be taken of the fact of the insolubility of the salts used, especially where the amount is large. This is obvious when it is considered that a homogenous mixture could not be maintained by the agitation of the liquid as caused by the bubbling of the alcoholic mixture through the mass in the esterification flask. The action of the catalyzers can be seen by comparing with each other the experiments given in Tables I and II. Evi- dently zine chloride present in the larger amount as a second catalyzer with the small per cent of hy ydrochlorie acid causes the esterification of benzoic acid with ethyl alcohol in largest amounts although copper or tin chloride present in molecular ratio for chlorine content are almost equally good. The pres- ence of more than two grams of water of crystallization with ‘the copper chloride in experiments (21) of Table I would appear to have produced no noticeable reduction in the amount of ester produced by the anhydrous salt taken in experiment (20) of Table I, where the yield is only 0-9 per cent better, although the time of action is decidedly longer. The chlorides of bismuth, antimony, and mercury in the higher condition of oxidation seem to be about equally efficient as catalytic agents. They are nearly as effective in their action as either zine, copper, or tin chloride. Calcium chloride present in the smaller amount seems to be without effect, as is seen when experiment (3) of Table I is compared with experiment (3) of Table Il. In larger amount it hinders esterification to a marked degree. Strontium chloride in the amounts taken seems to hinder esterification but not to such an extent as did the larger amount of calcium chloride. Barium chloride seems to hinder esterification slightly. The chlorides of lead, mer- cury in the lower condition of oxidation, manganese, and potas- sium at least do not assist esterification if their action is not entirely without effect. It was noticed that from the mercur- ous chloride small amounts of mercury distilled into the con- denser, indicating decomposition of the mercurous chloride under the conditions of esterification. Sodium chloride appears to hinder esterification slightly, lithium chloride hinders more, ammonium chloride still more, and aluminium chloride more than any of the chlorides studied here except calcium when present in the larger amount. It is worthy of note that each chloride, and, moreover, dif- ferent amounts of the same chloride, tend to show an individual and characteristic effect as a catalytic agent with a small amount 300 Phelps and Kddy—Kster Formation. of hydrochloric acid in the formation of ethyl benzoic ester from benzoic acid and ethyl alcohol. However, it seems to be true that certain of the chlorides may be gr ouped according: to their behavior as catalytic agents here. Although from the theoretical considerations some of the catalytic effects shown in this paper might have been predicted, such as the accelerating action of antimony and bismuth chlorides and the lack of action of potassium chloride, certain of the other effects could not, so far as we are aware, have been predicted. The action of cupric and mercuric chlorides in producing large amounts of ester or of lithium chloride in hindering esterification could not nave been predicted. Further, Claisen* in the study of catalytic effects in the formation of acetals from aldehydes and ketones, found that ammonium chloride was an efficient catalytic agent and that the alkali chlorides were without action. It becomes evident then that catalytic effects are not only specific and individual for different chlorides and for different amounts of this same chloride, but, also, each catalytic agent gives a characteristic effect, either positive, negative, or neutral, in each specitie kind of chemical change. All of the results given here were obtained under conditions of temperature ranging from 100 to 110°. What differences may be fornd at other temperatures will be determined and given later. It is obviously too early in the study of catalysis to complete our imperfect theory for such effects as are shown in the contribution to the study of catalysis given in this paper. * Berichte, xl, 3908. Relicf Map of the United States We have just prepared a new relief map of the United States, 48 x 32 inches in size, made of a special composition which is hard and durable, and at the same time light. The map is described in detail in circular No. 77, which will be sent on request. Price, $16.00. WARD’S NATURAL SCIENCE ESTABLISHMENT, 76-104 College Ave., ROCHESTER, N. Y. Waro’s Natura Science EsTaBLlisHMENT A Supply-House for Scientific Material. Founded 1862. Incorporated 1890. DEPARTMENTS: Geology, including Phenomenal and Physiographice. Mineralogy, including also Rocks, Meteorites, ete. Palaeontology. Archaeology and Ethnology. Invertebrates, including Biology, Conchology, etc. Zoology, including Osteology and Taxidermy. Human Anatomy, including Craniology, Odontology, ete. Models, Plaster Casts and Wall-Charts in all departments. Circulars in any department free on request; address Ward’s Natural Science Establishment, 76-104 College Ave., Rochester, New York, U.S. A. CON TEN T's:. Page Art. XXI.—Retardation of “Alpha Rays” by Metal Foils, and its Variation with the Speed of the Alpha Particles; by. S.2hay nore tes see) feos XXIL.—Notes on the Lower Paleozoic Rocks of Central New MexicosroyAVe dT. Lem. 5 ee 180 XXIII.—Kaersutite from Linosa and Greenland; by H. S$. WasHINGTON; with Optical Studies by F. EK. Wricur_.. 187 XXIV.—Geology of the Isthmus of Panama; by E. Howe 212 SCIENTIFIC INTELLIGENCE. Geology—Geology of the Adirondack Magnetic Iron Ores, D. H. NEwnanp: Geologishe Prinzipienfragen, E. Reyer: Die Entstehung der Kontinente, der Vulkane und Gebirge, P. O. KOHLER, 238.—Geological Survey_of Canada, A. P. Low: Geography and Geology of a Portion of Southwestern Wyoming, A. C. Veatcu, 239.—Hinfihrung in die Paléontologie, G. STEIN- MANN: Niagara Stromatoporoids: Occurrence of Hobocystis in Ontario, 240. Miscellaneous Scientific Intelligence—Publications of the Japanese Harth- guake Investigation Committee, 240.—The Physical Basis of Civilization, T. W. Hetneman: General Physics, H. Crew, 241.—Die Insektenfamilie der Phasmiden, K. B. v. WATTENWYL und J. REDTENBACHER, 242, SUPPLEMENT. Page Art. XX V.—On the Esterification of Malonie Acid; by I. K- Paetrs and EW. TitLorson, Jn. =. So eee 243 XXVI.—Concerning the Purification of Esters; by I. K. and: M. A. -Prmies and. HA. Hppy 22 22s 2 ee 253 XX VIT.—On the Conversion of Cyanacetic Ester to Malonic Kster ; by IL. K. Pnrnrs and E. W. Tirvotson, Jr. --. 257 XX VIII.—Researches on the Influence of Catalytic Agents in Kster Formation. On the Esterification of Cyana- cetic Acid ; by I. K. Pustrs and KE. W. Tittotson, Jr. 264 X XIX.—On the Preparation of Malonic Acid or its Ester from Monochloracetic Acid; by I. K. Puertrs and E. AV = HELLEOTSION dR: 2.2 oes ee oe 22k er XX X.—On the Preparation of Cyanacetic ‘Acid and its Ester from Monochloracetic Acid; by I. K. Poenps and E. Ws TIELOTSON, JB oo a ee ee 275 XX XIJI.—Researches on the Influence of Catalytic Agents in Ester Formation. Hydrobromic Acid and Zinc Bromide in the Formation of Ethyl Benzoate; by I. K. and M. A. Parips-and HE. A. Wppy i. Ve ee ee 281 XX XII.—Researches on the Influence of Catalytic Agents in Ester Formation. The Effect of Certain Sulphates on Benzoic and Succinic Acids; by I. K. Puutpes, H. E. Patmer and, Re SMILLIE. 6. oe 290 XX XIII.—Researches on the Influence of Catalytic Agents in Ester Formation. The Esterification of Benzoic Acid with Certain Chlorides; by I. K. and M. A. Puetps and i. Aa Hapa nee i a ese ee - iy Librarian U. S. Nat. Museum. — oT - VOU XX VI. OCTOBER, 1908. Established by BENJAMIN SILLIMAN in 1818. THE AMERICAN JOURNAL OF SCIENCE, Epitorn: EDWARD S. DANA. ASSOCIATE EDITORS Proressorss GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or CamsBrince, Proressors ADDISON E. VERRILL, HORACE L. WELLS, L. V. PIRSSON anp H. E. GREGORY, or New Haven, Proressor GEORGE F. BARKER, or PuinapELpun, Proressor HENRY S. WILLIAMS, or ItwHaca, Proressor JOSEPH S. AMES, or Battimorz, Mr. J. S. DILLER, oF Wasuinerton. FOURTH SERIES VOL. XXVI—[WHOLE NUMBER, CLXXVI_] No. 154—OCTOBER. 1908. NEW HAVEN, CONNECTICUT. 1908 THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET, Published monthly. Six dollars per year, in advance. $6.40 Aoxountries in the Postal Union ; $6.25 to Canada. Remittances should be made eithér by»money orders, _registered letters, or bank checks (preferably on New York banks). f= RARE CINNABAR CRYSTALS FROM CHINA. The first consignment of these rare specimens were sold on sight. We have now received a second consignment of all there are, as the mine is now filled with water. They come from Wanshanchang (Hamlet of Ten Thou- sand Hills), Tungyen Prefecture, Prov. of Kweichow, China. They occur in ordinary and interpenetrating twins of bright ruby red color. In size the crystals are from + to ? inches. The matrix is a pure white quartz, the crystals always occurring in cavities, with quartz crystals; prices run from $10, $15, $20, $25, $385, $50, $75, $100, $125 and $150. NEW ARRIVALS IN RARE MINERALS AND NEW-FINDS. | Anatase, with Hisenrose, Binnenthal; Altaite, N. Mex.; Atacamite, S. Aus. ; Albite with brookite and oetahedrite, Canton Wallis, Switz. ; Angle- site, Monte Poni; Argentine, San Miguel Co., Colo., and Freiberg, Saxony : Axinite, France; Apatite, Saxony; Benitoite, San Benito Co.,, Cal. ; Beryls, pink, Mesa Grande; Bindheimite, S. Dakota; Breithauptite, Andre- asberg: Bismuth, native, Cobalt, Ont. ; Cronstedtite, Pribram; Cassiter- ite, large crystals, Morbillan, France; Californite, new find, Tulare Co., Cal. ; Cabrerite, Laurium, Greece; Cerussite, Broken Hill; Crocoite, Berezow, and Tasmania ; Calcite enclos. copper, Calumet, Mich. ; Dycrasite, Germany; Domeykite, Coquimbo; Diopside with essonite, Piedmont; Diamonds in matrix, New Vaal River mine and Old Kimberley Mine ; Epi- dote, Prince of Wales and Alaska; Embolite, Broken Hill; Emerald in pyrite, Bogota, S. A.;. Gold, Transylvania, octahedron xls in quartz ; leaves in matrix from Verespatak, also from Siebenbergen, etc.; Herderite, Auburn, Me. ; Ilvaite, Elba; Jordanite in dolomite, Binnenthal ; Josephine- ite, Oregon; Libethenite, Hung.; Nagyagite, Nagyag; Neptunite, San Benito Co., Cal.; Niccolite xls., Saxony and Germany ; Opal, Barcoo River ; Pyrosmalite, Sweden; Parisite, Montana; Pseudobrookite, Hung.; Pent- landite, Ontario; Pyromorphite, Nassau, Germany; Pericline, Binnenthal ; Polianite, with pyrolusite, Brazil; Phenacite, Norway; Pyrargyrite xls., Hartz; Polybasite, Hung. ; Ruby in matrix, Burma; Ruby spinel in matrix, Burma; Strontianite, Hamm, Germany ; Sphene, Binnenthal ; Smithsonite, Kelley, N. Mex.; Sartorite, Binn.; Titanite on adularia, St. Gothard, Schwarzenstein and Tilly Foster ; Torbernite, Cornwall; Tetrahedrite, Cobalt, Ontario and Bingham, Utah; Tourmaline, pink and green, in feldspar with quartz; Uranothallite with uraninite, Bohemia; Uranite, Bohemia; Vanadinite, Kelley, N. Mex.: Vivianite xls., Leadville, Colo. ; Vesuvianite, Italy; Witherite, Northumberland; Wulfenite, Organ Mts., N. Mex. ; Zaratite, Lancaster Co., Pa. REMARKABLE CUT GEMS. Green garnets, aquamarines, zircons, sapphires, star sapphires, star rubies, chrysoberyl cats-eyes, opals, topaz, spinel, pink beryls, sphene, tourmaline, amethyst, andaluzite, star quartz, peridote, reconstructed sapphires and rubies, and other precious and semi-precious stones. Write to-day for our 10-page Bulletin. A.H, PETEREM, 81—83 Fulton Street, New York City. AU sto; AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES. ] Art. XXXIV.—Buried Channels Beneath the Hudson and ats Tributaries ; by J. F. Kemp. Tur Hudson river has afforded to previous observers prob- lems of more than ordinary physiographic interest. The dis- sected peneplain, which the Highlands about West Point present to those who look abroad from any of the neighboring summits, is one of the best exhibitions of this land-form, easily accessible to routes of travel. Although the Highlands appear from the surface of the river to be a range of mountains, from the summits themselves the group becomes an incised plateau forming part of the Schooley peneplain, first identified in New Jersey ‘to the southwest by W. M. Davis and J. W. Wood, Jr.* This peneplain probably marked the closing of a eycle of drainage fairly coincident with the Cretaceous period. An ‘uplift subsequently revived the streams and the Tertiary cycle began. The traces of the latter are still visible along the Hud- son in rocky terraces, which stand out with marked conformity as a series of shelves, best shown in the western bank and especially prominent in the cold season. A walk from Fish- kill to Peekskill on a crisp winter’s day, when the foliage no longer masks the relief, will serve to bring out many points not visible in the months of leaves. Having excavated the broader valley outlined by these shelves, the river was obviously again revived and eroded, within the older limits, the narrower channel of whose details we are just now gaining possession. They piece out in part a missing or fragmentary chapter in its history, the one which | relates to the time antedating the invasion of the continental glacier. As will be shown, they corroborate the previously * The Geographic Development of Northern New Jersey, Proc. Boston Soe. Nat. Hist., xxiv, 565, 1890. Also W. M. Davis, The Catskill Delta in the post-Glacial Hudson Estuary, idem., xxv, 318, 1892. Am. Jour. Sci.—FourtH Series, VoL: XX VI, No. 154.—Ocrossr, 1908. Pd ow 302 J. fF. Kemp—Buried Channels Beneath inferred elevation of the land which we had been led to assume from the general phenomena of ice accumulation and from the specific characters of the submarine valley of the Hudson, whose recognition even before 1863 by the late Professor J. D. Dana marks one of the many acute observations and infer- ences regarding the local geology which we owe to his tireless ACUIYIt ce eikat ofessor Dana, however, had but imperfect data and consequently an inadequate idea of the depths involved.* Mr. A. Lindenkohl of the U. 8. Coast and Geodetic Survey, and with more extended soundings, took up the question anew in 1885 and 1891. A canyon was demonstrated in the continental shelf which about 50 miles off Sandy Hook was 2400 ft. below the neighboring sea-bottom, there found at a depth of 420 ft. Beyond this point and along the course of the submerged channel, soundings of much less depth were met, and for some years ‘the inference was drawn that a bar, of inexplicable character and apparently too far out to be a ter- minal moraine, crossed the mouth of the canyon and filled it up. Subsequently more numerous soundings proved this apparent bar to be due to asharp southerly bend in the canyon, whose course had hitherto been southeast to east by south, and that it extended with increasing depth to the edge of the con- tinental shelf. These latter features have been especially brought out and emphasized by J. W. Spencer in a valuable series of papers discussing off-shore phenomena in the sea-bottom, and best summarized in this connection in the reference given below,t in which will also be found a review of earlier work. Dr. Spencer demonstrates the existence of the canyon down to 9000 ft. below the surface. This aspect of the subject will not be pursued further in this paper, the object being merely to remind a reader that these conditions exist off the mouth of - the Hudson and that they have an interesting connection with the phenomena of its land-channel. ‘The most obvious sugges- tion in explanation is the elevation of the land, yet the amount of elevation required is a bit staggering. We are reminded of the alternative view, not without its advocates, that the land may have remained stable while the ocean drew off to the southern hemisphere and by lowering the sea-level established equivalent drainage relations.§ * See the Manual of Geology, Ist ed., 1865, p. 441, where the depth at about 80 miles from Sandy Hook is given as only 720 ft. + Geology of the Sea-Bottom in the approaches to New York Bay, this Journal, xxix, 475, 1885. Notes on the Submarine Channel of the Hudson River, and other evidences of Post-Glacial Subsidence of the Middle Atlantic Coast Region, this Journal, xli, 489, 1891. {+The Submarine Great Canyon of the Hudson River, this Journal, Jan- uary, 1905, 1-15. Se Wi "Pearson contributed a series of papers upon this point to the supplement of the Scientific American, early in 1908. the Hudson and its Tributaries. 303 Others have been impressed with the possibilities of sub- marine erosion by a current along the sea-bottom, and have sought in this way to avoid the necessity of assuming an im- pr obable elevation.* The submarine channel is first and somewhat faintly discern- ible about 5 miles off Sandy Hook. From this point north it is submerged in later sediments and is unrecognizable. From Princes Bay on the Staten Island shore outward the strata are the soft and incoherent beds of the Mesozoic and Tertiary, but from Princes Bay northward to Cornwall-on-Hudson the hard metamorphic and plutonic rocks form the bottom; still farther north are the scarcely less resistant slates, sandstones, and lime- _ stones of the Ordovician and Cambrian. In these two portions erosion must have proceeded more slowly. Turning from the pre-Glacial Hudson for the moment, the post-Glacial work may. be briefly reviewed in order to make clear the state of our knowledge from this point of view. Upon the iater deposits much the most detailed work has been done. It was early recognized that a period of subsidence had followed the retreat of the ice sheet, making of the valley a quiet estuary in which the fine Champlain clays were laid down. Upon these and after an uplift the very prominent gravel and sand deltas were built up. Subsequent elevation | brought about their bisection and the exposure of the clays well above tide- level. F. J. H. Merrill has traced these and their relations from New York to Albany.t+ The local details of terraces, deltas, moraines, ete. have been elaborated in still greater detail by C. E. Peett of the Depart- ment of Geology at the University of Chicago, and by J. B. Woodworth under the auspices of the N. Y. State Geological Survey. Both these writers treat the interesting question of the old relations of Lake Champlain and the Hudson, but these later problems do not bear very closely on fhe points here to be elaborated. These concern the drainage relations in that crit- ical stage when the Glacial epoch was approaching, and they give us some insight into the attitude of land and sea during this and later time. The data utilized in this paper were gathered by the Board of Water Supply of the City of New York and in connection * This alternative is briefly discussed with citations in J. B. Wcodworth’s Ancient Water Levels of the Champlain and Hudson Valleys, Bull. 84, N. Y. State Museum, 71-72, 1905. + Post-Glacial History of the Hudson River Valley, this Journal, xli, 460, 1891. Origin of the Gorge of the Hudson River, Bull. Geol. Soc. Amer., x, 498, 1899. t Glacial and Post-Glacial History of the Hudson and Champlain Valleys, Jour. Geol., xii, 415-469, 617-660, 1904. $ Ancient Water Levels of the Champlain and Hudson Valleys, Bull. 84, _ N. Y. State Museum, 1905. 304 J. i. Kemp—Buried Channels Beneath with the new sources of water which are to be tapped from Esopus Creek in the Catskills for the rapidly growing popula- tion of the metropolis. The writer would express his acknowl- edgments to J. Waldo Smith, C.E., Chief Engineer, for permission to use the data in this way, and to Robert Ridg- way, C.E., Department Engineer of the Northern Department, within whose territory nearly all the ground here covered is embraced. From the Division Engineers, A. A. Sproul,-W. E. Swift, L. E. Brink, L. White and C. E. Davis, and from J. F. Sanborn in charge of the geological features and records, every facility has been received. - Alfred D. Flinn, C.E., Department Engineer of Headquarters, has written of the Storm King crossing.* The writer has constantly worked with his colleague, Dr. C. P. Berkey,t in the field and has dis- cussed results in the laboratory. The interpretations here given and the details of local geology are based upon th observations and inferences of both. The General Line of the Aqueduct.—The main reservoir for the new supply will be developed by a huge masonry dam which will cross and impound Esopus Creek at the Olive Bridge site, a few miles below Shokan in Ulster County. The dam is to be in the more open country southeast of the Catskills, which are in full view afew miles away. At this point the Esopus is in a deep post-glacial gorge in the Hamil- tont flagstones, which dip at a flat angle to the northwest and are cut into extremely regular blocks by a most remarkable series of joints. The master joints average N. 21 E.; the next in prominence, N. 71 W., while rarely there are others at N. 9W.and N. 8 E. The spillway of the dam will be at 580 ft., so that wher- ever in its course to the city the aqueduct crosses a. valley, the water must be conducted in a pressure tunnel. Since bed- rock tunnels for a clear cross-section of fifteen feet or more are far cheaper than steel pipes, it has been of prime import- ance to keep the aqueduct in solid rock, with sufficient cover wherever it dipped below grade; at the same time a tunnel whose bursting pressure is from within, rather than from * Explorations for Hudson River Crossing of the Catskill Aqueduct, New York City, Engineering News, April 2, p. 308, 1908. + Early in the development of the explorations the writer was appointed consulting geologist to the Board. About the same time Prof. W. O. Crosby received a similar commission and a year. later Dr. C. P. Berkey. While the writer has often worked in association with Dr. Berkey, our reports have been made in entire independence of those of Prof. Crosby, with whose - results the writer is not familiar. { The name Hamilton is here used in a general sense to include the Hamil- ton, Sherburne and Ithaca. The section embraces practically uniform sand- stones and shales, almost if not quite devoid of index fossils, and with no. sharp demarcation. It is quite certain, however, that the higher members are included. the Hudson and its Tributaries. 305 without, presents certain novel and interesting problems and makes solid rock a fundamental necessity. In the locations the geologist is of well-nigh indispensable service to the engineer. In the course of its line from the reservoir to New York, the aqueduct has to cross the following principal depressions: Rondout Cr. 160’ A.T.; Wallkill River 150’ A.T.; Moodna Cr. 90’ A.T.; Hudson River 0 A.T., which it reaches at El. 400+; Sprout Br. 145’ A.T. and Peekskill Cr. 60’ A.T. Before the final line was selected several tentative ones were explored, giving us the records of depressed channels not on the final line. At the outset wash-borings alone were used, but when later tested by the diamond and ealyx drills they were found to be entirely unreliable. On the basis of their records a bowlder might be taken for the bed-rock as easily as not. The sections subsequently plotted on wash-borings, there- fore, show merely that the bed rock is presumably deeper yet. To this extent they are, however, of value. The geological. section inevitably crossed is complex both in number of formations and in their structural relations. Beginning on the north in the flat Hamilton beds of the Devonian, the tunnel in passing beneath Rondout Creek pene- trates the full section of the Helderberg series as shown in fig. 4 together with the Shawangunk grit, in and west of the mountains of the same name. It passes beneath Bonticou Crag, three or four miles north of Lake Mohonk, and thence through or over Hudson River slates until it reaches the Archean granite of Storm King mountain, here thrust up on the slates by a reversed fault. The tunnel dips under the Hudson in the granite entirely, and rises on the east bank in the same rock. Thence it continues over or through the sedi- mentary gneisses, marbles, etc., of the Grenville,* but at Peekskill Creek also cuts the Poughquag quartzite and Wap- pinger limestone of the Cambro-Ordovician. These formations have ail dips from flat to vertical; are folded often in a violent way; and are faulted in a very complicated manner. Several of the tentative lines had more to do with the Wap- pinger limestone than the one finally selected and therefore this formation appears along the more northerly routes. The General Drainage Relations —The relations of the tributary streams to the Hudson north of the Highlands are in some respects peculiar. Those on the west bank present * After consultation with Dr. C. P. Berkey, who has done much detailed mapping in the Highlands, and after going over together the exposures both in this locality and in the eastern Adirondacks, this name, hitherto current in the more northerly region, is employed, in the belief that the formations are essentially equivalent. Kemp—Buried Channels. Beneath 7 . Stee 306 Ge dee ‘qutog ISOM ‘d'AA + SUL MLI0IG “YS + q uMo Suds “gq’g | V UMOYSUIIdG “Y's ‘ynopuoYy “y + 4ulog ssSeg ‘q'g ‘ etsdeeyxySnog ‘g ! z3[V¥g MON ‘g'N | Sanqureyy MON “H'N *SInqMoN “N ‘}ufog Auozg oT “aS * OAdqtoqry “I : uoyssury “yf sted ysty ‘WH $ Aopiny ‘A : SutpueT TPT -Ysty “Wf + LouUeysueg™ “q + esel[lA TPISIwO “O + 4ULOg Ueply “VY ‘pejse} SSUISSO1N OY} Jo SUOTZVOO] OYA MOYS 09 dew ]eIoUEy “] ‘DIA the Hudson and its Tributaries. 307 an apparent reversal of drainage—that is they come in from the southwest and turn a sharp corner so as to flow to the sea in a southerly direction. The lower Esopus, the Rondout, the Wallkill and the Moodna all conform to this rule, as will be seen from fig. 1, while on the east bank the streams enter from the northeast. The relations are due to the geological struc- ture. The strike of the rocks and the trend of the ridges are northeast and southwest. The Shawangunk ridge, the wonder- fully folded and faulted Helderberg strata and the Archean Highlands are particularly influential. The upper waters of Esopus Creek come across the flat opposing dip of the Hamil- ton until they strike the folded lower strata and then make the turn. Jn the upper waters the joints are the chief struc- tural influences, the stream on the whole adopting a resultant between the N. 21 E. and the N. 71 W. , although sometimes on one set, sometimes on the other. Tributaries on the West Bank. The Tongore Crossing of the Hsopus.—This is shown in fio. 2, which is drawn looking down stream or southeast. The stream at.this point flows nearly east, along the minor series of joints N. 71 W. It has cut a steep gorge in the flagstones and shales to the 320 ft. contour. On the south bank the bed-rock rises above the 550 ft. contour and is often Fie. 2 Fic. 2. The Esopus channel. The bed-rock is the Hamilton flagstones and shales. exposed, but on the north bank drift conceals all the rock. A series of borings has shown the cover to vary from 60 to 250 ft., and that a buried channel exists at 240 ft. A.T., or about 70 ft. below the present river. Its sides have gentle slopes and its profile is that of a much more mature stream than the pres- ent Esopus, from which it is distant about 2,000 ft. to the north- east. Several other crossings have also been explored, but this one will serve to illustrate the relations in altitude. The Hurley Crossing of the Ksopus.—In its southeasterly course the Esopus breaks through the Hamilton escarpment and strikes a broad open valley, with all the characteristics of an old lake bottom, located upon some well-developed valley 308 3. Ge J. F. Kemp—Buried Channels Beneath HURLEY CROSSING N.28W, “YD SNdOS ,90$ 3218” Lie) <2 eal ao Salenle Wap eleks as ena ea ees . 47 ort+i3b 2.0 Ne eo «2 ele E SE Fic. 3. The Hurley Crossing of the present Esopus and of the Pliocene Kripplebush. of pre-glacial drainage to which fuller reference is subsequently made: lt turns a right angle in this and passes northeasterly. Near Hurley station on the New York, Ontario and West- ern Railway, a series of borings was made across the valley with the results shown in fig. 8, On the northwest side is the abrupt escarpment of the Hamilton with characteristic shelf and talus outline. For nearly a mile to the southeast the sandy level extends and the first outcrops encountered are those of the Onondaga (old name Corniferous) limestone with a very flat northwesterly dip. The wash- borings, which are probably a fair indication of the bed-rock, reveal a surface which corresponds very closely to the dip of the strata. Apparently the pre-glacial stream followed down the dip of the limestone against the basset edges of the shales and sand- stones, sapping them until it had attained a depth of 68 ft. below the present sea-level, and showing that there must have been a correspond- ingly greater elevation of the’ land. In the closing stages of the ice epoch some barrier must have impounded the water in tis valley and have caused the accumulation of the sands. There are some very interesting questions raised by the relations of the pre-glacial Esopus and pre-glacial Zondout. The former after its north- easterly bend proceeds to the Hudson along the general prolongation of the latter, which turns at Rosendale through a high and abrupt ridge of the Helderberg strata, with quite precipitous sides, and joins the Wall- kill. The combined’ streams then reach the Hudson through an estuary with steep rocky sides and appar- ently much freshened up by the ice- sheet, if, as seems unavoidable, it is of er eater seological age. the Hudson and its Tributaries. 309 The High Falls Crossing of the Rondout Valley.—As shown in fig. 4, the aqueduct reaches the Rondout valley on the northwest side at 500 ft. grade. Beyond this point for over four miles it must utilize a pressure tunnel known as the Rondout Siphon. The borings reveal a buried channel on the northwestern part of the valley sunk in the bed-rock to 80+, or 270 it. below the present surface. As this is near the hamlet of Kripplebush we have called it the Kripplebush channel. The cores prove that we have a fault, probably a Fic. 4. HIGH FALLS OO MANLIUS o- HIGH. SHA WANGUN I< GRIT HUDSON RIvER SLATE S.S, SHALE SHAWANCGUNK Fic. 4. The Rondout Crossing near High Falls. B. Becraft limestone : N.S. New Scotland limestone; C. Coeymans limestone; H.F. High Falls shale ; H.R.S. Hudson River slate. The sections are continuous. normal one, upon which the stream was located at the time it was overwhelmed. It had worked down the Onondaga lime- stone against the basset edges of the Hamilton just as had the old-time stream at the Hurley crossing, and then it had been arrested and obliterated by the drift. We have no further records of this stream, but along the line of the siphon and after leaving the valley of the Kripplebush there is only a thin cover of drift, with frequent outerops before the next depression is reached. Then about two miles farther east- 310 J. F. Kemp—Buried Channels Beneath ward is the present Rondout Creek, now flowing on drift at the line. of the section. A little to the north, however, it swerves eastward and cascades over a ledge of Manlius lime- stone. About 400 ft. west of the point where the siphon passes below the present stream, the drill revealed an old, buried channel with a bottom at minus 10, or beneath some- what over 200 ft. of drift. The pre-Glacial stream had evidently followed down a dip slope of Manlius, sapping the edges of the Coeymans (labeled C in the figure) and New Scotland limestones until it also was obliterated by the drift. We think it had left a projecting ledge of Manlius as shown in the figure, because the drill passed from drift into limestone and then into drift again before it caught the bed-rock. The Manlius is the formation containing the waterlime beds, and its subdivision into the Rondout, Cobleskill, Rosendale and Manlius proper has not been attempted here. Somewhat con- - trary to one’s natural expectations, it (an argillaceous variety) is the formation containing the fissures and caves of the region, whereas one would be inclined to look for these rather in the Becraft, which is a very pure limestone. The ancient Rondout Creek moreover, somewhat strangely, seems to have followed the Manlius rather than the hard Shawangunk grit with its soft overlying High Falls shales, which were cut far- ther east. Of the further relations of this buried channel we have no records. The Hurley crossing showed a depth of minus 68, so © that so far as gradient is concerned it might have gone out this way. A mile and a quarter eastward there is a patch of drift in a little synclinal valley of High Falls shale, whose bottom stands at 160 plus, but the depression is of no great consequence. The Wallkill Crossings. —After passing through the Sha- wangunk ridge by a tunnel at grade, the aqueduct turns south- west ; along the surface and crosses the Wallkill valley near the little hamlet of Libertyville, three miles southwest of New Paltz. The entire tunnel will be in Hudson River slates. Before this line was selected, however, two lines of borings were made near Springtown about three miles north of New Paltz and several additional ones in or near the town, which are not here used, as they introduce no essential change in the conclusions. The two Springtown lines are rather less than a mile apart and are called, on fig. 5, Springtown A and Spring- town B. In the former a Wied Giawnel was found at minus 79 almost beneath the present river, which stands at plus 150 or about 229 feet above. The old channel is filled with drift. About a half mile westward a small depression was detected at plus 50, evidently marking a tributary and smaller stream, separated by a divide of less than 40 ft. the Hudson and its Tributaries. oe The Springtown B profile found the old channel of the Wallkill about a thousand feet east of the present one and just at sea-level. The old valley is broad and open with a fairly mature aspect. ‘The western channel is also shown but at 4000. ft. distance and at plus 65, or 15 ft. above the more northerly section. A divide of 150 ft. separated it from the Wallkill, and we cannot but infer that it was coming down to the larger stream from a source to the southwest. It is a striking fact that the Pliocene Wallkill dropped 79 ft. in the mile or less from Line A to Line B and that its valley narrowed appreciably. it may have been on softer slates in the southern section and, encountering the reefs of sandstone char- acteristic of the Hudson River series, cascaded over them to lower reaches on the north. Springtown A is the last record LIBERTYVILLE Fie. 5. Crossings of the Wallkill river near Springtown and Libertyville. which we have of it. The present Wallkill after combining with Rondout creek forms the Rondout river and enters the Hudson in the deep estuarine gorge at the city of Rondont, which has been previously mentioned. The rock bottom of this estuary must lie at a goodly depth below minus 79. The Libertyville section is six or eight miles southwest of the Springtown crossings. It found the Pliocene channel some 600 ft. east of the present Walikill and 120 ft. below it, or at plus 65. Therefore in the intervening stretch, while the mod- ern Wallkill in its meandering course over the drift-filled valley drops 25 ft. the ancient river descended 65 ft. and thus had a fairly steep gradient. It must have been feeling the effects of uplift, although the profiles do not indicate any notable incision. In this section the westerly tributary is not pronounced. Its Fia.. 6, 2 ,000T% JO AVAYBINI J. Lf. Kemp—Buried Channels Beneath The Moodna Se Fite. 6. source must either have been passed or else it still lies beneath the drift farther west than it was necessary to bore. The Moodna Crossing.—-Having passed the Wallkill valley by a siphon, the aqueduct reaches sutliciently elevated ground to make a long course to the south at grade. Its first depression is at Moodna creek, which enters the Hudson just north of Cornwall. The Moodna proper comes in from the west, but it receives an important tributary, Wood- bury creek, which drains the valley De- tween Schunemunk mountain and the Highlands. In the three miles before it discharges into the Hudson the Moodna has cut a deep al in exceedingly heavy drift, with huge bowlders of very impressive size and extremely trouble- some to penetrate with either the calyx or the diamond drills. The details of the Moodna crossing are shown in fig. 6. The present creek is in the drift but very near the southern emergence of slates, and at about the 100 ft. contour. For a half mile to the west the drills have shown a very even floor with a maximum depression at minus 59°2 beneath 360 ft. of drift. A divide then rises of sharp outline, beyond which is another old channel reaching minus 10. The surprising thing about this section lies in the fact that two or three miles away is the gorge of the Hudson with a depth below of minus 600 ft. We have therefore been slow to admit that there is not somewhere in this section an incised notch through which the ancient drainage of the country to the southwest must have been poured without leaving a hanging valley 550 it. above its master depression. Still, care- ful search and fairly close-set holes have failed to locate it. Somewhat the same relationships with the Hudson gorge are shown by other tributaries, such as the Wappinger, and Fishkill a and the Or oton, as will be later brought out. - ye ae the Hudson and its Tributaries. . 313 Rather more than half a mile southeast of the present Moodna, the siphon will cross the impressive reversed fault which has brought the Archean granite to rest upon the Hudson slate, as is clearly exposed at one siguificant locality. A small drift- filled channel has been met in the granite, but it is only worthy of passing remark. Thence to the Storm King crossing the aqueduct is located at grade. 1Diceey a Fie. 7. The Casper Creek Crossing. Tributaries on the Kast Bank. Casper Creck.—The most northerly of the eastern tribu- taries is Casper Creek, which enters the Hudson six miles below Poughkeepsie. It was encountered by the proposed Pegegs Point route and tested with wash-borings, but as only peat, sand, and at the bottom some gravel, were encountered the results are fairly reliable. They are plotted in fig. 7. The lowest point was minus 67. The section passed from Wappinger limestone on the north to Hudson River slate on the south. The section is perhaps half a mile from the Hudson. Bie. &:. Fic. 8. .The Wappinger Creek Crossing. Wappinger Creek.—This good-sized. stream enters the Hudson just south of New Hamburg. Two miles back it yields a fine water power by a series of cascades over Hudson River slates, with a fall of 60 or 70 ft. The water power supports the village of Wappinger Falls. Thence to the river it forms an estuary at tide level. At New Hamburg it hes along the contact of the slates and limestone, but while the limestone appears with steep dips in the north bank, the drill has shown slate beneath the water. There may be a faulted contact; the relations are obscure. As shown in fig. 8, one 314 aS. LF. Kemp—Buried Channels Beneath wash-boring reached 50 feet below tide, and of the three core borings, the deepest was minus 39. These comparatively shallow depths, taken very near the Hudson itself, are inter- esting parallels with the Moodna and Casper crossings, and are indicative of hanging valleys of somewhat striking altitude, and with appar ently abr upt drops to the gorge of the Hudson. Fishkill Creek.—Fishkill Oreek enters the Hudson just — south of Fishkill Landing, an important town on the river immediately opposite Newburg. Within two miles of its mouth it cascades over the Hudson River slates and is obyi- ously off the line of its Plocene channel. A proposed line ~ for the aqueduct, but one which was afterwards abandoned, crossed it just east of Fishkill village, a small settlement five miles back from the river and not to be confounded with Fish- kill Landing. The line crossed the creek in a direction a little east of south and at a flat area of meadow land where the stream was split into three parts. The profile is given in fig. 9. The entire section is drift-covered, but the two core- holes revealed the Wappinger limestone, as had been antici- idivel, 3) S FNS RU ere ee Fie. 9. The Fishkill Brook Crossing. pated. The deepest channel was found at —40, but strangely enough the drill after penetrating about 8 ft. of limestone met fine yellow sand, in which it continued for over 60 ft. until the hole was abandoned. This was interpreted as a crevice in the limestone filled with decomposition products. The conventional signs in fig. 9 are not intended to indicate the dip of the limestone, but merely its presence. The dip is unknown to the writer. A short distance to the southeast the limestone gives place to the Archean, but directly south it extends as a faulted block much farther into the Highlands. The. Fishkill crossing is much farther back from the Hudson than the Wappinger, but its depth is only 10 ft. less. Sprout Brook.—In its passage of the Highlands the aque- duct does not encounter any badly depressed area. Foundry brook, east of Cold Spring, has occasioned some drilling but has revealed no important physiographic data. Spring Brook is on the southern side and, with Peekskill brook, flows into : bar re the Lludson and its Tributaries. 315 Annsville cove just north of Peekskill. Along this line is the crucial area in tle interpretation of the geology of the High- lands, as has already been discussed by C. P. Berkey.* The crossing is about three miles from the Hudson, in a narrow and rather steep-sided valley. Archean gneiss forms the hills, but after a concealed strip on each side the drill Fic. 10. SPROUT BROOK GON Cin CR ts Cpl i Ie Vv Vv NW, Vv v v Vv Vv Vv v Vv Vv v v Vv Vv Vv v Fic. 10. The Sprout Brook Crossing. Sedimentary gneiss on the west ; marble in the valley ; granite gneiss on the east. revealed white marble in the bottom, beneath 125 ft. and less of drift. The bottom of the buried channel was caught at —8, as shown in fig. 10. This is less than the last two, but the stream is hardly as large. The stream evidently selected the easily eroded limestone in preference to the harder gneisses. Kiera, ore Ce ant fen - < Ol-- <---- =< Onn yO ae er FE Ot Cra ess oF “Gone. 2 Ses} PEPE TTH pee Fre Fic. 11. The Peekskill Brook Crossing. Peekskill Creek.— Kast of a steep divide from Sprout Brook lies the valley of Peekskill Creek, a somewhat larger stream and one in a broader valley. On the north side is the Hudson River slate at a steep angle and in an abrupt hillside. The present creek flows upon drift at its foot. Thence to the south for over a mile there are no exposures and the geology must * Structural and Stratigraphic Features of the Basal Gneisses of the High- lands, Bulletin 107, N. Y. State Museum, 361, 1907. 316 Jif. Kemp—Buried Channels Beneath be interpreted by the drill-records, of which there are a satis- factory number. Figure 11 illustrates the section interpreted in the most reasonable way in order to account for the great thickness of limestone, whose unrepeated thickness is known to be about 1,000 feet. Utilizing folds only, it requires a com- pressed sigmoid fold of limestone from whose anticline the quartzite has been pinched out and from whose syneline the slate. As against this a series of normal faults with southerly dip might be imagined but are less likely since the prevailing fault of this region is of the reversed type. The lowest channel is almost beneath the present stream and stands at +10, somewhat unexpectedly high. The Croton River.—From the Peekskill crossing the aque- duct bears away to the east and encounters so much high ground as to reveal little of moment from the depressions. But the earlier work in connection with the great dam across the Croton River gives us abundant data regarding this stream. Mica schist appears on the northwest bank and extends about half way across the valley. It is then succeeded by white marble to the south. The former would be interpreted by F. J. H. Merrill* as metamorphosed Hudson River slate; the latter as the recrystallized Wappinger. A possibly different view has been conservatively suggested by C. P. Berkey.t+ SUMMARY OF TRIBUTARIES. WEST BANK. Distance from Hudson Name. in miles. Contour. So pus 2.5 ie Seep: 25-30 + 240° Hsopus (Hurley) __..-- 13-18 — 68 IGipplebush. 22.22 224- 12-27 + 80 eOmClOwnG Ses fe sah eee 12 — 10 N allie spr. Ale. 2 12 — 79 RS Age 6 alg ra 13 0 i alle cece ant Vy +. 65 Mic ocitapee as Ue ssi 2 — 59 KAST BANK. Casper meee tion Woe 1/2 — 67 Wappingers. 8.) oo! 1/2 a UTA SONG a oo 2 a 6 — 40 Tey OP ROUGE pagae 0: aie eon 3 = Peek einem 84S Le 3 + 10 Croton! yes ee 2 — 20 * F. J. H. Merrill, The Geology of the Crystalline Rocks of Southeastern New York, 50th Ann. Rep. N. Y. State Museum; I. App. A. 21-81, 1898. +C. P. Berkey, Structural and Stratigraphic Features of the Basal Gneisses of the Highlands, Bull. 107, N. Y. State Museum, 861, 1907. the Hudson and its Tributaries. ad The buried channel was found at —20, while the present Croton flowed at +50, leaving 70 ft. of drift, mostly of bowl- der clay, between. The section is two miles from the Hudson. Summary.—F rom this tabulation it is evident that consider- ing the distances of the crossings from the Hudson, the Esopus at Hurley (or perhaps the Pliocene Kripplebush), and the Wall- kill at Springtown A, have eut deepest. The bed-rock at their mouths must be relatively far down. For those near the river Casper Creek is strikingly low for a stream that is decidedly smaller to-day than the modern Wappinger. Fishkill Creek with —40 at six miles back may well be much deeper at its Pliocene mouth, but its location has not been established, much less explored. Even giving all possible latitude to these depths, it still remains true, that as compared with the gorge of the Hudson, now demonstrated at the Storm King crossing, all these tributaries entered in the last stages of erosion, either just preceding or during the Glacial epoch, by hanging valleys of 500 ft. or more above the bottom of the main stream. The Hudson Crossings. The borings which cross this great river furnish naturally the most interesting data of all, and, as will appear, they show a surprising depth to bed-rock. The crossings begin about eight miles north of Newburg or seven miles south of the _ Poughkeepsie bridge. They are somewhat irregularly distrib- uted, but extend at the extreme limit, about two miles south of West Point. The one of greatest interest is the Storm King cross- ing, between Storm King mountain on the west and Break- neck mountain on the east. Of this we have the completest data, but there are core-borings available at Peggs Point, next to the most northerly crossing, and at Little Stony Point, although they are not numerous. All the rest are based on wash-borings, which, as already stated, are only of value in. showing that the bed-rock is lower yet. In some of the wash-borings an artesian flow of fresh water was encountered which spurted above the decks of the lighters carrying the drills, deluging the drillers with water and sand. There is thus a connection with the hills on the banks and beneath the silt. A few little shells have been yielded from beds forty feet below the bottom of the river. The Tuff Crossing.—This is situated a half-mile above Peggs Point, where the next crossing south was located, and runs diagonally across the river in a direction about N. 63 W. Its profile is shown in fig. 12. Hudson River slates are on the west bank, and the heavily bedded Wappinger limestone of the Clinton Point quarries on the east. The geological relations are presumably like those at Peggs Point, of which we have much Am, Jour. Sct.—FourtH Series, Vou. XXVI, No. 154.—Octoser, 1908. 23 318 J. ff. Kemp—Buried Channels Beneath fuller data. The wash-borings returned a fairly even profile for the supposed bed-rock, at depths varying from 210 to 236 below tide. Upon them, however, we can place no reliance. The Peggs Point Crossing. —This is situated a half-mile south of the last named and strikes directly across the river at one of the narrowest possible passages. A steep hillside of Hudson River slate forms the west bank, while a short distance back from the east bank are the Clinton Point quarries in the Wap- pinger limestone. Several lines of wash-borings were run across, giving distances to supposed bed-rock ranging from 139°5 to 256 in what might be esteemed the depths of the river. The records were sufficiently discordant to induce the sinking ‘of three diamond drill cores in the river and one on each bank, as shown in fig. 18. The deepest one caught the slate at 223 ft., a record not very different from some of the wash borings. The next easterly one revealed the limestone at only 92 tt. There is an unexplored stretch of 1,040 ft. between, which presumably contains a deep and rather narrow gorge in order to fall in with the records of the Storm King crossing, roughly ten miles south. Otherwise the great depth at Storm King is very difficult to understand unless we assign to the river a fall of over 375 ft. in ten miles, certainly a most unusual gradient. The borings on Casper Creek, which enters the Hudson just below this crossing, showed the bed-rock at — 69, as has been already stated. ; The New Hamburg Crossing.—Two miles south of Peggs Point the river narrows again between the point on which is situated New germans on the east bank and Cedareliff on the west. Only 2,300 feet intervene from shore to shore, but both banks have the Wappinger limestone. On account of an over- thrust fault, which is beautifully shown near the north portal of the New Hamburg tunnel of the New York Central R.R. and is again revealed by the drill, we know that the slates le be- neath itas shown diagrammatically in fig.14. Only wash-borings were made in the river bottom, but of these five different but closely adjacent lines were run, of which one typical case has been selected for the figure. W here taken beneath the portions of the river well out from the banks, the extremes were 130 and 263°5 with a general range from 195 to 255. Yet such wide differences were met as to destroy all confidence in the borings as indicative of the actual bed-rock. As against the maximum of 263°5 we may contrast the bed-rock in Wappinger Creek at 50 and less. Even this would indicate a hanging valley of over 200 feet, above the bottom of the Hudson. Danskammer Crossing.—A mile south of New Hamburg, on the west bank, is a point with a lighthouse known as Dans- kammer, apparently from its having been a center of merry CROSSIUN G, TUFF re 5 102) ©) J x ° 1°) + 4 2! the Hudson and its Tributaries. Hie: £3: Ut c oF me B 4 aa ‘uy b 3 Sth :< G 2 fe) ans. ae iw x :0 >) os ‘0 a a Tae 3 = = @ 5 : y a = ) 3 uw : S pace gatas area w a. ; Zz a" 7 No ~ 00° x - eh 3 ANY \ 319 \ QS Fie. 12. The Tuff Crossing. Fie. 18. The Peggs Point Crossing. Fie. 14. The New Hamburg Crossing. 320 hee eee Channels Beneath making in earlier years among the German-speaking inhabitants. The distance from bank to bank is 3,500 ft., and two lines of wash-borings were made. In the portions beyond the influence of the banks the depths range from a minimum of 133°2 to a maximum of 268°5, with a general range from 200 to 250.. Again such wide variations were found and such irregularities as to only justify the inference of a bed of bowlders of irreg- ular upper surface. The crossing passed from Wappinger limestone on the west bank to Hudson River slates on the east. The details are given im fig. 15. The Storm King Orossing.—Some six miles south af Dans- kammer the abrupt ridge of the Highlands crosses the river, and Paleozoic strata give way to Archean granite. The river widens in Newburg bay, so that no more lines of borings were considered. Between the great granite buttresses of Storm King mountain on the west and Breakneck on the east it narrows to 2,800 ft., and at this point it was finally decided to locate the siphon for a number of important reasons, based upon the western approach and the local geology. Detailed exploration with the diamond drill was at once begun, but proved an exceedingly difficult matter to prosecute because of the bowlders which were encountered at two horizons. One is about two hundred feet from the surface and is thin; the other is roughly four hundred feet down and is thick and trouble- some. The details are shown in fig. 16, in which the relations of a line of wash-borings to the actual bed-rock are brought out in an interesting manner. Besides the diamond drill-holes a shaft is being sunk on each bank for later use, as a part of the siphon, and for horizontal diamond drill-holes when suffi- cient depth is attained, to test the existence of fault-zones which vertical holes could not locate. Hole 16 at 300 ft. from the east bank is certainly in the bed- rock at about 200 ft. Hole 19 at 560 ft. out probably stopped near it, but the tools met an almost impenetrable series of bowlders of large size and the hole was stopped. Hole 10 is not. on the exact line of the others, but is 300 ft. nearly due south from No. 22. It is the most significant of all. It caught the rock at 608°6 ft., penetrated it for 8°8 ft. , bringing up a core of granite identical with that on the banks. When this had been attained after months of difficult work, the trag- edy of the crossing occurred. A river steamer, having passed the hole, suddenly became unmanageable from an accident, drifted down on the drill and caused the loss of the one hole which has as yet reached the bed-rock and, as we believe, pene- trated it rather than a bowlder. It would indicate, therefore, the bottom at a depth of 608°8 feet. GI CO the Hudson and its T; ributaries. ieee lo. Wigs 17s Fie. 16. Fig. 1o. See ere rae roe gar Se IC ICR MESS ese ee ae : ESSSSNSSSSS y sss | a \ “OUISSOLD JUIOg Ueptly oT, ‘ST ‘S17 ‘Zurssorg yuIog LU04g OTT] OUL “LE PIA “SUISssoH SULY UII0}g oY, “OT “OIA “SuIssorp rommIBysuBg oy, “CT “Ol en-— TS ONTAOE-HSYM Pr) >= WN ~~ peal ; ; i ae SS Drs ! : ae eee ‘i ' ' ! 1 | tease Wy fy pYypyyyy Wi, : Q ives 1, - g ’ VHT, Yi Wi) Wy i : { i { H 1 tet ~~ if), ousssiana i] /, 4 YO, . Why WM Mo pyyppylyyfZ ai GY OL i) » “Wy fy YUU UYU YY Li 8 Yi yl LULL if I ilile 0017 iy 0 4 MN LLL ‘LNIOG N3AGQUY NNO yy 2 ; Lalas) OGL, Lip LY Y Yip, Mili), “LY, YL fy fy VHA AAHAAHA ay (etek NIA NAA SLINYVUD A AAA NIA WHOLS PArndrda® initod RAK ADLIO WAVYD . / sic} a . C) : f aS Pe 52. Tov Dente SONISOSHSVM 40 3NITV #88 2s eeee OLS \ A OIms : 1 eO- yea d z Weoion’ ONys Bees 2fisaoo \oro:s: Meee KS. S|: PF. aba 5 ) ; OF... : (Ons peers as aor iet aaaae wal ae ‘ : re Ee ALINVUS kanye . : ~ ; B 5 ONIN WYHOLS ‘NLW RI WOINAVSYD AAANANAA NA OA AA A 3 * | ,Q0LE > f YIWWvVYASNYG 322 J. FF. Kemp—Buried Channels Beneath Hole 20 at 580 ft. failed to get through the bowlders, so that we only know that the bed-rock is still deeper. Hole 21, about 700 feet from the west bank, stopped in the bowlders at 475°3, so that we only know here that the bed-rock is deeper. It will be driven farther. Hole 18 at approximately 500 ft. from the bank was probably very near the bed-rock if not on it. The casing became bent and the hole was lost. The drillers believed that the bed-rock had been encountered at a shelving point so that the casing glanced off. This is not improbable, although a shelving bowlder is also a possibility. | If we believe that the great ice sheet found a V-shaped river valley in the Highlands and operated to change its section to the characteristic U-shape, then the bottom ought to flatten rapidly beyond hole 10, even if it had not already done so at this point. It is anticipated that future holes which are now being driven will find the bed-rock at depths not very much greater. We are justified in inferring that on the retreat of the ice this gorge was left with some 200 ft. of bowlders and sand on its bottom. Upon this foundation subsequent water deposited about 200 ft. of sand, gravel, and sandy clay. Either a slight readvance of the ice sheet or floating ice then yielded the upper, thin bowlder bed, after which only fluviatile conditions pre- vailed. If we donot credit the great glacier with much eroding power, then the continent must have been elevated decidedly over 600 ft. in order to provide a run-off. But if we believe that the ice-sheet operated to deepen this channel, then this amount of elevation is not absolutely necessary. The heavy, bottom bowlder-bed indicates that after the retreat of the ice fluviatile scour has not affected the bed-rock. The Little Stony Point Crossing.—About a mile south of Storm King a point juts out from the east bank called Little Stony. The river narrows to 2,360 ft., and a line of wash- borings was made with the results shown in fig. 17. No one of them reached 200 ft. Three diamond drill-holes were also sunk, of which the deepest, near the middle of the river, reached 322°2 ft. They all stopped in bowlders, so that to this extent the deep gorge is corroborated. The Arden Point Crossing.—Three or four miles south of the Little Stony crossing and about a mile below West Point is the Arden Point line, to which some exploration has been directed. The shores are in the sedimentary metamorphies of the Highlands, and on fig. 18 are called Grenville, using the name current in the Adirondacks and Canada. The river is only 2,120 ft. wide, so that the shortest section beneath its water of all those tested is afforded. Wash-borings, the only ones used, reached depths of 220 ft. and less. The line, how- the Hudson and tts Tributaries. 323 ever, was abandoned for other reasons before any core-borings seemed called for. Concluding Remarks.—The great depth to bed-rock at the Storm King crossing leads to some interesting lines of reflection. We have no reason to think that the river has ever done otherwise than flow down-grade to the sea along its present channel. Some suggestions have been made of its diversion to the Hackensack valley, but in this the writer and his colleague, Dr. Berkey, who is very familiar with the local geology, have no confidence whatever. For this region we believe in a rather abrupt elevation of the land in the closing Tertiary which brought about a deepening of stream channels to a point as much below the present as the depths of the exploring holes, above cited, indicate. The Hudson, however, obviously cut much more rapidly than its tributaries, and with this the temporary diversion of the drainage of the Great Lakes through the Mohawk may have had some influence. The _ ice-sheet served to still further accentuate the difference, and, as often appears along a trunk glacier, left the tributaries as hanging valleys. The drill-holes at Peggs Point prove that at this cross- ing the gorge must be relatively narrow, but since there seems no way of explaining the depths at Storm King by a gigantic pothole or exceptionally deep, local scour, or any other reason- able method other than water or ice erosion, the still undis- -covered gorge is inferred in the 1,040 ft. between the holes. Since the Storm King granite is the hardest and most resistant rock in the whole course of the river, if it has yielded anything unusual it ought to form a reef rather than a depression. Doubtless the thought will come to a reader, as to the char- acter of the Hudson valley opposite New York. Thus far only the records of wash-borings have been published, and of these the deepest is 300 ft. at a point 2,000 ft. off the bulkhead at 57th street.* Yet there is reason to anticipate something like 700 ft. or more to bed-rock, and the hope may be expressed that some future exploration for an engineering enterprise will give the actual determinations. Postscript.—Since July 15, when the above pages were completed, addi- tional data have been obtained at the Storm King Crossing as follows (com- pare fig. 16): Hole 22 has caught the granite at 507’, indicating that the profile of the eastern bottom flattens from hole 19 to this point, more than as sketched. Hole 20 is 626’ and is in fine sand and clay. Hole 21 had penetrated 4 to 5 feet in a granite ledge or bowlder when temporarily stopped by the collision of atow. Thusthe extreme depth has not yet been reached and appears to be beneath the middle of the river near Hole 20. August 31, 1908. * Cited by W. H. Hobbs in U.S. Geol. Survey Bulletin 270, p. 31, from Ch. McDonald, Engineering News, xxxiii, 159, 1895. 324 C. Barus—-Thomson’s Constant. Art. XXXV.—Thomson’s Constant, e, Found in Terms of the Decay Constant of Lons, within the Fog Chamber ; by Cart Barvs. 1. Lntroductory.—In the last paper,* an account is given of certain tentative experiments to determine Thomson’s elec- tron, by aid of the fog chamber and a separate well-leaded eylindrical electrical condenser. The results obtained for e agreed well with the accepted values. It was shown that the constants of coronas are determinable from purely optical consideration of diffraction and interference, and that the accuracy of the method may be enhanced by using the mer- eury lamp as a source of light for the coronas. There was, however, one grave misgiving; inasmuch as the distribution of ionization within the fog chamber varies in marked degree from place to place, for any given position of a sealed radium tube, and that the mean value assumed was in a measure oratuitous. The results seen in the fog chamber are a com- plication of the effects of primary and secondary radiations together with a very marked exhaustion displacement of the ions. The maximum ionization does not coincide, as a rule, with the position of the radium, and there is no reason why the ionization in the fog chamber should be quite. identical with the ionization produced by the same radium tube in the - electrical condenser, unless both are one in the same apparatus. This is the case in the experiments of the present paper. 2. Electrical condenser—fog chamber.—lt is therefore necessary to make the fog chamber itself an electrical condenser, and this is easily done if the chamber is cylindrical, by install- ing a tubular core of aluminum closed in the inside of the chamber and running axially from end to end. This core is charged to a definite potential and made the inner surface of the condenser, while the scrupulously clean inner wall of the glass chamber (to which water adheres easily) is the outer sur- face and put to earth. Finally the radium, contained in small sealed tubelets of aluminum, is placed within the length of the axial aluminum tube or core, in such a way as to make the ionization within the fog chamber uniform,—a condition vouched for in case of the oceurrence of uniform coronas on exhaustion, from end to end of the chamber. There are thus three currents to be determined. (1) The conduction current due to inevitable leakage between the condenser surfaces. This is made a minimum and nearly negligible in value, by keeping the aluminum core out of the *C. Barus: This Jour., xxvi, pp. 87-90, 1908. C. Barus—Thomson’s Constant. 325 condenser except when not in use and by sheathing it with an annular air space beyond the condenser. It is found experi- mentally, by direct measurement in the absence of radium. (2) The current resulting from the ionization of the room air near the fog chamber and on the outside of it, due to gamma rays. This is made a minimum by allowing the thin wire communicating with the electrometer to run axially away ° from the fog chamber; for the gamma rays, in spite of their penetrating power, are quickly reduced by distance. This eurrent is found in the presence of radium within the axial tube, by leaving all adjustments identically in place, but break- ing the metallic connection between the aluminum core and the electroscope, etc., by a hard rubber insulator. If an auxiliary condenser is used, the measurement (1) must be made without it, as otherwise its leak would be counted twice. Fortunately the conduction current is relatively quite negli- gible. (5) The current due to ionization within the fog chamber... This is found by deducting from the total current found on connecting the charged aluminum core and the elec- trometer, the two preceding currents. 3. Auxiliary condenser.—To vary the experiments to the extent that different speeds of leakage may be obtained, as well as to find the capacities of the electrometer and fog chamber, an auxiliary condenser must be inserted as a part of the electroscope. This condenser consisted in the present experiments of two plates of brass, having an area of 315°4™, and usually kept at a distance of °382™ apart by outrigged feet of hard rubber, which stood on a plate of glass. By put- ting small glass plates under these feet, this capacity could be varied at pleasure. The usual equation was corrected by aid ot the factor 1+ (d+din 16 Var (d + 0)/@ + On (d + 8) /0)/rV/ar, where @ is the area, d the distance apart and @ the thickness of the plate of the auxiliary condenser. Naturally a guard ring condenser would have been preferable for standardization, but none was at hand. To determine the very small capacity C of the electroscope- fog-chamber, two successive full charges from the lighting circuit, at a potential V=250 volts, were in turn imparted from C’ to the auxiliary of capacity C’. If V” be the potential observed after these two charges and S= V"/(V— V"), © C= 8S/(1+ V1+4+S). It is curious that this method of successive charges leads to complicated cubic, quartic, quintic equations, etc., which follow no simple rule. The ratios /? of the potentials after four and after two charges 2 = V"’”’ / V" is however still available. Apart from these complications, 326 C. Barus— Thomson's Constant. - the large deflections obtainable after many successive charges would, in the absence of conduction leakage in the condensers, make this method very satisfactory. In the definite measurements, however, almost the whole capacity may be placed in the auxiliary condenser, so that the capacities of the electrometer and fog chamber are of small ‘importance. Ratios of C’/C= 86/17, 30/17, 20/17, and others, were tried. 4. Method.—In the preceding paper the value of e found was ultimately dependent upon the velocity of the ions in the unit electric field. In the present experiments a value will be investigated, based on the decay constant b6=1:1 107°, of the ions. This method has the advantage that large core poten- tials are admissible in the electrical condenser, so that an ordinary graduated Exner electroscope suffices for the meas- urement of current. The small capacities of the instrument make it necessary to insert an auxiliary condenser, as other- wise the discharges are too rapid for trustworthiness. If @ is the number of ions produced per second per cubic centimeter by the radium placed within the condenser core, and /V the number of nuclei (ions) found when the core is free from charge, dn/dt=a—bNW’=0. Again if n is the number of nuclei found when the core is charged and 7 the corrected current observed, ¢ Thomson’s constant and v the effective volume of the fog-chamber-condenser, dn/dt=b( V*—n’*) —t/ev =(. Hence if the capacity of the system is C and V the corrected fall of potential per second e= OV/(bv(N*—n’)) Usually V is measured in volts, so that V/300 replaces V in the equation. It is obvious that V must be large enough to keep the current V constant, and the observations always show this at once. 5. Data disregarding external gamma rays.—The alu- minum foil electroscope made it convenient to use the high potentials of the electric lighting circuit (about 250 volts) for charging. The number of nuclei (ions) found in the exhausted fog chamber free from charge, at its central core, was V=474,000. The number of nuclei found in the exhausted fog chamber when the core was charged to 250 volts was n’=82,500. Hence about 391,000 vanished in the presence of the electrical cur- rent, the original apertures of the coronas being reduced from about 22 degrees to 18 degrees. The drop of pressure 6p/p='30 nearly, was taken high enough to catch all the ions, but not so high as to catch the vapor nuclei of dust-free wet air. O. Barus—Thomson’s Constant. 327 The amount of exhaustion was equivalent to the volume ratio v,/v=1:29. Thus the number of ions in the fog chamber at atmospheric pressure was V=611,000 per cubic centimeter for the uncharged core and n=106,000 for the charged core. Hence WV*—7’ is about 362 10°. The value of 6=1:1X10° is taken from Prof. Rutherford’s book. The source of light for the corona is part of a Welsbach mantle, as usual, and the old constants of coronas were used, since it is a part of the purpose of this paper to test those constants. The volume of the fog chamber was estimated at 51,000 °°". In the first experiments, the effect of the gamma rays penetrating into the air on the outside of the fog chamber was neglected and the data on using different condensers were as follows, all data being given in electrostatic units. ( denotes the capacity of the system, V the drop of potential per second, 2 the corrected current passing through the condenser fog chamber. C Vx 10° aX 10° ex 10" 1038 10 103 b°] 47 21 101 5°0 47 19 92 4°5 LT 40 70 3°4 the last observation being made without an auxiliary con- denser. The current ¢ was quite constant throughout the voltage interval (near 250 volts) of observation. Hence the effect of gamma ray penetration has seriously increased the leakage, and ¢ therefore appears too large, except in the last observation, where 2 is probably no longer measureable. 6. Further data.—In the following experiments the effect of the external gamma rays was eliminated as specified in § 2. The conduction current was usually quite negligible. The nucleations observed in the exhausted fog chamber were n= 82,000 and WV*=506,000, when the core was charged and un- charged, respectively. The exhaustion was again equivalent to a volume increase of v,/v=1'29. Hence in the fog chamber full of air the respective nucleations are V=653,000 and n= 106,000, whence V?—7’?=415 x10". The electrical measurements, if all data are given in electro- static units, may be summarized as follows: HG. Vio: Oye BX NO" 103 7°5 Ti 3°3 47 17°2 81 3°5 37 19°6 72 ; 3°] Li 33° 58 2°5 / 328 C. Barus— Thomson's Constant. Thus with a correction for the external gamma radiation, the data for ¢ show reasonable values, in spite of the simplicity of the experiment. It follows, therefore, that even in the ease of such large numbers of ions as occur in these experi- ments (over 500,000), both positive and negative ions must have been caught in the fog chamber and that the constants of coronas used heretofore are substantially correct. In case of the last value ex10°=2°5, for the small capacities of 17°, the aluminum leaves on the electroscope converge too rapidly for measurement, so that the air resistance may have produced an appreciable discrepancy. Hence both 2 and e are too small. No refinement has been attempted in these experiments, their chief purpose being to test the standardization of the fog chamber in terms of coronas and the degree to which positive and negative ions may be caught even at very high nucleation. One may note in conclusion that the currents of the order of ~=7 electrostatic units or 2°6 x10" amperes, are already quite within the reach of the sensitive galvanometer.* My thanks are due to Miss L. B. Joslin, for assistance throughout this investigation. Brown University, Providence, R. I. W. A. Drushel—Cobalti-Nitrite Method. 329 Arr. XXXVI.—The Application of the Cobalti-Nitrite Method to the Estimation of Potassium in Soils; by W. A. DRvUsHEL. [Contributions from the Kent Chemical Laboratory of Yale Univ.—exce. ] Ty a previous paper* from this laboratory it was shown that potassium may be estimated with a fair degree of accuracy by precipitating it as potassium sodium cobalti-nitrite in a solution acidified with acetic acid and oxidizing the precipitate with standard potassium permanganate. In the same paper the applicability of the method to the estimation of potassium in commercial fertilizers was shown by a series of experiments. In the method as previously worked out an excess of con- centrated sodium cobaltinitrite solution acidified with acetic acid is added to a neutral solution of a potassium salt, and the mixture is evaporated to a pasty condition on the steam bath. After cooling, the residue is stirred up with sufficient cold water to dissolve the excess of sodium cobalti-nitrite. The precipitate, consisting of K,NaCo(NO,),.H,O, is filtered on a rather close asbestos felt in a perforated crucible and well washed with cold water, or preferably with a half saturated sodium chloride solution. The precipitate and felt are trans- ferred to an excess of standard N/10 or N/5 potassium permap- ganate which has been diluted to about ten times its volume and heated nearly to boiling. If particles of the precipitate stick persistently to the walls of the crucible and cannot be removed with a spray of water, the crucible is put into the permangan- ate solution. After stirring for a few minutes the solution is gradually acidified with 5° to 20°™* of dilute sulphuric acid, and the oxidation is allowed to go to completion, a process which seldom requires more than five minutes. If no parti- cles of the yellow precipitate settle out on standing a minute, the oxidation may be considered complete. The hot solution is then bleached by running in a measured amount of standard oxalic acid, containing 50°™* of concentrated sulphuric acid per liter. The solution after bleaching is titrated to color with standard permanganate in the usual manner. In this process the cobalt in the molecule is reduced from the trivalent to the bivalent condition and not reoxidized, consequently from the molecule of the potassium sodium cobalti-nitrite we find 0:000857 grm. K,O equivalent to 1°™* of strictly N/10 potassium perman eanate. This factor of course must be corrected for any variation in the normality of the permanganate solution used. * This Journal, xxiv, 4383, 1907. 330 W. A. Drushel—Cobalti-Nitrite Method. For the extraction of the alkalis 10 grm. of dry soil are placed in an Erlenmeyer flask with 25° to 35%° of about 20 per cent. hydrochloric acid. The flask is thoroughly shaken and a small funnel is hung in its neck to avoid too great a loss of acid by evaporation. The contents of the flask are digested on the steam bath for 24 hours. From this point several methods for the final preparation of the sample were tried with satisfactory results, given in the . table. Since duplicate estimations were to be made by the gravimetric chlorplatinate method, for which it was necessary to remove the iron, aluminum, calcium, magnesium, phosphoric acid and ammonium salts, if present, from the soil extract, the following general procedure was found to be most expeditious. The soil extract was filtered through paper into an evaporating dish and the residue was washed with boilmg water until the filtrate gave no reaction for chlorine with silver nitrate. The filtrate was evaporated almost to dryness to remove the excess of hydrochloric acid as far as possible. The residue was dis- solved in about 200°" of water and, after heating to boiling, a little ammonium hydroxide and ammonium oxalate were added. The mixture was boiled a minute, settled, filtered and the precipitate was washed with hot water until a drop of the filtrate give no chlorine reaction. ‘The filtrate was concentrated, transferred to a 200° flask, cooled, and made up to the mark. After thoroughly shaking, 50° aliquots were pipetted off for the gravimetric and volumetric estimations. The aliquots were evaporated to dryness in platinum dishes, and gently ignited to remove the ammonium chloride. After cooling, the residue was moistened with dilute sulphuric acid and again ignited, gently at first and finally at the full heat of the Bunsen flame, to remove the last trace of ammonium present as the sulphate, and to destroy any organic matter which might be present. The residue for the gravimetric estimation was dissolved in a little water and a few drops of hydrochloric acid over the steam bath, and the estimation of the potassium was made according to the modified Lindo-Gladding method. To dissolve the residues for the volumetric estimations a little water and a few drops of acetic acid instead of hydro- chlorie acid were used. In the volumetric work approximately N/5 potassiuin permanganate was used, 26:08°™ of permangan- ate being equivalent to 50°° of exactly N/10 oxalic acid. From this ratio the factor for K,O was found to be 0:001642. In each case the potassium was precipitated as the cobalti- nitrite by evapurating off with 10°* of sodium cobalti-nitrite prepared according to the method of Adie and Wood.* * Journ. Chem. Soc., Ixxvii, 1076-80 (London). W. A. Drushel— Cobalti-Nitrite Method. ao Character Soil taken K,0 found No. of soil grm. Method erm. per cent. TP. Clay (1) 2°5 vol. 0°0028 Orrt 2 ai eG 0°0035 0°14 (3) ce gray. 0°0035 0°14 hh. Clay = () a Vv. 0°0100 0°39 (2) 3 6 00092 0°37 (3) o g. 0°0098 0°37 III Loam (1) . v. 0°0074 0°30 (2) % SS 0:0068 0°27 (3) ee g. 0°0075 0°30 EV Loam (1) e v. 0°0060 0°24 (2) Es a 0°0058 0°23 (3) $ g. 0°0058 0°23 V Gravel (1) v. 0°0042 0°17 (2) er g. 0°0045 CRIS) VI Gravel (1) og Vs 0°0047 0°19 2 os rs 0:0044 0°18 ts} a g. 0°0050 0°20 Clay (1) & v. 0:0048 0°19 VII Gravel (2) a g. 0:0046 018 (3) Vv. 0:0045 0°18 (4) “6 e 0:0040 0°16 (5) aS 0°0044 0°18 The following exceptions are to be noted to the general method previously outlined for the preparation of the sample. In I the excess of hydrochloric acid, the iron, aluminum and calcium were removed from the separate portions after aliquot- ing, and in (1) and (2) sodium carbonate was used instead of ammonium hydroxide and ammonium oxalate for the removal of the iron, aluminum and calcium. In V (1) and (2) bases other than the alkalis were removed in the separate aliquots by ammonium hydroxide and ammonium oxalate. In VII the aliquots were made directly from the hydrochloric acid extract. That of VII (2) was treated in the usual manner for the gravimetric estimation of potassium. The other aliquots of VII were evaporated to dryness and gently ignited to remove any ammonium chloride present and to char the organic matter. The residues were extracted with hot water and a little acetic acid, filtered and evaporated with sodium cobalti-nitrite in the usual manner. In this work the results are based on a small amount of soil (2°5 grm.) in each case because but a limited amount of each sample was available. A higher degree of accuracy may be secured by using 10 grm. of soil for each estimation instead of 2°5 grm. 332 W. A. Drushel—Cobalti-Nitrite Method. Summary. A weighed amount of dry soil is extracted with an excess of hydrochloric acid over the steam bath. The excess of acid is removed from the extract by evaporation. The bases which might interfere with the process are removed with sodium carbonate or ammonium hydroxide and ammonium oxalate. Ammonium salts and organic matter are removed by ignition. The residue is dissolved in a little water and a few drops of acetic acid, and the mixture evaporated with an excess of sodium cobalti-nitrite to a pasty condition, stirred up with cold water, and filtered upon asbestos in a perforated crucible. The precipitated potassium sodium cobalti-nitrite is washed with half-saturated solution of chloride, and treated with an excess of permanganate in hot dilute solution. The color of the per- manganate is destroyed by an excess of standard acidulated oxalic acid, and the excess of oxalic acid titrated to color with permanganate. Edgar— Estimation of Chromic and Vanadie Acids. 333 Arr. XX XVII.—The Llodometric Estimation of Chromic and Vanadic Acids in the Presence of One Another; by GraHAM Enear. [Contributions from the Kent Chemical Laboratory of Yale Univ.—cxci. | Tue difficulties which are met in the separation and gravi- metric estimation of chromium and vanadium have led to various attempts to accomplish the estimation of these elements in the presence of one another. In 1898 Hillebrand* proposed a method based on the fact that chromium may be accurately estimated when present in small quantities by comparison of the color of the solution containing an alkali chromate with that of standard solutions of potassium monochromate. If then the solution be reduced with sulphur dioxide and the excess removed by boiling, vanadium, if present, may be esti- mated by titration with standard potassium permanganate. This latter process, however, is open to the cbjection that if titration be made in the hot solution an appreciable amount of chromic salt is oxidized, while if the solution be cold the end point is more or less uncertain. If, however, a correction be made for the permanganate used up in partially oxidizing the chromium, the method is accurate for the small quantities of vanadium found in rocks. A method has been also proposed by Campagnet for the estimation of these elements in the same solution, in which the vanadium is estimated by titration with potassium permanga- nate after reduction with hydrochloric acid and subsequent — treatment with concentrated sulphuric acid to convert the oxychloride into oxysulphate. The solution is then boiled with an excess of strong potassium permanganate, the excess of reagent destroyed by the addition of a piece of filter paper, and the oxides of manganese filtered off, the filtrate containing the vanadium as vanadate and the chromium as chromate. Ferrous ammonium sulphate in excess is now added and this excess determined by titration with. potassium permanganate, the difference being of course the amount of ferrous salt used up in reducing the chromic acid, thus allowing the chromium to be calculated. The definiteness with which vanadic acid is reduced to different stages of oxidation by means of varied reducing agents makes it especially suited for processes of differential reduction, in which the substance to be analyzed is treated with two or more reducing agents whose action upon the constituents to be estimated is different. The latter are then calculated from the determination of the total reducing effect in each case. * Journal Amer. Chem. Soc., 20, 461-465, ¢ Bull. Soe. Chim. (8), xxxi, 962-965. Am. Jour. Sci.—FourtH Srrizs, Vou. XXVI, No. 154.—Ocroser, 1908. 24 3384. Ldgar—LKstimation of Chromic and Vanadic Acids. That vanadic acid is reduced under proper conditions to the state of tetroxide by hydrobromic acid has been shown by Holvescheit,* and that the reduction may be carried to the state of trioxide by hydriodic acid has been shown by Fried- ham and Euler,t by Gooch and Curtis,t and by Steffan.§ Chromic acid, however, is reduced to a chromic salt by both hydrobromie| and hydriodic acid. In 1895, Friedheim and Euler], in connection with their work upon the estimation of vanadie and molybdie acids, suggested that vanadic and chromic acids might be estimated in the presence of one another by a process based upon the differential reducing action of hydro- bromic and hydriodic acid: but inasmuch as no experimental data were given and as none have appeared on the subject in the thirteen years now elapsed, the present writer has considered himself justified in investigating the feasibility of such a process. The apparatus used will be briefly described. As reduction flask served a 100°™* Voit flask, to the inlet tube of which was sealed a small separatory funnel, serving for admission of acid to the flask and for the entrance of a current of pure hydrogen gas from a Kipp generator. To the outlet tube of the Voit was sealed a Drexel bottle, provided with a Will and Varren- trap trap, the two serving as absorption apparatus for the bromine and iodine liberated in the process. The experiments were carried out as follows: Portions of standard solutions of sodium vanadate and potassium bichro- mate were measured into the Voit flask, one to two grams of potassium bromide were added, and the flask connected with the absorption apparatus containing a solution of potassium iodide made alkaline with sodium carbonate or sodium hydrox- ide, and the whole apparatus filled with hydrogen gas. Fifteen to twenty cubic centimeters of concentrated hydrochloric acid were added through the separatory funnel and the solution boiled for ten minutes, the reduction having been found’to be always complete in that time. A slow current of hydrogen was maintained to avoid the “sucking back” of the liquid from the Drexel bottle. The apparatus was disconnected, the Voit flask placed in a beaker, containing cold water, and the alkaline solution in the absorption apparatus cooled by running water. The contents of the trap were washed into the Drexel bottle and the solution therein made slightly acid with hydro- chlorie acid. The lberated iodine was titrated with approxi- mately N/10 sodium thiosulphate and the color brought back by a drop or two of N/10 icdine solution, after the addition of starch. The results of this titration are given in the follow- ing table: *Tnaug. Diss. Berlin, 1890. + Ber. Dtsch. Chem. Ges., xxviii, 2067-2073. t This Journal, ii, 156-162. § Treadwell, Quantitative Analysis, p. 327. | Farsoe, Zeitschr. Anal, Chemie., 1907, 308-310. *| Loe. cit. Edgar—FEstimation of Chromic and Vanadie Acids. 335 (I) (11) V2.0; CrO3 NaeS203 Na.S.Os Error Error No. taken taken on on of as as N/10x1:031 = N/10x1°081 V,0; CrO; Expt. NaVO; K.Cr.0; grm. erm. em? em? grm. grm. (1) 01528 .-.. 16.20 16°22 tI) 4 Fae apie tne (ewe 015285." 022. 16°19 16°20 {ir} BY ou eee 23) OLE eee 21°59 21°58 ti) 1 cena Bhd ( 4) 0°1523 0°0685 36°08 16°22 +0°0002 —0°0001 (5) 0°1523 0°0685 36°10 16°20 +0°0000 +0°0000 ( 6) 0°1523 0°0085 . 36°12 16°17 —0°0003 +0°6002 ( 7) 0'1523 0°0685 36°07 16°20 +0°'0000 —0:0001 ( 8) 0°1523 0°1370 56°00 16°17 —0°00038 +0°0001 ( 9) 0°1523 0°1370 56°02 16°22 +0°0002 +0°0000 (10) 071523 0°1370 56°03 16°19 —0'0001 +0:°0001 N/10 x 0:992 N/10x 0:992 — Changed Stand. of Na2S.Os (11) 0°2031 0°1370 63°82 22°46 +0°0000 —0O:0001 (12) 0°2031 0°1370 63°80 22°48 +0°0002 —0:0003 (13) 0°1016 0°0685 32°00 11°25 +0°0002 +0:0001 (14) 0:1016 0°0685 31°92 11°24 +0°0001 —0:0001 (15) 0°1016 0°0685 31°90 11°25 +0:0002 —0°0002 (16) 0°0508 0:03843 15°95 5°63 +0°0002 —0°0001 (17) 0°0508 0°0343 15°95 5°62 +0°0001 —0-0001. Alkaline potassium iodide was again placed in the absorption apparatus and the latter connected with the Voit flask. The eurrent of hydrogen was turned on and, after all air had been expelled, the apparatus was disconnected momentarily, one or two grams of potassium iodide were added to the solution in the Voit flask, and connections made again. Through the separatory funnel ten to fifteen cubic centimeters of concen- trated hydrochloric acid and three cubic centimeters of syrupy phosphoric acid were added and the solution in the reduction flask was boiled to a volume of ten to twelve cubic centime- ters. The absorption apparatus was removed and cooled, hydrochloric acid was added and the liberated iodine titrated with approximately N/10 sodium thiosulphate. The results of this titration are given under II, Table I. It is evident that the iodine determined in the first titration corresponds to reduction of the chromic and vanadic acids corresponding to the equation, V,0,+2CrO,+ 8HBr=V,0,+ Cr,0O, + 81+4H,O0 while in the second case the iodine corresponds to a reduction 336 Hdgar—Lkstimation of Chromic and Vanadie Acids. of the vanadium tetroxide to trioxide as indicated in the equation, V,0,+2HI=V,0,+21+H,0 The second titration, therefore, determines the vanadiec acid present, and the difference between the first and second far- nishes the necessary data for the calculation of the chromium. From the small error shown in the results of Table I it is evident that vanadic and chromic acids may be accurately estimated in the presence of one another by the differential reducing action of hydrobromie and hydriodic acids under the conditions used above, the liberated halogen being absorbed in potassium iodide solution and the iodine titrated with sodium thiosulphate. FL. Ransome—Apatitic Minette. 337 Art. XXXVIIL.--An Apatitic Minette from Northeastern Washington ;* by FrRepreRiIcK Les~iz Ransome. In the course of a rapid reconnaissance along the Northwest Boundary in the summer of 19V1, numerous dark micaceous dikes were noted on both sides of the Columbia River, between the smelter-town of Northport, Wash., and the settlement of Waneta, at the mouth of Pend d’Oreille River, in British Columbia. These dikes cut a series of sediments of unknown age, possibly Carboniferous, consisting of steeply inclined dark slates, banded schistose limestones, and subordinate quartzites. Similar dikes were noticed for 12 miles eastward along the Pend d’Oreille and as far west as Kettle River, which crosses the boundary about 40 miles from Waneta. Not all of these dikes have been carefully studied, but most of them appear to be minettes or rocks closely related to this type. One speci- men from a particularly micaceous dike at the mouth of Sheep Creek, on the northwest bank of the Columbia opposite North- . port, has been analyzed by Dr. W. F. Hillebrand and found to be of rather unusual chemical character—enough so, it is thought, to warrant the publication of the analysis with a petrographical description. The rock is dark, greenish gray, by far the most prominent constituent in hand specimens being the closely crowded scales of biotite up to 4 or 5 millimeters in diameter. The other mineral constituents form in general a fine-grained matrix to the biotite and are not individually distinguishable without a lens. Parts of the rock, however, contain irregular light- colored streaks in which the feldspar is a little more distinct than elsewhere. Under the microscope the rock shows a poikilitic texture, large irregular areas of optically continuous feldspar being crowded with automorphic crystals of biotite, pyroxene, apa- tite and titanite. The feldspar is chiefly orthoclase, although much of it is not optically homogeneous and some evidently contains microperthitically intergrown albite. It is all more or less turbid with minute, brown, dust-like inclusions. The biotite is chestnut-brown in most sections with the usual strong absorption parallel with the cleavage. The axial angle is small and the interference figure shows no distinet opening of the lemniscate cross into hyperbolas. The pyroxene, which is automorphic in the prism zone with occasional terminal planes, is monoclinic with a maximum observed extinction of Zac of about 45°. It is colorless to very pale green, non- pleochroic and is probably an augite near diopside in compo- * Published by permission of the Director of the U. S. Geological Survey. 338 I. L. Ransome— Apatitic Minette. _ sition. The pyroxene is for the most part fresh, but some crystals have undergone slight partial alteration into green hornblende and calcite. The apatite occurs in stout prisms up to two millimeters in length and is noticeably abundant as seen in thin sections. It is included chiefly in the orthoclase and biotite, although the augite is not free from occasional prisms. Rather abundant titanite and a little magnetite and pyrite make up the minor constituents. ~ The chemical analysis of this rock by Dr. Hillebrand is as follows : Chemical Analysis of Minette. Oi ws ee ee PA Boos. jose ee PIB O SG po bn Ree setae 2D OOD Or. 0) oe cee ae a 04 Bie O) 2 cave Geant 8 4°06); i NIO) 2b 200s eee eee eee 02 MeO irae nies Te 404 UF iy Mins 2 ee eee 25 MoO. tues ow ek 8:65. (BaO. ve te 44 CAO ra tek eee amet ermal MeN) SrOe 2533 22 11 INO) se a dete 157 Li,O 0. ue a trace 1G a EE DESC SRR Ss 6°10 NOC Se ee 04 EO) ee eget ee 1°54 HeS,.. 2-0 2202 Oe |e ROY? Sa a 2°30 TOU Wee 2°36 100-01 DV OES oes ESE ae 502 Ikess-O'for Fl, Cl. Sea 11 WO Nae eee Wee 1:24 Phaadirces Uiaeive Sale 4-05 99°90 Ol ites ase Bani te eee AO In comparison with most minettes the analysis of the rock here described is low in silica and alumina, rather high in potash and titania, and remarkably high in phosphoric acid. This last feature, indeed, is the most noteworthy peculiarity of the rock. Out of the 2112 rocks of which superior analyses are collected in Washington’s tablest only 59, or 2°7 per cent, contain over 1 percent of P,O,. Of these only 10 contain over 2 per cent, and, of these 10, only 2 have over 3 per cent of P,O,, namely a pyroxene-apatite-syenite (orendase) from Finland, described by Hackman,t with 5-98 per cent of P,O, and an avezacite (avezaciase) from the Pyrenees described by Lacroix§ with 3°32 per cent of P,O,. | The norm as calculated from the chemical analysis and the place of the rock in the quantitative classification are as follows : * Not corrected for influence of V2Os. + Washington, H.S. Chemical analyses of igneous rocks. Professional Paper U.S. Geol. Survey No. 14, 1903 ; Superior analyses of igneous rocks. Professional Paper U. S. Geol. Survey No. 28, 1904. + Hackman, V. Neue Mittheilungen tiber das Ijolithmassiv in Kuusamo. Bull. commission Géologique de Finlande No. 11, 1899, pp. 36-37. § Lacroix, A. Les roches basiques accompagnant les lherzolites et les ophites des Pyrénées. Comte Rendu VIII Congres Géol. Int. 1900, Paris, 1901, p. 382. FL. Ransome—Apatitic Minette. 339 Norm. N 3° Or oan see Ory eairemane Bem. 5111 ING 2h TE : 2 6 ir OT a aaa Ween 57 5:67 Sali 43-07 F-30330" Portugare. N ate. So 8 9 eee = 1. Wyomingase. i=. 29:38 ST eee 6 ; Ap .-- 9°41 Fem.=51°11 Na OQ ™op= 2 Washingtonose. Oi 6-73 Pipe 6°05 CS Sena 4°56 HeOs 23°84 Bee. 104 Bes, -- 06 99°32 This norm, as shown above, places the rock in the salfemane class, in the lendofelic or Portugare order, in the peralkalic or Wyomingase rang, and in the dopotassic subrang. It is the first recorded representative of this subrang, and as it falls well within the classificatory limits of the division, washing- tonosé is suggested as an appropriate subrang name. It is worthy of note that the wyomingite, orendite and madupite described by Cross* from the Leucite Hills, Wyoming, are all comparatively rich (1:39 to 1°89 per cent) in phosphoric acid. For comparison the norms of the two rocks standing nearest to washingtonose in Washington’s tablest are given below. Wyomingite (Cross). Shonkinite ( Weed and Pirsson). ieee 8 rete 44:5 Oy Bae eee es te! 20% eer ere 8) jst) 105 Pan OY gs oe oy ae Saat pm 8°9 SNE eee Se tere led DAN TY en Us a SG begin SLT Bea) INC ies eye as 6°2 The es eins en 74 11) NO ae pained ta!) 1 Re Rates A on Se a 2859 RSE Bi Pig So AIS ee ea 79 (Qa ee ae ane a 14°8 Mee Ls Decpeneee eso ANT NYG os es Ae ele ee ein 6 Pe! eer ee tT LOPS JU oes eye ere eal em Ca ny) hea. 2 RE or) ot Mg 4°5 LO) A ee Nea fame a ez * Cross, W. : Ignéous Rocks of the Leucite Hills and Pilot Butte, Wyo., this Journal (4), vol. iv, p. 130, 1897. + Professional Paper U.S. Geol. Survey No. 14, p. 339. 340 Lf. L. Ransome—A patitic Minette. The wyomingite is in the perpotassic subrang and the shonk- inite in the sodipotassic subrang of the wyomingase rang. None of the minettes of which reliable chemical analyses are available falls near washingtonose in the quantitative clas- sification. The three rocks of this type included in Washing- ton’s tables contain from 50°81 to 52°26 per cent of silica. Consequently they are all in the dosalane class. Owing to their relative richness in silica and alumina as compared with the rock here described, their norms are more feldspathic and all are in the perfelic.or Germanare order. These norms, from the tables, are as follows: Norms of Minettes. I 16 III AS) he ES Se ys ae DASA Pi peeg anc est 1 95°67 er 41-4 AD sc Se eee Bet LO BEG. Gain pein earn 39°02 ala 8:4 EAM Ce ree ua tere 14S eae iene aot Oise 5 eee 1671 aN GrQe Ree a eae OPO Mi epiclhy (cutie Sarai tts S980 ee oe BT giyecompmaete okt ernie Lh Bibs ea Nei fae 5:3. 243 3°9 Thy, 2) ad ie SE Be: TPIODRS PSRs oe Sete Mer mer res Bn! (Oca mete, 8 ay aha SiO eee ea Van 56.2 oe Sie 14:7 Mig ce at wee aes Sc Ou pehuyprier pd ways Bed 3. lon eee 3° AT ape: oie Se ei ree iL pigs ocala 3° 9s ely oie iran 3°92 Ap Su ee eo ee es AD Se Nee. or aa bee ae 1°4 I. Augite-minette (Pirsson). Monzonose. W. T. Prof. Paper 14, p. 255. II. Soda-minette (Brégger). Akerose. W. T., p. 263. Ill. Augite-minette (Doss). Dopotassic subrang of andase. W. Aree pe 205. All of these norms contain considerable albite, the quantity reaching 23°6 per cent in the augite-minette (monzonose) from the Little Belt Mountains, Montana, -described by Pirsson (1), whereas in the washingtonose here described there is no norma- tive albite. The three norms just given also contain much more anorthite than that of the rock from the banks of the Colum- bia, but show no normative leucite. Other differences will appear in comparing these norms with that of washingtonose. The actual mineralogical composition, or mode, of the wash- ingtonose described can not be accurately calculated, as the compositions of the augite and biotite are not known. The abundance of lenads in the norm, conditioned by the low silica and high alkalies, suggests modal nephelite. None, however, has been detected, while it is certain, on the other hand, that EF. L. Ransome—Apatitic Minette. 341 the albite molecule is present in the feldspar. It is, therefore, to be concluded that the actual development of the abundant biotite in the rock leaves available sufficient silica to form albite rather than nephelite. A rough calculation of the mode, in which the pyroxene is figured as ideal diopside and the biotite as one having the ratios MgO FeO SiO, _ K.O a Ona and eee although not accurate enough to give the mode of the rock, shows that, when biotite is allowed for, the readjustment of the constituents locks up suflicient potash and disengages enough silica to raise the remaining lenads to feldspar. 342 Palache and Warren—Krihnkite, Natrochalcite. Arr. XX XIX.—A7rohnkite, Natrochalcite (a new mineral ), and other Sulphates from Chile; by Cuarues PatacuE and C. H. Warren. THE minerals briefly described in this paper* were sent to the Harvard Mineralogical Laboratory for identification by the Foote Mineral Co. of Philadelphia, whose manager, when the scientific interest of the material was pointed out, at once placed at our disposition all of the material in his possession with- generous permission to use whatever was necessary for the investigation. The collection comes from the mining district of Chuqui- camata in the Province of Antofagasta, Chile. It was obtained from exhausted copper veins and includes the following species: krohnkite, natrochalcite (a new mineral), blodite, brochantite, atacamite, chalcanthite, copiatite, botryogen, sideronatrite, halite and gypsum. Krihnkite. Krohnkite is the most abundant mineral in the collection and appears in three distinct habits, as follows: * A more extended crystallographic description of this material will appear hortly in Zeitschrift ftir Krystallographie. Palache and Warren—Krohnkite, Natrochalcite. 3438 Phase a.—Clusters of octahedroid crystals of the type of figure 1 but mostly in twin groups, the erystals firmly aggre- gated to a highly cellular mass, largely infilled with an earthy yellow iron sulphate which may be copiapite. These crystals reach a diameter of 1°5°" and are of a dull greenish blue color with smooth but lusterless faces. Phase b.—Single crystals and fibrous or acicular aggregates of pale blue. color, implanted on the white quartzose vein material. The crystals are slender prisms with the forms of figure 1 but with the prism largely developed and its planes much curved and facetted through the presence of steep vicinal pyramids. Single crystals reach a length of 4™. Phase c.—Solid crusts up to 2” in thickness of deep vitriol blue color, the crystals composing the mass often large and either short or long prismatic, with the forms of figure 2. In cavities on the surfaces of such crusts is a second generation of prismatic crystals of pale blue color, beautifully crystallized and showing the complex combinations of figures 3 and 4. Twin crystals of the type shown in figure 5 are also found on this deep blue material. The position adopted for the crystals differs from that given by Dana, front and back being interchanged. The axial ratio calculated from measurements on a number of crystals is Gane O22) bh 204357 8 = 56° -17'°20" The observed forms are as follows :—a(100), 6(010), (110), h(120), &(180), e011), d(021), f(031), ¢ (101), w(802), v(801), p11), 7121), g(111), s(121), w(211), #(221), 2(831), 7(551), 0(10°10°1), 7(232), n(132). (Tables of measured and calculated angles and combinations will be found in the paper cited above. Twinning.—The twin plane is the base, (001). Twins are either contact or interpenetrating, the latter resembling parallel growths owing to peculiarities of angles and distortion. 344 Palache and Warren—Krohnkite, Natrochalcite. Cleavage.— Cleavage is perfect and easy parallel to 0(010) and good but not so easily produced parallel to c(011). No trace of a prism cleavage as recorded by Darapsky could be detected. Hardness is a little less than 3, just scratched by the fingernail. Specific gravity is 2°06i (Warren), determined in absolute alcohol and calculated for water at 4°C. Optical Characters, determined by H. Eb. Merwin.—The principal indices of refraction, determined by means of the refractometer, are : a = 1°54387, 8 = 1-57715; y = 1600s tee sodinm light. 2V,, calculated from the refractive indices is 78° 36’; from observation of the acute optic angle in ei] 78°42’, The plane of the optic axes is in the plane of symmetry, with the acute bisectrix for yellow (ether-axis @) inclined 48° 45’ to-the erystallographic axis ¢ in the obtuse angle 8. The dis- | persion, as determined by the colored hyperbolas of imterfer- ence figures, is inclined. The acute bisectrix for blue is nearer ¢ than the bisectrix for red. The optic axes also are slightly dispersed, more for blue than for red, as indicated by broader color fringes on the hyperbola emerging nearly per- pendicular to ¢. Chemical Composition, with analysis by C. H. Warren.— Analysis of the very pure material available confirms the com- position of the mineral as given by earlier writers. CuSO,.Na,SO, +2H,0. Most of the water is given off below 150°. Small additional amounts continue to come off up to 350°, when dehydration is complete. The residue may be brought to complete fusion without further decomposition, yielding a bright green enamel. Per cent. Mol. Ratio. Theory. a ee SS, CiOe 325 "292 “98 23°49 Na,O 18°89 °304 F-O02 18°39 SO, 47°60 "595 2°00 A744 H,O TORT "095 2:00 10°68 Atacamite trace 100°46 100°00 Paragenesis—Krobnkite is the most abundant sulphate in these specimens and the first to be formed. Atacamite alone of the few associated minerals may be older, thin crusts of it sometimes lying between the krohnkite and the vein matrix. Crystals of kréhnkite also show occasional inclusions of copiapite, brochantite and atacamite; none of the other minerals mentioned above as occurring in the collection 1s found with krohnkite. Palache and Warren—Krohnkite, Natrochalcite. 345 Natrochalcite, a new mineral. Bright emerald-green crystals of what proves to be a new hydrous double sulphate of copper and sodium were found on several specimens. The crystals are either isolated or in closely adhering crusts upon the white vein matrix; in one specimen they are embedded in chaleanthite and doubly terminated erys- tals were obtained by carefully breaking away the enclosing blue vitriol. The mineral is monoclinic with a striking pyra- midal habit, shown in figure 6; the crystals attain a length of about 1™ and are generally attached to the matrix in such a way that portions of both prism and pyramid are developed ; an oscillatory striation parallel to intersection edge of these forms, due to the development of steeper pyramids between them, is generally well marked. A small basal pinacoid and the other faces shown in figure 7, but of minute size, are generally present. The faces proved to be of good quality in most cases and the measurement of six crystals, mostly very small and one of them doubly terminated, yielded satisfactory data for calculation of the elements. For this purpose 45 faces of seven forms were available. p= "8526 —— 1:065 = (arilecs ea vey Oleg from which was calculated G20 76 = 1493 0-314 8 — 61° 17’ 30” 346 Palache and War ren—Khrohnkite, Na trochulcite. The table contains the forms found, the angles calculated from — the elements and the observed angles with their range of vari- ation. It will be seen that the calculated and observed angles show a very satisfactory agreement. 3 Natrochalcite.— Table of Angles. Calculated. Measured. Limits. ¢ CR aa PS SSS SSS — No. p p d p p pag of ped Nb eter C1001) 90200) 82494905 00 FOS aA eee eee 28°39’ 23°49’ 3 6-010 00: 00: 90-00-00 00° 90 00 222) 2 eee 4 mz 110 88 41. 90°00 $8.41 90 00:.38°40'—382%43" eee 12 p 111 51 23 6247 5123 62 47 51 15-51 31 62°45 consol ® 112 59 35 50 10 59 85 50 09 59 28-59 44 50 07-59 12 3 u 221 45 45 78.58 4555 78 39 45 41-4601 73 14-74 11 5 w 831 48°34: 78 45 48-35 78 55.0.2 1 qg 111 19 16 5208 19 16 5208 1912 -19 20 52 06-52 10 7 w 521 39 54 7021 35 56 7017 29 50-30 04 70 13-70 21 5 The habit of the crystals is always dominated by the forms chosen as prism. and pyramid, mand p,; all other forms are developed with but small faces and often without the full number of their faces. Cleavage._—Cleavage is perfect and easy parallel to the base, (001). Hardness 4°5, scratching fluorite easily and not scratched by it. Specitie oravity 2°33, determined by Warren. Optical Characters determined by H. L. Merwin.—The principal indices of refraction, determined by means of the retlectometer, and the optic angle for sodium lght, are as follows: a = 1°6491, 8 =1°6555, y = 1:°7148. 2V,,, calculated from the refractive indices, 36° 52’; by observation of obtuse optic angle in oil, 36° 48’. The plane of the optic axes is in the plane of symmetry, the acute bisectrix for yellow being inclined to the crystallographic axis ¢ 12° in the acute angle 8. The acute bisectrix is the axis ¢; the mineral is therefore optically positive. Dispersion of the optic axes is strong, the acute optic angle for the strongest blue rays transmitted by cobalt glass being 3° greater than the corresponding angle for yellow. There is also a ) slight inclined dispersion of the acute bisectrix, that for blue lying ~ nearer ¢ than that for red. Chemical Composition._-The composition of the mineral may be expressed by the formula Na,SO,.Cu,(OH),(SO,), + 2H,0. Palache and Warren—Kréhnkite, Natrochaleite. 347 The water is given off gradually on continued heating above 150°. The mineral decomposes and gives off SO, between 350° and incipient redness. Before the blowpipe it decrepi- tates and fuses very easily (about 1) toa black bead. Gives off acid water in closed tube, fusing to a dark enamel. It is very slowly dissolved by water and easily by acids. The analysis which follows was made on less than one gram of material and is not wholly satisfactory to Dr. Warren, the analyst; lack of available substance, however, except at the expense of one of the two remaining specimens, made it seem well to publish it as it stands, subject to revision later should more of the mineral be discovered. Per cent. Mol. Ratio. Theory. CuO 41°95 "528 4-00 42°08 Na,O 8-44 136 1:03 8-24 SO, 42°10 "326 4-00 42°51 H,O 770 427 3-23 717 Insoluble rés. aft : Clfrom atacamite +05 100-94 100-00 Paragenesis.—Natrochalcite does not occur with kréhnkite in these specimens, but takes its place, bearing the same age relations to atacamite and brochantite which occur sparingly with it. As above mentioned, it is embedded in chaleanthite in one specimen. Llodite.—Bloédite was identified by the following analysis. It is a massive granular form of the mineral, white where not stained blue by chaleanthite or pale green by finely divided atacamite. It showed no trace of crystalline form. In one specimen it was accompanied by halite and krohnkite. Composition, analysis by C. H. Warren: Per cent. Mol. Ratio. Theory. MeO 12-00 300-100 11-48 Na,O 18°20 "296 0-98 18°56 SO, 47-49 593 1:98 47-90 HO 21-60 1:20 4-00 21°56 Insol.-atacamite and quartz, 50 99°70 © 100-00 leading to the usual formula, MgSO,.Na,SO,+4H,0. 348 Palache and Warren—Krohnkite, Natrochalcite. Of the remaining minerals listed on the first page as occur- ring in this material there is little to note of special interest. Brochantite is sparingly present in acicular crystals implanted on or surrounded by kréhnkite. Atacamite is in green and deep blue-black crystals of ordinary prismatic and tabular habits; also in erystals elongated parallel to the brachyaxis with nearly equal development of the forms m, e, 7 and n. Chaleanthite is in the form of granular crusts, copiapite and botryogen in granular masses of no distinct form and sideron- atrite in yellow needles. Halite in small cubes was present on | one specimen of blodite and gypsum is shown in a coarse granular form saturated with finely divided hematite and also in aborescent crystallizations of snow-white color, closely resem- bling cave-formations of calcite. Cambridge, June, 1908. F. E. Wright—Measurement of Extinction Angles. 349. Arr. XL.—On the Measurement of Extinction Angles in the Thin Section; by FrRep. Eugene Wricut. 1. The Measurement of Extinction Angles of Minerals in the Thin Section. The Adjustment of the Petrographic Microscope with Special Reference to the Measurement of Extinction Angles. 3. A Device for Holding Minute Crystals. ) 1. The Measurement of Hxtinction Angles of Minerals in the Thin Section. Tue petrographic microscope as an instrument has under- gone many changes and modifications since its introduction nearly forty years ago, but from the very first, each improve- ment has tended to increase its efficiency in such a way that the optical features of minerals in the thin section can be ascertained more readily and more accurately. In the hands of geologists, the microscope is merely the means to an end—an apparatus to aid in recognizing the minerals composing a given rock (mineral composition) and the relations of such rock components to each other (rock texture); and for such purposes the modern petrographic microscope is admirably adapted, especially since, as arule, approximate results only are required. But with the increased knowledge of rocks thus attained, the demand for data which are precise and quantitative in character rather than qualitative has become more imperative, with the result that, at the present time, one expects to find in a thorough petrographic investigation accurately determined optical con- stants of each of the rock-forming minerals examined, and in critical points, the probable. error of each determination given. This passage from qualitative to quantitative work implies consequences of profound importance; an additional burden is placed on the working geologist, and the time and energy re- quired for the investigation of a given problem is much greater under the present régime than formerly; at the same time, this transition indicates that, in one phase of geology, at least, the step from the lower first plane of preliminary reconnaissance work to the higher level of precise and detailed work is being taken. The optical properties which are made use of in the practical determination of minerals under the microscope are, briefly: refractive index, birefringence, optical axial angle, optical character, extinction angles, color and pleochroism. By means of these properties alone it is possible to ascertain the crystal system to which a given mineral belongs, and by a short pro- cess of elimination to determine definitely the mineral in ques- Am. Jour. Sc1.—Fourts Serres, Vout. XXVI, No. 1o4. —OcTOBER, 1908. 25 350 FEL Wright—Measurement of Extinction Angles. tion. This process has been carried to such refinement in certain instances, as in the isomorphous series of plagioclase feldspars, that it is now possible, from extinction angles alone, to deter- mine very closely the actual chemical composition of the par- ticular plagioclase in hand. For a given mineral plate in the thin section, the term extinction “angle signifies the angle between a known crystal- lographie dir ection (cleavage line, or trace of a crystal face on that plate) and one of its optic ellipsoidal axes or directions along which it extinguishes when these directions are parallel with the principal planes of the crossed nicols. In order to ascertain this angle satisfactorily one must be able not only to measure plane angles accurately, but also to locate correctly the position of the optic ellipsoidal axes of the particular crystal plate. The first condition is easily accomplished and demands no special comment, while the second requirement is extremely dificult to meet with any degree of satisfaction without great expenditure of time. Many methods have been suggested by which the position of the optie ellipsoidal axes of a given crystal section can be located more or less exactly, and all are based on the fact that when the optic ellipsoidal axes are parallel with the principal planes of the crossed nicols the plane polarized light normally incident from the lower nicol passes through the erystal plate without changing its plane of vibration. "In case the optic ellipsoidal axes of the plate do not coincide with the planes of the nicols, interference in general takes place and some light passes through the upper nicol. The different methods pro- posed have for their common object the rendering apparent the extremely small percentage of light which thus emerges from the analyzer when the angle of revolution of the crystal plate from its position of absolute darkness is very small. Before considering in detail the different methods for accom- plishing this result and their relative merits and defects, it will be well to treat the subject mathematically and to derive the formulas for the intensity of light with special reference to the subject of extinction angles. This treatment is given in some detail in the following paragraphs, since the deductions given later are all drawn from these fundamental equations. Theorétical. Mathematical—The phenomena of light are considered to be produced by periodic changes or disturbances in the ether, transverse to the direction of propagation. Different hypoth- eses have been proposed which assign different properties to this medium, but no one of the hypotheses yet suggested is FE. Wright—Measurement of Extinction Angles. 351 satisfactory in ail its details.* or the purposes of this paper, however, these disturbances may be considered vibrations of ether particles about positions of rest and in a plane normal to the line of wave propagation. Adopting this. view chiefly as a matter of easy expression, we may assert that in plane polarized light the disturbances or vibrations are confined to a plane, each particle vibrating then with simple periodic motion, to and fro, pendulum like, along a straight line. An equation which satisfies such a periodic vibration and which has been found to represent satisfactorily the ether disturbances, is the following: Say aT y = Sin a (¢—7,) (1) . 3 : Se AE 2rt in which @ represents the amplitude, T the periodic time, - the initial phase, ¢ the time which has elapsed at any given instant, and ¥ the distance of the ether particle from its posi- tion of rest. The velocity of the ether particle at any instant is given by Qa Ase Some a Cle wre COSH opine aE eae tae re) and if 72 be its mass, its kinetic energy is 1/2 mv’. 27a aie 9 accordingly the average kinetic energy during a complete vibration will be 1 T mor {r- ae af P2 modi = oe on ae dt 1@) () mar (°T dor Ma mr bat T a TS COS se (SI YG a (er | ns _f ( sl 4 fe ( :) pe [ le Aa Tv ( ) fl 1 TG, in. = Now, a vibrating particle possesses at any instant a definite amount of kinetic energy and also a definite amount of poten- tial energy, and the sum of these two amounts of energy remains constant throughout the vibration, so that as the kinetic energy in the particle increases its potential energy The above equation shows that v varies from 0 to * These theories, as well as the mathematical treatment of the same, are given in the standard text-books on light (Preston’s Theory of Light; P. Drude, Lehrbuch d. Optik, and many others; also Rosenbusch-Wiilfing, Mikrosk. Physiagraphie i, 1; and Duparc and Pearce, Traité de Technique Minéralogique et Pétrographique). 352 FE. Wright—Measurement of Extinction Angles. decreases, and vice versa. At the moment the kinetic energy of the particle becomes zero, the total energy is potential, and similarly, when the kinetic energy attains its maximum, the potential energy is zero. In other words, the average poten- tial energy is equal to the average kinetic energy, and the whole energy is twice the average kinetic energy, given in the above expression. A measure for the intensity of hght is the energy per unit volume of the vibrating ether, and if in the above expression for the kinetic energy m stands for the mass per unit volume, the whole energy or intensity will be measured by the expression, Qin a" oe In practice, only relative intensities are encountered. The relative intensities of two light vibrations of equal period (T) at a given point will be in the ratio of the square of their amplitudes : od Da, a 2H One ae I = re on ue ah a aia (2 ) In other words, the intensity of light of a given period of vibration (color) varies as the square of its amplitude (@) of vibration. This relation will now be made use in determining the rela- tive intensities of the plane polarized light waves which emerge from the upper nicol of the microscope, after having passed through the lower nicol and an intervening crystal plate in different positions. Disturbances in the ether which produce light phenomena are ascribed to the action of forces on the ether mass, and if two or more separate disturbances are simultaneously impressed on the same element, the resultant disturbance can be caleu- lated according to the principle of the resolution of forces on the assumption of direct superposition of the forces. If, in the case of plane polarized hght, two separate vibrations be imposed simultaneously on an element, the resultant vibra- tion will also be in that plane, and its amplitude, on the prin- ciple of superposition, is the algebraic sum of the amplitudes of the components. The mathematical expression for the resultant vibration of a particle simultaneously impressed by two periodic disturbances of the same period but differing in phase and amplitude, can be deduced from the equations of the separate vibrations. sin 2a (t—t RG be cA 2) and Y,—4a, sin mune ) Oia FE. Wright—Measurement of Extinction Angles. 358 The resultant displacement . any time ¢ is Y¥=Y, + Y,=4, mee (t—t,) +a, sin ; - (it) ; ar Go 1 1.) Qa ; Hf t.) = sin —t(a sagen a, Coe = os asin a, ainae fe 7.3 The pea sin ort Tes == S07 22 Sy = ay ) if A cos ¢,=@, cos, > vt +4, os 27 t, and : fe Th eae A sin ¢,=a, sin a ¢. +a, sin — t, By squaring and adding the last two expressions, we obtain 2 A’=a,'+a,'+2a,a, cos 7 (¢,—¢,) (3) In this expression r 2 bs, = (t¢,—t,) denotes the difference in phase of the two component periodic displacements and A the amplitude of the resultant vibration. In considering the effects which different crystals exert on transmitted light waves, it has been found, both in practice and theory, that these influences can be predicted accurately and satisfactorily by reference to a triaxial ellipsoid, the optical ellipsoid, the position and relative axial lengths of which vary in general with different minerals, and with the wave length of light employed. Thus the directions of vibration of light waves emerging normally from a mineral plate are parallel with the major and minor axes of the ellipse which a central diametral plane parallel to the given plate cuts out of the optical ellipsoid for the particular mineral and wave length used. The determination of the actual position of these directions in the plate is accomplished in polarized light by observing the relative intensity of the transmitted light for different positions of the plate parallel to the principal planes of the nicols. Light waves emerging from the lower nicol are plane polar- ized and their vibration is given by the equation Qxrt UWE i a On entering the crystal plate, this vibration is resolved into two vibrations in planes normal to each other. If @ (fig. 1) be the angle included between the major optic ellipsoidal exis 354. FE. Wright—Measurement of Hatinction Angles. of the plate and the plane of the incident vibrations, the equations for the resultant waves are Qat | a mae Each of these vibrations traverses the plate with a different velocity and the time required by the fast wave to traverse the plate of thickness d will be ¢, = d.a’, while the time required x—=u cos 6=a cos 6 dae ind y=u sin 0=a sin 6 sin: Tt iL. by the slow wave is ¢, = d.y' where a’ and ¥' are respectively the refractive indices of the two waves. On emergence, therefore, the oe for the periodic ae will be x’ =a Cos 6 sin - 7: 2 (ce ) and y’=a sin 6 ne - =e t— dy?) On reaching the upper nicol each of these vibrations is again resolved further into two component vibrations again normal to each other, one of which, however, is annulled by total reflection. If @ be the angle between the principal planes of the nicols, then the component vibrations emerging from the upper nicol are « f=’ cos (0—¢)=a cos(6—) cosBsin = Ejsodas on sin a — da") n=y sin (6—¢)=asin (O@—¢)sin sin ye (t—dy')=A, sin a (t—dy’) and the resultant amplitude A=£+7=A, sin a (t—da’) + A, ae 7 (t—ay') LE. Wright—Measurement of Hixtinction Angles. 355 ‘The intensity I’ of the emergent wave is then proportional to the square of the amplitude T A? Emer ar The equations (3) above, moreover, ay that AP=A?7+A,°4+2 AA, cos d (y'—a’') On substituting the values of A, and A, in this equation, and noting that ike a.) 9 COs ae A(y —a')=1—2 sin’ d (y we obtain A?=a’ [cos? ¢—sin 2(6—¢) sin 26 sin? 7 d (y'—a’)] (4) 1 2 and finally, Tasos °p—sin 2(8—¢@) sin 26 sin*rd (y'—a') (5) But the velocity of hight V, period of vibration T,and wave length A, are so related that VT = 2 and if we consider the velocity unity, then we may replace T by A and the equation (5) reads: 1 ; . «gE po = os 6—sin 2 (@—@) sin 26 sin” I d (y'—a’) This is the usual expression for the relative intensity of the emergent waves; it may, however, be br ought into more con- venient form for practical purposes. To save space, let sin’ = (y' — a’) = K, where K may have any value from 0 to + 1; then 1 2 T= 8 =? ik sin 26 sin 2(6—$) 21 =1+ cos 26—2K sin 26 sin 2(0—¢) (a) =1+4+cos 246—K (cos 2¢ (1—-cos 46)—sin 2¢ sin 46) _(b) =1+(1—K) cos 2¢ + K cos 2(¢—26) (6) For a given angle @ to find the condition that the intensity will be zero, the equation (@) of the foregoing can be changed to 21, =1+(1—2K sin’ 26) cos 26 + 2K sin 26 cos 26 sin 26 —=1+K, cos 24 + K, sin 2¢ (7) in which K,=1—2K sin’ 26 and K,=2K sin 26 cos 26. 356 FE. Wright—Measurement of Hxtinction Angles. If 31, = 0; then 1+K, cos 26+K, sin 26=0 or 1+ K, cos 26=—K, sin 2¢ squaring 1+2K, cos 2¢6+K,’ cos’ 26= K,” — K,? cos’ 26 abi /( K Kee k 9 paar 1 + se 1 2 2 pak pon PK Kae aa “KCK? i Kt us, Dee omer 2 K’— eee by VeRO a ; (8) In order that cos 26 have a real value, the expression K,*+ K *—1 must be zero or positive. But, K *=1—4 K sin’ 264+ 4K?’ sin* 26 K2= 4 K* sin? 26—4K? sin‘ 26. Accordingly, K?+K"—1= — 4K sin’ 26 (1—K) (9). The right hand of this equation is a negative quantity, and cos 2d can vhave a real value only when K,’+K,’—1=0, and _this condition is fulfilled a when (i) I= 0 Sir onssiine = Sr iyi ch) Olde 7 d (y'—a') =na (2) 01K I o7ssiin ee (yi—a'\=1, 1. @, ae (y—a')=(2+1)5 (3) sins) 20 0 eG eae The value for cos 2¢ then reduces to COs Be = = —K,=—(1—2K sin’ 26) For the three dierent cases the value of cos 2¢ becomes (1) cos 26 = — 1 1 e. 6 = @n+1), (2) cos 26 = — (1—2 sin’ 26) = — cos 46, i. e. 6=2x—26 (3) cos 26 = — 1 i.e. b= (2n + 1)G If the nicols be not crossed, therefore, it is not possible to obtain absolute darkness for a given section unless = 2n + 1 : phere 2 ws Be BUEN! - rs e., unless monochromatic light be used of such a wave length that the one wave is an odd number of half wave lengths ahead of the second, and in this case, d=7—20. If white light be employed, abnormal interference colors will appear because of the abnormal conditions, and at no point will darkness ensue. LE. Wright— Measurement of Kxtinction Angles. 357 The condition that the entire light be transmitted is 1 1=;>! or (1—K) cos 26+K cos 2(¢—20)=1 which is satisfied if both 6=0 and @=0. In case either ¢ or @ be given, the problem of finding the particular disposition of upper nicol or crystal plate for which the intensity of the transmitted light reduces to a minimum or maximum, involves the first partial differential quotients of the function 2I, (equation 6) after either @ or @. If ¢ be given, the point in question is determined by aK fH a 0 . 6=20 or d6=1—20 The second partial differential quotient shows that for the first value of @ the intensity is a minimum, while for 6=7—2@ the intensity isa maximum. ‘This relation is of importance in cer- tain of the methods described below. If @ be given and ¢ is the variable, o(21,) ona ip | . From this equation we find (1—K) sin 2¢—2 K sin 2 (6—26)=0 K? sin’ 40 (1—K)*+2K (1—K) cos 46 + K’ a complicated expression which for K=1 simplifies to sin’ 26= sin® 26=sin"40 and this equation is satisfied for d= 20 d= 7 — 20 It is of interest to plat the values given by the equation for different values of @ and @. 21,=1+(1—K) cos 26 + K cos 2 (6-28). eo) In fig. 2, curve V, the rate of increase in intensity of light is given for the special case of 6 = 0, where simply the upper nicol is turned and the crystal plate has no effect in the polarization of the waves from the lower nicol. In this case 21 = 1+ cos 26 (10) From the curve it is evident that the rate of increase is very slow at first, but rises rapidly and reaches a flexion point at 45°, after which the intensity increases with decreasing rapid- ity to its maximum value at 90°. iS) 8 Lf. LE. Wright—Measurement of Extinction Angles. dicey ae Ov O20" S040. 50 60" 0) aera Fie. 2.—Curves showing relative intensity of light emerging from upper nicol after transmission through polarizer, crystal plate and analyzer, the positions of the crystal plate and also the analyzer ranging from 0°-90°. The abscissa values refer to angular distances of the major ellipsoidal axis of the crystal plate and also of the plane of the analyzer. For curves I-IV, the nicols are considered crossed (¢= >) and the crystal plate alone to be 9 revolved from 0° to 90°. In curve 1, sin? a d(yi—a'!)=K=1>» in eurve if K=1/2; in curve 1, K=1/4; in curve 1V; K=0. ‘Curve V (shomemiae relative intensity of the emerging light for different positions of the analyzer alone ((=0, @ ranging from 0° to 90°). Curves calculated from the general formula J,=1/2 (1+(—K) cos 264K cos 2 (¢—26)). FE. E. Wright—Measurement of Kxtinction Angles. 359 In ease the nicols are crossed (¢= a the rates of increase for different values of K are given by the reduced equation 2], = K (1 + cos 40) = 2K sin’ 26 (11) which defines a curve similar in aspect to the foregoing except that ¢ is replaced by 2@ and the factor K tends to reduce all values proportionately. The curves I—IV of fig. 2 represent the relative intensities for values of K=1, 1/2, 1/4 and 0° respectively. The greatest possible intensity is thus attained when K=1, i. e., when the waves, aiter emerging from the crystal plate, are an odd number of half wave lengths apart (In opposite phase); the intensity is zero for all positions of the plate when K=0, i. e., when the distance between the two emergent waves is a whole number of wave lengths. In figs. 8-6, intensity curves are drawn showing the relative intensity of the emergent light for different positions of the erystal plate (9 usually 0°, 15’, 30’, 45’, and 1°) with the prin- cipal plane of the lower nicol, and for different positions of the upper nicol (¢ ranging from 88° to 92°). The heavy curve in each figure is the relative intensity curve of the crystal alone (nicols crossed, ¢ = 90°) and @ (ranging from —2° to + 2°) The narrow range of intensities only is considered, since in general it represents about the order of magnitude of the prob- able error of a single determination made in the usual manner. In each of the figures the unit ordinate division represents 025 per cent of the total intensity and the unit abscissa division HO Or arc, In fig. 3, K is considered=1 or sin? 2 - d(y'—a’)=1, which obtains when the one wave is any odd number of half wave lengths ahead of the second on emergence from the plate; in figs. 4, 5, 6 and 7, the relative intensity curves are drawn for K=3/4, 1/2, 1/4 and 0 respectively. . It is not a difficult matter to grasp the meaning of these curves, as the following example will show: let it be required to find the percentage of light which emerges from the nico] in the case of a mineral plate of such thickness and birefring- ence that for yellow light the faster waves after emerging from the plate will be precisely one haif wave length ahead of the slow waves (K = 1, fig. 3), the direction of extinction of the plate to make an angle of 30’ (@=30’) with the principal plane of the lower nicol, and the principal plane of the upper nicol to include an angle of 89° 10’ with the lower nicol (6=89° 10’) On the 30’ curve of fig. 3 the ordinate for 89° 10’ is -104 and the relative intensity is therefore -104 of 1 per cent of the total intensity. 360 FE. Wright—Measurement of Hxtinction Angles. BiGs te. Fic. 3.—Intensity curves for crystal plates making angles 0’, 5’, 10’, 15’, 20', 25’, 80’, 40’, 45’, 50’ and 1° with plane of polarizer. Analyzer revolved about axis from 88° to 92° with lower nicol plane. sin? (y'—a’)=K=1 Curves calculated from the formula I= 1/2(1 + cos 2 (¢ — 28) ). The F. E. Wright—Measurement of Extinction Angles. 361 Hie. 4. oL NA 153 Gna Bee th eee oe ANKE SOV AINE A PZ EUS ee OO O gge—gqr ats q[= q2' Fic. 4.—Intensity curves for crystal plates at angular distances of 0’, 15’, 30’, 45’, and 1° from plane of polarizer, the analyzer being revolved about Q7xt (a) =K =38/4, axis from 88° to 92° with plane of polarizer. sin’ Curves calculated from the formula 1,=1/8 [ 4+cos 2643 cos 2 (@—26) ]. As in fig. 3, the heavy curve indicates the relative ‘intensity of emergent light for different positions of the crystal plate ( ranging from 88° to 92°, or —2° to +2°) under crossed nicols (6= =i and K=3/4. Calculated from the formula I, = 3/8 (1+ cos 46) ———— heavy curve indicates the relative light intensity under crossed nicols (@=5) for different positions of the crystal plate (9 ranging from 88° to 92° or —2° to +2°)and K=1. Calculated from the formula 2 1, = 1 + cos 48. 362 FE. Wright—Measurement of Ketinction Angles. These figures 2-6 are well adapted to show graphically cer- tain facts which are evident from a mathematical consideration of the intensity formula. (1) If K = 0, which occurs when the one wave is any number of whole wave lengths ahead of the second, the erystal plate is dark and remains dark for all positions of revolution, as indicated by the heavy abscissa line Inne i), des Fic. 5.—This figure differs from the two preceding figures only in the value of K, whichis 1/2. The curves were calculated from the formula I, = 1/4 (2 + cos 26+ cos 2(@— 280) ). The heavy curve from the formula I, =1/4 (1 + cos 40). of fig. 7. (2) In case K = 1/4, fig. 6, the intensity comvemen erystal plate, coincides very closely with that for the revolving nicol. The extinction curves, moreover, for the crystal plate at different angles (@ = 15’, 30’, 45’, and 1°) with the principal plane of the lower nicoland for different positions of the upper nicol (@ ranging from 88° to 92°) are similar and lie close together, so that, in this particular case, methods involving the F. EL Wright—Measurement of Hxtinction Angles. 363 revolution of the upper nicol for the location of zero intensity directions are not greatly different in their degree of accuracy from those in which the nicols remain crossed and the erystal plate is revolved. Nevertheless, even in this instance the former are the more sensitive methods and results attained by their use are correspondingly more accurate. For K=1/2, fig. 5, the extinction curve for the crystal plate alone (nicols crossed and plate only revolved) no longer coincides with that for the upper nicol alone, but similar conclusions can be drawn as to the relative sensitiveness of the two methods, the one involvy- ing the revolution of the crystal plate (while the nicols remain crossed), and the second, the revolution of the upper nicol while the crystal plate remains stationary. The amount of light O gg IS Sa q2e Fic. 6.—Differs from fig. 5 only in K, which is 1/4. The curves are expressed by the formula I,=1/8[ 4 +3 cos 2¢ + cos 2(¢ — 29) | and the heavy curve by I, =1/8 (1 + cos 46) which is required to produce the sensation of light in the human eye is different for different persons. But for a given eye the limit of the actual sensation of monochromatic light is fixed for any particular instant and may be represented by one of the horizontal percentage lines of the figures. Let us assume that for a source of monochromatic light of definite intensity I, the limit for the sensation of light is ‘050 per cent of the 364 FL. Wright—Measurement of Kxtinction Angles. total intensity and represented by second horizontal line above the base line of fig. 5. Then the curve for the crystal alone shows that for all points below that line, 1. e., between 89° 04° and 90° 56’, the crystal will appear absolutely dark and on a single determination an error of nearly +1° may bemade. If, however, the erystal plate remain stationary, and the upper nicol be revolved through small angles from its normal, crossed position (6 =88° to 92°), it is evident from the figure that if, for example, the om ae IN, FALLEN BERREBD RS See ee ca SGEEE 0) O oe aie O ° 66 OF 90 Fie. 7.—In this particular case K is considered = 0 and the general formula reduces to ees (1 + cos 26) which is independent of 6. In other words, if the thickness of the plate be Qqrt such that sin? a (y'—a@')=0, or the emerging waves are any number of whole wave lengths apart, total interference takes place and the plate is dark under crossed nicols for every angle of revolution about its normal axis. crystal plate is 80’ distant from its position of true extinction and still dark under crossed nicols so far as the eye of the observer can detect, the differences in intensity between the field and crystal plate for different angles of revolution of the upper nicol (measured by the ordinate intercepts between the curve 0’ and 30’ of figure), are of such a character that at the point where the illumination of the field can just be observed (88° 43’) the intensity of illumination of the crystal is more than twice as great (106 per cent instead of -05 per cent), whereas - FE. Wright—Measurement of Extinction Angles. 365 on the other side, where first indications of illumination on the central plate can be detected at 91° 41’, the field is lighted up by ‘085 per cent instead of :050 per cent of the total intensity. These differences of intensity are of such a character that they ean readily be observed, and the sensitiveness of any method involving the revolution of the upper nicol while the crystal remains stationary is in this case at least twice as great as that for which the nicols remain crossed and the cry stal plate alone is revolved. Similar theoretical conclusions can be drawn from figs. 3 and 4. If white light be used and the upper nicol be revolved abnormal interference colors result. The rapid and pronounced change in interference colors near the position of crossed nicols, on a plate which is not precisely in the position of true extinction, is well adapted for use in the location of its ellip- soidal axes. With a given color of monochromatic light extinction angles should be determined on plates of such a thickness that K is about +1 (the two emergent waves are a whole number of half wave lengths apart). Thus if sodinm light be used the plates should show in white hght an inter ference color of about straw- yellow of the first order but not sensitive violet, since for this particular thickness the two waves are 589uu apart and the yellow waves are totally destroyed, with the result that the plate appears dark in all positions. It follows, furthermore, that a plate which is well adapted for determinations in one kind of monochromatic light may be useless for another color. It has been found that the insertion between the crossed nicols of specially cut piates and wedges of birefracting sub- stances, as quartz and selenite, is often well adapted to increase the accuracy of the measurement of the extinction angle on a given plate. The principle there invulved is that of the super- position of birefracting plates, the action of which is to pro- duce a resultant which differs from that of either component. It is possible to select a wedge or plate of such a character that the interference phenomena produced by it alone are extremely sensitive to the slight changes produced by a second crystal plate when it is not pr ecisely i in the position of true extinction. From the mathematical standpoint, the insertion of a second plate ‘involves a new set of conditions for the vibrating ether elements and the equations for the resultant are correspond- ingly more complex. Their derivation, however, is exactly counterpart to that for the intensity of a single crystal plate and the final result only need be given here. If the nicols be crossed ard @, be the angle which y,’ of the crystal plate of Am. JouR Scl.—FourrtH Series, Vor. XXVI, No. 154.—Octoser, 1908. 26 SHG 2 ae, Wright—Measurement of Hxtinction Angles. thickness d, includes with the principal plane of the polarizer and @,, the angle between y,' of the inserted plate or wedge of thickness d, and the polarizer plane, the relative intensity is given by the formula, and d,(y,'—a,') = T, and d,(y,'~a,.) ==) then i , : : ener I, = sin 2(0,—90,) sin 26, cos 26, sin = rT, re ; Specs + sin 2(6,—0,) cos 20, sin 26, sin p cs : ae a + cos’ (6,—-6,) sin 20, sin 26, sin = Ce — sin’ (6,—6,) sin 26, sin 26, sin “7 (T,—T,) From this formula the relative intensity can be calculated for any given values of (0,,@,) and T, and T,. In case the crystal plate is of such a thickness that . 2 e ry e sin. T, = 1 and at the same time the inserted plate is also i Heel of a thickness that sin — TT = 1, this equation reduces to _—— = 2 1 = sin 20 — 0) an expression for a curve similar in every respect to those of fi . 5 : : “ip g. 2, but which is zero for 6,=@, and also for 0, =— + @, and As ¢ ° “ TT reaches its maximum of 1 at 0, = = +6, It can also be shown that for a given increment of 0, as d@,, the ratio of the value of the function for (@, + d@,) to its value for @,-is greatest when @, is equal to 0,. If, therefore, the angle 0, be so chosen that the field is just illuminated, the change resulting from small angles of @, will be greater than for any other position of the inserted plate. In the Calderon method described below, the calcite plates are purposely so thick that they show the white interference colors of higher orders in white light, in which ease the thick- ness is so great that for a number of different colors through- out the spectrum the path difference of the emergent waves is a whole number of wave lengths, in other words, in the Calderon method it is permissible for practical purposes to con- sider the plate of such a thickness that for white light the © . TT Set) ° expression sin’ T, is unity, and that therefore the angle @, should be small in order to secure the best results, so small in fact that the illumination of the field is just visible. F. E. Wright—Measurement of Extinction Angles. 367 In several of the other methods cited below for the exact location of the ellipsoidal axes of a given plate, use is made of quartz plates or wedges, cut normally to the principal axis, which rotate the planes of vibration of normally incident, plane polarized light. For the purposes of this paper it is not neces- sary to enter into the mathematical discussion and theory of the rotatory power of quartz, but simply to apply the known laws of rotatory polarization as they were first proved experl- mentally by Arago and Biot on this mineral. A quartz plate perpendicular to the principal axis rotates the plane of normally incident, plane polarized waves, through an angle which is pro- portional to the thickness of the quartz plate and also approx- imately proportional to the inverse square of the wave length used. The rotation effected by two superimposed plates is moreover the algebraic sum of the rotations produced by each separately. By using, therefore, a properly constructed quartz wedge, it is possible to counteract exactly the effect, in plane polarized monochromatic light, of any crystal plate in any given position with respect to the nicols, by rotating the new planes of vibra- tion, determined by the crystal plate back to the original plane of the nicols. In the intensity formula (5), J, = cos’ ¢ = sin 26 sin 2(0 — @) sin’ +a (y' — a) this rotation affects the angle @ only, and y the nicols be crossed, then 1 T, = —sin’ 26sin’° 2 (y'—a') Equation (11, page 359) In all measurements of extinction angles, however, @ is a small quantity and in place of the sine we may use the ele itself without sensible error; accordingly, == KG. (18) This formula, which for cue angles @ states that the light intensity is proportioned to the square of the angle 9, will “be employed later in the description of a new combination quartz wedge for use in determining extinction angles. In certain other methods, convergent polarized light is em- ployed and the disturbing effects of an intervening crystal plate observed whose optic ellipsoidal axes are not precisely parallel with the planes of the nicols. The intensity formulae applying to such conditions are similar to those for plane polarized and the general deductions from the latter may be considered to apply to the phenomena in convergent polarized light. The methods involving convergent polarized light, however, have several 368 LL Ek. Wright—Measurement of Extinction Angles. important defects which render their general application eum- -bersome and unpractical. In the foregoing pages, the intensity formulae for light trans- mitted by crystalline plates under different conditions have been dey eloped and the attempt has been made to treat the subject in such a way that the results attained shall be directly applicable to the practical methods for determining extinction angles under the microscope. In the following sections, the different methods for accomplishing this end will be deseribed, with special reference to their general applicability and rela- 3 tive accuracy, and the conclusions reached in this mathematical part will be used constantly as criteria of fundamental import- ance. Methods. Extinction angles can be measured either in plane polarized light or in convergent polarized light; and in plane polarized light the exact location of the positions of zero extinction 1s fixed, either by observing relative intensities of monochromatic hght ‘under special conditions or by means of the interference colors resulting from the use of white light. In all measure- ments of extinction angles it is imperative that careful attention be given to the source of light, especially if monochromatic light be used. ‘The source should be as intense and steady and uniform as possible in order that the variation in the source of light itselt be not mistaken for actual differences in the micro- scopic field. The rays of light incident on the preparation should, moreover, be as nearly parallel as it is possible to obtain them. To meet these requirements satisfactorily requires both time and patience, but in: order to attain the best results they cannot be overlooked. The microscope, moreover, should be in perfect adjustment, the optical system should be accurately centered and the cross hairs in the ocular should be precisely parallel with the prin- cipal planes of the crossed nicols. The adjustment of the microscope is not a difficult task to accomplish if suitable apparatus is at hand, and will be discussed briefly in part 2 of this paper. Assuming the microscope to be in perfect adjustment and the source of light satisfactory, we may use any one of the following direct ‘methods for measuring the extinction angle of a particular crystal plate: Parallel Polarized Light. (1) Phe ordinary method, which consists in turning the erys- tal plate under crossed nicols until the position of maximum LE. Wright— Measurement of Extinction Angles. 369 darkness is attained. This method is in general use and is equally well adapted for white light and for monochromatic light. With it any degree of accuracy can be attained provided a sufticient number of measurements be taken to reduce the bable error. In applying this method it is customary to note not only the positions of maximum darkness attained by the erystal when revolved clockwise from a position of bright illumination, but also when revolved counter clockwise from such position. This was the method used by Max Schuster* in his classic measurements of the extinction angles of plagio- clase feldspars. He determined for each cleavage flake the position of zero extinction eighty times for clockwise revolu- tions of the plate and eighty times for counter clockwise revolutions, and averaged the two readings. His work in this line remains unsurpassed, even to the present time. To increase the accuracy of each determination on a crystal plate under crossed nicols, different schemes have been devised, all of which depend on the disturbing influence of the plate on inserted plates or wedges of bir efracting substances. Hach of the inserted plates or wedges i is constructed in such a way that the interference phenomena which it presents are markedly influenced by the slight disturbing effects from the crystal plate when it is not precisely in its position of zero extinction. Sensitive Tint Plate-—Plates showing this interference color (violet of the second order) are usually made of selenite or quartz and are under certain conditions very sensitive to the slight changes which the erystal plate produces when it is not precisely in its position of total extinction. As a general rule, the eye 1s more sensitive to slight differences in color than in intensity, and in certain cases “the sensitive tint plate can be used to advantage to increase the accuracy of the ordinary method. Its use is most effective on and practically limited to colorless plates showing low interference colors of the first order. Its efficiency is seriously impaired in the case of deeply colored minerals which veil the true interference color and also in thick plates of strongly birefracting minerals showing high interference colors, even red of the first order. It can, moreover, only be used with white light and accordingly cannot take cognizance of the dispersion of the bisectrices in the mono- clinic and triclinic systems. This method is therefore not of general application and can be employed to advantage only under specially favorable conditions. Bravaist-Stobert plate-—TVhis plate is also cut to show the sensitive violet interference color, and consists of two such * Tschermak’s Min. Petr. Mitth. v, 189, 1882. + Comptes Rendus, xxxii, 113, 1851 ; also Pogg. Ann. xevi, 897, 1855. } Zeitschr. Kryst. xxix, 22-24, 1898. 370 LL EL Wright—Measurement of Extinction Angles. plates in combination instead of a single one. A single sensi- tive tint plate of mica or quartz is taken and cut along a line at 45° with the directions of extinction; the one half is then turned through angle of 180° and the two halves reeemented as indicated in the figure. By this combination plate, which is placed in the focal plane of the ocular, the interference color is made to fall in the one half and to rise an equal amount in the second, thus doubling the sensitiveness of the single plate. Hire. 8; This plate is intended for use only in white light, but under cer- tain. conditions it may serve to good advantage mm monochro- matic light. 3 The Combination Wedge.*—On the principle of the Bravais- Stober plate, the writer has had a combination wedge prepared in which the interference colors range trom total darkness to green of the second order. This wedge was made by taking *This wedge was prepared with great care by Voigt & Hochgesang, Gottingen, Germany (cost, 48 mks.), and the writer desires to express his appreciation of the interest taken by the firminthe same. The compensation on different ends of the wedge, however, proved to be of unequal value, with the result that although the dark zero interference bands were precisely adjacent, the interference colors near the ends of the wedge did not coincide exactly. This defect could be eliminated by combining two quartz plates (45™™" long by 5™™ wide ana of such a thickness as to show interference color green-yellow second order, the ellipsoidal axis ¢ of the one to be parallel to the long direction, while in the second q is parallel), with two wedges of the same pitch (45™™ long by 5™™ wide and ranging in interference colors pale gray of the first crder to violet gray at the top of the third order, and like- wise the ellipsoidal axis ¢ parallel to the direction of elongation in one and q to the same direction in the second) ; the wedge of long direction ¢ to be com- bined with the plate of long direction q. In this manner the plate and wedge compensate in the center of the wedge and the interference colors rise to about blue green of the second order at both ends. e FE. EE. Wright—Measurement of Extinction Angles. 371 an ordinary combination wedge* showing the zero interference band exactly in the center and green of the second order on each end, cutting the same longitudinally in half parallel to the ellipsoidal axes; the edges were then polished and the halves again recemented, the one half, however, having been rotated first thr ough 180°, so that in the resultant combination by wedge the phase difference of the adjacent half at any point of insertion is always equal and opposite. By this method the principle of the Bravais-Stober plate is extended to cover interference colors from total darkness to blue green of the second order, and to allow the observer to select an interference color which, in combination with that of the mineral plate examined, is most sensitive. The low gray tints of this wedge (particularly the dark band region on both sides of which the interference colors rise and thus divide the field into four quadrants and produce an effect similar to that of the Bertrand ocular) have been found specially useful with minerals showing interference colors from red first order to blue second order. This wedge is held in a brass carriage, which in turn slides in the wedge holder shown in fig. 10, and is viewed by the Ramsden ocular. Calderon; Calcite plate-—This plate is also placed in the focal plane of the ocular and consists of two calcite plates placed side by side and so cut that the direction of extinction in each plate makes an angle of about 34° on opposite sides of the common line of junction. The plate is so thick that the interference color is white of the higher orders and when used alone without intervening crystal plate, lights the entire field under crossed nicols with a dull gray tone. If a crystal plate whose lines of extinction do not coincide with the principal nicol planes be then observed, the field appears divided into two unequally illuminated halves and only when the extinction directions coincide with the nicol planes is the intensity of illumination in both halves equal. Calderon claims an accuracy of +2’ with this ocular, but for a single determination and for general preparations the probable error is considerably larger *Compare F. E. Wright, Tschermak’s Min. Petr. Mitth. xx, 233-306, 1901 ; also Jour. Geol., 33-35, 1902. + Zeitschr. Kryst., ti, 70, 1878. —The calcite twin plates of a Calderon ocular from R. Fuess in Steglitz were tested by the writer and found to be inaccu- rately ground. The plate was 3:18™™ thick and cut at an angle of about 45° with optic axis. The extinction angle in each half of the plate was measured in convergent polarized light by means of the dark bar in the center of the field and found to be + 4°4° on the one half and 3°2° on the other. Extinc- tion angles measured with this ocular, using the junction line of the plate as the line of reference, would therefore be out 0°6° from this sourcealone. The field of the ocular is. moreover, small and unfavorably lighted because of the thickness of the plate and of the wide dark junction line across the center of the field, which in turn disturbs the exact matching of the halves of the field. 372 FE. Wright— Measurement of Extinction Angles. (10’-15’). The principle on which this method is based is evi- dent from the intensity formula, for in case the ellipsoidal axes of the plate do not coincide precisely with the principal nicol planes, they make unequal angles with the optic ellipsoidal axes of the calcite (in the one half, this angle is 34° + @ and for the second 84° — @) and this produces at once a caeed difference in intensity of illumination. Quarter-undulation plate of H. Traube.t-—This plate con- sists of two adjacent quarter-undulation mica, plates so cut that Hire. 9) the optic axial plane of each includes an angle 33° with the common line of junction and for slight deviations of a crystal plate from its true position of extinction, the two halves appear unequally lighted, and only when the crystal i is precisely in its position of zero extinction do the halves show the same intensity of illumination. Twinned Selenite plate.—The use of a twinned selenite plate has been recommended recently by E. Sommerfeldtt for the accurate adjustment of the ocular crosshairs to the planes of the nicols. But the same twins can be made to serve admir- ably in the measurement of extinction angles. The extinction angle which the ellipsoidal axis makes in each plate with the twinning plane is 873°, and if the twinning line on such a plate be turned to the diagonal position with the crossed nicols, the extinction angle on ‘each side of the nicol measures 45°— B75 = 74°, but in the opposite halves different optic ellipsoidal axes are adjacent the principal nicol plane. The net result of this arrangement is a change in intensity dependent not only on * Neues Jahrbuch, 1898, i, 251. + Zeitschr. f. wissensch. Mikroskopie, xxiv, 24-25, 1907. EF. EL Wright—Measurement of Extinction Angles. 373 Hires 10; Fic. 10.—Mikroskop-polymeter of Voigt and Hochgesang carrying wedge holder H and also device described below for holding small crystals. The wedge holder in turn carries a cap nicol, K, and a Ramsden ocular in the focal plane of which the wedge, W, isintroduced. The crystal-holding device con- sists of the following parts: S, the stand; C, graduated vertical circle with 1° divisions: A, centering plate; B, ball and socket adjusting device; D, crystal holder; L, plano-concave lens for holding drop of liquid of same refractive index as crystal and held by stand, E. The mikroskop-polymeter is a useful instrument in many ways but for accurate work the mechanical workmanship on the same leaves much to be desired. The fine adjustment screw is practically worthless for even approxi- mate readings, the accurate adjustment of the different circles to the common optic axis is not possible, and even if once accomplished does not remain in adjustment: the divisions on the different verniers are inaccurate; many other defects have been felt by the writer and several of them remedied in the workshop of the Geophysical Laboratory.— A stop diaphragm M has been introduced into the upper tube, also an optic axial angle reflector. R (compare F. E. Wright, this Jour., xxii, 19-20, 1906); a holder for the combination wedge is shown at T, which fits on the objective supporting arm and is revolvable about the axis, an arrangement which has proved advantageous in place of revolving the microscope stage when the interference plates are inserted.—The fine adjustment screw P has been lengthened so that it can be reached from either side of the microscope. The base of the stand at N has been milled out to allow space for the reflector when the condenser lens is lowered. 3874 FL he Wr ight—Measur ement of Extinction Angles. the angle but also on the different compensations of the path differences in the two plates, and if white light be used this results in a rapid change in interference color in the two halves if the crystal be only a small angular distance from its position of true extinction. The writer has had such a plate cut showing the sensitive tint and also a wedge, so that on insertion different interference colors, or intensities in monochromatic light, can be used for which the eye under certain conditions is most sensitive. These plates, as well as the preceding, are inserted in the focal plane of the objective and the junction line serves for the vertical crosshair. For such plates the Ramsden positive ocular has been found by experience to be best suited and a specially con- structed holder convenient. Artificiully twinned Quartz plate.—Still another. advan- tageous arrangement can be had by cutting on a polished quartz plate parallel to the principal axis a ver rtical edge making an angle of about 3°-6° with the principal axis. The quartz plate is then divided transversely to the polished edge and the polished edges cemented together, thus producing an arti- ficial twin whose two halves extinguish at equal and oppos- ite angles from the common line of junction. Such plates may then be ground to a thin plate showing either the sensitive tint or dull gray of the first order or to wedge form, thus increas- ing the range and usefulness of the device. All of the preceding plates, the Bravais-Stober, the Calderon and the Traube, the selenite twin plate and the quartz combina- tion plate wedges of the last paragraph, can be made somewhat more sensitive by dividing the field into quadrants instead of halves, after the example of Bertrand in his rotatory polarizing quar tz plates described below. Bi-nicol ocular.—In the practical application of these differ- ent types of plates the angle @ has been small (2—4°) and found to furnish good results, but in each case there is a par- ticular angle @ which is best adapted for the observations; the limit of sensitiveness of different eyes introduces, moreover, a variable element of such wide range that the angle @ cannot be calculated and fixed once for all. In order therefore to have control over all angles @ and thus in each instance to be in position to proeure the best possible conditions, the writer has had constructed the following ocular attachment. The principle of construction of the apparatus is apparent from the figure, and need’ not be expressed at length at this point. The light after passing through the lower nicol and the erystal plate reaches the lower reflecting prism pair of this ocular and passes thence through appropriate nicol prisms (Thompson prisms) or bir efracting plates of exactly the same BOL. Wright—Measurement of Extinction Angles. 375 character, and after total reflection in the upper prism pair is again brought back to the common field of vision and viewed by the Ramsden ocular and upper nicol. The nicol prisms or birefracting plates fit in collars and can simultaneously be revolved about the axis and in opposite directions so that the angle @ can be made to vary from +90° to —90° in each plate and at any instant @ of the first plate is equal and opposite to @ Innes 1. TET | ame | | | Y \ 1 te Fie. 11.—Bi-nicol ocular; consists of the following parts: Two pairs of reflecting prisms, P, and P», (ground specially for the purpose by Steeg & Reuter of Homburg v. d. Héhe, Germany); two tapering revolving brass holders, M, into which either two Thompson prisms, N, or birefracting plates are introduced. These conical brass carriages are revolved in opposite directions and through equal angles by means of the worm thread S and grooved wheels T, the angle of revolution being read off directly on the head H. The bi-nicol ocular fits in the microscope as an ordinary ocular and into it in turn a Ramsden ocuiar is introduced and above this the cap nicol. of the second plate. In the mechanical construction of this apparatus, special care has been taken to make the angular movements of both plates exactly equal and opposite. With this revolving bi-nicol ocular, it is thus possible to allow any proportion of the light incident on the crystal plate 876 EL. Wright— Measurement of Extinction Angles. to pass through the upper nicols by simply revolving the same and at every instant to state what percentage is passing through. The particular angle of revolution @ for which the intensity of light transmitted is best snited for the maximum sensitiveness of the eye of the observer can be readily ascertained, and the actual position of extinction for any given mineral plate be determined by its use. By means of the bi-nicol ocular, the adjustment of two crossed nicol prisms ean also be tested accu- rately and easily. This ocular, although serviceable, suffers from one defect which it is difficult to overcome satisfactorily, namely, the depolarizing of the total reflecting prism pairs on light waves transmitted when the planes of ‘the revolving nicols are not parallel with the planes of the polarizer and: analyzer. Asa result a certain amount of false light is introduced into the field and tends to veil the sharp contrast of the two halves and thus to decrease the sensitiveness of the instrument. Bertrand plate—In place of birefracting plates, whieh introduce an entirely new set of conditions in the path ‘of heht waves and which complicate the expression for the relative intensity correspondingly, Klein* and Bertrand} have used the rotary power of quartz plates, cut normal to the principal axis, on the plane of polarization of normally incident, plane polarized waves. As shown above, the total effect of such a quartz plate in monochromatic light i is me to increase the angular distance @ in the intensity formula (6). This power of rotation of quartz varies with different wave lengths and with the thickness of the plate. If white light be used, inter- ference colors result. A quartz plate 7-50" thick shows the sensitive tint under crossed nicols and can be used to good advantage in measuring extinction angles, since for slight deviations of the crystal plate from the position of true extinction its interference color rises | or falls. ‘To inerease its sensitiveness, Bertrand combined two plates (2™" in thickness) of right-handed with two plates of left-handed quartz, so that each right-handed plate is adjacent to a left-handed p slate. This plate is inserted in the focal plane of the ocular and the sharp junction lines serve as crosshairs. The Bertrand plate can be used in monochromatic light, pro- vided for the particular wave length used, its angle of rotation is not a multiple of 7, in which case darkness ensues and the observed effect is nil. By revolving the upper nicol it is pos-. sible in white light to bring out the sensitive interference tint over the entire tield covered by the Bertrand plate, and in such a position a very slight turn of an intervening erystal from its position of true extinction is sufficient to disturb this equality * Neues Jahrbuch, 1874, p. 9. + Zeitschr. Kryst., i, 69, 1877. pa a ty met +i FL EL Wright— Measurement of Extinction Angles. 377 of interference color and to divide the field into four quad- rants, the opposite sections of which are similarly colored, while adjacent sections are differently colored. The Bertrand plate is best adapted for use in white light, although it is possible to use it in monochromatic. light pro- vided its thickness be correct for the particular wave length employed. The quartz half shade plate of S. Nakamura.—tin a recent paper,* S. Nakamura discussed ifhc problem of the sensitive- ness of the half shade system and arrived at practically the same conclusions as those noted above. He suggests the use of a double quartz plate of -4™™ thickness instead of 3°5™™ or 7™™ thick as in the Bertrand ocular, and by actual tests finds the theoretical deductions valid and the plate useful. The thickness of -4™™ is equivalent to an angle (90—@) of about 867° on each side of the junction line; under certain con- ditions of illumination this angle is undoubtedly the best, and with the plate the accuracy of the measurements thereby attained equal to that of any of the other measuring devices. Bi-guartz wedge-plate.—It is possible, however, to construct a combination wedge of quartz plates of sucha character that any angle otf rotation from 0° to any other value, positive or negative, can be had on insertion of the wedge, thus adapting to wedge form the advantage of the revolving bi-nicol ocular. This has been accomplished by combining two plates of quartz cut normal to an axis and of specified thickness, the one of right-handed, the other of left-handed quartz, each with a wedge of quartz of the opposite sign of rotary polari- zation, as indicated in figure Bae The effect of this combination is to produce zero rotation in each half wedge where plate and wedge have the same thick- ness and as the w edge is inserted or drawn out from this point of zero rotation the angle of rotation increases proportionately and in a positive sense on one side of the junction line of the combination and in a negative sense on the opposite half. This combination wedge, which is introduced at the focal plane ot the ocular, divides the field under crossed nicols into two halves, the intensity of color of which at any instant is equal, pr ovided no intervening crystal plate is present or is rendered inactive * Centralblatt f. Mineralogie, 1905, 267-279. Compare also J. Macé de Léepinay, Jour. de Phys. (2), iv, 267, 1885;(3), ix, 585, 1900. Unfortunately this paper did not come to notice until after the manuscript of the present article had been sent to the press and the mathematical discussion by S. Nakamura, which considers the problem from a somewhat different standpoint, could not well receive the analysis and recognition which otherwise might have been given it in the general theoretical part. + Made for the writer by Steeg & Reuter of Homburg v. d. Hohe, Ger- many. Cost, 100 marks. The accuracy of this wedge was tested by the writer and the grinding found to be exceptionally perfect. 378 EE. Wright—Measurement of Extinction Angles. by the parallelism of its ellipsoidal axis with the principal planes of the nicols. So soon as the erystal is turned even a very small angle ont of this position, the intensity of illumina- tion of the two fields is no longer equal. By inserting or withdrawing the combination wedge, the most advantageous angle of rotation in the two fields can be procured so that the difference in intensity between the two halves is most apparent.: In effect this wedge is identical with that. of the bi-nicol ocular described above, is much simpler in con- Fie. 12. Fie. 12. Bi-quariz wedge-plate. Inthe plate-wedge ground for the writer the two quartz plates are 30™ long, 6™™ wide, and °30™™ thick. The wedges are 5™™ thick at the one end and ‘O™™ at the thick edge. Cement- ing material is Canada balsam whose refractive index is 1°54, while w for quartz is 1.544, a difference so slight as to render inappreciable the exceed- ingly slight deviation of the waves caused by the slight wedge surface of the wedge. This inclined surface is mounted next the Canada balsam and care is taken by inserting a thin glass strip at the thin end to make the upper and under surfaces of the completed wedge parallel. The thickness of the wedge is ‘9™™. At the one end the rotation is +1°1° ; at the thick end, +3°2° for sodium light, while at -85™™ from the thin edge the rotation is zero in both halves. For a wedge-plate of an angle of rotation 0° to 10° the follow- ing specifications are suitable: length 50"™, width of each half 6™™, total width of wedge 12™™.; thickness of plate, 4™™; thickness of wedge at thin end ‘30™™ ; at thick end, ‘80™". In such a wedge the point of zero rotation is 5™™ from the thin end. At the thin end the rotation is +1°1°; at the thick end, +9°9°. In the article following this, specifications for a wedge with rotation of 15° at the thick end are given. In preparing the wedge it is necessary that the edges be ground and polished in order that the central division line (fig. 11) be as sharp as possible. The two halves are eventually cemented side by side with Canada balsam and any disturbing influence thus eliminated which might arise from total reflexion on the sides. struction, and requires no adjustment ; the one condition which must be fulfilled for satisfactory results is that the wedge be not tilted on insertion but that the optic axis remain always parallel with the optic axis of the microscope, otherwise ais- turbing birefringence phenomenaappear. The wedge carriage should, therefore, slide in an accurately fitting holder such as FE. Wright—Measurement of Extinction Angles. 379 shown in fig. 10 above, which was constructed in the workshop of the Geophy sical Laboratory. Methods involving revolution of upper nicol—In all of the preceding methods the nicols have been considered crossed and the erystal plate revolved. The intensity formula shows, however, that the relative intensity is dependent not only on the angle 6 of the crystal plate but also on ¢, the angle between the principal planes of the nicols. It was shown in ‘the general mathematical treatment that this method is in general at least twice as sensitive as the method based on the revolution of the erystal plate under crossed nicols. The mode of application of this method to any particular crystal plate is obvious and consists simply in placing the crystal under crossed nicols in its position of apparent “trae extinction and then obser ving, either in white or monochromatic light, the changes which occur on revolving the upper or lower nicol through small angles with its normal position. In case the crystal is actually in its position of true extinction, the erystal and field attain their position of maximum darkness simultaneously and show the same increase in its intensity of illumination; if, however, the crystal be not in its position of true extinction, but a small + angle, as 30’ distant, then for a position of the nicol +2° from its normal position, the crystal plate will appear lighter than the field; and vice versa for the nico] —2° from its normal position the erystal plate will appear darker than the field. This method is extremely simple in manipu- lation and does not require special apparatus, but seems not to have been applied before to the measurement of extinction angles. Weinschenk,* in describing the adjustment of the nicols in the microscope, uses the interference phenomena which occur under these conditions, but does not appear to have applied conversely the principle to the practical deter- mination of the optic ellipsoidal axis in a given erystal plate. To double the sensitiveness of this method of revolving the upper nicol, the bi-nicol ocular attachment of fig. 11 can be used. By this device alone, without the upper nicol used in the above methods, the two halves of the field in the ocular preserve the same intensity of illumination at every instant, provided no disturbing crystal plate intervenes. If the posi- tion of the latter does not coincide precisely with its true position of zero extinction, the two halves of the field appear unequally illuminated and by revolving the nicols that position of the nicols can be found for which the effect is most pro- nounced for a given angle @. In its effect the bi-quartz wedge plate is identical with the revolving bi-nicol scheme, and has the advantage of requiring * Zeitschr. Krystall. xxiv, 581-583, 1895. 380 FF. L. Wright—Measurement of Hatinction Angles. no adjustment and of not suffering from the false light of depolarization noted above. Convergent polarized light.—Two methods have been pro- posed which require convergent polarized light and are based on the change in aspect of symmetrical interference figures caused by the intervening crystal plate when it is not pre- cisely in the position of zero extinction. The idea underlying the methods is that the eye can detect more readily slieht changes in the shape of a symmetrical interference ficure than proportionate changes in intensity or color. Theoret- ically, this principle is excellent, but its practical application to mineral sections is less satisfactory. The first method of this type was proposed by Kobell in 1851, who used a plate of calcite normal to the optic axis as his test plate. The micro- scope was arranged for convergent polarized light and the erys- tal plate with the calcite test plate above it placed ‘on the microscope stage and turned until the interference figure appeared perfectly normal and undistorted. Practically, the following objections apply to this method. The optical system of the miscroscope requires changing each time to meet the new conditions ; during the observations the er ystal itself is lost sight of, and in the case of minute crystals or crystals with undulatory extinction this is a serious drawback. Moreover — it is tacitly assumed that in the crystal plate itself for direc- tions other than the normal to its surface of the crystal plate the planes of polarization remain parallel, which in general is only approximately true even for small fields which inelude only a small angle with the normal. In the Brezinat method a more complicated interference figure is produced by two calcite plates cut at a small angle with the optic axis and cemented together one above the other in such a way that the optic axes of the two are in the same plane and at equal angles with the normal. The interference figure from such a combination is notew orthy because of a dark vertical bar through the center of the field. A slight revolution of an intervening crystal plate displaces this bar noticeably, but the same objections noted in the Kobell method apply with equal force to this method,.with the result that neither method is made use of at the present time by working petrologists. In fact, both these methods were suggested before the petrographic microscope had been introduced. The relative sensitiveness of the different methods.—The term position of extinction means practically that position of a birefracting plate for which heht waves are transmitted with- out changing their plane of polarization and for which no light * Pogg. Ann., xcv, 320, 1859. + Described in Schrauf’s Lehrb. d. Phys. Min. ii, 219-220, 1868. \ F.. E. Wright—Measurement of Hatinction Angles. 381 passes the upper nicol, i. e., the field is just as dark as though no crystal plate were there. A revolution of the plate thr ough a very small angle from its position of true extinction allows an equally small percentage of the total amount of incident light through the upper nicol and the field is very dimly illu- minated. For agiven angle of revolution, the actual amount of transmitted light can be increased only by i increasing the orig- inal souree of light. Since, however, it is not possible to increase the intensity of such a source indefinitely, and the human eye is sensitive only to a certain limit, the position of actual extinction can only be determined within a definite degree of exactness. By means of the above devices, however, certain phenomena are introduced which increase the accuracy of such a determination, even though the field of original illumination remains the same. That method or device is obviously the best for which the probable error of a single determination under the same conditions is the least. In comparing the relative accuracy of the methods described above it will facilitate the presentation to assume definite conditions and then by means of the theoretical intensity curves (figs. 2-7) to test the results attainable by the different methods under the most favorable conditions. Let it be assumed that under the conditicns of experiment the eye of the observer is of such sensitiveness that he is able to detect ‘05 of one per cent of the total ight intensity; in other words, he can just detect the difference between the dark field of the microrcope under crossed nicols and a erystal . section turned at such an angle as to allow -05 of one per cent of the total intensity through the upper nicol. For all posi- tions of the crystal, then, for which the intensity of the emerg- ent light is less than -05 per cent, the crystal will appear abso- lutely dark. The heavy curves in figs. 8-7 indicate the relative intensity of illumination of a crystal under crossed nicols for all positions of its major ellipsoidal axis from 88° to 92° or —2° to +2° with the plane of the polarizer; in fig. 3 there is an interval of 38’ at least on each side ot the true extinction position for which the eye is unable to detect any interference illumination. The possible error on a single determination under the most favorable conditions is in this case at least + 38’ while for fig. meiner 44 aor doy > (KG) Ey do tor fic. 6) se 1° 17" ; while for the K = 0 the erystal is dark for all positions. In any crystal, therefore, the conditions are most favorable when the plate is of such thickness that K = 1 or the emergent waves are half a wave length apart (in opposite phase). Con- versely, having given a crystal plate, not all wave lengths are best adapted for extinction-angle measurements. If yellow Am. Jour. Sci1.—FourtsH Series, VoL. XXVI, No. 154.~OctTosrEr, 1908. Ot 382 FL Ek. Wright—Measurement of Extinction Angles. sodium light be used, a plate showing the sensitive violet inter- ference tint is worthless since for that tint the path difference is 572 wp, nearly a whole wave length of Na hight (589), and for this difference K = 0 and fig. 7 applies. “Lf sodium hight be used, then plates should be. chosen for which the phase ain difference of the two emer ging waves is , br oht yel- low of the first order or pure yellow of rhe oe order or ereen yellow of the third order, ete. This is an important consideration and applies to all methods involving the inten- sity equations. The visible spectrum includes wave lengths ranging from about 400up, and T0O0upm, the interference color in white heht ranges from about red of the first order to blue of the second order ; in short, the sensitive interference tint region of the Newton color scale as determined by G. Quincke.* For this interval the distance between the emergent waves is not far from a whole wave length for the major part of the visible spectrum ; in other words, the phase difference is such that K is a small fraction not greatly different from zero and the intensity curves for practically all wave lengths will be cov- ered by 5-7. These are, however, the least favorable for showing difference in intensity and such plates are, therefore, the least suitable for the measurement of extinction angles by methods based on intensity differences. On the other hand, plates showing interference colors gray to yellow of the first order are best suited for such measurements. If the methods involving interference tints be used, however, these objections do not hold with equal force. Experience has shown’ that in ease the mineral plate does show red or blue interference tints of the first and second orders the best determinations ean be made either by the method of revolving the upper nicol or by the bi-quartz wedge plate, and the true position fixed by noting the absence of abnormal interference colors on revolving the nicol very shghtly or inserting the wedge. After this digression on the most suitable sections for the measurement of extinction angles, fig. 3 may again be con- sidered and the relative accuracy of the different methods under the same conditions of experiment deduced. The heavy curve indicates that for the assumed limit of sensitiveness ‘05 per cent of the total intensity, an error of at least +38’ on a single determination is possible by revolving the crystal plate alone under crossed nicols. On the other hand, if the crystal plate remains stationary and the upper nicol alone is revolved, the other intensity curves of fig. 3 are valid, each curve indicating the intensity of illumination of -Poge, vA nil) Cxaxd x muni loo on Ste ee Balti’ F. E. Wright— Measurement of Extinction Angles. 383 the crystal plate for a specified angular distance from its posi- tion of true extinction during the revolution of the upper nicol from 88° to 92°. These curves indicate that the probable error with this method is less than half as great as in the preceding method, for if the crystal be only +15’ distant from its position of true extinction, differences in intensity can even then be detected on revolving the upper nicol. The changes in intensity of illumination of the microscopic field on revolution of the analyzer are indicated by the 0’ eurve, while for the crystal plate the 15’ curve is applicable. At 88° 43’ (fig. 3) the field is just beginning to show detect- able ieneaiindin (05 per cent of total intensity), while for the same angle the crystal is illuminated with 0-97 per cent of the total intensity, nearly twice as great and easily notice- able. In this position the erystal plate appears, therefore, decidedly lighter than the field. On the other side of 90° the crystal plate passes its limit of light sensibility under the assumed conditions at 91° 42’, while for the same angle the microscopic field is illuminated by ‘97 per cent of the total intensity ; in this case the field is appreciably brighter than the crystal and the difference can be readily detected by the eye. li white light be used, these differences are accentuated by the abnormal interference colors which appear in the er ystal plate when it is not precisely in the position of true extinction. This method of revolvmg the upper nicol has the advantage, furthermore, of not being dependent on the accuracy with which the nicols are cross sed, since all data are referred at once to the plane of the analyzer. It is not, however, so advanta- geous in very weakly birefracting or deeply colored mineral plates. Dee. The sensitiveness of the latter method can, moreover, be doubled by devices which allow the phenomena on both sides of the 90° position to be observed simultaneously. This is the end striven for in the ocular plate, the Bravais-Stéber plate, the Calderon plate, the Traube plate, and accomplished most effectively by the new circularly polarizing bi-quartz wedge and also by the bi-nico] ocular, though less satisfactorily. In each of these last two devices the ‘plane of polarization of the incident waves is turned through equal angles on both sides of the junction line of the two parts, so that the field appears equally hghted throughout, while if the crystal plate be not in its position of true extinction, it will appear lighter than the field in the one half and darker in the second. Sinee, however, there is an angle best suited under the given condi- tions to show these differences most clearly, it follows that the best results can be had with a plate or apparatus in which the 884 FE Wright—Measurement of Extinction Angles. angle @ can be varied at wili. This is true of both the cireu- larly polarizing wedge plate and the bi-nicol ocular, and by their use the probable error of the extinction position of any erystal plate is at least one-fourth that of a determination after the usual method by revolving the crystal plate under crossed nicols. Experience has shown that with favorable sections, extinction angles can be determined by the use of the bi-qnar tz wedge with a probable error of less than +10’ on a single trial. Still another method for obtaining the most favorable con- ditions of experiment with a given plate is that suggested on page 374 with the artificially twinned quartz Hee The two halves of this wedge extinguish at a small angle (as 3°) on oppo- site sides of the line of junction, and by inserting the wedge that particular interference color, or phase difference if mono- chromatie light be employed, can be produced for which the given angle of revolution (8°) is the best. This wedge, how- ever, is less favorable than the circularly polarizing bi-quartz wedge, since its twinning line must be inserted precisely parallel with the plane of the polarizer, while with the cireu- larly polarizing bi-quartz wedge the rotation of the planes of polarization of transmitted waves is entirely independent of the line of junction of the adjacent halves. In the preceding pages, special emphasis has been placed on those methods for measuring extinction angles which are of general application and which are based on intensity differ- ences. The other methods, which are of limited application, and can be used only in white hght on favorable sections, depend on differences in interference colors produced by slight deviations of the crystal plate from its position of true extine- tion. Although these methods are serviceable in many instances, their application and the results obtained thereby are so dependent on the conditions of experiment that they are difficult to treat satisfactorily ina general way. Experience has shown that they are not more sensitive than the other methods and usually much less so. This is true both of the selenite sensitive tint plate and of all combinations of the same. Experimental Tests.—To test the different methods under different conditions, different mineral plates were chosen and the position of true extinction on each determined by the dif- ferent methods under precisely the same conditions of illumin- ation with white hght.—On an anhydrite plate showing white interference tints of the higher orders the possible error of a single determination by revolving the crystal plate under crossed nicols was found to be about 1-1°; ; by revolving the upper nicol alone, ‘4°; by inserting the quartz wedge, about O-1°.; by, using (tte Calderon ocular, about °5°; by means of red F. EB. Wright—Measurement of Extinction Angles. 385 the Bertrand ocular, about 0°1°; with such a plate the sensi- tive tint plate is of no value since the interference color of the anhydrite plate itself is so high that the violet of the inserted plate has no effect and the differences in intensity which occur, do so in astrongly lighted field and are not easily discernible— Similar measurements were made on an apatite plate parallel to 1010 and showing the interference tint, red of the first order. The possible error of a single determination of the position of true extinction on turning the crystal plate alone under crossed nicols was found to be +*9°; on revolving upper nicol about -2°, accurate because of abnormal interference colors which appear when the plate is distant only a slight distance from its correct extinction position; on inserting the bi-quartz wedge plate 0°2° to 0°3°; with the Calderon ocular, about 0°3°; with the Bertrand ocular, about 0°3°; the sensi- tive tint plate is again of no value since the interference color changes com paratively slowly as crystal is revolved.—A sec- tion of nephelite parallel to 1010 and showing the interference color, yellow first order, gave the following results: On revolving the crystal plate alone, possible error -4°; on rotat- ing upper nicol, less than 0-1° ; with bi- -quartz wedge less than 0-1°; Calderon: ocular about 0°2°; Bertrand ocular less than 071°; sensitive violet plate still of very little value as a med only slight changes in color for large angles of revolution of plate. On a plate of colorless gehlenite of very low interference color, dull gray, first order, the sensitive tint plate proved as satisfactory as any other and more so than the method of turning the crystal plate under crossed nicols or of revolving the upper nicol or the Calderon ocular. The Ber- trand ocular and the bi-quartz wedge plate proved about as favorable, the probable error being slightly less than 0°5°.—A plate of strongly pleochroic tourmaline was also used and the following results obtained: Probable error of determination on rev olving crystal plate alone, about 16°; the method of revolving upper nicol is of little value because of deep natural color of mineral and consequent ey to match fields ; with the bi-quartz wedge plate 0°3°; Calderon ocular, about 0-4°; Bertrand ocular, about 0°5°. The sensitive tint plate is useless because of strong natural color of mineral which veils the true interference colors. The results of these tests show that the theoretical deduc- tions from the general equations are in general valid, but that in certain instances other factors, as natural color and very low birefringence, become dominant and tend to render some of the methods less sensitive and to favor the use of other, in general less suitable methods. The bi-quartz wedge plate, however, seems to apply in all cases with equally favorable . 386 LL. Wright—Measurement of Hxtinction Angles. results and to equal in sensitiveness any of the methods, whether of local or of general application. 2, The Adjustment of the Petrographic Microscope with Special Reference to the Measurement of Hxtinction Angles. A petrographic microscope in perfect adjustment should satisfy the following requirements: (1) Its optical system should be accurately centered ; (2) the axes of revolution of all revolvable parts, whether stage or ocular, should coincide with the optic axis of the micr oscope ; (3) the principal ae of the nicols when crossed should be precisely 90° apart; (4) the crosshairs of the ocular should be parallel with the principal planes of the nicols. Of the four conditions, the first two can be accomplished without difficulty, and with the adjustment screws fitted on every petrographic microscope. The last two, however, require special appliances for accurate adjustment, and Shout these can be effected only with difficulty. The test usually applied in ascertaining the correct position | for crossed nicols is that of the Bertrand ocular. A cap nicol is used over the ocular and turned until the field of the Ber- trand ocular shows uniform intensity of illumination through- out. This can be accomptished readily and with an error of less than +15’ if strong illumination be used. The cap nicol is then revolved. thr ongh an angle of 90°, the lower nicol removed in its carriage, ‘the upper “nicol inserted and tested by the cap nicol in its new position and adjusted: until it is actually crossed. The Bertrand ocular, however, furnishes only one angle of rotation for the emerging waves, and allows of no variability in this angle to meet “different conditions in the best way possible. This can be accomplished, however, by use of the circularly pone bi-quartz wedge plate described above or the bi-nicol ocular. With the bi-quartz wedge plate the cap nicol is unnecessary and crossing of the micro- scope nicols can be tested directly. For this purpose, parallel hght should be used and the entire lens system, both conden- sor lenses, objective and ocular, removed from the microscope ; parallel incident rays are then allowed to fall on the reflector of the microscope (either sun rays or strong white light or the rays from a Nernst filament or are light emerging at the focal point of a large lens). The parallelism of the incident rays 18 necessary and an important factor, since with the thick quartz plates a slight deviation from nor mally incident and parallel hght produces disturbing inter ference phenomena. The bi. -quartz wedge-plate in its metal casing may then be placed on the microscope stage and with it the accuracy of the crossing of the nicols tested directly, just as the position of zero extinc- tion of a mineral plate is tested. The error of such a deter- mination should be considerably less than 10’ of are. F. E. Wright—Measurement of Extinction Angles. 387 -The final step in the adjustment is the alignment of the crosshairs of the ocular with the principal planes of the nicols. Many methods have been suggested for this purpose which may be used to advantage. A mineral showing good cleavage or lines of growth and parallel extinction (fakes of anhydrite or crystallites of quartz) serve well for the purpose. ‘These are first placed in the position of zero extinction (determined accurately by means of bi-quartz wedge plate), and the cross- hairs of the ocular brought to par allelism with the er ystal edge or cleavage line. The chief difficulty in this method lies in the fact that it is exceedingly difficult to obtain suitable ~ material. E. Sommerfeldt* has recently suggested the use of a twinned plate of selenite of sensitive tint. The plate is turned until under crossed nicols the interference colors in the two halves are of precisely the same tint, in which position oS twinning line parallels the principal planes of the nicols. This method is sensitive and satisfactory, especially if, instead of a plate, a wedge of the material be used, with which the interference eolor can be changed until that particular tint, for which the observer’s eye is most sensitive, covers the tield. In this con- nection care should be taken to select a selenite plate in which the twinning line is perfectly straight. In the selenite plate, however, the angle which the adjacent Ba. axis in each half makes with the twinning line is 374°, a very large angle and not so well adapted to show slight deviations from the true position as a twin of smaller symmetrical angle. In this connection experiments with plagioclase lamellze were tried but abandoned, since it was found by experience to be exceedingly difficult to procure suitable material for the pur- pose ; a simpler method can, however, be used, which accom- plishes the same purpose more readily. On a thin plate of quartz, cut parallel to the principal axis, an edge surface, making an angle of 4°—6° with the principal axis, is first ground and polished. The plate is then cut in half at right angles with the polished edge surface and the two halves placed with their polished edges side by side and cemented with can- ada balsam, thus producing an artificial twin of any angle which may be selected as most suitable and best adapted for fixing the crosshair in the ocular. The twin is finally ground thin and polished either to the sensitive tint or a pale gray of the first order or in wedge form.+ Like the selenite twin plate, it is placed on the stage of the microscope and revolved under * Zeitschr. fiir wissenschaftliche Mikroskopie, xxiv, 24-25, 1907. + Such a plate and also wedge were prepared for the writer by Voigt & Hochgesang of Gdéttingen, and have proved satisfactory in every respect. The twinned selenite plate of sensitive tint cost 5 marks; the wedge (I-III order interference colors), 21 marks. 388 L.A. Wright—Measurement of Hautinction Angles. crossed nicols until its halves show equal intensity of illnmi- nation, in which position the line of junction fixes the direc- tion for one of the crosshairs of the ocular. By use of the bi-quartz wedge plate and the artificial quartz twin plate or wedge, the adjustment of the nicols and also of the crosshairs in the ocular is a matter of only a few moments, and the method followed is theoretically and practically more accurate than the other methods for adjustment which have been suggested. 3. A Device for Holding Small Crystals for the Purpose of Measuring Extinction Angles in Zones ; also for Measuring the Optic Axial Angle of such Fragments directly. In measuring the extinction angles of certaim minute artificial pyroxene and other crystals in the prism zone, the writer has_ had occasion to use the following holding device or fin mei which has proved both convenient and practical.* (Fig. 16 on stage of microscope.) As indicated in the figure, it consists of two parts, a holding or clamp device for the ‘crystal itself, and a universal ball and socket joint and centering plate for adjusting and centering the crystal, and a vertical circle for reading any specified angle of revolution of the crystal. A small crystal (1-2™" in length) thus held and adjusted is immersed in a drop of liquid of the same refractive index and thus the disturbing phenomena of refraction and total reflection eliminated. The liquid drop is contained in the concave side of a planoconcave lens of 6™™ diameter (ground for the purpose by the Scientific Shop ot Chicago), which in turn is held by an adjustable and support- ing arm. Tn work with artificial preparations particularly it is often desirable to measure extinction angles on certain minute faces, or optical axial angles, of small fragments too minute for the optic axial angle apparatus and not suitable for measure- ment by one of the microscopic methods,t and this device has been made to fill that want. After adjustment and immersion in a liquid of refractive index 8 the optic axial angle can be read off directly on the vertical circle of this apparatus. Summary. The measurement of extinction angles of minerals in the thin section is one of the most common methods of petro- graphic microscopic practice, and at the same time one of * This apparatus was made in the workshop of the Geophysical Labora- ‘tory at the instigation of the writer, and can be constructed by any good mechanic. + Compare F. E. Wright, Measurement of the Optic Axial Angle of Min- erals in the Thin Section, this Journal (4), xxiv, 317-369, 1907. FE. Wright —Measurement of Extinction Angles. 389 the least satisfactory when accurate results are desired. It is an exceedingly easy matter to measure, with one trial only and on favorable sections, extinction angles with a probable error of +1° to 2°, but to do so within 1° is a very different mat- fer. In-.the “foregoing pages the problem is first discussed theoretically, and the mathematical equations covering the different methods applied to extinction angle measurements derived and discussed briefly. The methods for this purpose may be grouped into two classes,—those of general application and those of limited applicability The first class may again be subdivided into two classes ; either (1) the erystal is revolved between crossed nicols about its position of true extinction, or (2) the crystal remains stationary, and the accuracy of its position of true extinction tested by revolving the upper nicol or by inserting one of several different optical devices to increase the sensitiveness of the test under the prescribed conditions of illu- mination. These devices include the Calderon ocular, the Bertrand ocular, the Bravais-Stober plate, the Traube plate, also twinned plates and wedges of selenite, artificially twinned plates and wedges of quartz, the circularly "polarizing bi-quartz wedge plate and the bi-nicol ocular, all of which are described briefly above. Of these devices the last two are the most uni- versal and can be so used under any given conditions of illu- mination ‘that the phenomena observed are the most sensitive possible to attain by devices of this type. On comparing the relative sensitiveness of the different methods under the same conditions, it is found that the method of testing the position of true extinction for the crystal by rev- olution of the upper nicol is, on colorless mineral-plates, at least twice as sensitive as that of simply turning the crystal to its position of apparent maximum darkness under crossed nicols.. Under the same conditions the methods requiring the use of one of the several plates or wedges mentioned above are at least four times as sensitive as the ordinary method. With the exception of the last two devices, however, these different plates'do not furnish equally sensitive results for the different conditions of illumination which may arise. In accurate work adjustable sensibility is of prime importance, particularly if a given device is to be of general application. These require- ments are best filled by the bi-quartz wedge plate, by means of which the angle of rotation can be varied from 0° to any desired anele.- Vhe two halves of this wedge rotate in opposite direc- tions, and on insertion that angle of rotation can be secured for which the contrast in the intensity of the halves of the field is most striking for a slight deviation of the crystal from its true position of extinction. 390 LE. Wright—Measurement of Hatinction Angles. An application of the bi-quartz wedge plate, together with an artificially twinned quartz-plate or wedge, to the accurate adjustment of the petrographic microscope is considered in outline in the second part above, and a method of procedure for accomplishing the same indicated. In part 8 a simple device for holding and rotating small erystals for the purpose of determining extinction angles or for measuring optic angles directly, is described briefly, such an apparatus having been found peculiarly useful in work with artificial crystals. Geophysical Laboratory, Carnegie Institution of Washington, Washington,’ D. C., June, 1908. FE. Wright—Bi-quartz Wedge Plate. 391 Art. XLI.—The Bi-quarte Wedge Plate Applied to Polar- imeters and Saccharimeters ; by Prep. EuGENE WRricur. * In the preceding paper on the measurement of the extinction angles of mineral plates in the thin section, the general theory of the relative intensity of illumination of the field for different positions of the nicols and also of the crystal section is devel- oped. The different methods for crossing the nicols accurately and also for determining the position of total extinction for a erystal section are considered with respect to their sensitiveness, and the conclusion reached that the best results are obtained by dividing the field into parts in which the planes of polariza- tion are inclined at equal but small angles (90-¢ of the general formula) to the line of division between them (half shade system).+ For given conditions of illumination and sensitive- ness of the observer’s eye, there is always a certain angle (90-¢) for which the phenomena observed are most sensitive to slight movements from the position of true extinction. To reach the maximum efficiency of the half shade device, this angle (90-¢) should therefore be adjustable within the limits prescribed by these conditions. The conditions for maximum sensitiveness have been worked out for polarimetric purposes with sufficient care and accuracy by a number of investigators, { and several instruments have been designed in which high precision is possible, provided only that light of a certain (fel and intensity is available and that the substance under investigation (sugar, for example) permits just the right quantity of “this light to pass through it to fall within the rather narrow limits in which the observer’s eye is most sensitive. In practice, these conditions have proved somewhat exacting, and considerably greater elasticity in the adjustment of the optical system of the instrument with a view to service under a greater variety of experimental conditions, without loss of accuracy, is desired. Mr. Frederick Bates$ of the Burean of Standards has recently successfully designed an instrument in which this feature has received attention. His polariscope follows the Lippich system with the addition of sets of gear wheels for revolving the large polarizer and analyzer simultaneously in such a way that the angle of revolution of the polarizer at every instant is twice that of the analyzer. He has also demonstrated that the small angle (8) between the * The author is indebted to Dr. Arthur L. Day of this laboratory for several important suggestions in the preparation of this paper. +S. Nakamura, Centr. f. Min., 1905, p. 267-279; P. G. Nutting, Bulletin, Bureau of Standards, lii, 249, 1906. t See Winkelmann, Handbuch der Physik (2), vol. vi, pp. 1362-3. § Bulletin of the Bureau of Standards, iv, 461, 1907. 392 F. E. Wright—Bi-quarte Wedge Plate. normal to the principal plane of the analyzer and the bisector of the angle between the nicols of the Lippich polarizer can be readily compensated by a slight shift of the zero in the quartz compensator scale. This apparatus is easily superior to the ordinary quartz compensating polarimeter and is of more general application, but it is mechanically difficult and expen- sive to build and its adjustments are rather sensitive to wear and tear. It therefore occurred to the writer that the bi-quartz wedge plate described in the foregoing paper might serve the same purpose much more simply and therefore be of some prac- tical utility in polarimetric and saccharimetric measurements. By means of this wedge the plane of vibration of waves from the analyzer is made to rotate from 0° to any specitied angle by varying the thickness of the wedge employed, the rotation of the one half being right-handed and the other left-handed. Since the problem of saccharimetry from the practical standpoint reduces to the determination of the exact angle of rotation of the plane of polarized light after its passage through the sugar solution, or, in brief, to the accurate setting of the analyzer or the compensating system with the analyzer, it is apparent that the bi-quartz wedge plate is directly applicable. The mode of application is the same whether a monochro- matic or a white light source is used. If monochromatic light is employed a feasible arrangement of the apparatus is shown in diagram in fig. 1. The monochromatic light, properly restricted in wave length and as intense as possible, is polarized hy the nicol prism P. On passing through the solution T its plane of Syl is rotated and the angle of its rotation is determined by revolving the analyzer (fitted with an accurately divided degree circle C) until darkness ensues. The exact position of extinction is then found by inserting the bi-quartz wedge, which offers opportunity for a finer adjustment, and ascertain - ing that the intensity of illumination of both halves of the wedge is precisely equal for all positions of the wedge plate.* * The wedge plate should be mounted in a metal frame arranged to slide like the quartz compensator in accurate grooves. The quartz wedge plate vsed in the examination of extinction angles in crystals is 30™™ long, 12™™ wide, and is made up of wedges 0°3™™ thick at one end and 0°d™™ at the other, which are underlain by two quartz plates each *30™™ thick. The rotary angle range in each half is from —1°1° to +5°2°. For general polarimetric work, in view of the limited sensitiveness of the eye and the difficulty in obtaining homogeneous illumination of sufficient a intensity, itis desirable that the angle 5 should be capable of being increased to 15°. Toaccomplish this in a single wedge it would be best to make it 50™™ Jong, 10 or 12™™ wide with the thin end of each wedge 35™™ and the thick end 1:10™™ in thickness, combined with quartz plates 0:4™™ thick. This gives a pitch to the wedge of 1:5: 100 0r8°. The point of zero deflection or the position of the black band will fall 3°3™™" from the thin end and the deflection on the thick end will be + 15° for sodium light. The total thick-’ ness of the wedge plate is 1°5™™, Canada balsam should be used for the cementing material and care must be taken to have the wedge surfaces adja- cent to the Canada balsam and also the two surfaces of the complete wedge plate parallel. LIE. Wright—Bi-quartz Wedge Plate. 393 Moreover, if the wedge is inserted horizontally and at such a point that its effect on the plane of polarization of transmitted light is precisely zero, a straight, black, vertical band appears in each half of the field similar to the bands in the Babinet compensator (fig. 2). Dy means of this band, the true position jebeGen lye Fic. 1. Proposed arrangement of parts in polarimeter, using bi-quartz wedge plate as sensitive device. P, polarizer: T, sugar solution tube; W, bi-quartz wedge; C, degree circle of analyzer; O, Ramsden ocular; A, analyzer. Although the bi-quartz plate wedge may or may not be attached to the revolving circle mechanism, it seems preferable that its carriage should remain in one plane. The ocular O may be used either in front or back of the analyzer A, and may be a single acromatic lens in place of the positive ocular. Observations can also be made without the aid of the ocular and thus an increase of light intensity gained. of the analyzer can be found with great accuracy, for the set- ting is thus made to depend upon the exact alignment of two black bands and the photometric principle of comparing two dimly lighted fields is for the most part eliminated. Another advantage of this system lies in the fact that the boundary lines between the halves of the wedge can be made of knife-edge sharpness without the disturbing division line in the center of the field produced by the total reflection and con- 394 LE Wright—Br-quartz Wedge Plate. sequent depolarization of light waves on the edge of the small prism of the Lippich system. The edges of the two halves of the bi-quartz plate are first polished and then cemented with Canada balsam, which has practically the same refractive index (1:°540 compar ed with 1: 544), so that no appreciable total reflee- tion with the accompanying depolarization oceurs. This is the same result which was successfully attained by Brace,* who lie, 2. expended an extraordinary amount of care and ingenuity in an effort to free the field from this disturbing dividing line; but his instrument is not of such mechanical construction as to lend itself readily to ordinary laboratory use. Still another favorable feature of the wedge plate scheme for adjusting the sensibility of a polarimeter is “the position of the accumulator or magnifying lens of short focal length, which is necessarily of greater light intensity than the “telescopic ocular usually used to view the polarizing system. iG 3}. O Fie. 3. Proposed arrangement of parts in quartz compensating polariscope using bi-quartz plate wedge as sensitive device. P, polarizer: TT, sugar solution tube; Q, quartz compensating system; W, bi-quartz wedge plate ; O, ocular; A, analyzer. As in fig. 1, the ocular O may be used either in front of or back of the analyzer ; a single achromatic lens may be substituted for the Ramsden ocular, or the magnifying lenses may be done away with altogether. If a quartz compensating system be used and a white light source, the arrangement outlined in fig. 8 might prove advan- tageous. The quartz compensating system has been adopted on 7D) bebrace. hil, Mac.(6) ave ole Oa, FTE Wright—Bi-quarte Wedge Plate. 395 most commercial saccharimeters in order that white hght may be used, taking advantage of the fact that the rotatory dispersion of sugar solutions is approximately that of quartz. According to the scheme of fig. 3, plane polarized white light emerges from the polarizer P, passes through the sugar solution T, the different wave lengths being rotated through different angles and thence through the quartz compensating system Q, where they are again united and reduced to a common plane of vibration, —that of the original polarizer. The quartz com- pensating system is inserted until the field when observed through the analyzer appears totally dark. As in the preced- ing case, this condition is,verified with the greatest sensitive- ness by inserting a bi- -quartz plate wedge Ww. Up to this point, the advantages offered by the bi-quartz plate wedge are mainly those of simplicity, both in construc- tion and in manipulation, and of greatly decreased cost, without any corresponding sacrifice of accuracy or sensitiveness. Its sensitiveness is adjustable within any limits likely to arise in usual polarimetric work ; it is equally adaptable to the mono- chromatic and to the quartz compensating systems; it is possible with it to do away in considerable part with the photo- metric principle by the use of Landolt’s bands; and finally, the line of division between the hemispheres is so narrow as to be practically invisible as an independent line. Furthermore, in so far as it avoids the small prism of the Lippich polarizing system, it also avoids an error which is inherent in this ‘system, and incidentally also in the Bates system, due to the loss of light in one half of the field in passing into and out of this superposed prism.* In his analysis of the problem, Bates has shown that the small angle 6 between the normal to the principal plane of the analyzer and the bisector of the angle, between nicols of the Lippich polarizer, can be allowed for by a slight shift of the zero of the quartz compen- sating scale for which provision is made in his apparatus. A slight additional correction of the same character made at the same point will also serve to correct for the loss of light by reflection at the end surfaces, and by absorption within the small prism. Supposing this to amount to 10 per cent of the total intensity, the situation can be summed up as follows: If abe the angle between the two prisms of the Lippich polar- izer, I the intensity of the light from the large polarizer, then *Tt will be recalled that the Lippich system consists, either for monochro- matic light or with the quartz compensator and white light, of a large nicol in front of which is placed a small nicol covering one-half the field of the former. The plane of polarization of this nicol makes a small angle with the plane of the large nicol. In the matter of the loss of light at the surfaces and by absorption within prisms, the system is therefore not symmetrical with respect to the line dividing the fields. 396 I ki. Wright—Bi-quartz. Wedge Piate. the intensity I’ of the light emerging from the small nicol will not be precisely I’ = I cos’ a, but I’. = 9 I cosa If the two fields are matched to show echal intensity, the plane of the analyzer will not in general coincide with the bisector of the angle a, but will include with it a smail angle 6 (fig. 4). = An equation from which the angle 6 can be figured when a is given has been derived by Bates, but his expression as noted above does not take cognizance of the loss of light in the extra prism and accordingly requires slight modification. Let OP, = direction of plane of large polarizer OP, = direction of plane of small polarizer a = angle between two polarizers; angle between the normal (OC) to the plane of analyzer AA,, and I, intensity of light in the field covered by the large polarizer alone when viewed through analyzer; and I, intensity of light of the tield covered by the small polarizer. The angle P,OA angle between large polarizer and analyzer is T a eEONe =a 48 while PIO AU eae, : ets) From the intensity formula of the preceding article : T= cosa OFA stay ( — 5) = L, "0 cosa cos’. < (POA 9 ‘costa sim € = 5) The condition of matched fields is es OG FE. Wright—Bi-quartz Wedge Plate. 397 > 2 a a 9 . 2 a sini ote -9-€0s: a.Ssim x 23 SANT 4 Gy A = eR: == (ee sin = cosd—cos= sin §=41/°9 cosa sin % cos 6+4/°9 cosa cos > sin 8 =— = or sin © cos 8 (1 — V9 cos «) = cos 5 sin 5 (1 + /9 cosa) = 0 1l— ‘9 COS io. — ee 1 + 4/-9 cosa = or ess 1°05409 — cosa | a J° = 7-05409 + cosa” 2 The formula of Bates expressed in this notation reads ) SE. cos « oo cosa’ a and differs from the above only in the constant in the fraction. For the sake of comparison the angles § have been figured for a = 1° to 15° by both formulas and vlna in the e.g ing table: = 6 (new formula) 6 (Bates formula) SE Re Ap ape ey Sieur ee 0’ 0’ i Og Bee Sate Fabs [. 2! QO! 9° 1a eae Rial ts a ar ite ae ar 0’ z4 ny le aa S ge, Ce lg aa AL a 4’ tw 4° “aS ps patie Se an Balan pana as je =e es Wappen i 10’ 9! 5 El ath hs pie 9 Sak ga ok Aa gs 14’ 4! Fae: ule Pande a A get Pea paca ley 6! Ragen ee een rte) am Wea SN IN LOY Taye Spee? ease Were FS Lt: ao! 14’ pees ee eee ED! 19’ [Rs es A ie dO ea 43 AS CET YOR Sek Gi mS 12 oon Rg ae a hs L203: AQ’ VACHE Re ASS DP aE Ce er ee fad 53 TOSSA Gar Sh ee OE ee E30! bog! This table as well as a discussion of the two formulae shows that the angle 6 is increased in every instance by reason of the loss of light by reflection in the small nicol prism: These differences can be allowed for upon the scale of the quartz compensator. The operation for any angle a consists in first adjusting the zero of the quartz compensator scale with respect to the analyzer without intervening sugar solution Am. Jour. Sct.—Fourts Series, Vout. XXVI, No. 154.—OctToBer, 1908. 28 398 FE. Wright— Bi-quartz Wedge Plate. until the illumination of the two halves of the Lippich polarizer system is the same. After the introduction of the sugar solution, the quartz compensator is inserted until the original equal intensity of the halves is restored and the angle of rota- tion derived directly from the compensator scale. This being the case, it would considerably simplify the con- struction to allow the analyzer to remain rigidly fixed and to revolve the two nicol prisms of the Lippich polarizer system, thus eliminating the complicated gearing of the Bates ‘polar- iscope, which mechanically i is an exceedingly difficult piece of apparatus to construct and to operate without lost motion. The revolution of the polarizing prisms in equal and opposite angles can be accomplished either by means of the worm thread device adopted in the binocle ocular of fig. 11 of the preceding article, or by a grooved arm into which pins connected with the sup- porting collars of the nicols fit and slide as the arm is inserted. Mechanically these devices are not difficult of construction and the angle of revolution can be read off accurately. In the foregoing pages the bi-quartz plate wedge is suggested as a simple and effective basis for the construction of a polar- imeter of adjustable sensibility in which the error from the asymmetry of the Lippich system, together with all the serious complications of mechanism, are completely avoided without loss of accuracy. Such a plate has been constructed and successfully applied to the exact location of crystal extinctions, but unfortunately pressure of other duties has prevented the writer from actually constructing a saccharimeter. Through the courtesy of Dr. Bates of the Bureau of Standards, however, an opportunity was given to test the wedge on a large and accurate standard polarimeter illuminated by homogeneous green light from a mercury quartz-glass arc. Its performance was entirely satisfactory, minute displacements of the analyzer from its position of true extinction being readily detected. Geophysical Laboratory, Carnegie Institution of Washington, Washington, D.C., July 6, 1908. Chemistry and Physics. 399 Se EE NEI EEG ENTE LLIGEN CE. I. CHEMISTRY AND PHYSICS. 1. Determination of Phosphorus in Phosphor Tin.—GEMMEL and ArcusBuTt have devised a new method for making this some- what troublesome determination. They place two to five grams of the sample in a 500° Jena flask fitted with a tap funnel and delivery tube. The tap funnel has a two-way stop-cock, allowing gas or liquid to be introduced as desired. ‘The absorption appa- ratus consists of three Drechsel bottles, the first two charged with bromine and water, the last with bromine water only. The air is first removed by passing carbon dioxide, then 50 to 100° of con- centrated hydrochloric acid are introduced, and the contents of the flask gently heated and finally boiled until the substance is dissolved. Finally a current of carbon dioxide is again passed through the apparatus in order to drive forward any remaining traces of hydrogen phosphide, the liquids in the absorption apparatus are transferred to a beaker, evaporated to small volume, and the phosphoric acid is precipitated. with magnesium mixture in the usual manner. It is to be observed that any arsenic present in the substance would pass over with the phosphorus, but the authors did not find this element present in the samples that they examined. It was found that the actual evolution and absorption occupies less than twenty minutes, so that the process is shorter and simpler than those usually employed: Upon comparison with the methods consisting in the fusion, with “hepar sulphuris ” and with potassium cyanide, of the products of treatment with nitric acid the new method was found to give much more concordant results.—-Jour. Soc. Chem. Indus., 1908, 427. H. L. W. 2. Complex Calcium Salts.—By the action of cesium sulphate solution upon gypsum, D’Ans has prepared a very stable double sulphate Cs,Ca,(SO,),.. The calcium alkali double sulphates now known, of which the first is the mineral syngenite, are as follows: K,Ca(S0O,),.H,0 elite K,Ca,(SO,),.H,O (NH,),Ca(SO,),.H,0 (NH,),Ca,(SO,), (NH,),Ca,(SO.)..H.O Rb,Ca(SO,),.H.O Rb,Ca,(SO,), sans ioe CO,Ca,(SO,), pf Weare D’Ans has prepared also two interesting triple salts analo- gous to polyhalite, Ca,MgK,(SO,),.2H,O. These are the salts Ca ,Cu(NH,),(SO,),.2H,O and Ca,CdK,(SO,),.2H,O, which were prepared by boiling solutions of thé proper salts with gypsum. It is probable that a considerable number of compounds analogous to polyhalite and to krugite, Ca,MgK,(SO,),.2H,O, may be pre- pared, and it is the author’s intention to continue his investigations in this direction.— Ber tchte, xii; 1776. IS OOD Daye 400 Scientific Intelligence. 3. Radio-activity.—MAarcKwa tp has recently delivered a lecture before the German Chemical Society giving a good account of what has been done in the field of radio- activity, and discussing the prevailing views in regard to this subject. While this lecture contains little that is new to those who are familiar with the literature of the subject, it will be useful to those who desire a general knowledge of this new branch of science. The author gives a somewhat novel view of the enormous ener gy involved in the transformation of the radio-active elements by saying: “It was the dream of the alchemists to transmute base metals into noble ones. ‘The radio-active substances teach us that, if this | process could be achieved, there would either be obtained at the same time so much energy that in comparison to it the value of the noble metal would be insignificant, or on the other hand the consumption of energy w ould render the ennobling of the metal practically uneconomical.”— Berichte, xli, 1524. He We 4. A Simple Method for Determining Vapor Densities.— BiackMAN has devised an apparatus for this determination, which consists of a sealed tube in which is placed a capillary tube graduated in millimeters, closed at one end, and supplied with a thread of mercury at the other ena in order that it may serve as a manometer. It is evident that when this system is heated with nothing but air within and without the manometer the pressures — will continually balance each other ; but in the operation of deter- mining a vapor density a weighed amount of volatile substance is placed in the sealed tube with the manometer, and when this substance is volatilized by heating it exerts a pressure which can be measured by the movement of the thread of mercury in the capillary manometer. Then, when the two temperatures and the volume of the apparatus are known, the vapor density of the sub- stance can be calculated from the pressure produced by it. The calculations are somewhat complicated, and the execution of the operation does not appear to be as simple with this apparatus as with that of Victor Meyer ; but the method is interesting in its novelty.—Zevischr. physikal. Chem., \xiii, 48. HH. L. W. lie the telescope and total reflecting prism. By this device the path of the illumi- nating light was the same as the path of the reflected light which reached the observer, which served to give plenty of iliumination for the scale without increasing the size ot the openings beyond what was required to see the actual expansion and to measure it. The illumination was provided by a singie incandescent lamp of 100 candle pow er with a spiral filament of stock type A. L. Day and J. K. Clement—Gas Thermometer. 429 giving an intense and concentrated illumination. It was mounted behind the furnace some 20° distant from the open- ings, and was so screened that its heat did not reach the Fie. 6. Fic. 6. The expansion apparatus. A photograph of a furnace and acces- sory apparatus for the determination of expansion coefficients over long ranges of temperature. The illustration represents a later form of appara- tus than that described in.the text. The difference lies mainly in the increased length; the present furnace is arranged to take bars 50°" long instead of 25™ as described. optical parts of the apparatus save in the two beams which enter the furnace for the illumination of the bar. The temperature of the bar was determined at first with one thermoelement and afterward with two, which entered the furnace tube from opposite ends in such a way that their hot 430 ALL. Day and J. K. Clement—Gas Thermometer. junctions could be bound together and moved freely along the _ bar and in contact with it, in order to give a double reading of the temperature at any point desired. In this way we obtained the actual distribution of temperature along the bar corre- sponding to each determination of its length. To complete the system, a standard brass bar was prepared of the same size and shape as .the platin-iridium bar under investigation, but with silver surfaces let in at the ends to carry the divisions. This bar was compared at 20° C. with the standards of length at the Bureau of Standards, and served to establish the absolute distance separating the cross- hairs before and after each set of observations. The method of procedure was now substantially as follows: The standard brass bar was placed in position in the furnace at the temperature of the room. All the necessary adjustments to secure good illumination, to bring the cross-hairs parallel to the scale divisions, and to bring the lines into sharp focus, were then made once for all, and these adjustments were never again disturbed until the series was completed. The field of the microscopes included 5™ of the bar, but only the three scale divisions bounding the 2™™ nearest to the fixed cross-hair were used. Toward the close of the investigation, for an important reason which will presently appear, only the two bounding divisions of the single millimeter which included the fixed cross-hair were read and all the observations which had been made outside this limited region were rejected. Readings were made from left to right in each microscope and then repeated in the reverse direction to obviate errors from the micrometer screw. The temperature for this measurement was determined with a glass thermometer thrust into one end of the furnace tube adjacent to the bar and read before and after the series of micrometer readings. This observation served to establish in absolute measure the distance apart of the fixed cross-hairs of the microscopes. The brass bar was then removed and the platin-ridium bar corresponding to the gas thermometer bulb inserted in its place in the same relative position. It is necessary here again to emphasize the fact that all further adjustment must be made with the bar and not with the optical parts of the apparatus. Having brought the bar into exactly the same position with respect to the telescopes which the brass bar previously occu- pied, and having introduced the thermoelements in such a way that their hot junctions were free to travel along the bar from end to end witbout disturbing it, a second series of observa- tions at the temperature of the room was made in the same way as before. This yields the absolute length of the bar at room temperature in terms of the standard brass bar. dhe furnace is then ready for heating to the temperatures desired. A. L. Day and J. K. Clement— Gas Thermometer. 4381 In the determination of the high temperature scale carried out at the Reichsanstalt in 1900, four observations of the expansion of the. bulb material (250°, 500°, 750° and 1000°) were deemed sufficient, and it was not thought necessary in onr earlier observations to increase this number materially. We therefore began with a 200° interval. After the observation at the temperature of the room, the bar was accordingly heated to 200° C. and sufficient time (about 30 minutes) allowed for the temperature to become constant throughout the furnace, after which a temperature reading was made at the middle of the bar with each element. Observations of length were then made in the same order as before upon the pair of lines adjacent to the fixed cross-hair in each of the microscopes, followed by a second temperature reading at the middle of the bar. After these observations of length and before any change was made in the temperature, nine consecutive pairs of observations were made of the temperature distribution along the bar, first at the center, then on the left section at 5,10 and 12°" out from the middle, then the center repeated; then upon the right section with similar intervals, and again the center—all with both elements. By this means an accurate measurement of the temperature along the bar corresponding to the length measure- ment just completed was obtained. The whole procedure was then repeated at temperatures of 400, 600, 800 and 1000° C.,* after which the furnace was allowed to cool over night and the length of the bar at the temperature of the room again deter- mined. Immediately following this an observation of the _ brass bar was made in order to establish the fact that the dis- tance separating the cross-hairs had not been accidentally dis- turbed by the manipulation of the furnace during heating. At 800° and 1000° the bar is self-luminous to a sufficient extent to enable measurements to be readily made without out- side light, but it was deemed advisable to use the outside light in the same way at these temperatures also. In passing from outside to inside illumination, the lines are at first dark on a bright ground, and then bright on a dark ground, a change to which the eye accustoms itself only with considerable difficulty. The measurements were therefore much more uniform when outside light was used throughout. The measurements of the temperature at once encountered the difficulty that the exposure of the thermoelement in the presence of iridium at a temperature of 1000° contaminates it by an amount sufficient to cause a small but cumulative error. This exposure was necessary with the apparatus as we had * Subsequently, when we had reason to suspect an irregularity in the rate of expansion, these observations were repeated every hundred degrees and then every fifty degrees in the region between 600° and 1600°. 432 A. L. Day and J. Kh. Clement—Gas Thermometer. arranged it, and there was therefore nothing to do but to make the time of the exposure as short as possible, and by the use of two elements fastened together and extending out of the furnace at opposite ends, to so arrange the conditions that any contamination, if sufficient to affect the temperature, would become immediately apparent. As W. P. White of this laboratory has shown in a recent paper,* the most critical portion of a thermoelement is not the portion along which the temperature is constant, but the region where the element passes from one temperature to another. In our furnace, for example, the region of exposure to constant temperature could give rise to no error of reading however much the element might be contaminated in that region, but if a contaminated portion of the element were at any time to come into the region lying between the end of the bar and the outside of the furnace an immediate difference in its reading should become evident. It was therefore arranged that the junctions of two elements should be bound together so as to record the temper- ature of the same point within the furnace and that whenever this combination of two elements was moved toward one end of the bar or the other, that a greater length of one of the elements should be exposed within the furnace than of the other. If there is contamination a difference in reading between the two elements will be immediately conspicuous. In the earlier observations comprising this investigation, only one element was used, and by way of control at the close of a long series of observations a second element was introduced in the manner indicated above. It then became immediately evident that the first element had become contaminated and that the observations made with it were affected to a degree which could not be established after the observations themselves were over, and which therefore necessitated the rejection of several entire series. This misfortune may serve to emphasize the necessity of using more than one thermoelement in all cases where it 1s possible to do so. Three other dithiculties were met with which proved to be sources of considerable inconvenience, and which serve in greater or less degree to place limits upon the accuracy attain- able in this particular apparatus, The first was the temper- ature gradient along the bar, of which mention has already been made. Earlier observers have sometimes been content in similar cases to heat a bar with the electric furnace and to make their measurements upon cold projecting ends, that is, under conditions such that the actual temperature along the bar varies from the temperature of the room to a maximum near the middle of the bar. The resulting temperature to * Walter P. White, Phys. Rev., xxvi, p. 535, 1908. A. L. Day and J. K. Clement—Gas Thermometer. 483 which a given measured length is then referred, is an integral of a temperature range which varies all the way from that of the room to some point considerably higher than that for which the length measurement is recorded. This situation seems to us to “comport badly with the accuracy otherwise attainable in measurements of this kind, if not to violate fundamental definitions. Unless the expansion coefficient can be treated as linear, such a determination is obviously only an approxima- tion. Furthermore there is ample precedent for anticipating inversions in an alloy of this character such that the expansion coefficient of the material below the inversion temperature would differ considerably from that above it. An integration, therefore, in which the temperature range is large may well overlap two physical states 1m such a way that the length measurement loses all significance. We have not been able to establish the fact that such an Inversion exists in the 10 per cent platin-iridium alloy within the temperature range over which these measurements were made, although there is an obyious break in the continuity of the expansion, of small magnitude, which recurs with some persistence, as can be seen from the tables which follow (pp. 487 et seq.). Supposing such an inversion to exist, it would of course fol- low that the expansion would be a discontinuons function of the temperature, a separate expansion coefficient would require to be determined above and below this point, and the two would not bear any necessary relation to each other. If such a situation exists in the present bar, the difference is so small as to be negligible for our present purpose, but the plain indi- cation of an irregularity led us to appreciate the necessity of maintaining the bar as ‘nearly constant in temperature as pos- sible during the length measurements in order to enable us to interpret the measurements intelligently. The problem of accomplishing this result gave us consider- able anxiety. As has been stated above, the scheme of making optical measurements directly upon the bar without multiply- ing device of any kind necessarily involves an opening in the furnace coil opposite each end of the bar, and a consequent cooling of that portion of the bar which is ‘opposite the open- ing. The amount of this cooling, which is greatest at the highest temperatures, reached a value of about 4 per cent in the first furnace coil which we wound. The temperature dis- tribution along the bar is measurable with any accuracy desired by moving the thermoelements about, or its effective average can be determined by direct integration with a platinum resist- ance thermometer of equal length, stretched parallel to the bar. We chose the former method on the ground that it yielded more information, and then sought in addition to diminish the 434 A. L. Day and J. K. Clement—Gas Thermometer. irregularity as much as possible for the reason given above. Accordingly, another furnace coil was wound with the turns closer together near the openings. This changed the temper- ature ovadient considerably without materially improving it | (see Furnace II seq. ), after which a third coil was prepared with still closer windings, which proved to be considerably overcompensated and was rejected. In all, we made five sep- arate trials of this kind, in the last two of which (Furnaces IIT and IV ) a thick- walled iron tube was substituted for the por- celain furnace tube in the hope of gaining increased uniform- ity of temperature through the increased heat conductivity of the tube itself. This arrangement succeeded better, but we found it impossible to so arrange a winding that. the temperature opposite the openings was uniform with that at the middle of the tube for all temperatures between 0 and 1000°.* A winding which gave good results at the lower temperatures gave insufficient compensation at the higher ones. The obvious possibility of reaching a uniform distribution by subdividing the coil into sections in each of which the current could be independently varied was not tried on account of the cumbersome manipulation required, and in part also because the results which we obtained with considerable differences in the gradient appeared to agree very well among themselves. The temperature carried out in the tables in each case repre- sents the integral of the nine pairs of readings described above. The actual error which enters into an observation from the variation in temperature opposite the openings is therefore the error in establishing this integral, which can hardly be greater than 1° C. or 1 per cent. | It will probably occur to other experimenters, as it did to us, that this difficulty with the exposed ends of the bar is due in part to the unavoidable air currents circulating through the small openings, and that these ought to be checked by the introduction of windows. We made two attempts to reach the difficulty in this way, first using quartz windows set in the opening of the furnace tube itself and therefore heated with the tube; and second, by the use of glass windows set in the water jacket and therefore outside of the heated zone. The quartz windows behaved very well until high temperatures were reached, when they become displaced by the unequal expan- — sions in the apparatus, thereby causing displacements in the apparent position of the lines of the scale.. When the windows * A considerable part of the difficulty in correcting the irregular furnace temperature was due to the instability of nickel wire at the higher temper- atures. The oxidation is so rapid that a favorable arrangement of the wind- ings, when obtained, does not give uniform results for more than one or two series of observations. It is our purpose to abandon it in favor of a nickel- chromium alloy or pure platinum. yaar A. L. Duy and J. K. Clement—Gas Thermometer. 485 were removed to the colder parts of the furnace in order to avoid this displacement, sufficient water vapor condensed upon them from within to obscure the field, so that the window scheme had to be entirely abandoned. The second considerable difficulty to be encountered was due to the effect of the outside illumination of the divisions of the bar in a field of rather high power (about 25 diameters). Consider the bar to be illuminated by a beam of light from a fixed source (which remains constant in position while the bar expands) and the light received through the telescope into the eye to be reflected from the polished parts of the bar surface between the rulings. For reasons which appear in the adja- cent figure (fig. 7), this reflected light does not show the lines to be equally* displaced after expansion. The reason for this is Hires 7. Fic. 7. Showing how the lines appeared displaced after expansion. Ac- tual expansion, m to point indicated by the arrow. Apparent expansion, m to n. plain after a brief consideration. If lines are ruled with a sharp tool upon soft platinum metal which is afterward polished to remove the burr left by the cutting tool, the effect is to round off the two edges of each cut to a greater or less extent, and thereby to present approximately cylindrical bounding sur- faces to the incident light. The apparent boundary of the line will then be defined by the reflection of this light from the cylindrical surface into the telescope. Now, if this cylinder be moved laterally in the direction produced by the expansion, the light will be reflected from a different point on the *The small expansions of the millimeter sections themselves have been taken into account, although not explicitly mentioned in this discussion. 436 A. L. Day and J. K. Clement—Gas Thermometer. eylinder and will therefore show the line in a somewhat differ- ent apparent position from that which would be produced by the expansion alone. The drawing is purposely exaggerated to show exactly the character of this optical error. It was our habit in beginning these observations to select three appropri- ate lines upon each end of the bar, and to make all the meas- urements on these, whereupon it was found by a careful exam ination of the results that the displacement of the three lines after expansion differed systematically by a measurable amount and in a manner which could not be accounted for by the movement of the bar. This difference was very puzzling for a long time, but was finally traced to the source described, and this inference verified by actually moving the bar about in the field in various ways without changing the temperature. The consequence of this discovery was to compel the rejection of all measurements made upon lines other than those’ imme- diately adjacent to the fixed cross-hair in the center of the field. The number of observations at each end was therefore reduced to two, but the agreement of the results was very considerably increased thereby. The third and most serious dificulty of all amounts to an essential limitation of the material itself and is therefore not dependent upon the method of measurement. It is the failure of the bar to return to its initial length after heating. In this particular bar, 25°" in length, we actually found dit- ferences between the lengths before and after heating of the order of magnitude of -02™", which varied from one series of experiments to another according as the bar happened to be cooled rapidly or slowly. This quantity is some fifty times larger than the smallest magnitude we could measure, and inasmuch as it depends only upon measurements at the tem- perature of the room, is readily accessible. It is of course immediately obvious that this constitutes a limitation upon the accuracy of gas-thermometric measurements in a bulb of this material, but in this very particular the behavior of platin- iridium is enormously more favorable than that of any of the other materials (porcelain, fused silica) which have yet been applied to this purpose. Although this limitation of platin- iridium would not therefore alone be sufficient to deprive it of continued usefulness for the gas-thermometer, yet when combined with the contaminating action of the iridium which distils out of the alloy at all temperatures above 900° in sutf- ficient quantities to eventually destroy the accuracy of the thermoelement, it has led us to abandon the iridium alloy for the future, and to substitute an alloy of rhodium. This study of the irregularities present or possible in the expansion of the bulb was pursued much more persistently than is usual in an investigation which is but incidental to a A. L. Day and J. K. Clement—Gas Thermometer. 487. much larger one, on account of the unexpected values obtained. The expansion of pure platinum as determined by Holborn and Day* is given by the equation, . A = (8868 ¢ + 1:329 7) 107° while that of platinum, containing 20 per cent iridium, made in the same furnace at the same time, gave A= (8198 ¢ + 1°418 ¢) 10~°. We had expected, as Holborn and Day assumed in their cal- culations in 1900, that the expansion of the 10 per cent alloy ought to fall approximately between the two. When it there- fore became apparent that our observations were leading to a value for the 10 per cent alloy which was of the same order of magnitude as that hitherto found for pure platinum, we were for a long time quite unwilling to accept the result. After the close scrutiny of the apparatus and its limitations. described above, all of which, either singly or in combination, appeared totally inadequate to account for the unexpected expansion coefficient obtained, there remained the single pos- sibility that some confusion had arisen in the preparation of the bar; but Doctor Hereeus, who made the bar, would not admit this possibility. Even then, it was deemed wise to make a chemical analysis of the bar itself, and this was done by E. T. Allen of this laboratory, with the result that the iridium content was found to be 10°6 per cent. There appears therefore no further alternative but to accept the irregular variation of the expansion with the percentage composition as characteristic of platin-iridium, following the well-known example of the iron-nickel alloys. The observations follow: In Furnace I. Temperature Distribution along the Bar. Left Middle Right (Corrected 12cm 1(Qjcm jem Temperature) jem 10cm 12cm +10° | +11° | + 7° Heir — 4° | —18° | —15° +12 +13 + 7 511°2 — 2 —15 —24 +10 +23 +13 700° — 5 —21 — 30 +28 +33 +17 1044°1 —11 —3l —46 * On the Expansion of Certain Metals at High Temperature, this Journal (4), xi, p. 574, 1901. Am, Jour. aE Se ie SERIES, VoL. XXVI, No. 155.—Novemper, 1908 1 438 A. L. Day and J. K. Clement—Gas Thermometer. Expansion. : Corrected A/L Date na Temperature Observed Calculated eogeee < Dees sO, 190722) se 28-7 a. 002635 002661 — 26 Lhe 004871 "004879 — 8 712°9 007051 "006994 +57 Dee. SOO W ewe 700:0 ‘006878 006855 +23 866°6 008653 "008677 — 24 Jan. DO OSE a 504:°0 "004812 ‘004805 + 4 504°4 004813 004810 + 3 690°0 °006763 ‘006748 +15 689°4 "006755 "006742 +13 856°5 008600 ‘008565 “+35 856°4 008610 008564 +46 1044-1] "010616 010699 — 83 1043°8 "010635 "010695 —60 X _ total expansion Equation used for the “calculated” IL. initial length expansions, A=(8869°5¢ + 1°3192¢°)10~° In Furnace II. Temperature Distribution along the Bar. Left Middle Right (Corrected 12m 10cm Hem | Temperature) jem 10cm 12cm ZED es Fo 0° 294:0° —4° | —12° | —27° DS) — 6 — 2g 392°0 —6 —20 —4] —28 — 9 — 4 491°0 19 —17 —33 — 30 —13 — 5 592°5 0) —10 — 30 — 34 —15 — 5 695:°0 +4 — 3 == 27, 855 | 17) —97 | -795°0 EB) ee a at — 52 —21 — 9 | 894:0 +9 + 6 —12 —5l —2]1 —10 994:°0 +8 + 8 —14 *Tnasmuch as the expansion-cofficient which is here being determined itself enters into the determination of the temperature, the two quantities are not independently variable. The temperatures given above are, there- fore, based upon tentatively assumed constants which have been chosen about where the final values were expected to come. The assumed data are these : Zine melting: point = 225-22 -4 5-456 419° Silver ‘‘ ea Rae Denes Tee cc thee 960 Copper ‘‘ COON Ges See UR ee Pee 1085 With actual temperatures 1° higher or lower, the expansion coefficient would not be affected by an amount equal to one-tenth of one per cent in any part of the curve. The assumed values are, therefore, amply exact for the purpose. — A. L. Day and J. K. Clement—Gas Thermometer. 439 Expansion. Corrected A/ 1a Dat Obs.—Cal. a Temperature | Observed Calculated : P Feb. 25, 1908. _- 294°0° "002679 "002692 —13 392°0 "003665 "005638 +27 491°0 "004660 004619 +41 592°5 "0056382 "005651 = 1s 695°0 "006657 "006719 —62 795°0 "007741 "007788 —A47 894°0 "008848 "008871 —23 994:°0 ‘010086 "009991 +95 Equation used for the “ calculated” expansions, A=(8778°6¢ + 1°280127)10~ In Furnace III. Temperature Distribution along the Bar. Left Middle Right (Corrected ’ 12cm 10:2 Hem Temperature) jem 10)cn 12cm — 2° | — 1° | — 0° 297°9° — 1° — 9 — 7° — 6 — 3 — 1 397°3 — | —= 9 — 8 — 9 —. 6 = 496°3 — | — 5 — 9g —13 —10 = 3 594°3 + 1 — 3 — 7 —16 " —12 — 4 646°9 + 2 — 2 — 6 —16 —1]2 — 4 646°6 + 2 — 2 — 9 —17 —13 — 4 -697°0 + 2 0 — 4 —19 —14 — 9d 747°8 + 4 + 2 — 2 — 23 —17 — 6 796°3 =— + 3 6) —27 —20 — 7 846°2 + 6 + 5 + ] —26 ——2Z0 — 8 897°2 + 6 + 8 + 4 — 29 —23 — 9 946°6 + 7 +11 + 8 ie 95 9 1001°5 Seg A SN I ge eae | LXPAnsion. | = 2/L | Corrected Dat | — ae | Temperature Observed Calculated IU eee Apr. 6, 1908 ...| 297-9° 002770 | 002759 | 411 1 393 003739 | 003730 +9 496°3 "004720 004723 — 3 594°3 "005714 "005732 —18 646°9 "006267 0062838 —16 646°6 "006262 "006280 —18 Apr. 8, 1908 ---| 697-0 006800 006815 —15 747°8 "007346 "007360 —14 796°3 "007897 "007888 + 9 846°2 "008445 "008437 + 8 S972 "009013 °009005 + 8 946°6 "009579 "009561 +18 1001°5 °010206 "010187 +19 Equation used for the “ calculated ” expansions, A= (8874-4 + 1:28892")10-° 440 A. L. Day and J. K. Clement—Gas Thermometer. In Furnace IV. Temperature Distribution along the Bar. Left Middle Right Wesese || (Corrected 120m 10cm jem Temperature) Rem 10cm 12cm — 3° | — 2° ON 299°1° — 1° | — 4° — 5° — 9 — 3 0 399°2 — | — 4 — 6 — 9 — 5 — 1 497°0 — 1 = — 7 —14 — 9 — 2 598°3 0 — 3 — 6 —16 | — 9 — 3 648-0 + 1 — 1 — 5 —19 —12 — 4 709°5 + 3 0 — 4 —21 —15 — 4 748°7 + 4 + 2 = = —25 —18 — § 79971 + 6 + 4 ae — 30 — 22 — 7 846°1 + 7 + 6415 =e SW) —22 — 8 900°4 + 8 + 9 + 4 —36 — 26 —10 949°6 +10 +138 + 7 —36 ..| —27 —Il1 1000°5 +12 +17 +11 Expansion. Date Ha@orrected ae Obs.—Cal. | Temperature Observed Calculated Apr 17, 1908 es) 299oIe ‘002763 002755 + 8 | 399°2 "003750 "003730 +20 497°0 "004697 ‘004708 —1 598°3 "005702 "005748 —46 648:°0 "006265 "006268 = 3 709°5 "006889 006921 —32 T48°7 007344 "007348 ae 199-71 “007897 "007890 + 7 846-1 "008423 "008407 +16 900°4 °009018 “009011 + 7 949°6 °009585 ‘009566 +19 1000°5 °010160 "010146 +14 Equation used for the “ calculated ” expansions, A=(8814'1¢ + 1°326027)10~° The mean of the equations derived from the observations in the four furnaces, each weighted according to the number of observations in that particular series, is A= (8841¢ + 1° 3062°) 10 which is the equation used to compute all the gas-thermometer observations. A. L. Day and J. K. Olement—Gas Thermometer. 441 This interpolation formula is a simple equation of two coefficients obtained by the method of least squares, giving equal weight to all the observations. Inasmuch as no one of the differences between observed and calculated values reaches 1 per cent in value, this form of equation, which has been frequently employed for the pur- pose, is perhaps as well adapted to represent the experimental data as another. After it was discovered that the bar after heating did not return to its initial length, but varied within considerable limits from one heating to another, it became apparent that if the contraction upon cooling was not uniform, the expansion on reheating was probably also irregular to the same degree, and that the room temperature observations could not be expected to follow this or any other simple equation very consistently. That such irregularities exist: and attain such magnitude as seriously to limit the power of any simple formula to reproduce the expansions over the whole range will be immediately apparent from an examination of the columns of differences (Obs.—Cal.). It. is more directly observable in the experimentally determined values of the expansion between 0 and 300° taken from the four series which have just been given. Measured Expansion in Millimeters between 0 and 300°. Wee a0. WOO fet ete! ey a ne 0-687™™ ens. 1908) cus ee .- 0°681 TBAT Uike TOR pAG) OG ie teehee Dies ts Mee Spee 0700 A evit LO0S. rr. eI 0:696 By way of experiment we tried an equation of three coeffi- cients on the last two series, both of which contain observations at 50° intervals, omitting in each case the room temperature observation in which the irregularity in the expansion itself chiefly appears, and found it possible to reproduce the meas- ured behavior of the bar in the region from 300° to 1000° with differences less than one-fifth as large as those recorded in the tables above. There is, therefore, abundant evidence that the uncertain region is confined to the lower temperatures and that the higher temperatures have so far offered no serious difficulty or irregularity, either in measurement or convenient represen- tation. The expansion measurements over the entire range from 0° to 1000° are probably in error by about 0°5 per cent, most of which is directly attributable to these irregularities in the behavior of the metal at the lower temperatures. In the gas thermometer this corresponds to about 0°25° at 1000°. This uncertainty in the expansion of the metal at low tem- peratures appears in the gas thermometer data as a difference, 442 A. L. Day and J. K. Clement—Gas T. hermes from one day’s observations to another, in p,,—the pressure of the gas at 0°,—which will be found to vary irregularly within narrow limits. The Pressure Coefficient of Nitrogen.—A number of deter- minations of the pressure coefficient of nitrogen, under differ- ent initial pressures, were made by observing the pressure inside the bulb when it was immersed alternately in ice and in steam, and with the following results : __' Pioo — Po Initial Pressure a = 00m No. of Observations leet 0°003665 4. 550 005668 5 744 "003670 6 985 "003673 12 Chappuis has obtained the following values of a: Initial Pressure pepe eee 100 529mm. 0:00366811 Traveaux et Mémoires : du Bureau International oe pe es des Poids et des Mes- ae 95 00364477 ures, vols. 6 and 12. ety 9) : { Computation of Results—The tormula for the constant volume gas thermometer may be written in the form, SNE Pe PBs 1 + at 1 + at, 1 Shot are ee 1+ at’, 1+at’, In this equation : V..:= welime ofcbullbiati® 20s see ye eee 195°547°°. V = volume of bulb at 7°. p, = initial pressure, 1. e., pressure when bulb is at 0°. p = pressure at temperature of 7°. v, = portion of “unheated space” enclosed in furnace (in which temperature varies from the tempera- ture of the bulb to that of the room)--_-__---- O-L61°°. v, = portion of “ unheated space” outside of furnace .. 0°128°. ¢, = estimated mean temperature of v, when bulb is at 2. ¢’, = estimated mean temperature of v, when bulb is at 0°. t, = temperature of v, when bulb is at @°. t’, = temperature of v, when bulb is at 0°. a = expansion coefficient of nitrogen under constant volume, 8 = linear coefficient of expansion of platinum-iridium alloy. at pis OPN v, NG Writing A 3es Tia cer Ty and GBB) sing 1 +a, A. L. Day and J. K. Clement—Gas Thermometer. 448 the equation may be transformed into a more convenient form for computations : : D 1+ 56t tana cee ° Pee Ae = B 0 Here 38¢ represents the correction for the expansion of the bulb and A — = B is the correction for the unheated space. In computing p the mercury columns were corrected in the usual manner for temperature and latitude. Gas Thermometer Measurements.—Table I contains some of the earlier results, which were obtained after the temper- ature gradient along the bulb had been partially corrected. During this series of observations, the temperature difference between the middle and either end of the bulb varied between 50 and 150 M.V. (5° to 15°-). As it was impossible entirely to eliminate the gradient with the arrangement of coils in use at this time, the heating currents were adjusted so as to have the gradients toward the top and bottom of the bulb of opposite sign and of nearly equal value, thereby materially reducing the magnitude of the correction to be applied. Before beginning this series of observations and again after its completion, the thermo-couple ‘ P” was calibrated by deter- mining its electromotive force at the zinc and copper melting points. From the results which follow it will be seen that the electromotive force of the thermoelement, at the temperature of melting copper, has been lowered 15 M.V. (1°2°) through iridium contamination during the series of measurements. 1906 Zinc: — Copper. 72) 3g LRU Remnag pea F.* 3398 10488 M. 3398 10483 3398 10486 Moy sleet) 2s F. 3396 10469 M. 3398, 10472 3397 10471 After these observations (Table I) the furnace was rebuilt. In place of the heating coil of platinum-iridium alloy, a coil of pure platinum was substituted. At the same time the arrangement of the two auxiliary heating coils was so modi- fied that by proper adjustment of rheostats, the gradient along the bulb could be reduced to 0°5° or less. In order to eliminate, as far as possible, any error due to the contamination of the thermo-couples with iridium, the couples * F = Freezing point, M = Melting point. 444 A. L. Day and J. K. Clement—Gas Thermometer. Tape I. Tnitial pressure, 302°3™™, a= 0°003665. Average temperature difference between middle and either end of bulb, 10°. Equation used for “ calculated ” temperatures, T = 50°19 + 0°11176E — 1:289 x 10-°H”, 1906 ie een Op 5, Temperature) Temperature) T (obs.)— (mm) ec noeolte Observed | Calculated £ (cal.) Ap]. 30. 30°209 BOS IIS 396°8° 397°8° —1°0 | 4738: 550°4 550'8 — “4 6232: 696°5 696°6 — -] (a02s 797°6 T97°5 amie we 8428: 900-2 900°5 — 3 9547° 998°4 9996 —1:2 11004: 112375 1123°9 — 4 May 1, 30°220 May 2 POG = 368°6 369°8 —1°2 | 4944- ole? Dy Tle2 — 0 6448: TAGS (Ane? + °6 IBS S 818°6 818°2 + ‘4 8094: 869°7 870°3, — ‘6 9040: 954°5 955°1 — °6 9750° 1016°4 WOW es — °9 10583°t 1089°9 1188°5 se ils, May 3} 30°229 May 4| 3928: 470°0 469°3 +. °7 | 5186: 595°2 Hse + 2 5944: 669°1 668°9 + °2, 6725° 743°8 743°4 -_ -4 T755° 839°7 839°3 + °4 eee ONE 973°1 974-1 =i oe 10135° 1049°9 1050°4 — °5 May 5) 30-230 May 7 3918: ¥ 468°0 468°'3 — °*3 6926° - . 762°8 762°4 + *4 8910" 943°4 943°6 — 2 10631° 1092°8 1092°6 + 2 1264" EO ip habe met +1°6 May 9| 30°252 | were calibrated from time to time by metal melting point determinations. Columns 7 to 10 of Table II contain the * Observations below 415° were not used in computing the parabola. + Temperature fell 2° during observation. — > Temperature in Degrees, A. L. Day and J. K. Clement—Gas Thermometer. 445 — E.M.F’s of the standard thermo-couple “ W” for these eali- brations. As a check against accidental errors of observation, all observations were made in pairs, with an interval of from five Fie. 8. J SSS eek eee eee eee eee i eee se ee ee eh eels bee feo | JS eae ee ae ae eee ees JSS ae Rae SaaS eee eee ea eae _ Se RSS anaes ee aaa eee ee J JS See Sees es ae eee aaa eee ees PEELE EEE EEE Ppt a te a ee ee al I ae be ae oe 1100° 1000° |_| (sol gs BE eo Oe ap FR | 2h SSeS aS SES ese se 2a aR eee SSS SSS et 900° 800° ie Ba Sh PEE ae Bs ea a _ JSS Se EMS es eS aes ea eee ees see ee 600° Se =eS Sn SEZ i Ea dF | Se _ JS Se Aes eer eos ee wes ae ees _ ene 2 eee See ae eee Ree eee eee JU Se Anes ae eae ees eae eee eee ttt ep 7240 es Ja PY cas SS A es J Zh 2 SRS ORE Sea Ree eRe As eee eee Zoos RE Sa Saas eee ae eae ee eee eee eae 2 Sea SE Sea ae Sse aaa eee eee sees 500° 400° 4000 5000 6000 7000 8000 9000 10000 ——> MV. Fic. 8. Curves showing the systematic differences (much magnified) between the observations and the parabola which best represents them. For the dotted curve the ordinate (above and below the full curve) is 1 div. = 0-2”. to ten minutes between. The constants of the equation: E572 + 01124998 — 1°35512 « 10-°E 300° 446 A. L. Day and J. &. Clement—Gas Thermometer. TABLE 3 Thermo-couple ‘‘W.” Initial pressure, 287°5"™. a=0-003665. Maximum Equa- temperature difference between middle and either end of bulb, 0°5°. tion used for ‘‘ calculated” temperatures, T = 61°72 + 0°112499K — 1°35012 x 10-* KE? E.M.F. THERMO-COUPLE ‘* W ”’ Initial [P°™™-\Opserved) on (op No. Date Pressure es Temper- ae 5) Po. | ature +} Ca ae 2 M.V. Zine 1907 Feb. 4 3403 Mch. 6] 28°755 1 ge 6 3967" 414°49°| — 0°64° 2 ee 6 d371° 414:96 | — ‘a9 3 Ho 6 d002°* | 4381°76 | — 40 4 a 6 30386°* | 4382°25 | — -82 es 6 | 28-739 4) a 7 3416° 419:96 | — °28 6 “s 7 3417° 420°00 | — ‘dl- 7 fe 7 do71* | 486°23 | + 06 8 cs i d974°* | 436°40 | — -08 | 9 oh a 3789°* | 458°55 | + 03 10 oe 7 3793°* | 458°94 | + O01 on 9 | 28°734 11 es 12 3309°* | 408°51 | — °63 12 ef 12 3310°* | 408-71 | — ‘538 18 a 12 3430° 421-50 | — ‘1d 14 ie 12 3429° 421°50 | — °04 15 os 12 3040°* | 432°83 | — 15 16 ms 12 3040°* | 432°89 | — 09 s 13 | 28°785 ‘“ 14 F, 3405 M. 3404 ce 14 | 28°735 1g os 15 10430°5 | 1078-02 | + 31 18 is 15 10419° |1077°05 | + °31 +e 15 | 28°7389 ie 16 19 ie 20 9671° |10138°17 | + °22 20 a 20 9664- |1012°48 | + -:08 21 ee 20 10446° |1079°96 | + -95 22 ae 20 10426: |1078:26 | + 93 23 eS 20 10484° |1079°02 | + 1:02 24 i 20 9895°5 |1032°75 | + °49 25 AS 20 9895" |1082°70 | + 48 eae F. 3405 M. 3405 Be 23 | 28°755 Apr. . 2/3 #3 22 | 28°751 26 # 23 3383" 416714 | — -70 27 dg 23 3384° 416°26 | — °64 28 é 23 4393° 520°07 | + 29 29 es 23 4388° 519°71 | + -44 30 J 24 4095° 489-79 | + 11 31 ee 24 4097- 489°89 | + ‘01 Silver Gold Copper F, 9046 F. 9046) 10214 10214 10461. 10461. 10463,» 10462. 10497. 10456. 10454. 10454. * In computing the constants of Equation (1) the observations marked * in Table II were omitted in order to equalize the intervals between points. + The column of ‘‘ calculated” values has been omitted-from Table II for The column of observed temperatures and the convenience of the printer. the column of differences together contain all the information required to reproduce the calculated values obtained from the equation above. A. L. Day and J. Kk. Clement—Gas Thermometer. 44.7 _., |/Ehermo-| | “ Initial couple Observed T (obs) ate |Pressure} {, Ww | Temper- | — T (cal) Po. | yy vy. | aturet | \Apr. 24 | i 5073-51 S8007) 4... -40° fas‘ 24 LeaO re Om aS Oe dist. SA as 24 | | 6104° | 688:°44 | + ‘52 es 24 | 8°756 | 6099°5 | 687°84 | + °35 ce 95 a 26 | 4700-5 | 500°81 | + °23 bees 26 4698° | 500°65 | + 32 hess 26 | 6962- | -769°49 | +. °28 pane 26 | 6957°5 | 768°86 | + °03 faye 26 8020°5 | 866°45 | — °40 Se 26 | 8029= .) -S67-99_ )-, .-37-| fees 27 5648- | 644:34 |} 4 -44 yeas 27 | 0646° 644-24 | + 55 bess 27 | 9036°° | 957°20 | — 42 NOE as te (9004-5. | 9908°31 | —_ 739 rs 29 Momo. fea otO0 sa — 10286 ae 29 | 3716° 450°76 | — °29 Ee 29 10190°5 | 1057°32 |— -10 ae 29 8-755 |10188- |1057:°09 | — -11 Weert 29 May 1/10 ge 13 | 3480: AD AO =o 4G fee 13 | 3438°5 | 422°06 | — “47 ae 13 | 67715 |. 75162 |.4+ 18 e 13 | 6769°5 | 751°36 + °18 cis 14 | | 5845°5 | 663°62 | + -59 af 14 | — 9848- 663°33 | + :06 or 14 | 9274: O7Sst0 Me coo pt 14 | 9281: 978°82 | — 28 ‘et 16 4907° 571°49 | + °37 hea 16 4906- O71°37 | + 739d oy 16 | (242° 795-42 | + 05 ry 16 7243° 7935 °46 pease 16 7240: 795°24 | + -06 Kaas 16 10188- |1056°91 | — :29 | eS 16 10177- | 1056-15 | — -12 = 17 6338° 710°27 | — -08 ‘3 17 | 6341: PLO 1Oeleee ee LE Bs aSr¢ 8272: 888°96 | — °63 | ae ay 8277: 889°50 | — °d4 tyes 17 9024: 955°74 | —~ °82 | paste ay | 9015- 955°01 | — °76 } rier 18 3891- 468:56 | — °38 ple 18 3894°5 | 468°94 | — °35 aie 18 | | 7730°- | 839°99-| — ~.-41 | ais 18 | 7725° | 839°46 | — °45 | Eire 18 S741-- | 930-700) — --84 ie 18 O40 a4) 950-62 | —-" “85 Pt 19 | 28°763 | ** 20/22 | | és 23 K.M.F. THERMO-COUPLE ‘‘ W” Zine | Silver | Gold Copper | | | | | | | | | | Ip, 3400 9046 | M. 3400. 9050 | | | | | F. 3398 9040 10212 M. 3398 9040) 10216 | IF. 9042) | M. 9042 | eae 448 A. L. Day and J. Kh. Clement—Gas T hermometer. were computed by the method of least squares. In computing these constants the observations marked* in Table II were omitted in order to equalize the intervals between points. B comparing the T(obs.)—T (caleul.) values (Table II, column 6) of the various pairs of observations, it will be seen that any two values at the same temperature agree within 0°1°. With one exception, the differences between observed and calculated temperatures are all less than 1°. The average difference is 0-37° and the probable error of a single observation is 0°25°. The foregoing table (Table II) contains a complete series of 76 observations, without omission, covering a period of more than three months in time, in the order in which they were made and with the control melting points through which the constancy of the thermoelements was assured. If we now regroup these observations in the order of increasing tempera- tures and combine the pairs referred to above, the relation between the observed and calculated curves appears in a new and interesting light. (Table III.) The average difference in column 5, Table III, is somewhat smaller than in Table IL. The most notable feature, however, of these differences is their sytematic variation. Below 500° the observed values are less than the calculated ones; from 500° to 800° the observed values are greater, from 800° to 1000° the calculated values are greater and above 1000° the observed values are greater. If the calculated temperatures be taken as the ordinates of a curve of which the E.M.F’s of element “ W” are abscissas, the resulting curve will be a parabola, slightly concaved downward (fig. 8). 7 | Tf a the differences in column 5, Table III, be plotted on their proper ordinates, measuring upward from this parabola when the difference is positive and downward when it is nega- tive, the second curve will cross the first in three points, form- ing two loops of about equal length and area. In fig. 8 the dotted curve represents the observed temperature curve, with the deviation from the parabola plotted on an exaggerated scale. From this diagram, as well as from the figures in column 5 of Table III, it is apparent that a simple equation of the second degree is no longer quite competent to express the electromotive force of the thermc-couple as a function of temperature with the full accuracy of the measurement. In their paper on the electromotive force of metals of the platinum group, Holborn and Day* state that the “relation between the thermoelectric force and the temperature in metals of the platinum group could be represented, within wide limits, with an accuracy of +1:0° by a function of the second degree.” The results of our experiments are represented by a function + This Journal (4), viii, 46, 1899. A. L. Day and J. K. Clement—Gas Thermometer. 449 of the second degree, between 400° to 1100°, with an accuracy or -+0°5°. The differences between our observed temperatures and the temperatures calculated from formula I, however, cannot be attributed wholly to observational error on account of their systematic variation. For greater accuracy than 0:5° it will therefore be necessary to use a different equation— perhaps an equation of four parameters. Taste IIL. (1) T = 51°72 + 0°112499EH — 1°35512 x 10-°E? } _Thermo-couple Temperature Gpepation ae Fos pa eal 1 ne Observed | Calculated 1, 2, 5, ? 13, 14, | | 26, 27, 50, ol 3407-7 418°97 419°35 — 38° 46, 47 3717° 450°88 451-16 — 28 [8 Re ie 3892'8 468°75 — 469°13 — “38 30, 31 4096° 489-84 489°80 | + 04 28, 29 4390°5 019°89 519-53 + 36 36,3% . | 4699°3 500°73 500°47 + 26 58, 59 4906-5 1:43. |- 9571-08 + °30 D2, 33 | d073° 587°99 087756 + °43 42, 43 5647" 644°31 648°80 + ‘ol d4, 39 | 5846°8 663°48 663-16 + °382 34, 35 | 6101°8 688-14 687°73 + -°41 65, 66 6339°5 710-49 | 710°45 + °04 02, 09 6770°5 70144 | = 751°24 + ‘1d 38, 39 | 6959°8 76918 | 769-06 + °12 60, 61, 62 7241°7 795°37 795°39 + -02 73, 74 T727°5 839°71 840°15 — “44 40, 41 | 8024°8 867-22 867°25 — 03 67, 68 | 8274°5 889°23 889-82 — 39 75. 76 8740°5 930°66 | 931°90 — 84 44, 45, 69, 70 | 9032-2 956°68 957°28 — 60 56, 57 9277°5 978-45 78°78 — ‘32 19, 20 | 9667°5 1012-80 1012°67 + 13 24, 25 9895°3 1032-73 1032-25 + °48 48, 49, 63, 64 | 10185°9 1056°87 1057-04 — ‘17 17, 18, 21, 22, 23 | 10430-0 1078°31 1077-68 + 63 The numbers in column 1, Table III, correspond to those in column 1, Table Il. The figures in-columns 2 and 3 represent the means of the corre- sponding figures of Table II indicated by the numbers. After the series of observations represented by Tables II » and III, the bulb was evacuated and refilled with nitrogen under a somewhat higher initial pressure, p, = 325™™. With this filling, the results contained in Table IV were obtained. 1 450 A. L. Day and J. K. Clement—Gas Thermometer. TaBue IV. Thermo-couple “ W.” Initial pressure 325™™, a= ‘003665. Equation used for calculated temperatures same as in Tables IT and III. T=51-72 +0'112499K — 1:35512 x 10~°EH’. Thermo- TEMPERATURE - T (obs.)— 1907 couple “ W” : T (cal.) Microvolts Observed Calculated ie June 3 4025: 482°10° 482°58° — ‘48° 4026° 482°15 482°68 — *d3 5009° 981°29 581°23 + °06 5017°5 08222 582:07 + °15 5978" 675°32 675°81 == oY) SOT: 675°07 675°43 — °36 June 4 6238'5 700°88 700°81 + °07 6239: 701°05 700°S6 + ‘19 June 6 6992° THEANO HOOT + °03 6988: alice (EHO + 02 May 29 7486°5 818°65 818°00 + °65 7T495° 819-00 818°78 + °22 June 4 7954: 860°52 860°81 — °29 7952° 860°47 860°63 — °16 June 3) 8488°5 908°47 909°03 — °56 8488°5 908°57 909°03 — "46 June 4 “=O OMG: 955°30 955°86 — °56 9011: 954°91 955°42 — *5] June 5. 9474: | 99532 995°91 — “59 | 9480: 995°86 996°43 — ‘57 June 6 | 9975° 1038°82 1089°07 — 25 | 9940: 1035°65 LO35522 + *43 10197°5 1058'd50 1058°53 — ‘03 Column 5, Table IV, contains the differences between the observed temperatures and the temperatures calculated with the equation used in the previous series (Table ITT) T = 51°720 + °11250K—1°35512 x 10-°H’. The average difference is 0°33°, and the maximum difference 0°65°. The agreement between this series of observations and the preceding one is so extraordinarily close that not only the same equation serves for both but the differences “ obs.—cale.” A. L. Day and J. K. Clement--—Gas T. hermometer. 451 are of the same order of magnitude and similarly distributed, thereby confirming the conclusion that the observations are more accurate and consistent, within this temperature region, than the interpreting parabola. Metallic Melting Points.—By way of establishing perma- nent records of these observations, the usual procedure was adopted of determining various metallic melting points which occur within the range of the temperatures investigated, with - the thermoelements which had been directly compared with the gas thermometer. The metals chosen for this purpose were gold, silver, copper and zinc. Metal melting points were given the preference over pure salts which have been repeatedly suggested for this purpose, (1) on account of the greater sharp- ness of the melting point, (2) on account of their general availability for such determinations, and (38) because of the now very generally established custom of comparing the results of different observers through the medium of these standard melting points. In choosing the materials for such determinations, two not altogether concordant standpoints must be recognized; (1) the materials used must be of absolutely known composition and of high purity in order to give the melting point determina- tions a positive significance ; (2) the same materials in the same purity must be easily obtainable by other investigators in order to enable the results to be conveniently utilized by others if desired. The metals used in this investigation were from various sources which will be specified below. Each has been very carefully described and analyzed by Dr. E. T. Allen of this laboratory, whose report is printed in full. We pre- pared none of the metals ourselves. Those which were used were purchased from firms who may fairly be expected to supply the same nominal quahty to any other investigator who may care to use them, but it must be emphasized in this con- nection that metals furnished under the same description by the same dealer at different times have not always proved of uniform purity and probably cannot at present be expected to do so. ‘he variations in the thermal behavior of the different samples is not great, never amounting to more than 1° in our experience ; but we are of course unable to offer any guarantee that the same metals obtained in future will remain within this limit, nor is the dealer’s guarantee at present a sufficient protection. As the situation now stands, the errors in the gas thermo- meter measurements are rather smaller than the differences between the melting points of different samples of a given metal obtained at different times from the same dealer and of the same (nominal) purity. This may serve to emphasize more than 452 A. L. Day and J. KB. Clement—Gas Thermometer. ever before the desirability of some provision, preferably by the Bureau of Standards, for standard metals, the uniform purity of which ean be absolutely depended upon, in terms of which such constants can be expressed. In the absence of such a provision, it is difficult to see just how to make the gas scale conveniently available for general use in its full accuracy. This is furthermore a matter of considerable importance in view of the extended extrapolation to which the gas scale is frequently subjected by the use of thermoelements or other- wise. Supposing the metal melting points to be capable of reproducing the temperature curve correct within 1° at the copper point (1081°), the extrapolation to 1500° may easily remain uncertain by at least 5° in the hands of different indi- viduals using the same functions for the extrapolation. _ Nor is this the only difficulty to which this use of standard melting points may lead. On account of the systematic errors attending the application of an equation of the second degree to the gas thermometer observations, to which reference has been made (p. 448), the usual standard melting points are not competent to reproduce the gas scale exactly. A curve of the second degree, developed by least square solution from seventy- six observations which show systematic deviation, cannot be reproduced with only the three observations which chance to represent the standard melting points. This will be imme- diately apparent, though on a somewhat exaggerated scale, if we locate on the dotted curve in fig. 8 the tempera- tures corresponding to the zine (419°5°), silver (957°9°) and copper (1080°9°) points and then undertake to reproduce the ‘calculated ” curve from them alone. . As this is the method almost universally used for the pur- pose, it is worth while indicating by a special case exactly where it leads. Column 38 of the table below contains the actual melting points of four of the purest metals interpolated from gas thermometer observations close by. Column 4 contains the same points computed from Equation lL If TABLE V. Temperatures Memineratiues Element extrapolated Gain cared EW? from nearest I Difference observation Fine fsa 3403° 418°48 418°87 —'39 Silver ls. 9046° 957°90 958°50 —°60 Goldicinves! 10214: 1059°26 1059:'42 —'16 Copper spect 10461° 1080°92 1080°29 + °63 A. L. Day and J. K. Clement—Gas Thermometer. 458 now, following the usual practice, we take the melting temperatures of zinc, silver and copper and pass a parabola through these, we obtain | T=55'51 +0°110789 E— 12197 (10) °K? (Equation II) instead of Equation I. And if, by way of illustration, we recompute all the temperatures of Table III, using this new equation and compare the results with those obtained from Equation I, we have an illustration (Table VI) of the effect of Taste VI, Equation I T=51-72+0-112499 HK—1°35512 (10)—*°K? Equation IT = T=55-51+0-110789 E—1-2197 (10)-°E? No. of Termperauare Obs.-Cale.| Obs-Cale. Observation Ob Caleulated | Calculated I II | served I II 125.6, 13, 14 eae o0ol | 418-97" | 419°35° | 418°97° | —-+38° 0:00 46,47 450°88 | 451°16 450°46 | —-28 + 0°42 Wa 12 468°75 469°13 468°31 == iG + 0°44 30, 31 489°84 489°80 488°84 +04 +1:00 28, 29 519°89 | 519°53 518°42 + +36 7 36, 37 550°73 550°47 549°21 + +26 +1°52 58,59 571°43 571:08 569°73 +35 =n 32.33 587°99 | 587°56 586°15 +43 + 1°84 4943 644:31 | 643°80 642°94 S55 | LL Oy 54,55 663°48 663°16 661°58 ERD +1:90 34, 35 688'14 | 687°73 68611 +-4] DB 65, 66 710°49 710:45 708.84 4-04 Se TRS 52,58 75)°44 751:24 749-70 to +1°74 38, 39 769°18 769-06 767°50 +'12 | +168 60,61,62 | 795°37 | 795°35 793°85 aE -()2 E152 73,74 839°71 | 840-15 S38-S0e 44 +0°91 A ale 2 le 86 7°22 867-25 866°04 == )% JE ISIS, 67,68 889°23 | 889°82 888°72 a9 +0°51 75, 76 | 980°66 | 931°50 930°68 | —°'84 — 0°02 44,45,69,70 | 956°68 | 957°28 956-68! | ——"60 0:00 56,57. | 978-46 | 978-78 | 978-35 | —-32 EOP) 19, 20 | 1012°80 1012°67 | 1012-57 45413 + 0°23 24,25 | 1032°73 | 108225 | 1032°37 + °48 + 0°36 48, 49,63, 64 | 1056°87 | 1057°04 | 1057°45 esol T —0°58 17,18,21,22,23, 1078°31 | 1077°68 | 1078°31° | +°63 0:00 Am. Jour. Sci.—FourtH Series, VoL. XX VI, No. 155.—NovemsBeEr, 1908. 454. A. L. Day and J. K. Clement—Gas Thermometer. interpolation of this character even when used with the best experimental data which we have obtained. The error amounts to 2° in the region 600—700°. The solution of this difficulty is obviously to obtain additional fixed points and thereby to reduce the interval for interpola- tion but we have so far found difficulty in obtaining suitable substances. The metals which melt in the desired region either are not obtainable in uniform purity or easily become oxidized or otherwise contaminated during the manipulation necessary for a melting point determination. Suitable eutectic mixtures may eventually offer a solution of the difficulty. The Metals Used.*—After some investigation, it was found that we could obtain in sufficient quantity, silver, copper and zine which ranged in purity from about 99°94 per cent to 99-997 per cent, and gold which was probably still purer. These figures do not include oxygen nor carbon (except in case of the silver), for, since the melting points of the metals had to be determined in carbon crucibles, it is evident that the pres- ence of these impurities would have no significance for present | purposes. Of course it would have been possible to prepare these metals, or at least the zine and copper, in still purer condition, but it was not thought to be worth while, since it is improbable that the most refined of present-day methods could safely determine any difference between the melting points of the chemically pure metals and those actually used. The Gold.—About 250 grams of “ proof gold” were obtained from the Philadelphia Mint. It was prepared by Mr. Jacob Eckfeldt. A sample of gold prepared in a similar manner by Mr. Eckfeldt was used by Prof. Mallet mm his determination of the atomic weight of this metal. The method of purification is given inthe Am. Chem. Jour., vii, 73, 1889. Prof. Mallet found no systematic difference between this gold and two other samples, one of which was obtained from the Mint of England, and the other of which was prepared by himself. In view of these facts, it was evidently unnecessary to analyze the old. : Regarding the methods which were used in the analysis of the silver, copper and zine, there will be no need of giving all the details, especially where accurate methods of procedure are well known, but in view of the very small quantity of impurities estimated, some explanation and some conclusions as to the accuracy of the data will be presented. Most of the work was done in a new laboratory under exceptionally favor- able conditions of cleanliness; large samples, generally 100 grams, were taken for analysis and the reagents were subjected to rigid examination. Separations were always repeated, in *By EH. T. Allen. A. L. Day and J. K. Clement—Gas Thermometer. 455 some cases many times, and filtrates were not rejected until they had been reduced to small volumes and had been proved free from ee elements looked for. Of course, the accuracy of such work is most satisfactorily tested by synthetical methods. Mylius and Fromm,* by using a preparation of metallic zine in which they could find no impurities, were able to detect gualitatively as little as 0°1"8 of lead, cadmium or mercury, in a solution containing 40 grams of zine, without difficulty. Quantitatively, I have never found greater varia- tions than ‘002 per cent in duplicate determinations of the heavy metals in silver, copper or zine, with a single exception which was rejected, and some of the figures agree closely in the ten thousands of a per cent. The Silver.---This metal as well as the gold, was prepared by Mr. Eckfeldt at the Philadelphia Mint. A block weighing about 100 grams was cut from a larger brick with a hard cold chisel, and after cleaning, transferred to a large casserole of Berlin porcelain and dissolved in a slight excess of nitric acid. During the operation the dish was covered with a watch glass. A small black residue was now filtered off on the felt of a large porcelain Gooch crucible, washed and dried. The asbestos of the felt was previously heated to redness. The residue was then laid in a porcelain boat which was slipped into a combustion tube containing copper oxide and heated in a eurrent of oxygen. The outflowing gas was passed through a very dilute standard solution of barium hydroxide, 1° = 0:97™8 of CO,, in which a decided white precipitate appeared at once. The excess of baryta was then titrated with standard acid. A blank determination previously made gave no precipitate in the baryta water. This determination is of no importance as regards the melting point of the silver, since the metal had to be melted in graphite, but considering the source of the silver and its unusual degree of purity, the determination may be of some interest. What remained of the residue after the carbon was burned, was extracted with aqua regia. The solution was evaporated to dryness, taken up with hydro- chloric acid and the gold precipitated by sulphur dioxide. The filtrate from gold gave a slight black precipitate with hydrogen sulphide. This precipitate weighed only 0:1™ after it had been glowed in a small porcelain crucible, but it remained black, dissolved in a few drops of aqua regia which left a yellow stain when evaporated, and gave a very strong rose color when dissolved in water and tested with a drop of potassium iodide,—all characteristic of platinum. It was sus- pected that a trace of platinum might exist in the acid used to dissolve the silver, but a blank test on the same quantity of * Zeitschr. anorg. Chem., ix, 144, 1895, 456 A. L. Day and J. K. Clement—Gas Thermometer. reagent proved the contrary. The silver solution was now diluted to several liters and precipitated with hydrochloric acid. The filtrate was evaporated in porcelain to a small volume and in this the remaining impurities were sought for by well-known methods. Only lead and iron and the merest trace of copper were found. A blank determination was made for iron. Found in the silver + reagents, 0013 per cent; in the reagents, ‘0002 per cent; leaving -0011 per cent in the silver. For the estimation of sulphur, a separate portion of 38 grams was taken, the silver was removed in the same manner, and the filtrate evaporated to dryness in porcelain. The small residue was then evaporated again with hydrochloric acid to decompose nitrates. The final residue was dissolved in a small volume of water acidulated with hydrochloric aeid, filtered to remove any silver chloride which might have escaped precipitation and precipitated with barium chloride. Found 1-478 BaSO,, while the same quantity of reagents gave 0:4™8 BaSO, : S = ‘0004 per cent. Analysis of Silver. WAN SHIMON 2 i Sagem weit pee hal au none Sb Bia bike iy eee cee 6¢ Sr Dr) OE Re Sea bal 6¢ ACU fe tet Geo ea sgt as "0005 | ie RG alee eS ee eS ‘0001 Opes re ot eRe te merest trace Bier) 225 ee ae een OMe ds opener 2 AES pe hae <9 ‘0008 hot Vic eek a See none Cd Sy 1 Bg eG Ok aCe (1 Tn PN CEE Ne aaah ep ee 66 ING ee ea ee Co ere tele is Cpe tars (45 | eng cae SA hse Aes ‘0011 S the US Canes Pa aes eee QO O4 Cee. Soest, SOUS} 0032 % The Copper.—The copper was of the form known as “ cop- per drops cooled in hydrogen” and was obtained from Eimer and Amend of New York. Not all copper of this brand is equally pure. The sample analyzed was a portion of a 25 |b. lot. The method followed in the analysis was essentially that of Hampe,* in which the copper is separated from the impuri- ties by precipitation as cuprous thiocyanate. A 100-gram por- *Lunge, Chem.- tech. Methoden, ii, 202. Chem. Ztg. 1691, 1898. A. L. Day and J. K. Clement —Gas Thermometer. 457 tion was placed in a large casserole of Berlin porcelain, dis- solved in nitric and sulphuric acids and the solution then evapo- rated to drive off the excess of nitric acid. This troublesome operation can be greatly facilitated by the use of a crown burner, though as dilution and evaporation have to be several times repeated, small losses are difficult to prevent. Duplicate determinations, however, proved that they were entirely neg- ligible as regards the small percentage of impurities. The sul- phate of copper was now dissolved in water and diluted. A little HCl was added, and after standing, the solution was fil- tered. The residue left on the filter was extracted with am- monia to remove silver chloride and the remaining part of it was treated with aquaregia. There was still left a little silica, from the porcelain dish in which the copper was dissolved. The solution obtained by aqua regia after the nitric acid was entirely driven out by hydrochloric acid, was tested for gold by sulphur dioxide. There was no precipitate in the cold even after long standing, though evaporation caused the pre- cipitation of about half a milligram of black metal. This remained black on heating, dissolved only partially and with difficulty in aqua regia, and with sulphuric acid and ammo- nium nitrate gave a faint blue color. These tests indicate irid- ium, though there was too little to identify with certainty. The rest of the solution which had been tested for gold was precipitated by hydrogen sulphide and the precipitate was fil- tered, washed and burned in a porcelain capsule. It formed a yellow chloride with aqua regia, gave a precipitate with am- monium chloride and a very strong test for platinum with potassium iodide. This platinum did not come from the acids used to dissolve the copper, since the same quantities were very carefully tested by hydrogen sulphide after nearly the whole portion had been driven off by heating in porcelain, and found to contain not a trace. The solution containing the copper was then warmed and saturated with sulphur dioxide. After standing, a further portion of silver was precipitated, filtered off and washed. It was then dissolved in a little nitric acid, precipitated again as chloride and added to the main portion of the silver chloride, which was dried at 130° and weighed. The solution still containing the copper was diluted to about 8 liters, and from it all but a small portion of the copper was precipitated by a standard solution of potassium thiocya- nate, 1° of which was equivalent to about 50™€ of copper. The thiocyanate was proved to be free from heavy metals by a test with hydrogen sulphide. The small amount of iron which it contained was separated before the solution was stand- ardized, by the addition of a little ammonium alum followed 458 A. L. Day and J. Kh. Clement—Gas Thermometer. by ammonia. The solution was allowed to stand and then filtered from iron and alumina. The precipitation of the cop- per was done very gradually with constant shaking to avoid carrying down the impurities, and after long standing was fil- tered. The filtrate was concentrated to asmall volume in porce- lain. A small additional precipitate which came down in this process was worked over with care to avoid any possible loss of impurities, especially lead, though no metal but copper was found in it. The filtrate was then examined as usual. seen aes Nine ik os oie "0020 °0083 The Zine.—This metal was obtained in the form of sticks from the firm of Eimer and Amend. The method of Mylius and Fromm was followed for the principal impurities.+ 100 grams were dissolved in nitric acid. The solution was then diluted and ammonia was added until the zine at first precipi- tated was entirely redissolved. Then enough hydrogen sul- phide was added to throw down all the impurities of the hydrogen sulphide and ammonium sulphide groups together with considerable zinc. The precipitate was filtered off and further separations were made as usual. The platinum metals and gold were not looked for as it was thought quite improbable they would be present, but arsenic and antimony were sought for by Ginther’s method.t This consists in the volatilization of the hydrides of these metals which are separated from the hydrogen which forms at the same time by passing the gas through silver nitrate solution. A special form of apparatus was used which consists of a 1 liter round-bottom flask with long neck 35"™ wide at the top. *Two separate determinations. + Zeitschr. anorg. Chem., ix, 149, 1895. t Lunge, Chem.-tech. Methoden, ii, 322. Zeistchr. analyst. Chem., xx, 503. 460 A. L. Day and J. K. Clement—Gas Thermometer. This is closed by a glass stopper in which are sealed a small glass tube passing to the bottom of the flask and serving to fill the flask with hydrogen and to replace the gases formed in the experiment; a dropping funnel through which the acid used to dissolve the zine is mtroduced, and lastly, an upright outlet tube surrounded by a small condenser. The outlet was connected with a wash bottle containing a solution of silver nitrate. As pure zine dissolves with difficulty in dilute hydrochloric acid, the metal was reduced to the form of shavings by the aid of a lathe. Fifty grams of these shavings were introduced into the flask, the air in which was at once replaced by hydrogen. Dilute hydrochloric acid was then let Analysis of Zine. AG ee fe 89 oe ae eee ea, NOME BhOUNIG No), te eee ey 002 Sh ss See ee Not looked for Au BEI NT Re) ITN pO ae ae (<4 BPG es, ae ae el Teemccn el (<3 AGC ae ca na Shere ek eee NOME |) Gilg ae A gas es aks ed ac IPD: 2 Cer ie ee eee 051 CU ORES Oe eae "004 ING ph pee hae ae he None CO Ae ee ee OG Aah ie ae a ea "006 Shih eae ee eee None S a nh al 0d 0 Ee igre on lle oe 6¢ °063 in through the dropping funnel. The solution was facilitated by warming. At the end of the operation, the gas in the flask was driven out by pure hydrogen. The silver nitrate solution which contained a black precipitate was then filtered. The antimony in the precipitate was determined by dissolving it in nitric acid with the addition of a little tartaric acid, precipitating the silver with hydrochloric acid, evaporating the filtrate to dryness on the steam bath and precipitating by hydrogen sulphide. The precipitate was dissolved in a few drops of ammonium sulphide, the solution filtered into a small tared porcelain capsule, evaporated, decomposed by nitric acid and weighed as Sb,O,. After separating the silver from the first filtrate which contained the arsenic, 1t was evaporated to dryness, reduced with sulphurous acid and precipitated by hydrogen sulphide. None was detected with certainty. If this solution had been tested by Marsh’s method, no doubt a trace would have been found, but as its quantity was of a different order of magnitude from the other impurities it - was not thought worth while to make the test. Giinther deter- Ae: Day and J. K. Clement—Gas Thermometer. 461 mines sulphur at the same time with arsenic and antimony, by interposing between the generator and the absorption cylinder which contains the silver nitrate another cylinder containing potassium-cadmium cyanide which absorbs all the hydrogen sulphide and according to him retains no arsenic and antimony. Sinee a solution of this cadmium compound is always alkaline, it was thought safer to take a separate portion of zinc for the estimation of sulphur, silver nitrate being used as the absorp- tion reagent. The small precipitate was examined for sulphur by dissolving in nitric acid and proceeding as usual. Found 0-47 BaSO,. Blank gave 0°3™% BaSO,,. The zine was tested for silicon in the same way as the copper. : Redeterminations with other metal samples.—The zine, silver and copper melting points were redetermined in 1908 with other metal samples from the same sources as before and serve to show the accuracy which may be expected in random samples of the same (nominal) purity. The gold was not redetermined for the reason that no second charge was avail- able, nor indeed was a redetermination deemed necessary.* Complete analyses of these samples have not been made, but such tests as were undertaken serve to show that the copper was even purer than that of which the analysis is given. The results of these determinations are shown in the following table: +. Thermo-Couple 1908 Metal eRe Wee SAS Temperature March 10_._-... Zine 3404: 418°58 Miareh £2... 3S . - Silver in CO 9055° 958°78 pepe WO 2 eee Copper in CO 10476: 1081°5 Following are the most probable values of the metal melting points : LINC BN I eo we 418°5°+ 0O°1° PME ety e at pe et 958°3 + 0°5 Ol vcesps cae es sate a 1059°3 + 1:0 CGD PCiiee hae ox Serge 1081°0 + 0°5 Summary. The gas thermometer problem at the present stage -of its development has become primarily a problem for experimental study with two definite purposes, one to increase the accuracy - * Sometime after our work with gold had been completed a reéxamination of the charcoal which had been used to cover the surface of the metal during melting, very unexpectedly yielded iron. The gold was also found slightly contaminated with iron. The gold point here offered is, therefore, no longer entirely above suspicion and will now be repeated as soon as a fresh charge can be obtained. The probable error is accordingly given much larger than the original measurements indicated. + Determinations by R. B. Sosman. 462 A. L. Day and J. K., Clement—Gas Thermometer. of the measurements, the other to increase their range. The application of the gas laws is no longer subject to serious question. The progress of recent years has given us electric heating in place of gas and the consequent possibility of con- trolling the temperature with great certainty and exactness. Tt has also given us the metal bulb with a definite and measur- able expansion coefficient and capable of holding the expand- ing gas without loss. It has discovered a gas which does not diffuse through the bulb or react with it chemically, which does not dissociate within the limits of practicable measure- ment, and of which the expansion can be expressed with reasonable certainty in terms of the Kelvin thermodynamic scale. It has discovered the source of the errors in the ther- moelements and a way to avoid them. In 1904, Prof. Holborn of the Reichsanstalt increased the range of this scale as far as 1600° C., the probable error of the new portion being 10°. Simultaneously with this effort, work was begun at the Geophysical Laboratory in Washington with a view to increasing the accuracy of the scale, first over the existing range (to 1150°}, and later, as much beyond this point as it should prove possible to go. Temperature measurements between 250° and 1150° have now been made and form the sub- ject of the present paper. The particular points to which we have given the most attention are the following: (1) To provide a uniform temperature along the bulb by a suitable arrange- ment of the heating coils. (2) To enclose the furnace in a gas-tight bomb in which the pressure outside the bulb can be maintained equal to that within for all temperatures. This offers three distinct advantages: It provides against the deformation of the bulb through differences of pressure within and without in the region where the bulb material becomes softer. By using the same gas within and without, there is no tendency to diffuse through the bulb wall. It enables the initial pressure to be varied within considerable limits, thereby increasing both the scope and sensitiveness of the manometer. The sensitiveness in our instrument with this arrangement was about three times that of the Reichsanstalt. (8) The expan- sion of the bulb material was determined with great care and is probably accurate within 1/2 per cent. (4) The unheated space between the bulb and manometer has been reduced until the total correction in this hitherto uncertain region amounts to less than 5° at 1100°. An error of 5 per cent in the deter- mination of its volume or temperature distribution is, therefore, practically negligible. It is probable that these changes serve to reduce the aggregate error of the gas thermometer in the region of 1100° to about one-tenth the magnitude which existed at the time of the establishment of the present scale. Furthermore, and most important of all, these refinements are not limited to this temperature region. It is therefore A. L. Day and J. &. Clement—Gas Thermometer. 463 reasonably probable that the gas scale can be extended to 1500° or 1600° with a proportionately small error in its absolute value. The immediate future of the present investigation will be to undertake this extension. The interpretation of these measurements in terms of the melting points of readily available substances encounters certain difficulties. The melting point of pure salts is not sufficiently sharp and is somewhat difficult of interpretation. The metals which have commonly been used for the purpose are not obtain- able commercially in sufficiently uniform purity to guarantee an accuracy within 1° at the higher temperatures. This is too large an error for the interpretation of the gas thermometer scale in its present refinement. No effort has been made to prepare metals in our own laboratory of exceptional purity for the reason that such metals would not be available for ene use and would therefore be of little service. We have accordingly adopted metals which are carried per- manently in stock by dealers (whose names are given in con- nection with each) from whom the same metal in a nominal quality equal to that which we used can readily be obtained. We have analyzed these with extreme care to show the exact content of the sample supplied tous. We have duplicated the purchases ourselves, and have found no errors greater than 1° in their melting point determinations. Another difficulty arises from the fact that the melting points of the purest metals available for use as constants in reproduc- ing a high temperature scale (zinc, silver, gold and copper) are distributed in such a way that, although they may be located upon the gas thermometer scale with a probable error not greater than 0°5°, the calculation of a similar curve passing through these points does not suffice to reproduce the scale with this accuracy. In the region midway between zinc (418°9°) and silver (958°5°) the error of interpolation may amount to 2° even with metals of exceptional purity. Extrapolation is even more uncertain. This can be avoided by locating intermediate points which are equally trustworthy, if such can be found. We have not been fortunate enough to find points which fulfil these conditions satisfactorily but hope that we may yet be able to do so. _ As the matter now stands therefore we have succeeded in perfecting the constant volume gas thermometer until the aggregate error afiecting the measurements between 300° and 1150° appears not to be greater than 6°5°, but we are not yet able to offer adequate assurance that our scale can be reproduced by another with this accuracy. This matter will re- ceive further attention in a later paper. Geophysical Laboratory, Carnegie Institution of Washington, August, 1908 464 W. Duane—Range of the a-Rays. Art. XLITL—On the Range of the a-Rays; by Witi1am Duane.* Tuer researches of Madame Curie, of Bragg and Kleeman, and of Rutherford have shown that the a-rays abruptly lose their powers of ionizing gases, of affecting a photographie plate, and of producing phosphorescence after they have pene- trated several centimeters of air or an equivalent thickness of other substances. Further, Rutherford has found that near the point where it loses these powers the a-narticle still possesses sixty per cent or more of its initial velocity. Several years ago I made some experiments to determine whether the charge carried by the a-particle could be detected beyond the limit of its ionizing power, or, possibly, off to one side of its range. The results were negative. Recently l have taken up the research again at the laboratory of Madame Curie of the University of Paris, with more and purer radium, and with the additional purpose of investigating the power of the a-rays to produce secondary rays, and the transformation of the kinetic energy of the a-particles into heat. Figure 1 represents the arrangement of the apparatus. A is a cylindrical box of brass 3°8™ i long and 3:3" in diameter. A round hole (1°8°" in diameter) in the bot- tom of the box is covered with a very thin sheet of mica, B. The mica weighs only 2 milligrams per square centimeter, and is rein- forced on the inside by a grating of fine wires. It is so thin that the a-rays can pass through it eas- ily, and strike the plate C, which is connected to an electroscope or B electrometer and serves as an elec- trode. In order that a magnetic R field may be produced parallel to Cepmechinnaeu: to jpumips the plate B and to the mica window, L, to electrometer or electro. the apparatus is placed between the scope ; EH, to earth ; B, to bat- poles of an electromagnet; and in Pony pee earn order to produce an electric field between the electrode and the window, the ring D, to which the mica is fastened, is insulated from the sides of the box by wax, and connected to the pole of a battery. The method of procedure follows: A very small quantity of radium chloride was dissolved in water and recrystallized twice * Abstract of notes presented to the French Academy of Sciences. Comptes Rendus, 11th and 25th of May, 1908. Currents (arbitrary units) W. Duane—Lfrange of the a-Rays. 465 in succession at an interval of several hours, in order to free it from most of its emanation and induced activity. Finally it was dried upon a flat sheet of platinum. The platinum was then held horizontally below the mica window at different distances from it, and the ionization currents between the win- dow and the electrode were measured by a quadrant electro- meter, the window being at a potential of 88 volts above the electrode. To make the rays that entered the box parallel to each other a set of fine glass tubes (not shown in the figure) was fastened between the radium and the window with their axes vertical. Curve 1, figure 2, represents the ionization current as a function of the distance from the radium to the bottom of the box. It is evident that most of the ionization in the interior of the box disappears if the radium is removed to a distance From the window greater than about 2™. Wie. 2. 332 SSR Eee aes me PN Eee ae bah - SARASSLBAS A 2a as ea® ele teed Zz Ws aa Ge ge mail | tty ae ae8 a Ea a Uo at ta | ? Volts per second (proportional to current) Distance from radium to window. ~ Distance from radium to window. ~ The currents for distances greater than two centimeters are due to a small amount of emanation and induced activity remaining after or having accumulated since the final crystalli- zation. ‘The a-rays from these, as is well known, have greater ranges than have the a-rays from radium itself. The ioniza- tion due to these a-rays of longer range is well shown by curve 2, which represents the currents due to radium (a smaller amount than before) that had been left two days in a dry state, and which therefore contained considerable amounts of emana- tion and induced activity. In order to measure the positive charge of electricity car- ried by the a-rays, I exhausted the air from the box by means of a mercury pump, producing a high vacuum of less than 0001" 466 W. Duane—Lange of the a-Rays. of mereury as measured by a McCleod gauge. This was to prevent the charge being neutralized by the ionization of the air in the box. I then measured the current flowing toward the electrode by means of a Wilson gold leaf electroscope, using much more radium than before (about 2 milligrams of pure radinm chloride). For these measurements the set of fine glass tubes between the radium and the window was not used, and the window was kept at zero potential. When the a-rays pass through the mica and when they strike the metal electrode they produce slow-moving secondary rays. In order to suppress these a magnetic field was produced parallel to the surfaces of the electrode and window. That the magnetic field stopped all the secondary rays was proved by the fact that increasing its strength from 2400 gauss to 3600 gauss did not alter the current flowing to the electrode. That there was no appreciable ionization current in the interior of the box is shown by the fact that with the mag- netic field a difference of potential of several volts between the window and the electrode did not alter the current from one to the other. Curve 3, figure 2, represents the currents of electricity car- ried to the electrode by the a-rays. It is evident that the greater part of the charge carried by the rays does not pass through the mica and reach the electrode, if the radium ts more than 2°" from the window. This is the same critical dis- tance asfound before for the ionization. By reason of the form of the curves near the limit it is difficult to estimate the exact length of the range, but we*can say that approximately the charge of the a-particles and the ionization produced by them stop at the same pornt. | Curve 4 represents the currents due to radium in which the emanation and induced activity had been allowed to accumu- late for over two days. The currents for distances greater than 2°" are due to the charges of electricity carried by the rays from the emanation, and radium A, and C. In order to determine whether or not the a-rays lose their power of producing secondary rays at the point where the charge and ionization stop, the currents without the magnetic field were measured with the radium at different distances from the window. Under these conditions the current of electri- city carried to the electrode is to some extent masked by the secondary rays produced at the lower surface of the electrode and the upper surface of the window. If a is the current car- ried to the electrode by the a-rays, s, the negative charge car- ried per second by the secondary rays away from the electrode, and s, the negative charge carried per second to the electrode W. Duane—Range of the a-Rays. 467 by the secondary rays from the window, the total current toward the electrode is t=O){S ——S, The secondary rays s,and s, are stopped by the magnetic fields, leaving only the current a. If the window is charged positively the electric field from it toward the electrode stops part (if strong enough all) of the eurrent s,, and if the window is charged negatively some or all _of the current s, is stopped. To determine how strong, the field must be in order to stop all of the secondary rays, the radium was placed 1°5°" below the window, and the currents toward the electrode measured by the Wilson electroscope when the window was charged to different positive potentials. The following values were obtained : Potential + 0 2. 4 9°5 14 D5 84a le 46S Sat 170r -volitss Current toe LO ee A235 a fe Oro 0n 92. 9°85 1070 «100 It is evident from these figures that after a potential of some 70 volts is reached the current is not increased even if the potential is doubled: i. e., 70. volts stops all of the second- ary rays coming from the window. If the field is reversed about the same potential, 70 volts stops all of the rays coming from the electrode. The above readings were taken without a magnetic field. If a magnetic field of 2800 gauss is produced the current is 1-6 on the same scale, and is increased only a few per cent even by 170 volts. We now have a means of studying the secondary rays com- ing from the surface of the electrode alone, s,, for by charging the window to a potential of + 70 volts or more the current s, is suppressed. Then the difference between the currents with and without the magnetic field is s,, the secondary rays from the electrode. To determine whether or not the power of the a-rays to produce the secondary rays ceases at the point where the charge and the ionization stop, the 2™* of radium chloride were freed from emanation and induced activity as before, and placed at different distances below the window. The window was charged to a potential of +85 volts, and the currents for each distance were measured, first without and then with, the mag- netic field. The difference between the two currents, repre- senting the secondary rays from the electrode cut off by the Volts per second (proportional to current) Distance from radium to window in cm. 468 W. Duane—Range of the a-Rays. magnetic field, is represented in fig. 3, curve 5, as a function of the distance of the radium from the window. It is evident that the greater portion of the secondary rays stop when the radium is removed further than about 2" from the window ; and this is about the same distance as found before for the charge of a-rays and the ionization. For better comparison the currents measured in the same series with the magnetic field, representing the charge carried by the a-rays, are plotted on curve 6, fig. &. Owing to the form of the curves it is difficult to determine the exact point where the effect of the a-rays from the radium itself ceases. The'effects due to the small amount of emanation Volts per second (proportional to current) Charge of secondary rays. Charge of a-rays. and induced activity are always appreciable. If there is any difference, the curves seem to indicate that the power of pro- ducing secondary rays is appreciable a littie further away from the radium than is the charge. I do not think, however, that we can tell with certainty. This point is of great importance in connection with the idea advanced by J. J. Thomson, that at a certain velocity the a-particle attaches to itself an electron, which neutralizes its charge and therefore changes its properties. I have tried the above experiments with polonium, but the currents were too small to be satisfactory. I hope to be able to repeat them with a more active preparation of polonium. Distance from radium to window in cm. C. H. Warren—Alteration of Augite-limenite Groups. 469 Arr. XLIV.—Wote on the Alteration of Augite-Ilmenite Groups in the Cumberland, R. 1., Gabbro (Hessose) ; by C. H. WARREN. [Contributions to the Geology of Rhode Island.—No. IIT | In a recent paper* descriptive of the geology and petrog- raphy of Iron Mine Hill, Cumberland, R. L., it was shown that the basic, titaniferous rock rhodose (cumberlandite) was closely associated with a strongly metamorphosed and altered gabbro, whose areal extent and general characteristics were there described. The rocks taken together formed a roughly circular area of basic igneous rock entirely surrounded by highly metamorphosed, ancient sedimentaries and granite intrusives. The close association of the two rocks, and the fact that both are characterized by a large content of ilmenite, naturally led to a detailed study of the gabbro in connection with that of the rhodose. As a result, it was found that, while the metamorphism and alteration of the gabbro pre- sented many of the common characteristics of such rocks, cer- tain mineralogical changes had taken place, which if previously observed have not been adequately described, and it is the object of the present article to call attention to them. The unaltered gabbro.—Although the gabbro is nowhere exposed in an unaltered condition, a study of its less highly altered forms shows clearly that it was originally a rather coarse (millimeter) grained gabbro containing, beside abun- dant augite and accessory apatite, an unusual amount of ilmenite in the form of large irregular grains occurring in close associ- ation with augite. The grain of the rock seems to have been generally uniform throughout, although an occasional coarse- grained, almost pegmatitic, development has been noted, as well as a fine (aplitic?) phase. The optical properties of the feldspar indicate a plagioclase of about the composition Ab, An, or a little more acid. The textural relations of the feld- spar to the augite and ilmenite is diabasic. In habit the feld- spar is strongly tabular on 010, the crystals averaging perhaps almost 1™ square and from 2 to 3™™ in thickness. The ore grains, although now more or less reduced in size by alteration, are still large and abundant, many of them averaging as much as 4 or 5™™ in diameter. Their distribution seems to have been fairly uniform. Locally, in several places, the grains have been noted larger in size and much more abundant. On treating a polished surface of the rock with hot hydro- _ chloric acid the ore grains are but slightly attacked, and no such interesting structure (intergrowth of magnetite and ilmenite), * This Journal, vol. xxv, Jan., 1908. Am. Joug. Sci.—FourtH SrErRiges, Vout. XXVI, No. 155.—Novemeer, 1908. 33 470 OC. H. Warren—Alteration of Augite-Llmenite Groups. as was noted in the case of the ore in the rhodose, was observed. The grains react strongly for titanium and possess a magnetic susceptibility like that of ilmenite. Alteration, as will be pointed out beyond, often develops the characteristic reticulate structure of ilmenite. There can be no doubt, there- fore, that the ore in this rock is ilmenite and if any magnetite existed as an original constituent 1t was very inconsiderable in amount. Occasional flakes of reddish brown secondary mica are present. No indications of any pyroxene, other than augite, or of other original ferromagnesian mineral have been found. Chemical composition and classification.—For the purpose of showing the chemical character and of classifying the rock quantitatively a single chemical analysis has been made on material taken from a specimen showing relatively the least amount of alteration. The results are as follows — Analysis of Gabbro (Hessose). Si0, 45°27 TiO, 207 Al,O, 18°30 Orthoclase ‘O11 sal 65°93 Fe LO: 3°30 Albite — "055 fem rs 31°53 a 2°0, Dosalane (2) FeO 10°13 Anorthite 104 MgO 4:08 Corundum ‘006 Q or I, 0 1 Gee ; CaO 7°32 Hypersthene -068 : ic Gee ee a Na,O 3°64 Olivine 064 : (5) K,O 1:07 Ilmenite "034 , MnO 86 Magnetite. -02) Ss N20 8) Co,Ni tr. Apatite 008 CaO 130 (4) PO: 1:27 KO “A S 08 Ce why: H,O 2-08 NO 380 The rock is therefore perfelic, docalcic and presodic, its co- ordinates in the quantitative system being 2,5,4,3 and may be called a Hessose. It may be noted that the ferrous iron and titanium are both high, and that although the potash amounts to over one per cent no orthoclase has been detected with the microscope. Much of the potash is probably now present in the biotite and sericite. General megascopic character of the rock.—The entire rock mass has been subjected to more or less severe dynamo- and hydro-thermal metamorphism. The changes thus effected are CO. H. Warren—Alteration of Augite-Llmenite Groups. 471 now somewhat obscured by superficial alteration, but it seems clear that their intensity varied quite irregularly, being stronger toward the northern and northwestern contacts, and in inde- finite zones through the body of the mass. Specimens showing the least metamorphism and but little superficial alteration may be found in the north-central part of the area along an old car track, where the surface of some of the ledges has been blasted away. The rock here is of a dark, greenish brown color, and breaks with a more or less distinct cleavage, owing to a rudely parallel orientation of the tabular plagioclase crystals. The plagioclase is dark brown in color and is beautifully _striated. Between the crystals are irregular patches of dark green or brownish green, finely crystalline, secondary silicates in which are very generally embedded lustrous grains of ilmen- ite. The green material can be seen to penetrate to some extent the feldspar substance. Polished surfaces, looked at with a good hand lens, serve to show the texture admirably. Outcrops in the southeastern part of the area show about the same degree of metamorphism, although superficial alteration has gone further. On exposed surfaces the feldspar becomes chalky and retreats, leaving the black ore grains and their matrix of secondary silicates, now of a dull, pale green color, standing out in relief. ree, More severely metamorphosed phases may be recognized by the fact that a portion of the feldspar has changed to a dull white saussurite, a change that becomes complete in the more extreme types. The latter are also characterized by the diminished nmnber and size of the ore grains and by the general loss of the original texture. Types representing these stages may be collected a little east of the exposures of the least altered type above alluded to, and from the ledges on the north and south of the railroad track. ‘Toward the northern border of the area and along portions of the high ridge that forms the eastern outcrop of the gabbro toward the Iron Mine Hill, the rock has lost, so far as its macroscopic appearance is concerned, almost every vestige of its original texture. It shows an indistinct schistose structure, and has a mottled, greenish white appearance. Chlorite, a little sericite and an occasional remnant of ilmenite and feldspar are the only minerals that can be distinctly identified, although the dull white dense groundmass, especially in its weathering, is sug- gestive of a feldspathic composition. At the contact with the granite and schists on the north, the gabbro has been sheared into a fissile green schist. At the extreme northeast extension of the outcrops the rock has become a greenish white, fissile schist. 472 0. H. Warren—Alteration of Augite-Limenite Groups. Microscopic characteristics of the gabbro—least metamor- phosed types.—In addition to this being the least altered of any of the types examined, it may be designated as one whose chief distinguishing features are a large development of secondary biotite and the absence of leucoxenic alteration of the ilmenite. In thin section the feldspar substance is com- paratively fresh although occasional patches of sericitic and saussuritic material may be noted, and the crystals are generally characterized by the presence of a brown pigment. The ecrys- tals show abundant evidence of mechanical strains. The augite appears to have been the first constituent. that vielded to alteration, and although by far the greater portion of it has gone over to secondary minerals, occasional crystals, may still be seen in intermediate stages of alteration. In the thin sections studied, the dominant mode of alteration (if not indeed the only one in this type) is to a more or less confused fibrous aggregate of hornblende. Under low powers this has a cloudy appearance and is of a light yellowish brown color. With higher powers the fibers have a sub-parallel, also some- what divergent arrangement, and possess in general a pale green or yellowish green pleochroism. The change to this sub- stance begins about the edges, along cleavages or fractures, and encroaches in. a quite irregular manner on the augite sub- stance. The’amphibole often assumes a distinetly greenish or bluish green color next the plagioclase. Instances may also be noted where the whole aggregate has a more compact tex- ture while at the same time the color is a deeper brown. Small crystals of a yellowish to reddish brown, massive horn- blende may also be seen, which are undoubtedly secondary and strongly suggestive of a direct formation from the augite. The change of the augite to the fibrous form is followed by a further and more important change, which often begins before the original augite has entirely disappeared and in which are involved reactions with constituents from the ilmen- ite and plagicclase. The space originally oceupied by the augite as well as a part of that occupied by the ilmenite and plagioclase becomes filled with an aggregate composed of hornblende, biotite and particles of ore, the latter largely residual although possibly to some extent secondary. Where augite grains unaccompanied by ilmenite have suffered altera- tion, biotite is characteristically lacking. This hornblende is strikingly different in appearance from the fibrous form previously described, and consists essentially of an aggregate of small prismatic crystals and fibers exhibit- ing a distinct, though not very strong, bluish green pleochroism. Associated with this is a variable amount, often considerable, of hornblende in the form of prismatic crystals of relatively C. H. Warren—Alteration of Augite-Llmenite Groups. 478 large size and possessed of a strong, bright blue pleochroism, parallel to C. Near the borders of the aggregates they are more abundantly developed, in fact they not infrequently form a distinct border of blue prisms jutting out into the plagioclase. Crystals of hornblende having a yellowish to reddish brown pleochroism parallel to C occur with the blue and appear to have been formed in the same way, indeed the same crystal may show a blue pleochroism in one portion and a brown in another. Both the blue and the brown variety may also be seen, changed, probably by a bleaching process, to a nearly colorless variety. Within the aggregates there appears to be some actinolite. The biotite is of a light brown color and its formation is clearly connected with the presence of ilmenite. It occurs in part in the form of flakes or shreds mingled with the horn- blende. Its amount varies, being most abundant in the neighborhood of the ilmenite, where it not infrequently prac- tically displaces the hornblende. Its most striking mode of occurrence is, however, in the form of a clearly marked border lying between the hornblende and the plagioclase. These borders as a rule practically surround the entire altered area. Their width, relative to that of the area enclosed, varies considerably in different cases but is always large (perhaps from 1/5 to 1/12). The border consists essentially of narrow shreds or flakes, orientated perpendicularly to the contact with the feldspar, into which they penetrate quite irregularly. Blue hornblende (rarely brown) prisms frequently make their appear- ance in these biotite rims, and, in places, as noted above, may actually constitute a rim themselves. Needles and stouter prisms of hornblende often accom- panied by biotite are abundantly developed along fractures and cleavages in the feldspars, while minute isolated horn- blende crystals are to some extent disseminated in many of the plagioclase crystals (gewanderte hornblende). A small amount of calcite and limonite is also present. Both biotite and hornblende have been noted lying directly in contact with the ilmenite; of the two, biotite seems, how- ever, to be much the most intimately connected as regards formation with the ilmenite. It sometimes forms a continuous . erystal about a portion of or even an entire ore grain. In such cases the biotite often contains a considerable amount of included material in the form of minute grains, for the most part opaque but sometimes feebly translucent. Some which show a reddish color are probably rutile. These particles are undoubtedly residual since every gradation between biotite with only a small amount of included matter and that contain- ing a core of massive, unchanged ilmenite, may be seen. It 474 OC. H. Warren— Alteration of Augite-Llmenite Groups. is in fact generally evident from the relations of ilmenite to the secondary silicates, that a very considerable portion of the former has disappeared through reaction. It is to be par- ticularly noted that leucoxenic alteration of the ilmenite is absent in this type. Neither have the oxides of titanium been observed to more than a trifling extent. The appearance of characteristic aggregates of secondary hornblende and biotite with residual ilmenite is shown in the accompanying photograph. | Summary of the mineralogical changes.—After the first change of the augite to the fibrous (or compact) hornblende there appears to have been a profound readjustment of the various constituents of the hornblende, ilmenite and surround- ing plagioclase, accompanied by reerystallization and more or less transportation, possibly removal, of material. The par- ticular chemico-mineralogic changes which may be noted here are: 1. The formation of a blue, soda-iron or soda-aluminum amphibole molecule from the soda (and alumina?) of the plagio- clase and from the ferric iron of the ilmenite. The brown C. H. Warren—Alteration of Augite-llmenite Groups. 475 variety may well be due to the entrance into the molecule of a different proportion of ferric iron and titanium. 2. The for- mation of biotite, for which both ferrous and ferric iron were derived from the ilmenite, alumina and alkalies from the feldspar, and perhaps magnesium from the augite or fibrous hornblende. The titanium, originally combined with the iron used up in the formation of biotite, probably also enters into this latter mineral; at all events, leucoxene or other titanium. bearing minerals are absent. The water given off by the rock on intense ignition reacts acid, indicating the presence of fluorine. This is probably present in the hornblende and biotite and it is possible that it may have originally been derived from emanations from the adjoining granite intrusives, which are believed to be later than the gabbro and are known to contain considerable fluorite in places. The effect of even an exceedingly small amount of fluorine in promoting changes like those recorded here would be unquestionably great and is worthy of consideration. The occurrence of secondary biotite and hornblende about magnetite or ilmenite associated with augite in a manner which appears somewhat similar to the one described here, has been noted very briefly by Mr. 8. Allport in an article “On the Metamorphic Rock Surrounding Land’s End Mass of Granite, Tolearn.”* Mr. Allport describes brownish spots having the mode of occurrence of magnetite and containing centers of magnetite ‘surrounded by innumerable minute flakes of brown hornblende or mica, while a short distance is usually green—a fact clearly indicating the diffusion of ferric oxide.” Again, “ Augite was abundant and has been converted into a brown granular substance.” Dr. G. H. Hawes, in “The Mineralogy and Lithology of New Hampshire,t descr: and figures mag- netite or ilmenite grains surrounded by folie of biotite radially arranged in fan-shaped aggregates with the iron oxide as a nucleus, as of very common occurrence in the “syenite”’ near Jackson, N. H. He expresses the opinion that biotite may be very commonly a secondary product of this kind. Again, Wadsworth recognized it as a secondary mineral developed irom magnetite or ilmenite and the surrounding feldspar and more rarely from pyroxene, in some of the gabbros described by him from Minnesota. Dr. Wadsworth also notes its occurrence as a secondary mineral from titanic iron ore and plagioclase in the basic titaniferous rock from Taberg, Sweden, and in the * Quart. Jour. Geol. Soc., xxxii, p. 420, 1876. Also referred to and figured by Teall (British Petrography, plate xvii). + Geology of New Hampshire, C. H. Hitchcock, 1878, Part IV, p. 205, fig. 6, plate xi. ¢ Gtol. and Nat. Hist. of Minn., Bull. No. 2, St. Paul, pp. 65 and 90, 1878. 476 OC. H. Warren—Alteration of Augite-Ilmenite Groups. closely similar rock, cumberlandite (rhodose), from the present locality, a fact to which the present writer has more recently again called attention. Its occurrence has also been noted else- where, particularly in certain European “ Flascher gabbros.” Its development as a secondary mineral jointly from the con- stituents of ilmenite and plagioclase does not, however, appear to have received much attention in standard works on Petrography, although its importance as bearing on the presence of biotite in metamorphic rocks is obvious. The effect of superficial alteration of this type is a gradual change of the biotite (to some extent.the hornblende) to chlorite accompanied by the formation of epidote. Kaolin, calcite, and limonite also develop. Second type of altered gabbro.—This type may be character- ized, in distinction to the above, as one in which the ilmenite has suffered a leucoxenicé alteration as well as a biotitic and hornblendic one, and in which the augite, besides a passage to the fibrous aggregate, shows a direct change to a more or less compact green, or brown hornblende. Specimens of this type have evidently suffered more severely from shearing. The feldspar crystals are frequently crushed and are very generally filled with saussuritic material, in addition to secondary silicates more directly derived from the alteration of the augite and ore. The saussuritic material consists essentially of epidote and zoisite with some muscovite and hornblende. The change of the augite to the same fibrous aggregate as previously described may be seen clearly and is unquestionably a common one inthis type. Another change, sometimes in the same crystal, to a semi-compact pale, bluish green hornblende, which assumes a deeper blue color where it is in contact with the plagioclase, is of frequent occurrence. This hornblende examined with high powers shows an indistinct reedy structure and is filled with minute crystallites of other minerals, iron oxides and epidote chiefly. Irregularly throughout its mass, somewhat divergent fibrous patches may be noted which in color and appearance seem to be identical with fibrous amphi- bole formed directly from the augite. These may result directly from the alteration of the green hornblende or perhaps simultaneously with it from the augite. A massive reddish to yellowish brown hornblende may also be seen forming directly from the augite. The brown hornblende very often passes sharply in a bright blue variety and also intoa practically colorless mineral which seems also to be an amphibole in its character although its double refraction is abnormally low. The fibrous secondary hornblende formed directly from the augite, the green and the brown hornblende, as well as the colorless amphibole, all suffer a further change into a confused C. H. Warren—Alteration of Augite-Llmenite Groups. ATT ageregate of hornblende prisms exactly lke that noted in the previous type. Biotite has also formed in the same manner but both the hornblende and the biotite, particularly the latter, have suffered a more general distribution through the rock and are less closely confined to the place of original formation. Much of the biotite and hornblende has now ‘suffered further alteration to chlorite. Leucoxene is abundantly developed about the ilmenite and is plainly of later origin than the biotite and hornblende. It is distinctly crystalline and has the characteristics of titanite. This leucoxenic alteration has brought out with great clearness the reticulate structure of the ilmenite. In comparing what have here been called the two types of alteration, the writer is led to conclude that the former is one brought about under conditions of deep-seated, hydrothermal action but unaccompanied by extreme shearing and crushing. In the second type generally similar changes obtained for a time but were succeeded by others induced by more severe local dynamic action. The latter changes are: The forma- tion of leucoxene; a greater tendency for the augite to pass directly into a compact or semi-compact hornblende ; increased saussuritization of the feldspar ; and a more general distribution of the secondary hornblende and mica throughout the rock. More highly metamorphosed types.—F urther metamorphism and alteration of the gabbro presents little that is novel, and a very general statement will suffice. The microscope reveals increased crushing and saussuritization of the feldspar, a com- plete change of the hornblende and biotite to chlorite and epidote, an almost complete alteration of the remaining ilmenite to leucoxene, and the development of some sericite. Some- secondary quartz and recrystallized feldspar may also be noted. A more or less schistose arrangement of the constituents becomes evident and the outlines of the original structures are less and less distinct, until in extreme phases, from near+the northern contacts, the slides show little more than a schistose mass of finely crushed feldspar mingled with secondary products. Laboratory of Mineralogy and Petrology, Massachusetts Institution of Technology, Boston, Mass. 478 T.. Holm—Studies in the Cyperacee. Art. XLV.—Studies in the Cyperacee; by Turo. Horm.— XXVI. Remarks on the structure and affinities of some of Dewey’s Carices. (With 24 figures drawn from nature by the author.) Amone the circa eighty Carzces, which Dewey described, there are some which have proved very troublesome to caricolo- gists. The diagnoses are not always so complete or exact as they might have been written ; or the material on which certain species were founded was not quite mature, thus the reader does not always receive a very clear impression of tle most essen- tial characteristics of some of these species, even if they may be perfectly distinct and valid. In such cases, where we cannot depend entirely upon the diagnosis, the examination of Dewey’s own specimens may, sometimes, be helpful. But unfortunately the material left by Dewey is not only small, but it contains, moreover, specimens which are not all in conformity with his diagnosis, and such specimens must consequently not be looked upon as his original, those on which he founded his new species. The best set of Dewey’s species is in the herbarium of Kew; these specimens were named by Dewey himself and presented to Boott. There is, furthermore, some material at present incorporated in the Gray herbarium at Cambridge, which is very valuable so far as we are able to distinguish between those plants that were parts of his original specimens and others, which he simply identified as beg identical, but which, sometimes, are very different species. Dewey did not work with types, he worked with species, and naturally expected that his species were to be identified by means of the diagnoses. It would be very unsafe and unjust to give prefer- ence to the specimens in case of determination, instead of to the diagnosis. When Dewey’s specimens do not agree with the ‘diagnosis, we may feel sure that they were not correctly named. Much confusion has arisen from the attempt of certain authors to identify species by means of “ supposed types,’ especially when a critical examination of the diagnosis necessarily must convince us that said specimens were not the original, not the one on which the species was established. The interpretation of Allion’s Carex fusca and bipartita is a good example of the result of this kind of verifying old speci- mens,* but several other cases might easily be recorded.t * This Journal (4), vol. xvi, p. 145, Feb., 1908. + The following note, copied from a letter received from Mr. Clarke, may be of interest to the reader: ‘‘ Willdenow did not work from types but from small packets (now largely sorted into different species in the Berlin Herb.). You can see this, because Kunth repeatedly cites Willd., folio 2 or folio 8, for species which he sets up as new (or separate). Feb. 24, 1902.” T. Holm—Studies in the Cyperacee. 479 Now in regard to Dewey’s species, it is our intention to demonstrate that some of these have been misunderstood ; that some of these are not so difficult to identify, if we give prefer- ence to the diagnoses, and not to the specimens extant. It appears to the writer that the reprinting of the original diag- noses may be necessary, inasmuch as some of these are not known to several caricologists, who have no access to the earlier vol- umes of this Journal, in which they were published. But what- ever importance may be attached to the present supplemental notes on these species, we must not forget to mention that we owe much information to a prolonged correspondence with the late cyperographer C. B. Clarke of Kew, who was so very familiar with the large herbarium of Boott and many others. The modern method of identifying species by means of “ sup- posed types’ was, according to Clarke, a most dangerous experiment, since in particular respect to Carices neither Dewey nor Boott worked with types. In a note on Carex Tolmiei Boott,* Clarke has shown how very difficult it may be, sometimes, to reach an exact conclusion even from a work so excellently written and illustrated as that of Boott: Lllustra- tions of the genus Carex.+ The species of Dewey which we intend to discuss are: C. petasata, C. Barbare, C. magnifica, C. Schottii, C. petricosa and C. mirata. Of these C. petasata has been suppressed entirely, as will be shown in the subse- quent pages; C. Barbare, C. Schottii and C. mirata have either been referred to other plants, or have been merged into each other as synonyms; C. magnifica was never described, but mistaken for C. Sitchensis Prescott, while C. petricosa was known only from very immature specimens. With the exception of (. Barbare of which there is no material in Kew, but in Cambridge, we have had the opportunity to compare the others with authentic specimens, authentic to the full extent of the word, since they were in accordance with the original diagnoses; this material was made accessible to the writer through the kindness of Mr. Clarke. In regard to @. Barbare a young, but nevertheless quite complete, specimen named by Dewey in the Gray herbarium was loaned to the writer; beside that we succeeded in finding some mature specimens of this very rare species in the collections of Mr. Parish, now deposited in the U. S. National Museum. * Journ. Linn. Society, vol. xxxv, p. 403. + ‘‘ Boott has named in his own hand many Carices in the herbarium and X takes these as ‘authentic’; but they are, in very many cases, authen- tically wrong. In his herb. propr., Boott usually pasted down, all mixed together, 3, 4 or even 7 collections on one sheet. His figures often include utricles from several ecltsekions: to show his idea of the. range of variation ‘in each species.’ If, however, the utricle varied really to the degree he depicts it, it would be of very little use in diagnoses of species” (C. B. Clarke in litteris, Oct. 11, 1899). 480 7. Holm—Studies in the Cyperacee. Carex petasata Dew. According to Professor Bailey deve. p. 52) “the original sheet of this species is in Herb. Torr. It contains three plants: C. lagogina Wahl, C. festiva Dew., and C. Liddoni Boott, to all of which Dewey’s description will equally apply.” For this reason Professor Bailey does not think that C. petasata ‘‘ean be pressed into service,” and although as he states him- self, ““C. Presliz is not clearly accounted for,” and “ the original does not exist, either in the coilection of Presl or Steudel,” he nevertheless adopts Steudel’s name “in lieu of any other.” It may be that the specimens in Torrey’s herbarium were mixed, but there are several good examples ot C. petasata in Boott’s herbarium, received from Dewey and authenticated by his hand, and Mr. Olarke has informed us that these specimens are not mixed. Consequently there is. no reason why the name petasata should not be retained for this species, and the diagnosis written by Dewey* reads as follows: “Spicis distigmaticis androgynis, inferne staminiferis subqua- ternis ovato-oblongis cylindraceis subsessilibus approximatis ; fructibus lato-lanceolatis utrinque acutis rostratis vel acuminatis ore bifidis subalatis, squama lato-ovata subobtusa longioribus. Culm 4-8 inches high, erect, slightly scabrous, triquetrous, striate; leaves shorter below, upper one about as long as the culm ; spikes androgynous staminate below, oblong cylindric, about 4, short pedunculate, approximate, brownish; fruit broad lanceolate, acute at each end, acuminate or rostrate, compressed, bifid and slightly winged, convex above; scale ovate, obtusish, tawny, broad, shorter than the fruit. Found on the Rocky Mountains.” Characteristic of the species is, thus, the ovate-oblong spikes, which are short peduncled and approximate, but not sessile, forming a head; moreover the broadly lanceolate perigynia tapering at both ends, and narrowly winged. Frequently the - spikes, especially the lower ones, are somewhat remote, very distinctly peduncled and subtended by setiform bracts, thus resembling C. pratensis Drej..—The perigynia are light brown and faintly, though very distinctly, veined (about six veins on the outer face). It would consequently be very unjust to con- sider the specimens in Torrey’s herbarium as being Dewey’s type. In the first place because Dewey did not work on types, and secondly because his diagnosis by no means applies to the three species which Professor Bailey found in said herbarium. This may be readily seen from the fact that so far as.concerns C. lagopina, the perigynia of this species are ovate to almost round, and never winged; in C. festeva the spikes form a * This Journal (1), vol. xxix, p. 246, 1836. T. Holn—Studies in the Cyperacee. 481 dense head, and the perigynia vary from ovate to suborbicular, broadly winged and prominently veined; in C. Lzddonz the very heavy spikes and large perigynia, which are many-nerved and with prominently serrulate- -winged margins, make this species very distinct from C. petasata. In regard to the geographical distribution C. petasata has been found in Alaska, but only in a few localities; furthermore and apparently fre- quent in the mountains of Washington, Oregon and Idaho, in the Alpine regions; it occurs, also, in British Columbia, Alberta, Assiniboia, Vancouver Island, Montana, Wyoming, Utah and Colorado; it seems to be rare in Colorado, and is there confined to the highest peaks above timber-line. Carex Barbare Dew. Through the kindness of Mr. M. L. Fernald the writer has had the opportunity to examine probably the only specimen extant from Dewey’s own material of this very rare species, deposited in Win. Boott’s herbarium, now incorporated in the Gray herbarium at Cambridge. It is an immature specimen, but labeled by Dewey himself: “ C. Barbare, Santa Barbara, New Mexico,” and the principal characteristics are easily rec- ognized to be in conformity with the diagnosis, as this was written by Dewey.* Subsequent authors have not, however, paid due attention to the original diagnosis, for instance, the peculiar structure of the squamee, and the result has been that Carex Barbure of to-day comprises several distinct species, among which Dewey’s C. Schott and C. dives nob. While thus the specimen, which we have examined, is immature, the diagnosis plainly shows that Dewey based his description on more perfect and mature specimens; if not he would have said so, for he was careful enough to state in his diagnosis of C. Schottéi that the perigynia were either wanting or We ge: To deal with immature specimens of Carices especially is most difficult task, but, in the present case, we have a well written diagnosis beside a specimen labelled by the author him- self. It seems very strange that so much confusion should arise in regard to the identity of this species, since the group of OCarices to which it belongs is rather poor in representatives, and, as stated above, so very distinct from that to which the other plants belong, which erroneously have been referred to C. Barbare. From reading Dewey’s diagnosis there is abso- lutely no doubt that his Curex is a member of the orasta- chye, and related somewhat to C. Schotti and C. magnifica, which he distributed under this name though without append- ing a diagnosis. Most of the species which subsequent authors * Emory’s Report U. S. and Mex. Bound. Survey, p. 281, 1858. 482 LT. Holm—Studies in the Cyperacee. have named C. Larbare are members of the grex Microrhyn- che, and more or less closely related to Prescott’s C. Sttchen- sis. It seems a strange coincidence that this species of Prescott should suffer the same fate as C. Barbarw, to become so entirely misunderstood for many years, although the diagnosis plainly shows us that it was not intended for the very char- acteristic C. magnifica of Dewey. When thus modern eari- cographers consider C. LBarbare to be a close ally of C. aquatilis Wahlenbg., we are now in the positicn to state that it is not by any means related to this, neither to this particular species, nor to any of the other members of the ALicrorhynche. Inasmuch as the real C. Barbare seems to be a very rare plant (not represented in any of the large herbaria at Kew) it might be appropriate to reprint the original diagnosis, and to give an account of the confusion into which the species has fallen. Dewey’s description reads as follows: “ Carex Bar- bare Dewey: spicis staminiferis terminalibus 2 raro 3 erectis cylindraceis, suprema longe pedunculata, inferiore breviore illi contigua, infima sub-elongata; pistilliferis 3 longo-cylindraceis, 2-4 uneialibus gracilibus, superiore apice staminifera brevi- bracteata erecta, inferioribus, longioribus, subremotis, subrecur- vis basi laxifloris brevi-vaginatis foliaceo-bracteatis, omnibus nigro-purpureis, perigyniis distigmaticis oblongis obovatis api- culatis ore integris, squama oblongo obovata dorso pallida mucronata brevioribus; culmo erecto glauco longe-foliato vagin- atoque. Danks of streams, Santa Barbara, California ; Parry.— Culm 16-20 inches high, erect, with long leaves towards the base and long leafy bracts above, glaucous; spikes 3-6, eylin- dric, slender, blackish purple; staminate terminal 1-3, com- monly 2, the upper nearly two inches long, peduneulate, the lower sessile, contiguous and shorter, the third longer than the last and more remote; pistillate 3, long cylindric, 2-4 inches long, slender; the upper staminate at the apex, short-bracteate, erect ; the lower longer, subremote, subrecurved, loose-flowered at the base and short-sheathed; perigynium oblong-obovate, short-rostrate, entire at the orifice, stigmas 2, pistillate scale oblong-obovate, on the back pale, and the nerve extended into a mucronate point, making the end of the scale sometimes emarginate. The locality gives the name of the species,” Some mature specimens of C. Barbare Dew. have been col- lected by Mr. Parish in San Bernardino Mountains,* and by examining these we noticed that the perigynia exhibit several very fine nerves, which must have been overlooked by Dewey ; otherwise these specimens showed exactly the same habit and structure of squamee as Dewey’s own specimen. It is to be * Southern California, alt. 3000 ft.,S. B. Parish, No. 3276, deposited in the herbarium of U. 8. National Museum. T. Holm—Studies in the Cyperacee. 483 remembered that the minor structure of the perigynium in Carex is seldom noticeable in the dried state, but readily visi- ble in fresh material or by soaking the dried specimen in boil- ing water and alcohol. In pointing out some of the most striking characters of this species we might mention: the long and slender, blackish purple spikes, the oblong-obovate, mucro- nate squamie, sometimes emarginate, which are longer than the faintly nerved, oblong-obovate perigynia; furthermore the short beak with entire orifice. Let us now consider some of the other plants which formerly have been referred to C. Barbare, but which we believe are distinct from this. There are, for instance, in the Gray herba- rium some specimens, and very well represented, of a Carew, which are named C. Garbarw in Dewey’s own handwriting; they are from Hayden’s collection and the localities are given as: “Lake Fork, 6,000 ft. above the sea, also on Madi- sons aver,” -“ and near Fort! (the name written very indis- tinctly), high on Rocky Mountains.” cos. 6: Fo = V a sin @ Sees a tee Pie eT ab 3')2 con Fi 201) isinig mm, 7 of 506 LH. A. Bumstead—Lorentz-FiteGerald Hypothesis. The acceleration along the radius vector is J,sm 6 + f, cos d = NE (1 — By? (1— 8’ cos’) 2 CR The acceleration ee to the radius vector is SiG = 8)? sin pcos ¢. If we take the earth as a numerical example, this perpendic- ular acceleration is very small. Its maximum value will occur when the earth is at the extremities of the minor axis of its orbit ; at this point COS|@ = € ; sind = ="a/ i See b) Jn COs & — fae WM ie where € is the eccentricity of the orbit. Taking e = 1-7 x 10~ and 6 == L0e® we find 2 . We é 1 Acceleration al Be eA? a toe £46 : ee TS (1 B)? [1 2°93 x 10 | Ve a\t 71. “Q—10 ma" (18) 2 [1:7 3< ake ] Acceleration perpendicular to r= I am not sufficiently familiar with the details of astronomical calculations to be able to say with entire confidence whether or not such an acceleration perpendicular to 7 could be detected. It seems, however, unlikely. The maximum effect is of the same order as would be produced by a perturbing body at a distance equal to that of the sun, and whose mass was only that of the earth. The perturbation, moreover, would 200,000 be periodic, vanishing at perihelion and aphelion and acceler- ating the earth’s motion in one-half the orbit, re it the other half. When the sun is also moving, the problem becomes more complicated. For the present purpose it will be sufficient to obtain the order of magnitude of the acceleration perpendicular to the radius vector. Let v be the velocity of the sun, and u that of the planet relative to the sun. Then the force on the planet is (E)=E+(v+u x where E is given by equation (1), in which 8 is now the ratio of the velocity of the sun to the velocity of light, and @ is the angle between the radius vector and the sun’s path. The mag- nitude of Hl is given by equation (2). The force E is along the radius vector; the force (v + u) X His normal to the resultant path of the planet. Let W be the angle between 7 and the tangent to the resultant path of the planet, then A: Bumstead—Lorente-F; iteGerald Hypothesis. 507 F, = E cos w F,=Esny+|(v+u)xH| in which the term enclosed by vertical lines represents the magnitude only of the vector. HZ is perpendicular to the plane containing rand Vv; let w, be the component of u in this plane and let w be the resultant of vw, and v. Then |(v +u)xH | =vH =~; E sin 0 and F, = E (sin wy + +i sin 6). Dividing F, and F, by the longitudinal and transverse masses respectively, we obtain for the accelerations, 3 crs Osea oe Te 3 a E cos ¥ Ce eee (Bh 7 wo. a a K (sin wy + V2 sin 6) The acceleration perpendicular to the radius vector is 1 2) 2 ; J, cos W — f, ny =—*) E (@° sin w cosy + =7zsin 6 cos y). Recent estimates make the sun’s velocity about 20 kilometers per second, so that @* = 0°-45x10~; its direction makes an angle with the plane of the earth’s orbit of about 55°.. When ¥ is perpendicular to the plane containing Vv and the normal to the plane of the orbit, cos w is nearly zero; it must in fact be less than e (the eccentricity of the orbit) even in the favorable case when the minor axis falls in this position ; with the major axis in this position it will be zero. In this position, therefore, the acceleration perpendicular to the radius vector cannot be as much as twice that which was found for the sun at rest. When r is in the plane containing v and the normal to the plane of the orbit, 0 = 55°, Wh < 55° and w =v. So that the © acceleration perpendicular to the radius vector will be less than iret td am ) E f° sin 110° a 2 that is its ratio to the acceleration in the direction of the radius will be less than 14107. In order to be quite certain that astronomical facts are not in conflict with the principle of relativity, it will doubtless be necessary to make detailed comparisons between observation 508 H. A. Bumstead—Lorentze-PitzGerald Hypothesis. and calculation based upon this hypothesis. The small magni- tude of the departures trom the Newtonian law, of which more or less rough estimates have been given above, render it prob- able that there would be no serious lack of agreement. This probability is strengthened by a ealeculation published some years ago by Lorentz.* In this he found the secular variations of the elements of the orbit of Mercury due to the substitution of electro-dynamic forces for the strictly Newtonian force. The variations in the angular elements amounted to only a few seconds of arc in a century and the change in the eccentricity to 0:000005. He did not, it is true, take into account the effects of variable mass, which had not at that time become prom- inent even in electrical theory. The introduction of electro- magnetic mass will, in general, tend to diminish the effects of the sun’s motion and to exaggerate the effects of the motion of the earth relative to the sun. But from a comparison of the theoretical accelerations in the two cases, it does not appear that the variations could be increased enough to produce a sensible discrepancy. [| Note added in Proof, Oct. 12. Since the above was written, two papers have come to my knowledge which bear upon this question. A. Wilkens (Phys. Zeitschr., vii, p. 846, 1906) has introduced electromagnetic mass in the ordinary Newtonian equations and has calculated the resulting secular variations in the elements of Mercury, Venus, the Earth, Mars, and Encke’s comet. In all cases the variations are within the limits cf accuracy of the observations. F. Wacker (Ibid., p. 300) considers the case when both force and mass are electromagnetic and, upon applying his equations to Mercury, finds for the motion of its perihelion a value less than one-fifth of that which is at present unaccounted for. The changes in the scales of length and time which would be introduced by the principle of relativity could affect these results very little; so that it seems quite certain that our present observational knowledge of gravitation is not sufficientiy exact either to exclude the general application of the principle or to supply evidence in its favor. | * Amsterdam Proc. II, p. 571, 1900. Chemistry and Physics. 509 SCIENTIFIC INTELLIGENCE. I. CyeEmistry AND “PuHysics. 1. Utilization of Atmospheric Nitrogen. read before the Faraday Society, Dr. AtBerT FRANK has given an interesting account of recent progress in the preparation of nitro- gen compounds from the nitrogen of the air, particularly in the form of calcium cyanamide, to which the commercial name nitro- lim has been given. He discusses the production of Chili salt- peter and the ‘approaching exhaustion of this source of nitrogen supply, and gives interesting data in regard to the production of ammonium sulphate in the gas industry. He indicates the increasing demand for nitrogenous compounds for agricultural fertilization, and shows that the synthetic products are assuming great importance. He states that the production of nitrates by the oxidation of atmospheric nitrogen is making excellent pro- gress in Norway, but it is his opinion that the Norwegian salt- peter will remain the only direct competitor of the Chilian variety, on account of the cheapness with which electrical energy can be obtained in that country, which has unrivalled resources of water power. He states that calcium cyanamide has been found to be a satisfactory nitrogenous fertilizer, and gives an account of its preparation. The atmospheric nitrogen is first concentrated by the fractional distillation of liquid air by the Linde method. The remaining oxygen is then removed by passing the gas over heated metallic copper. The nitrogen is then absorbed by finely ground calcium carbide in a heated retort, according to the equa- tion CaC,+N,=CaCN,+C. The product contains from 57 to 63 per cent of calcium cyanamide, giving about 20 to 22 per cent of nitrogen. The product is used directly as a fertilizer, and ammonia can be prepared from it very readily by the action of steam upon it. Works for the manufacture of this product have been started in many localities, most of which are in Europe, but one has been started on the Canadian side at Niagara Falls, and another in Japan. It is estimated that at the end of the present year works for the production of 45,000 tons of nitrogen by the cyanamide process will be in oper ation.— Chem. News, X¢Vii, 289 and 303. BH. b, W: 2. The Action of Radium Emanation on Solutions of Copper Salts.—Last year the sensational announcement was made by Ramsay and Cameron that they had observed the production of alkali metals, particularly lithium, in solutions of copper salts which had been subjected to the action of the radium emanation. These results appeared to be of so much importance that Mpme. Curie and Mpiiex. Grepitscu have attempted to reproduce them. In the first place they placed a solution of copper salt in a little 510 Scientific Intellagence. glass flask into which a large quantity of emanation was introduced and allowed it to decay there spontaneously. In four such experi- ments lithium was detected in the solutions, while blank experi- ments in which no emanation was used gave no indications of lithium. ‘The experiments were then repeated with every possible precaution. It was found to be extremely difficult to get chemical products free from lithium. It is present in distilled water, and almost all reagents ; and if a reagent does not contain it, and is allowed to remain some time in a glass vessel, it is then found to contain traces of the element. Water which had been distilled from platinum and gave no test for lithium upon evaporation, gave a distinct spectroscopic test for that element after it had stood for 24 hours in a glass flask. It was found that fused quartz also contains lithium, and therefore platinum was selected as the material in which the careful experiments were carried out. As a result of these experiments the investigators were unable to find any indication of the production of lithium by the action of radium emanation upon solutions of copper salts, and therefore they could not confirm Ramsay and Cameron’s results.— Comp- tes, Rendus, exlvii, 345. H. L. W. 3. The Formation of Mists in Presence of Radium Ema- nation.—MpmeE. CURIE showed some time ago that the presence of radium emanation causes the condensation of saturant or non- saturant water vapor as well as other vapors. This condensation is manifested by the formation of a persistent mist, visible by the light of an electric are. Upon further study it is the author’s opinion that chemical compounds capable of absorbing water vapor until drops are formed are produced under the action of the emanation. The mists are persistent and may last more than a month, and they disappear gradually as the emanation decays. . With pure water and air charged with emanation, a slight mist lasting some days is observed. If the air is replaced by carbon dioxide, no persistent mist is observed. But if instead of pure water a mixture of equal weights of water and sulphuric acid is used, a dense, persistent mist is obtained. A very persistent mist is produced in the presence of concentrated sulphuric acid and carbon dioxide. It was found that a flask containing water and air charged with the emanation gave a much more dense mist when closed with a rubber stopper than when sealed up, and when a piece of sulphur was placed in the air in a similar sealed flask the amount of mist was increased, while traces of sulphuric acid could afterwards be detected in the water. Mists which were very intense at first and lasted more than a month were observed with petroleum ether and with carbon disulphide in the presence of air charged with emanation. Anhydrous ether in presence of carbon dioxide and the emanation also gave a persist- ent mist. Experiments with certain solids showed that in the presence of emanation, iodine and carbon dioxide as well as cam- phor and air gave dense mists of long duration. A mist may be observed with actinium in the presence of water and carbon Chemistry and Physics. pit dioxide. The emanation must be fairly concentrated at first to produce the mist, which, however, may persist for a month when the concentration of the emanation has been reduced about 200 times.— Comptes Rendus, exlvii, 379. H. L. W. 4. The Preparation of Argon.—FiscHeR and RineE have worked out a new method for producing argon from the air, which can be carried out on a relatively large scale at a moder- ate cost. The novel feature in the process consists in the use of calcium carbide for absorbing oxygen and nitrogen at a single step. This is used in the form of a powdered mixture of 90 per cent of calcium carbide and 10 per cent of calcium chloride heated in an iron vessel to 800°C. The absorption of these gases is complete after sufficient circulation, practically with the forma- tion of calcium oxide, carbon, and calcium cyanamide, accord- ing to the equations O,+2CaC, = 2CaO+4C, and N,+CaC,= CN,Ca+C. The authors prepared eleven liters of atmospheric argon in the course of two days by means of their apparatus, and showed its purity by making determinations of its density. They make the statements that the atmospheric air contains 0:937 per cent by volume of “ noble gases,” the so-called crude argon ; that the density of crude argon is 19°94 compared with oxygen as 16; that crude argon itself consists of 99°75 per cent by volume of argon, and 0°25 per cent of a mixture of helium, neon, krypton, and xenon, in which neon with the density 10 predom- inates, so that crude argon is somewhat lighter than pure argon, the density of which is 19°95.— Berichte, xli, 2017. H.L. w. 5. The Chemical Analysis of Iron; by ANDREW A. Bratr. Seventh edition. Pp. xix, 327, 108 figures and 5 tables. Phila- delphia, 1908 (Lippincott & Co.).—This excellent handbook appears in its seventh edition, the first edition having been issued in 1888 (cf. (3), xxvi, 387). The fact that its admirable. charac- ter has been fully recognized by those using it is well shown by the frequent revisions called for. The present edition contains a description of some new analytical processes concerning the separation of vanadium, molybdenum, chromium and nickel in steel; and further an account of the volumetric method for nickel. The methods for gas analysis have been revised, as also the subject of atomic weights ; the table of factors for the latter have been recalculated from the values for 1908 given by the International Committee. 6. Decomposition of Water Vapor by Electric Sparks.—It has been suggested that the decomposition of water vapor in the case of thunderstorms may explain certain phenomena in those storms. Messrs. A. Horr and EK. Hopkinson conclude from their experiments “that when electric sparks pass through water vapor or carben dioxide the separation and arrangement of the decom- position products is not an electric phenomenon but results from gaseous diffusion. The hypothesis of electrolysis in liquids is therefore inapplicable.”— Phil. Mag., July, 1908, pp. 92-110. eos 512 Setentifie Intelligence. 7. Reflection from Glass at the Polarizing Angle.—Lord Rayueies concludes a study of this subject with the remark “that even a recently repolished surface, which may exhibit but a small ellipticity, is in a highly complicated condition. Grease itself may be comparatively inoperative optically on account of its index approximating to that of the glass. But why varying degrees of moisture should make so little difference is not appar- ent. Surface phenomena generally offer a wide field for investi- gation, which might lead to results throwing much needed hght upon the constitution of matter.”—Phil. Mag., Sept., 1908, pp. 444-447, 8 fas: 8. Himission of Electrons from Glowing Metallic Oxides.— Fevix Jenrzscu refers to the work of J. J. Thomson and to that of Professor O. W. Richardson on the general subject of the. emission of electrons from glowing bodies, and finds that the oxides arrange themselves in respect to rise in potential according to their electromotive series, that electropositive substances hold more free negative electrons than the electronegative. The work which the electron has to do in being thrown off is greater with electropositive substances. This work is performed only at the surface of the substances. The velocities of the electrons was found to be in agreement with Drude’s theory and with the observation of Lenard on the photo effect. The paper of Jentzsch contains a comparison between the energy of electron emission and the radiation energy.—Annalen der Physik, No. 11, 1908, pp. 129-156. Jie 9. Lhe Winetic Energy of the Negative Electrons Emitied by Hot Bodies—J. J. Thomson has stated that the carriers of nega- tive electricity emitted by hot bodies are electrons. Prof. O. W. Ricuarpson has assumed that the transitional energy of the elec- trons inside the metal has the same value as that of the molecules of a gas at the same temperature as that of the metals, and that the translational kinetic energy of the electrons outside the metal possesses the same value. Professor Richardson’s paper embodies the result of an investigation of the portion of the kinetic energy which depends upon the component of the velocity normal to the emitting surface. What is determined is the value of $mz° where m is the mass of an electron and wu is its component of velocity perpendicular to the surface from which it is emitted.— Phil. Mag., Sept., 1908, pp. 353-376. Fis Il. GEOLOGY anp MINERALOGY. 1. Die Entwicklung der Kontinente und ihrer Lebewelt, ein Beitrag zur Vergleichenden Erdgeschichte ; by THEopoR ARLDT. Pages 730, figures 17, and 23 maps. Leipzig, 1907 (Wilhelm Engelmann).—In this large and detailed work the Principal of the Realschule at Radeberg, Saxony, presents the history of con- tinental development and their biota past and present. Geology and Mineralogy. 513 So great an undertaking cannot be adequately reviewed here and the reader will be informed only as to the manner in which the study is presented. The first 30 pages summarize the methods of paleogeography according to petrographic, paleontologice, plant and animal data. Following these are 371 pages giving the biogeography of geologic organisms, a study of the greatest value to all desiring to know the regional and subregional distri- bution of extinct floras and faunas. - ii OL GREET TE Ps = sa hit ia 37 color; they are translucent, or in some cases transparent. The matrix is a pure white quartz, the crystals occurring in cavities with quartz crystals. The rare beauty of these crystals led the writer to urge his correspondent in China to scour the country for more and larger specimens. A second consignment of all that could be found has just arrived, and two of these remarkable speci- mens are Shown in the figures, reproduced from photographs, actual size. It is stated that the mine from which these were taken is now filled with water, and will not again be worked. Am. Jour. Sct —Fourrs Sremse Vor, XXYVI No 155 —Noververe 1908 518 Scientific Intelligence. III. Botany. 1 Grays New Manual of Botany. A Handbook of the— Flowering Plants and Ferns of the Central and North-eastern United Stutes and Canada. Rearranged and extensively revised by Bensamin Lincotn Rosrnson, Asa Gray Professor of System- atic Botany, and Merrirr Lynpon FERNALD, Assistant Professor of Botany, in Harvard University.. Pp. 926. New York, 1908. (American Book Company.)—Sixty years have elapsed since the first edition of Gray’s Manual of Botany was. published. From time to time during that long period the treatise has received careful editorial attention, and necessary additions have been incorporated. Under the limitations of stereotyped pages, some of these additions have been, of course, rather unwelcome, and have found their place sometimes in supplementary pages of new issues. The last thorough revision before the present one was undertaken after Professor Gray’s death. The work was very satisfactorily done by the late Dr. Sereno Watson and by Professor J. M. Coulter. Numerous important changes were made after the most careful deliberation, and the decisions proved acceptable to the majority of working botanists. But in the eigh- teen years which have passed since the publication of that sixth edition, great advances have been made all along the line in Systematic Botany, and it has been obvious that a new edition of the Manual is imperatively demanded. For some years this revis- ion has been in progress at the Herbarium, where the first edition was prepared. The Curator of the Gray Herbarium, Professor Robinson, and his aid, Professor Fernald, have given to the task a great part of their time and the most loving care. Serious difficulties confronted them. In the first place, the accumulation of material of late has gone on with a rapidity which threatened to carry the size of the volume beyond the limits of convenience, so that it could not longer be called properly a “ handbook.” But by the exercise of much skill, the revisers have kept the book within reasonable bounds, and have given it essentially the form and size of the sixth edition. ‘The second serious difficulty con- sisted in the absolute necessity of bringing order out of the cha- otic condition of nomenclature. ‘This order has been measurably secured by a consistent adherence to the Vienna agreement, which is justly acknowledged as International instead of provincial. But the synonyms which have found a place in other systems have here been placed within reach of the student. This part of the work has obviously demanded the exercise of the greatest care, and this it has received. A third difficulty, promptly met, was the complete change, -amounting almost to inversion, in the sequence of the natural families. This change has erown out of arecognition of affinities between plants, which compels a general re- arrangement. It is perhaps not too much to say that such a re-arrangement would have been unwise in 1890 when Drs. Watson and ‘Coulter issued Miscellaneous Intelligence. 519 the sixth edition ; it would manifestly have been unwise to fail to make this change to-day. Many extremely perplexing questions of a minor character have been well and skilfully met by the authors of the present revision. One of these is the selection of illustrative helps. These are incorporated in the body of the page, and are not too numerous to be confusing. They are, for the most part, excellent and telling. Another difficulty, and the last to which we shall now refer, was the discrimination between forms in polymorphic genera where such differences can be made to appear as specific instead of varietal. The multiplying of these forms under the name of spe- cies has introduced a question of the first magnitude. Of course, one cannot expect to satisfy everybody even by compromises, but such compromises seem to be demanded now and then. The revisers, who may well be called the authors of the present edition of Gray’s Manual, have shown great ability in managing these perplexing matters, and are to be congratulated on their SUCCESS. G. L. G. TV. MiscerztaAngous Screntiric INTELLIGENCE. 1. Carnegie Institution of Washington.—Recent publications of the Carnegie Institution are given in the following list (continued from p. 100): No. 39. Handbook of Learned Societies and Institutions. America. Pp. villi, 592. No. 75. The Fossil Turtles of North America; by OLiver Perry Hay. Pp. iv, 568, with 113 plates; 4to. No. 85. (Massachusetts). Index of Economic Material in Documents of the States of the United States. Massachusetts, 1789-1904. Prepared for the Department of Economics and Sociology of the Carnegie Institution of Washington; by ADELAIDE R. Hasse. Pp. 310, 4to. No. 87. Volume I, Parts I, IL. The California Earthquake of April 18, 1906; by AnDREw C. Lawson, chairman. In collabo- ration with G. K. Gitpert, H. F. Rem, J. C. Branner, and others. Pp. xvii, 451, with 146 plates, 25 maps, 15 seismo- grams; 4to. Report of the State Earthquake Commission. In two volumes and Atlas. No. 89. The Old Yellow Book. Source of Browning’s The Ring and the Book, in complete photo-reproduction with trans- lation, essay, and notes; by Cuartes W. Hopety. Pp. cclxii, 345, 4 plates. No. 94. The Structure and Life History of the Hay-Scented Fern; by Henry SHOEMAKER Conrap. Pp. 56, with 25 plates. No. 95. Papers of the Station for Experimental Evolutions, No. 10. Inheritance.in Canaries; by Cuarxies B. Davenport. Pp. 26, 3 plates. No. 99. Botanical: Features of North American Deserts; by Daniet Tremprty MacDovear. Pp. 111, 62 plates. No. 101. The Variation and Correlations of certain Taxo- nomic Characters of Gryllus ; by Frank EK. Lutz. Pp. 63. 520 Screntifie Intelligence. 2. Ricerche Lagunari ; in charge of G. P. Magrini, L. De Marcut, and T. Gnersorro, under the auspices of the Reale Istituto Veneto di Scienze, Lettere ed Arti. Wo. 8, Osserva- ziont Mareometriche, Lungo il litorale e in Laguna ( Biennio 1906-1907), 50 pp. and 3 figs. Wo. 9, Impianti Mareografici Eseguiti, 17 pp. and 4 photos. Wo. 10, Operazioni Geodetiche Fondamentali per il Rilievo della Citta e Laguna di Venezia, 64 pp., 2 photos, and 1 fig. Venice, 1908.—The reports of the study of the lagoons of Venice (see this Journal, xxi, 407, xxiii, 397, xxv, 89) are continued in the three bulletins listed above. The work has now advanced to the point where the velocity, direction of propagation, and physical character of the tidal wave are approximately determined. ‘The records show also a rather uniform wave of translation and a second tide more or less undetermined. The stations of observation have been increased in number and some of them reiocated until now there are three in the lagoon of Malamocco, ten in the lagoon of Ven- ice, three in the lagoon of Chioggia, and one each in the Het of Murano and at Caorle. Of especial assistance to the committee in charge of this inves- tigation has been the action of the city of Venice in undertaking detailed geodetic work in the region about the city, including the establishing of a new base line. The plans include the prepara- tion of a large scale topographical map of the entire district. H. E. G. 3. Beitrdge zur Chemischen Physiologie und Pathologie, herausgegeben von F. Hormetster. XI Band. Braunschweig, 1908 (Fr. Vieweg und Sohn).—This volume concludes the inde- pendent existence of Hofmeister’s Deitrdge, which henceforth is to be merged with the Biochemische Zeitschrift, edited by Pro- fessor C. Neuberg of Berlin. Professor Hofmeister will enter the editorial board of the latter journal. The noteworthy con- tributions to physiology contained in the first ten volumes of the Beitrige have been referred to in these columns from year to year. The final volume forms no exception in point of merit. Among the forty or more papers mention may be made particu- larly of HE. Friedmann’s extensive studies of the katabolism of carboxylic acids in the animal body ; Embden’s investigations on the genesis of the acetone bodies ; Wiechowskvi’s observations on the formation of allantoin in metabolism; Baer and Blum’s experiments on acidosis; and, as usual, numerous contributions on proteins and their derivatives. L.. Bev 4. Canada’s fertile Northland; edited by Ernest J. Caam- BERS. Pp. 139, with 6 tables and 5 maps in pocket. Ottawa, 1907 (Government Printing Bureau).—How extensive are the natural resources yet undeveloped in the vast northern area of the Dominion of Canada is well brought out in this volume. The information is given in the form of evidence presented to a Committee of the Senate by a considerable number of persons. This is classified as follows : the territory of Ungava ; the region west of Hudson Bay; the navigability of Hudson Bay; the climate of Canada. A series of large maps accompany the report. New Circulars. 84: Eighth Mineral List: A descriptive list of new arrivals, - rare and showy minerals. 85: Minerals for Sale by Weight: Price list of minerals for blowpipe and laboratory work. 86: Minerals and Rocks for Working Collections: List of common minerals and rocks for study specimens; prices from 1% cents up. Catalogue 26: Biological Supplies: New illustrated price list of material for dissection; study and display specimens; special dissections; models, etc. Szxth edition. Any or all of the above lists will be sent free.on request. We are constantly acquiring new material and publishing new lists. It pays to be on our mailing list. Ward’s Natural Science Establishment 76-104 Cottecr AVE., RocuEster, N. Y. Warns Naturat Science EstaBlisHMent A Supply-House for Scientific Material. Founded 1862. Incorporated 1890. DEPARTMENTS: Geology, including Phenomenal and Physiographic. Mineralogy, including also Rocks, Meteorites, etc. Palaeontology. Archaeology and Ethnology. Invertebrates, including Biology, Conchology, ete. Zoology, including Osteology and Taxidermy. Human Anatomy, including Craniology, Odontology, etc. Models, Plaster Casts and Wall-Charts in all departments. Circulars in any department free on request; address Wards Natural Science Establishment, 76-104 College Ave., Rochester, New York U.S. A. CONE EN Ss : Page Art. XLIJ.—Some New Measurements with the Gas Ther- mometer ; by A. L. Day and J. K. Cusmenr ___.____- 405 XLIII.—Range of the a-Rays ; by W. Duang .__.--=--_-- 464 XLIV.—Alteration of Augite-IImenite Groups in the Cum- berland, R. I., Gabbro (Hessose) ; by C. H. Warren __ 469 XLV.—Studies in the Cyperaceez. XXVI. Remarks on the structure and affinities of some of Dewey’s Carices ; b SES Hootie Sb ao ee as ne eae ee -. 478 XLVI.—Applications of the Lorentz-FitzGerald Hypothesis to Dynamical and Gravitational Problems; by H. A. BUMSTEAD:.. 2 24 Sol oe oe ee eS SCIENTIFIC INTELLIGENCE. Chemistry and Physics—Utilization of Atmospheric Nitrogen, A. FRANK: Action of Radium Emanation on Solutions of Copper Salts, MpMs. Curre and MpLur. GLEDITSCH, 509.—Formation of Mists in Presence of Radium Emanation, Mpme. Curie: Preparation of Argon, FiscHER and RINGE: Chemical Analysis of Iron, A. A. BLatr: Decomposition of Water Vapor by Electric Sparks, A. Hott and E. HopxKtinson, 011.—Reflection from Glass at the Polarizing Angle, RAYLEIGH: Hmission of Electrons from Glowing Metallic Oxides, F. Jenrzscu: Kinetic Energy of the Negative Electrons Emitted by Hot Bodies, O. W. Richa&psoy, 512, Geology and Mineraiogy—Die Futwicklung der Kontinente und ihrer Lebe- welt, ein Beitrag zur vergleichenden Erdgeschichte; by T. Arnupt, 512.— Archhelenis und Archinotis, H. v, TneRinc, 513.—Camarophorella, a Mississippian Meristelloid Brachiopod, J. E. Hype: Geology of Pike County, R. R. RowLey: Annual Report of the State Geologist of -New Jersey, for the year 1907, by H. B. KUmmert : Geological Survey of Canada, 514.—Mission scientifique au Dahomey, H. HuBert.—Fossil Turtles of North America, O. P. Hay, 516.—Beautiful Cinnabar Crystals from China, AOH. PETEREIT, 517. Botany—Gray’s New Manual of Botan), 518. Miscellaneous Scientific Intelligence.—Carnegie Institution of Washington, 519.—Ricerche Lagunari: Beitrige zur Chemischen Physiologie und Pathologie, F, HormMeIsTER : Canada’s Fertile Northland, E. J. Chambers, 520. ‘Librarian U. S. Nat. Museum. Boa DECEMBER, 1908. || —__ Established by BENJAMIN SILLIMAN in 1818. J | THE | | \ AMERICAN \ : , JOURNAL OF SCIENCE. ‘Eprrorn: EDWARD S. DANA. ASSOCIATE EDITORS Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW anp WM. M. DAVIS, or CAMBRIDGE, L. V. PIRSSON anp H. E. GREGORY, or New Haven, Proressor GEORGE F. BARKER, oF PHILADELPHL, Proresson HENRY S. WILLIAMS, oF ItHaca, Proressor JOSEPH S. AMES, or BALTIMORE, | k Proressors ADDISON E. VERRILL, HORACE L. WELLS, | f Mr. J. S. DILLER, or WASHINGTON. | FOURTH SERIES VOL. XXVI-[WHOLE NUMBER, CLXXVI] No. 156—DECEMBER, 1908. NEW HAVEN, CONNECTICUT. eh eT THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET. | en sae ee Published monthly. Six dollars per year, in advance. $6.40 to co Postal Union ; $6.25 to Canada. Remittances should be made either b registered letters, or bank checks (preferably on New York banks). | 1908 ee | | | Rare Cinnabar Crystals from China. Of the second consignment of these beautiful Cinnabar Crystals, which were described and illustrated in the November Number of this Journal, there still remained at present writing eight specimens. Of course some of these may be sold before this advertisement reaches you, but if you want one of these specimens before they are all gone, it is necessary to send in your order immediately. These eight specimens are among the choicest of this remarkable consignment. Their prices range from $10.00 up to $125.00. Write for illustrated pamphlet. BEAUTIFUL ZEOLITES, FROM THE NEW ERIE R. R. CUT AT BERGEN HILL, NEW JERSEY. We have added to and enriched our stock of these unsurpassed zeolites. Write us and we will send you a box on approval, or better, call and see them. NEWLY DISCOVERED RARE AND SHOWY MINERALS .AND NEW FINDS. ‘We have received this month quite a number of new consignments con- taining new discoveries both as to minerals and localities. Also some of wonderful beauty and extreme rarity. Space will not permit an extended description here, but we will mail list on request. SUITABLE FOR CHRISTMAS GIFTS. We have secured for the Christmas trade a number of large and small lots of Cut Gems, Cameos, Antique Mosaics, Opal Carvings. On account of the hard times both in Europe and in this country, we secured these at un- precedented low prices and will sell them likewise. Don’t miss this chance, as it may never come again. We name a few below :— Garnets, green and red; Aquamarines; Zircons. all shades: Sapphires, all shades ; Star Sapphires and Star Rubies ; Chrysoberyl, Cats-eye ; Spirels, all shades ; Topaz, pink, blue, brownish and golden color; Pink Beryl; Sphene ; Tourmaline, all shades; Amethyst, Siberia, royal purple color; Andalusite; Star Quartz; Peridote; Opal matrix, Mexico and Australia ; Precious Opal, Australia, Mexico and Hungary ; Hyacinth ; Turquoise, Mex- ico and Persia: Kunzite; Reconstructed Rubies and Sapphires ; Emeralds ; Opal Carvings, such as pansies, vine leaves with bunches of grapes, and other small Opal novelties ; Antique and Modern Cameos; Antique Mosaic and other semi-precious stones. POLISHED MINERALS. Lapis-lazuli, Chili, Persia and Russia; Malachite, Russia; Californite, Tulare Co. and Pala, Cal.; Petrified Wood, Arizona; Williamsite, Pa.; Ser- pentine, Pa.; Bloodstone, HE. India; Moss Agate, E. India and Lake Supe- rior ; Quartz with Rutile, Madagascar and N. Carolina; Smoky Quartz, White Quartz and Amethyst, from Japan, cut in form of a crystal; Chrysoprase, Cal. and Germany; Jade, China; Tiger-eye, Africa; Onyx, Mexico; Rose Garnet; slabs, Mexico; Jasper, all known localities ; Silver with smaltite, Cobalt, Ont.; Opal, Queensland and N.S. Wales; Agates, from all known localities, very picturesque ; Tourmaline, beautiful sections, from all known localities. Let us know your wants, and we will send them on approval to you. AsoH. PETEREIE, 81—83 Fulton Street, New York City. THE AMERICAN JOURNAL OF SCIENCE [FOURTH SERIES.] Arr. XLVII.— The Preparation of Urano-uranic Oxide, U,O,, and a Standard of Radio-activity ; by H. N. McCoy and G. C. AsHMAN. Iris obvious that for the comparison of the activities of radio-active substances a standard or unit of activity is of prime importance. One of us* has proposed to take as this unit the activity due to one square centimeter of a thick film of U,O,, of sufficient thickness to have maximum a-ray activity. It was shown that such a standard could be reproduced easily and apparently with quite definite activity. We have now studied the problem of the preparation of a standard of activity more fully, with the results recorded below. The chief points of importance are : First, The complete removal from uranium of radium and other active as well as inactive impurities ; Second, The preparation of an oxide of perfectly definite composition ; | Third, The preparation of uniform films of the oxide. Fourth, The activity due to the 8 rays. Material from three distinct sources was used : (A) So-called chemically pure uranyl nitrate from the firm of C. A. F. Kahlbaum. This was practically free from all ordinary impurities. The radium content, which was deter- mined by means of the emanation, amounted to 2°5 x:0-° of the equilibrium quantity. A portion of this nitrate, recrystal- lized from water twice, constituted sample A. It was not tested for radium again, as the original amount would increase the activity less than 0°01 per cent. (B) An old sample of uranyl acetate, which contained a con- siderable amount of sodium as the chief impurity. The amount * McCoy, Phil. Mag., xi, 177, 1906. Am. JouR Sci1.—FourtH SERIES, Vout. XXVI, No. 156.—DercremsBer, 1908, 37 . 4 | | | 522 McCoy and Ashman— Urano- Uranic Oxide. of radium present was very small: 2x 10—* of the equilibrium quantity. After conversion into nitrate and two erystalliza- tions of the latter from water, no trace of impurity could be found by chemical tests; the material formed sample B. (C) The third sample was prepared from the uranium extracted from 29 g. of pitchblende from the Wood Mine, Colorado and purified as previously described, (Joc. cet.) The process in brief consisted in treating the nitric acid solution of the mineral, after removal of silica, with an excess of ammonium carbonate solution to remove iron, etc., and the filtrate with ammonium sulphide to remove copper, lead, etc. The erude uranyl carbonate, obtained by boiling the last filtrate, contained two percent of the equilibrium quantity of radium. ‘This was removed by three precipitations of barium sulphate in the solu- tion, the first precipitate taking out 95 per cent of the radium , then present. The U,O, made from this material three years ago seemed to be very pure and had, as then stated, an activity within 0°15 per cent of that of another sample of U,O,, which was supposed to be, and probably was, of great purity. A large portion of this material had been kept in solution as ammonium uranyl carbonate from May, 1905 to November, 1907. During this time a very small precipitate had formed. The filtrate from this precipitate was boiled; the uranyl car- bonate so formed was called sample OC. 3 By decomposing uranyl nitrate at a temperature below red- ness, the orange oxide, UO,, is obtained; this loses oxygen at a higher temperature, giving U,O,. This latter oxide is not perfectly stabie, but loses oxygen slowly when very strongly heated, as first observed by Zimmerman.* From considerations based upon the phase-rule, for the three- phase system, U,O,, UO, and O,, the partial pressure of the oxygen is a function of the temperature; consequently for a fixed pressure of oxygen (say that in the atmosphere) there must be a definite temperature at which the three phases can exist in equilibrium; above this temperature U,O, will pass into UO,; below it, the dissociation will not take place. Since the temperature of a solid contained in a crucible heated in the flame of a blast-lamp is far from uniform throughout the mass, we have used an electric muffle which gave perfectly definite temperatures, which were accurately measured by means of a platinum-rhodium pyrometer. In one experiment 7 g. of sample A, purified uranyl nitrate, was converted at a moderate. temperature into the orange- colored trioxide. This was placed in a platinum crucible and heated in the electric muffle. The crucible was loosely covered, allowing free access of air. After constant weight had been * Ann, cexxxii, 276, 1885. MeOoy and Ashman— Urano- Uranie Oxide. 523 TABLE I. Duration of Heating. Temperature. Weight of Oxide. 60 min. 560° C, 3°698 35 600 3°697 75 600 to 700 3°693 Reduced to UO, in hydrogen, over Bunsen flame (37554) The dioxide was reheated, in air, in the muffle. 30 700° 3°691 And reduced to UO,, as before (3°552) and again heated in the air, in the muffle: 30 570° 3°690 This experiment was repeated several times with perfectly similar results. reached, the U,O, was reduced to UO, by the usual analytical method, heating over a Bunsen flame in a stream of hydrogen. The reduced oxide was again heated in air in the muffle. It absorbed oxygen rapidly and changed to U,O,. Table I gives the details of the experiment. ; The reduction of an oxide of uranium to UO, and weighing in that form is a standard analytical method. Therefore the composition of the product obtained by the reduction may be considered as known. The mean weight of the UO,, 3°553 g., corresponds to 3°693 g. of U,O, The fact that the product formed in air at 700° has the same composition whether formed by the decomposition of the trioxide or by the oxidation of the dioxide shows that this temperature is below that at which the U,O, can lose oxygen in contact with the air. There is, therefore, no danger that heating for any length of time will decompose U,O, at 700°. | Nine portions of U,O, were prepared from samples A, B and C,and made into ten standard films. Films nine and ten were made from the oxide prepared from sample A as shown in Table I. The U,O, for film thirteen was made as follows: A portion of sample C was heated in the muffle, in air, for thirty minutes at 700°; weight 1:°2134 9. This was then reduced by hydrogen and again heated’ in the muffle, in air, for thirty minutes, at 700°; weight 1:2133 @ Each of the remaining -seyen portions of U,O, was prepared separately by heating the corresponding sample for forty to sixty minutes, in air, in the electric muffle, at 700°. The method of preparation of films for activity measurements has previously been described in detail.* Seven films, Nos. 9 to 19 inclusive, were made in cireular tin dishes 7°15™ in diameter, with rims 0°8™ high; films 23, 24 and 25 were made on flat copper plates, 7-00 in diameter. * McCoy, J. Amer. Chem. Soc., xxvii, 391, 1905 ; Phil. Mag., xi, 176, 1906 ; McCoy and Ross, J. Amer. Chem. Soc., xxix, 1698, 1907. 5o4 McCoy and Ashman— Urano-Uranic Oxide. The activity measurements were made in a gold-leaf electro- scope having an ionization chamber 19°5°™ square and a dis- tance of 85° between the film and the electrode of the gold-leaf system. The scale of the micromoter microscope cor- responded to potentials between 576 and 473 volts, the fall of potential across the scale being 103 volts. Our standard films, which had about 0°020 g. of oxide per sq. cm., were all sufti- ciently thick to have the maximum a-ray activity. The 6-ray activity was small, but varied with the weight of the film. It was also evident from the nature of the B-rays that the observed activity of the latter must depend upon the size of the ionization chamber, which in the case of our electroscope, though large enough to get the maximum effect of the a-rays, was insufficient for the B-rays. We have determined the -ray activity for our electroscope in the following manner: The standard films of U,O, were covered with one to four pieces of aluminum foil 0:0048™ thick and the resulting @-ray activity measured. The first layer of foil cut off all the a-rays and a portion of the B-rays. By graphical extrapolation of the curve having 6-ray activity and number of foils as coordinates, it was found that this portion amounts to 9°2 per cent of the effective S-ray activity of the uncovered film. The further data for the films made from samples A, B, and C are given in Table II, the activity being expressed in terms of an arbitrary unit. TaBLeE II. Films in tin dishes 7:15°™ in diameter with rims 0°8™ high. a and 3 Sample No. Weight Activity (6 Activity* a Activity a Gps ee 9 0°816 10080 0°0108 0°9972 DNS on ete 10 0°807 1°0075 0°0100 09975 | Byes oe atee el Ve) 0°820 0:9990 0:0020 0:9970 } OVAh Gwe poe menue IS) 0°753 170005 0°0020 0°9985 ict caren 13 0°920 1:0040 0°0110 0°9930 Cire 14 0°892 1:0010 0:0110 0°9900 Cpopera te Lb 0°793 1:0070 00154 0°9916 Films on flat copper plates 7:00 in diameter. A eae 23 0°694 (SS 0°0296 10815 ACs gies 25 0°625 1°1080 0°0262 10818 Bi See 24 0°629 1:0950 0°01338 1°0817 These results show that the method of making U,O, gives a product of definite activity. The somewhat low activity of *The activities were measured before the maximum amounts of UX, which had been largely removed in the process of purification, had again accumulated ; for this reason the (-ray activity of a film is not a definite function of its weight. McCoy and Ashman— Urano-Uranice Oxide. 525 sample C may be due to a trace of impurity; but the value is sufficiently close to the mean for A and B to show that the method is satisfactory. . We next made a new set of films on flat copper plates which were cut on a lathe and were almost perfectly circular. These plates differed slightly in size, however, for which reason the diameter of each plate was measured (in two or more directions) with a comparator, capable of giving results accu- rate to 0-°001™. The U,O, for these films was prepared in a single portion from sample A. On account of the large quantity of material used it was necessary to continue the heating at 700° for about three hours in order to obtain con- stant weight. ‘The activities are shown in Table III; the last column gives the corrected activity for exactly 7° diameter. TABLE III. a+ Corrected No. Weight Activity 6 Activity a Activity Diameter a Activity a>. O-6812 1°:0800 00263 1°0537- 7:0168 1:0487 28a... -0°8776 1°0788 0°0347 10441 6°9912 1°0467 Zoe" 074996 10861 0°0300 1°0561 70174 1°0509 290= -. 07770 1°0812 . 0°0288 170524 70174 1°0472 aeae O° 7801 1°0935 0°0326 1°0609 70301 1°0519 ae -O'070 10737 0°0270 1°0467 6°9969 1°0476 eae O-Tor7 1'0796 0°0321 1:0475 6°9948 1°0491 34a... 0°9107 10891 0°0364 1°0527 6°9948 1°0543 Mean, 1-0495 The activities in Table III are also expressed in terms of an arbitrary unit, about three per cent greater than the unit used for Table II. One of us (Joc. cit.) has proposed to take as the unit of radio-activity, the activity due to one sq. cm. of such films of U,O, as those to which Table III refers. In terms of such a unit, the activity of each film of Table III is represented by its area. | Several years ago, Rutherford and McClung* determined the ionization current of layers of U,O, of different thickness, when placed between parallel plates sufficiently far apart for the complete absorption of all the a-rays in the air between the plates. The saturation current per sq. cm. of surface was 4:0 X10~-“ampere, for a weight of 0-0189 g. of U,O, persq. em. This weight is great enough to give the maximum a-ray activ-- ity provided the material is in the form of a perfectly uniform film ; but this was not the case in Rutherford and McClung’s experiment, as the oxide was merely “dusted” on the plate and therefore the observed current was far below that fora * Phil. Trans. A., exevi, 52, 1901. 526 McCoy and Ashman— Urano- Uranie Owidle. perfectly uniform film.* We have made a determination of the saturation current in absolute units for the a radiation of our standard of activity. To do this we made use of a stand- iielea be ard condenser of the concentric spheres type, in conjunction with a gold-leaf electroscope. See fig. 1. The electroscope case consists of two rectangular compart- ments, made of sheet brass 1°5™ thick. The ionization chamber, A, is 19°5°™" square and 14™ high; the gold-leaf chamber B is 8 by 10™ and 12°7™ high. The gold-leaf system is insulated by an amber plug, D, and carries as its lower end the circular electrode, E, a brass plate 14°™ in diameter. F is * McCoy, J. Amer. Chem. Soc., xxvii, 395, 1905. McCoy and Ashman— Urano-Uranic Oxide. 527 the charging device, which is connected with a three-point key, that keeps F earthed, except at the moment of charging. G represents a pair of glass windows, through which the motion of the leaf is observed by means of a micrometer microscope. The door, which slides upward, is at H. Lisa metallic support of variable height, for the film. L is a flat metal plate 18°" square with a 2° hole in the center; it is supported by a pair of brass rods at two diagonal corners and may be raised or lowered. When this plate, L, is placed about 8™™ above the electrode, E, the electrostatic capacity of the electroscope (without the condenser, C), is about five times as great as it is without L.* The standard capacity, C, consists of two concentric spheres — of sheet zinc.t The outer one is soldered to a brass block, J. having a 5” hole, through which passes a 1™™ brass wire soldered to the gold-leaf support; a very fine wire of spring brass makes contact with the inner sphere. The latter is sup- ported by three amber pins, KK; each pin is threaded into a small brass ring, so as to be adjustable; the brass ring is carried by another ring of vuleanite, which insulates it from the sphere and enables the brass ring to be used as a guard ring. It was found, however, that the insulation was sufficiently good without the use of the guard rings. The upper half of the outer sphere is detachable; this arrangement allows the inner sphere to be introduced or removed readily. The whole apparatus is surrounded by a wooden case, surmounted by a glass bell-jar. This serves to keep the temperature uniform inside the gold-leaf chamber and so avoids air currents which greatly diminish the accuracy of the activity measurements. The condenser spheres were made of spun sheet zinc. The radii were calculated from the weight of water required to fill each and the weight of the zinc of the smaller sphere. The exterior radius of the smaller sphere was 6°298™; the interior radius of the larger 7:590™. The capacity of the condenser * We have found that the observed activity of a given film is much more nearly constant when the capacity of the electroscope is increased by means of the plate, L; the reason for this is doubtless two-fold. First, the much slower movement of the leaf permits greater accuracy in timing; and sec- ond, the natural variation in activity during a fixed interval is a smaller fraction of the whole activity, the longer the interval. By increasing the time of discharge five-fold, the fractional error due to natural variation of activity would be reduced to less than half that for the more rapid discharge. See Geiger, Phil. Mag., xv, 539, 1908, and Meyer and Regener, Ann. Phys., xxiv, 757, 1908. +The mode of combination of electroscope and condenser is a modifica- tion of that suggested by Prof. Millikan; Electricity, Sound and Light, p. 301, Ginn & Co. 1907. (o'6) 2 Or MeCoy and Ashman— Urano-Uranic Oxide. ‘ See K.8.U. It is considered that when the inner sphere of the condenser is placed in posi- tion the calculated capacity, 38°83 E.S.U., is added to that of the system when this sphere is absent, but all else arranged as shown in the figure. Calling the capacity of the electroscope alone ¢, and with the condenser c,, then the ionization current mee — 2 Mite SARE to the two ends of the seale and z, and @, are the times of discharge for the same film for the changes of potent Dp, and p, respectively. Therefore (C, = GN t, P, l, P,—4P, The quantity ¢c,—c, is the capacity of the condenser=36°83 E.S.U. We used film No. 30, Table 3, placed at a distance of 3°6™ below the electrode of the gold-leaf system. Experiment showed that this distance was entirely sufficient for the pro- duction of the maximum ionization current; at a distance of 4:5°™ or more the current was slightly smaller, owing, of course, to partial recombination of the ions. As the closely agreeing means of several determinations, the times of discharge of the electroscope with and without the condenser were 243-9 and 420°9 sec. respectively, for the uncovered film. The times when the a-rays and 9-2 per cent of the @-rays were cut off by a sheet of aluminium 0:0043™ thick were 10,860 and 12,435 sec. respectively.* From these data it follows that the times of discharge for the a-rays alone were 249-9 and 436-4 sec. respectively. The fall of potential across the scale was 103°05 volts without the capacity and 102°63 volts with the capacity, the measurements being made with an electrostatic voltmeter. From these data it is found that c, = 48°84 E.8.U. and ¢, = 36°83+48°84 = 85-67 E.8.U. was, therefore where p, and , are the potential differences cor- = 1 Therefore the a-ray ionization current 7 = ap = 2250. 1 10-* amp. Film No. 30 is 7-0168°™ in diameter; its area= 38°67 sq. cm. Therefore the current per sq. cem.=5°79 KX 10>" amp. ‘The total activity of 1 g. of uranium? is equal to that of 796 sq. em. of a thick film of U,O,. Therefore the total a-ray ionization current of 1 g. of uranium=4°61 & 107" amp. * These times represent the 90°8 per cent of the @-ray leak plus the natural air leak and the leak across the insulation. The latter is somewhat greater when the condenser is attached to the electroscope; this makes the time 12,435 sec. shorter than would otherwise be expected. + McCoy and Ross, loc. cit. MeCoy and Ashman— Urano-Uranic Oxide. 529 It is well known that the potential gradient required to pro- duce a saturation current increases with increasing activity, the recombination of the ions being greater for a given poten- tial gradient the more intense the ionization. It was, there- fore, possible that the observed current for a standard film was below the maximum on this account. We made the following experiment to throw light on this point. Films of U,O, were made in the usual way on a pair of semi-circular plates, made by cutting an ordinary 7™ plate into halves. The activity of each half-film was measured separately and compared with that observed when the two were placed side by side to make a circular film. The sum of the separate activities was 0°33 per cent greater than that of the two together. The experimental error of the activity measurements did not exceed 0-05 per cent. The experiment shows that appreciably greater recombination of the ions takes place when the two plates act simultaneously, due to the more intense ionization. It follows from this, that the ionization current calculated above for one sq. cm. of 1G is somewhat smaller than that which actually would be observed for a film of unit area. However, the error which thus arises is eliminated in the calculation of the 1onization current of unit mass of uranium or thorium; since in such-a case the specific activity is calculated from the value based on the activity of an infinitely thin film* which would produce a vanishingly small ionic concentration. Consequently the ions would suffer no recombination and therefore the calculated ionization current is that which would be produced by all of the ions formed. Boltwoodt+ has determined the relative activity of radium and uranium by direct comparison of the activity of a minute known quantity of radium with the activity of very thin films of known weight of U,O,. It was found that radium (free from its products) is 1 30 x 10° times as active as an equal weight of uranium. Itutherford{ found that the ionization current of a thin film of 0-484 mg. of pure Rabr,, free from its active products, was 8-4 x 107° amp. as measured by a sen- sitive See ees: Considering half of the a-rays to have been absorbed by the plate carrying the film, this is equivalent to a current of 5-94 x 10“ amp. for 1 g. of pure radinm. We ‘have found the total ionization current of 1 g. of uranium to be 4°61 x 10-" amp. Therefore the a-ray activity of radium (free from its products) is 1°29 x 10° times that of an equal weight of uranium, a result which isin good agreement with that found by Boltwood. * McCoy, Jour. Amer, Chem. Soc., xxvii, 402, 1905. + This Journal, xxv, p. 296, 1908. t Phil. Mag., x, p. 207, 1905. 530 McCoy and Ashman-— Urano-Uranice Oxide. Summary. 1. Uranium is easily freed from all other radio-active sub- stances. 2. Pure U,O, of perfectly definite composition is readily obtained by heating any lower or higher oxide of uranium in air at 700° | 3. Uniform films of U,O,, 7°" in diameter, weighing 0°6 to. 0°8 g., have definite and constant a-ray activity and are there- fore recommended as standards of radio-actiwity. 4. The a-ray saturation current for such a standard film was measured in absolute units and from the results the total ionization current for 1 g. of uranium, in an infinitely thin film, was calculated. Kent Chemical Laboratory, University of Chicago, Sept., 1908. — Wright—Telemeter with Micrometer Screw Adjustment. 531 Art. XLVIII.—A Telemeter with Micrometer Screw Adjust- ment ; by Frep. Eucenn Wricut. Durtne the past few years many different devices have been suggested for measuring the distance to a distant object by merely sighting at it, and some of these, particularly the stereo- comparator of Pulfrich, have proved serviceable. Three or four years ago, in connection with geological field work involving considerable topographic sketching, the need of such an instrument was keenly felt by the writer and the following apparatus devised. ‘The apparatus is simple in construction and sufliciently accurate for the purposes for which it is intended. It appears, moreover, to be constructed on a prin- ciple not heretofore applied to telemeters, and may, therefore, be described very briefly. Fic. 1. The principle of its construction®* is illustrated in fig. 1. Light waves from a distant object strike the two telescopic lenses L, and L, (both 50™ focus) and after transmission are reflected from the two right-angled prisms P, and -P, to the reflecting prism pair P,, and thence to the ocular O. The incident rays are not precisely parallel and do uot converge to the same point in the focal plane of the ocular. They can be made to do so, however, by moving the prism P, back parallel with itself by means of the micrometer screw M until the two points coincide and merge apparently into one (indicated by the cross in front of the ocular). The angle between the inci- dent light rays from objects at different distances is different, but by moving the micrometer screw the two images resulting therefrom can be brought to coincidence. | Conversely, having once calibrated the micrometer screw readings for a number of distances, it is not difficult to mter- * The two test telemeters which have thus far been constructed on this principle were made in the workshop of the Geophysical Laboratory and can be duplicated by any good maker of instruments. 532 Wright—Telemeter with Micrometer Screw Adjustment. polate and to draw a curve indicating the distance away of any object in terms of micrometer screw readings. ‘The equation expressing the relation between micrometer screw readings and distances away of Obs is derived below (page 534), The optical parts of the instrument are the following: Two achromatic plano-convex telescopic lenses, 50™ focal length and 25™" diameter; two right angle reflecting prisms P, and P,, 18™™ length of side; a reflecting prism pair P, con- sisting of two right-angled reflecting prisms, the larger one of 20" edge and the smaller one of 14"™ edge, its hypothenuse Fie. 2a. el ere c: face fitting the side of the larger prism closely. The hypothe- nuse surface of this smaller prism is silvered ; at its center, a round circular portion 3 to 5"™ in diameter has been removed as in the Abbé camera lucida reflecting prisms. The two prisms are cemented with Canada balsam, the light from the larger prism reaching the ocular through the central part of the field, while that from the smaller prism is reflected by the peripheral silvered margin. Experiments with this type of double reflecting prism have indicated that the instrument should be so built that for aver- age distances the round aperture in the center of the field should be situated near the focal plane of the ocular O. Other reflecting devices were tried in place of the prism pair P., two of which are shown in figs. 2a and 2b, in which plane parallel glass plates are used. The four elas plates of fig. 2a are very difficult to adjust accurately and for the pur- pose special adjustment facilities had to be constructed. The device of fig. 2b consists of two glass plates and the light rays follow the paths indicated by the arrowed lines, With the glass plate devices the images are superimposed, while with the Wright—Telemeter with Micrometer Screw Adjustment. 533 prism pair P, the image in the central part of the field forms the continuation of that in the margin. By placing the reflecting prism, however, so that it is at some distance from the focal plane of the ocular, the two reflected images can likewise be superimposed. This was done on a trial test by replacing in fig. 3 the prism P, by a total prism pair, thus ’ bringing the light from both L, and L, to a common path and to reflection in a single reflecting prism at P, in place of the prism pair there indicated. The unfavorable features of the glass plate reflecting devices are chiefly the great loss of light and consequent dim field, and the double image from each lens which results from reflection from the two sides of each glass plate. For these reasons the permanent use of glass plates in this connection hardly seems feasible. Several other reflecting prism devices were tried, but that of fig. 1 and its modification in fig. 3 have thus far proved most satisfactory. The arrangement of the different parts is indicated in fig. 1; L,, L,, P, and P, are rigidly fixed and stationary, while the ocular O and the prism P, are movable,—the ocular for focus- ing purposes and the prism P,, by means of the fine micro- meter screw M,, for the purpose of measurement. It is imperative that the construction of the instrument be rigid throughout. In the trial instruments thus far used, the material has been either thick hard wood or a brass cylinder, and of these the brass cylinder is undoubtedly the more practi- eal. The three reflecting prisms P,, P, and P, are supported on their hypothenuse sides by brass blocks faced with cork, and these in turn are adjustable on a brass plate. By this method, the centering and adjusting of the optical parts can be accomplished at any moment accurately and with little trouble. The base line of the instrument to which all measurements are referred is the distance L, L, and its length should remain unchanged at all temperatures. Unless the instrument is made of some non-expansible material, as invar steel, however, this condition cannot be fulfilled, but for practical purposes the minute changes in length which the slight temperature variations produce may be neglected, since the instrument itself is not one of exceeding accuracy. Assuming’ the distance L, L, of fig. 1 to remain constant, the theoretical accuracy of the instrument for different dis- tances is not difficult to ascertain. -In fig. 1, let L, L,=a, then in the triangle CDA the side CD = aa Since the dis- tance of the object is always great with respect to the base of the instrument L, L, the angle, L,-object-L,, is small and the triangle DCA may be considered without sensible error sim- ilar to triangle L,-Obj.-L,. Accordingly 534 Wright—Telemeter with Micrometer Screw Adjustment. = Da (1) Let distance of object L,-Obj.=y and movement of micro- meter screw AD=a, then | y a EES Seed UW Gags 56) Y a 2.2 Q’ 22x Y= To find the relative accuracy of the instrument for different Fre. 3. distances y for a given small increment of a, equation (3) may be differentiated. ot tel (3) Since absolute values only are considered and not the fact that ~ and y are counted in opposite directions, the negative sign of equation (8) may be disregarded. On substituting the value of « from (2) in this equation, we obtain | 9 2 dy = 7. da | (4) Gi which states that the sensitiveness of the instrument decreases with the square of the distance and increases with the square of its length. . In the trial telemeter of fig. 1, the distance L, L,=a 1s 935° or about 1 meter. The micrometer screw reads to Wright—Telemeter with Micrometer Screw Adjustment. 535 ‘005"" and actual tests with the instruments show that changes produced in the field by a movement of -01™™ of the micro- meter screw can be readily detected. dx is therefore :01™™ or ‘00001 M, and equation (4) reads dy=2'y* °00001 or dy=:00002 . y’ At a distance of 50 meters, therefore, the probable error of the instrument is ‘05 M, or -1 of one per cent; at 100 M, -2 M of -2 of one per cent; at 1000 M, 20 M, or 2 per cent. In fig. 3, a slightly ‘different disposition of the reflecting prism is shown which for a given base line is twice as accu- rate as that of fig. 1, the entire base line a being used to pro- duce deflections of « instead of 7 asin fig. 1. For this modi- fication the equation reads, therefore, and dy, = z ae a The form of the prism pair P, in fig. 3 is slightly different from that in fig. 1, but can be ger ound with equal ease. ' From the diagrams it is apparent that the images from L, and L, do not form in precisely the same planes and are theo- retically, therefore, never in perfect focus at one and the same time, except for objects at an infinite distance. Experience has shown, however, that if the instrument be adjusted for ordinary distances, this defect is not serious, especially if a low power ocular or magnifying lens of 2 to 5 focal length be chosen. The eee image produced by the instrument as shown in the figures can be made upright by the use of a Rochon prism .pair directly in front of the ocular. The inverted image, however, is not a serious defect and equally good results can be obtained without the use of the extra prism pair, which encumbers the instrument and adds another adjustable part to-be looked after. Equation (3) shows that the accuracy of the instrument increases with the square of its length. It seems entirely feasible, therefore, to construct an instrument one or two meters in length on the principle of fig. 3, with which dis- tances of points within a radius of one or two kilometers or miles can be read off directly with considerable accuracy, thus accomplishing stadia measurements from the transit station without the aid of a rodman. Geophysical Laboratory, Carnegie Institution of Washington, Washington, D. C., June, 1908. 5386 Wright—Kxplanation of Interference Phenomena. Arr. XLIX.—A Device to Aid in the Explanation of Inter- Ference Phenomena, by Frep. Kugene Wrieut. StupEnts of crystal optics, on taking up the subject of bire- fringence,; frequently encounter difficulty in forming a clear conception of the exact course of the ight waves through the erystals and the resultant interference phenomena when polar- izer and analyzer are used (crossed nicols). The small apparatus of fig. 1 has been found serviceable as a model in this connection, and facilitates to a certain degree the explanation of several of the phe- nomena of plane-polarized light. The device consists essentially of a brass rod divided into three parts, a, b, c, which are so connected that each one is revolvable for itself about the common axis; into each section, moreover, longitudinal slits have been eut and plates of thin, transparent celluloid inserted. The celluloid plate A represents the plane of vibration of lhght waves emerging from the lower nicol of the microscope; B is a celluloid model of the erystal sec- tion with its ellipsoidal axes at an angle of 45° to the plane of vibra- tion of the lower nicol; C, and C, represent the planes of the two waves emerging at right angles to each other from the crystal plate, the heht waves ©, being a definite distance ahead of C, as a result of the une- qual velocities of the two waves in their passage through the crystal B. On entering the upper nicol, these two waves are again reduced to the common planes of vibration D, and D,, the waves vibrating along D,, however, being destroyed by total reflection and those along D, only passing through. By the use of this model, it is not difficult to prove: (1) that two waves emerging from a refracting crystal at a distance apart of one or more whole wave lengths (phase difference zero) interfere mutually when reduced to the common plane of vibration, D,; (2) while two waves one-half wave length apart (in opposite phase) mutually strengthen each other when reduced to the common plane of vibration, D,, of the upper nicol; vice versa, if the plane D, be considered, the phenomena are exactly reversed—tfacts which are difficult to represent clearly without the aid of some such model. Geophysical Laboratory, Carnegie Institution of Washington, Washington, D. C. rad T. D. A. Cockerell— Descriptions of Tertiary Plants. 537 * Arr. L.—Descriptions of Tertiary Plants, [1 ; by T. D. A. CocKERELL. Tue plants discussed below are all from the North American Miocene. They represent a flora containing many genera at that time widely spread over the Holarctic Region, but in later times driven southward, and to-day existing in much Rie: 2: Fie. 1. Geaster florissantensis. Ese el Fie. 2. Pinus sturgisi. more limited areas; some in Asia, others in various parts of America. They show very clearly that many of the ostensibly endemic genera of various regions may well have originated elsewhere, and are merely making their last stand where we now find them. Am. Jour. Sct.—Fourts Series, Vou. XXVI, No. 156.—Drcemper, 1908. 38 5388 7. D. A. Cockerell— Descriptions of Tertiary Plants. FUNGI. Geaster florissantensis sp. nov. Fig. 1. Diameter of “star” about 56™™, the segments about eight in number, five being visible, var iable i in form, the largest 20™™ long and about 114 broad at base, but one next to it only about 7 broad ; color dark brown, texture apparently leathery, with- out any sion of venation. Florissant, at a new station on the hillside not far from 20 (CHASE. Cockerell, 1908). It occurs on a slab with numerous remains of Typha lesquerewar Ckll., Ulmus hillie Lx., and other plants. The appearance is exactly that of a modern Geaster in the expanded condition, and the irregularity of the segments is unlike that of any calyx known to me. Geaster is, of course, common in Colorado to-day. GYMNOSPERMS. Pinus sturgist sp. nov. Fig. 2. Leaves in bundles of threes, apparently entire-margined, about 175™™" long and 13"™ broad, very straight, sharp-pointed. Two fibrovascular bundles are very distinct, being preserved as white lines. In all respects, the plant agrees very closely with the lving. P. teda L., of the Eastern and Southern States. Florissant; the type from Station 13 B (diss Gertrude Darling, 1908), but the species was also found, less weil pre- served, at various stations in 1907. ‘The species is dedicated to Dr. W. C. Sturgis, of the School of Forestry at Colorado College, in recognition of his contributions to Colorado botany. The fossil species of Pinus from Florissant must now be con- sidered to be three in number at least, separable as follows : Leaves in bundles of five ..-...-- P. wheeleri Ckll. (doubtfully recorded as P. palcostrobus (Ett.) Heer, by ao eames im punehes-ot-three: 222 52225). oe ee ee V; Weayesabouteliva"™ long 5.2 ee = alee is CkIl. Reaves-anouty 70"™ long {22 eee eee P. hambachi Kirchner. I formerly sunk P. hambachi under P. florissante Lx., which was based on a cone, but it must be restored, at least pro- visionally. Heyderia C. Koch. This genus, once widespread, is restricted to the Pacific coast region of North America (/leyderia decurrens (Torrey) C. Koch) and China (A. macrolepis = Libocedrus macrolepis Benth. and Hook. = Calocedrus macrolepis Kurz). At Sas ee { T. D. A. Cockerell—Descriptions of Tertiary Plants. 539 Florissant, Colorado, it is represented in the Miocene by 7. coloradensis Ckll., while in the ,Miocene of Europe, at Radoboj, Heyderi ia salicornioides (Libocedrus salicornioides Heer) is very well preserved. Other species, supposed to belong here, are from the Upper Cretaceous of Greenland and the Miocene of Spitzbergen. ANGIOSPERMS. Ailanthus americana sp. nov. Fig. 3. Samara about 38”™ long, 9 broad; seed 6™™ long and a little over 4 broad, placed with its long axis about 15 degrees from iDinsi. 4k. Fic. 3. Ailanthus . 1G. 4. Quercus knowltoniana. americana. FIG Quer tonia axis of samara; venation of wings well preserved, agreeing with that of A. glandulosa L.; apical part with a thickening along one side, as in Lesquereux, Oret. and Tert. Floras, oie ds a Florissant, Station 13 B, 1908. Type at University of Colorado. Adlanthus (wrongfully called Alzanthus in Knowlton’s Cat. Cret. and Tert. Pl.) is at present confined to Asia, with three species. It is well represented in the Tertiary beds of Europe, and is credited with two American Tertiary species, one from the Miocene of Oregon, the other from the Green River beds of Wyoming. The Oregon species is very distinct from ours ; that from Wyoming is based on a supposed leaflet with a remarkably long petiole, which seems to be doubtfully of this 540 7. D. A. Cockerell—Descriptions of Tertiary Plants. genus. However, Lesquereux figures with his A. donge- petiolata a samara, which he says “may not represent the fruit of the same species,” but which is evidently very much like that from Florissant. The seed is more transverse, how- ever; the venation is not shown. Quercus knowltoniana sp. nov. Fig. 4. Acorn-cup 30™™ diameter; scales in about 10 rows, triangular, from about the fifth row sharp-pointed, but the more basal ones broad and angled rather than pointed; no visible marginal fringe. Florissant (d/rs. Charlotte Hill). WHolotype at Yale Univer- sity, Cat. No. 1005. I had retained this curious fossil for months, hoping to be able to determine it, but failing to recognize its relationships. Dr. I’. H. Knowlton recently visited my laboratory, and upon showing the fossil to him, he at once recognized what it was. Now that the fact has been pointed out it is so evident that the specimen is an acorn-cup that I do not understand my obtuseness on the subject. The species recalls the recent @. macrocarpa Michx., the cups of which grow to an even larger size. I have no leaf from the shale that I can refer to it. The cup was evidently widely open and shallow, not partially closed as it is in Q. lyrata. Fossil acorn-cups have been found in the Miocene of Europe (Q. palwocerris Sap., Y. subcrenata Sap.). Rosa ruskinigna sp. nov. Fig. 5. Represented by a bud about 16™™ long, and six in diameter. Hypanthium subglobose, no doubt producing a practically spherical fruit, covered with minute spines; sepals with very large and thick-stalked glands or gland hairs on the basal half, these very much larger than the spines of the hypanthium ; apical portion of sepals long, with three or four large lobes on each side. Florissant, Station 138 B (W. P. Cockerell, 1908). By the character of the hypanthium this is evidently related to Ltosa cherokeensis Donn., but the sepals are strongly lobed. Such a rose would have trifoliate leaves, and these should resemble those of /?. Azlliw Lx., at least to a considerable degree. As, however, it is impossible definitely to connect the bud with the leaves of /2. Allie (we have not found the latter), I give the former a distinctive name; dedicating it to John Ruskin, whose copy of Lindley’s “* Rosarum Monographia,” with many marginal notes, is in my library. T. D. A. Cockerell—Descriptions of Tertiary Plants. 541 Hydrangea florissantia Ckll. Rhus rotundifolia Kirchner, Trans. St. Louis Acad., viii, p. 184, is the same thing. The name rotundifolia was much earlier used in ZZ ydrangea by Rafinesque. Kirchner’s type is, I believe, in the U. 8. N ational Museum. Sambucus newtoni sp.nov. Fig. 6. Leaflet (doubtless a lateral one) about 132™™ long and 26 broad ; texture thin, this and the venation exactly as in living Fig. 6. Fic. d. Rosa ruskiniana. Fic. 6. Sambucus newtoni. species of Sambucus; form parallel-sided, rapidly narrowing apically to a sharp point, very much as in S. arborescens Nuttall ; margin with exceedingly minute denticulations, 4 to 5 in 5m. and even these evanescent on the basal half. Florissant, Station 13 B (George Newton Rohwer, 1908). The best side shows all but the base; the reverse lacks the apex, but shows nearly all of the base, which is substantially as in S. arborescens. This is the first American fossil Sambucus ; in Europe the genus is represented by flowers in amber. Lomatia acutiloba Lx. is on the same slab as Sambucus newtont. 542 T. D. A. Cockerell Anona spoliata sp. nov. Fig. 7. Leaf apparently thick, oblong, entire, the blade 40™™ broad, and probably over 80 long (apex missing), the base broadly rounded, the midrib and petiole stout, the latter short, only about 9°™ long. Venation pinnate, the secondaries arising from the midrib at an angle mostly little less than a right angle, but varying in this respect, and gently curving upwards, ter “minating in submarginal arches connecting their tips, and variously enclosing areas of different shapes. Between the principal lateral veins are small and hardly noticeable ones, not ~ ie 7 Fic. 7. Anona spoliata. proceeding far from the midrib. In the shape of the leaf, the short petiole, and the venation, this is almost exactly lke the living Anona glabra L., of Florida. In one place two of the principal secondaries unite, as they sometimes do in A. glabra. Florissant, Station 13 B (Geo. WV. Rohwer, 1908). Sabina linguefolia (Lx.) Ckll. occurs on the same slab. Anona robusta Lx., from the Laramie (2) at Golden, Colorado, is a similar species, differing, however, in the character of the submarginal venation, which does not show the large enclosed areas. The resemblance of A. spoliata to A. robusta is, there- fore, not nearly so close as to A. glabra. The European A. me D. A. Cockerell—Descriptions of Tertiary Plants. 548 elliptica Unger, from the Miocene of Radoboj, is close to A. spoliata in respect to the submarginal venation, but very different in the cuneate base, the leaf being very like that of Crescentia latifolia. Juglans leonis n.n. Juglans californica Lx., Mem. Mus. Comp. Zool. vi, 34, pl. ix, x (1878). Miocene of California. (Not J. californica S. Watson, Proc. Am. Acad., x, 349 (1875).) Rhus mense vn.n. Rhus- metopioides Lx., Mem. Mus. Comp. Zool. vi, 31 (1878). Miocene of California. (Not “2. metoproides Turez., Bull. Soc. Nat. Mose., xxxi, 1, 468 (1858).) Salix merriaméi n.n. Salix elliptica Lx., Mem. Mus. Comp. Zool. vi, 10 (1878). Miocene of California. (Not S. elliptica Sleich., Ser., Ess. Saul., 44; cf. Steud., nom. (1841).) Lizyphus microphyllus Lx., and Magnolia lanceolata Lx., of the California Miocene, also bear preoccupied names. Weinmannia dubiosa Ckll. We found this at Stations 13 B and 14, at Florissant. The leaflets vary from five to seven. Robinia brittont sp.nov. Fig. 8. Represented by a leat, scarcely at all different from the living L. pseudacacia L. Five leaflets are preserved. Leaf- lets about 22™™ long and 94 broad, very briefly mucronate at apex, and with short petiolules about 2™™ long, which are as usual opposite, the pairs about 14™™ apart. From the first pair of leaflets to the insertion of the leaf is only 12™". The shortness of the petioles agrees best with /?. viscosa Vent., but the shape of the leaflets accords better with /?. pseudocacia. Florissant, Station 138 B (Jlelford Smith, 1908). Dedicated to Dr. N. L. Britton, who has contributed so much to our knowledge of American trees. Lobinia is to-day confined to America, but it is found fossil at Giningen and other European localities. Menyanthes coloradensis sp.nov. Fig. 9. Represented by a crown bearing five leaves, in form and appearance exceedingly like the living J/. trifoliata L., but 544 7. D. A. Cockerell— Descriptions of Tertiary Plants. two of the leaves are entire. One of the basal leat-sheaths, curled backwards, is well preserved, and exactly as in J, trifoliata. The whole plant is much smaller than J/. trafo- liata ; the petioles of the better-developed leaves only about 25m long, with leaflets about 30" long, and 9 to 10 broad. The prominent lateral or secondary veins are irregular, less numerous than in J/. ¢tr7folcata, and more or less strongly arched, with the concave side upwards. The entire leaves are broad-lanceolate to ovate, the largest being over 15™™ broad. Fie. 8. Fie. 9. Fic. 8. Robinia brittoni. Fig. 9. Menyanthes coloradensis. Florissant, Station 13 B (Geo. WV. Rohwer, 1098): also one found at the same place by Miss Gertrude Darling. Menyan- thes is to-day a monotypical genus of Holarctic “distribution. In the fossil state it is known, principally from capsules and seeds, from Greenland, Spitzbergen, and Central Europe. The occurrence of entire leaves on the fossil is of mterest in view of the fact that the allies of J/enyanthes are entire-leafed. Iasked Dr. L. N. Britton whether he had ever seen entire leaves on the living species: he replied that no such had ever come under his observation. The seedling of Jlenyanthes seems not to have been described. a Wright—Three Contact Minerals. 545 Arr. LI1—On Three Contact Minerals from Velardena, Durango, Mexico. (Gehlenite, Spurrite and Hillebrand- ite); by Prep Eugene WriGaHt. Dvurine the summer of 1907, a geologic examination of the Velardefia mining district in Mexico was made by Mr. J. E. Spurr, assisted by Mr. G. H. Garrey. Several of the thin sections of the material there gathered were sent to the writer for examination, and in one of these a mineral with peculiar optic properties was observed. At the suggestion of the writer an adequate collection of the rock from which the thin section had been cut, was then made by Mr. Garrey, and in this col- lection the three minerals to be described below were found. Two of these minerals proved to be new mineral species, sili- cates of interesting composition, while the third, gehlenite, is apparently novel for this continent. All three are contact minerals, formed near the junction of altered limestone and intrusive basic diorite and their relations to the contact and conditions of formation have been carefully studied by Messrs. Spurr and Garrey. As the results of their extended investi- gation will soon be ready tor publication it has not been deemed necessary to consider in detail in this present paper conditions of occurrence and formation of these minerals and their relations to the ore deposits in general. The chemical analyses of the three minerals and their specific gravity determinations were made by Dr. E. T. Allen of the Geophysical Laboratory, and to him the writer is deeply indebted for the courtesy. Gehlenite.* This mineral occurs in massive granular aggregates, usually dark gray or gray-black in color, from minute inclusions of magnetite and other particles. Rarely small pieces of gehlen- ite of amber-yellow color and free from magnetite inclusions were observed. The grains are rounded in outline and not suitable for crystallographic measurement. The physical and optical properties, however, are similar to those recorded for gehlenite from other localities, and the chemical composition also agrees as well with the prescribed formula as the analyses of the type material. Crystal system, probably tetragonal, judging from the cleay- age, which is imperfect after 001 and much less well marked after a prism. Fracture, uneven and irregular, conchoidal to splintery. Hardness, between 5 and 6, about 5°5. Luster, * Type specimen from contact aureole of the Terneras intrusion, Velardefia, Durango, Mexico. 546 Wright—Three Contact Minerals. resinous to greasy. Translucent to transparent in thin flakes ; in large masses, sub-transparent to opaque. Streak white to pale gray, the gray probably due to fine magnetite inclusions. In the thin section, the gehlenite appears weakly birefract- ing with comparativ ‘ely high refractive index. Maegnetite inclusions are abundant and often show crystal outline. The magnetite also occurs, fillmg cleavage and fracture cracks in the gehlenite and evidently was precipitated both before, dur- ing and after the crystallization of the gehlenite. In certain of the sections the magnetite crystals showed a distinet tend- ency to an arrangement. parallel with the first and second order prism faces. Round earthy spots also occur filled with earthy matter and are apparently of secondary origin, although they may possibly be weathered original spherulites of some mineral earlier than the gehlenite. In the thin section the basal cleavage is well marked and after it the crystals are often developed. in thick tabular form. In thick slides the interference color becomes intensely yellow, reminding one somewhat of the peculiar yellow interference tints of certain epidotes. In convergent polarized light the interference cross is wide and uniaxial. Optically negative. On one section the bire- fringence was measured at w—e = 00055. The refractive indices were measured directly on an Abbé Pulfrich total refractometer in Na light and found to be » = 1°666 + -003 and e = 1°661+:003. These values were obtained by using a polished plate of the granular material, and the refractive dices could not be determined under such conditions with an accuracy greater than + *008. The specitic gravity at 25° was determined on two samples by pycnometric methods at 38:°029 and 3:049 with an aver- age = 3°039. Part of this variability is probably the result of differences in relative amounts of inclusions. On uncovering a thin section and treating the exposed sur- face with weak hydrochloric acid and then, after thorough rinsing, with a solution of fuchsin, the eehlenite was found to have gelatinized slightly. This fact was corroborated by a chemical test with powdered material, which was found to gelatinize readily. The following chemical analysis does not agree with any simple formula and a comparison of other gehlenite analyses indicates that under the term gehlenite a solid solution series between several different end members is probably included. Compared with the other analyses, the Velardefa gehlenite is somewhat lower in silica and magnesia, and higher in alumina and lime, but otherwise very similar, and is essentially a calcium aluminum silicate. Wright—Three Contact Minerals. 547 CHEMICAL ANALYSIS. sf la 2 3 4 See re 26°33 "A359 DOTS 28°59 31°40 Ors eo 03 "0004 1.0 ee ree OS 2722 22°02 99°32 22°32 20 Se 1°43 0089 3°22 MeO “50 0070 1°82 0°3 0°03 Win Og oo. ‘01 0001 eee 0°50 0:96 eee Ss 9°44 "0605 3°88 ia 10°02 BOS. 89755 "7050 37°90 36°76 30°92 AO) ee. ol "0085 LPS 0°40 al ERO ‘10 0009 bodies 0°21 0°12 PieOe 8 2: 1°85 0103 1:28 eae dS a mon Oe 3°25 3°85 CO none eats le Jae age 100°27 1°5047 99°90 100°18 100°79 1. Gehlenite, Velardefia, Mexico. HK. T. Allen analyst. la. Molecular proportions of 1. 2. Gehlenite, Monzoni, Rammelsberg, Mineralchem., 1875, 604. 3. Gehlenite, Falkirk, Sweden. Edg. Jackson in Bauermann, Journ. Iron and Steel Inst., 1886, i, 88. 4. Gehlenite, Clarence, J. H. L. Vogt, Stud. Slagger, Stock- holm, 1884, 138. Before the blowpipe thin slivers of this mineral melt down with difficulty to dark, non-transparent beads, give a pronounced calcium flame reaction and glow intensely. In the hand specimens, gehlenite occurs either practically alone except for magnetite inclusions, or together with spur- rite, yellow garnet and calcite. Later veinlets consisting chietly of calcite were noted occasionally, cross cutting the specimens. So far as the writer has been able to ascertain from the avail- able literature, this occurrence is the first recorded for gehlen- ite on this continent. Spurrite.* This mineral, like gehlenite, occurs in granular masses which at first glance might be mistaken for crystallized marble, especially since the cleavage faces frequently glisten in the sunlight like those of calcite. No crystals were observed and the only goniometric measurements possible were made on cleavage fragments. Two cleavages were observed, the one good and the second much less perfect. The reflection signals from these faces were not of equal value and the cleavage *Type specimen from contact aureole of Terneras intrusion, Velardefia, Durango, Mexico. 548 Wright—Three Contact Minerals. angle could only be obtained approximately ; the best average of the results is 79° with a probable error of at least + 4°. Fracture uneven to splintery. Brittle. Hardness, about is Luster, vitreous to resinous. Color, pale gray with tints of blue or yellow to colorless. Transparent to translucent. Streak, white. On the hand Specimens a weathering or alteration crust, consisting chiefly of finely divided carbonate, occurs not infre- quently. Im the thin section the spurrite is well defined optically and is excellent material for optical work. From the relations of the optic properties to the erystallo- graphie it is highly probable that spurrite is monoclinic and that the cleavage faces are parallel with the orthodiagonal (} axis). If the good cleavage plane be called the basal pinacoid, the optical orientation is apparently the following: 6= 4; @:¢t, very small and possibly zero, the cleavage cracks not being sufficiently perfect for decisive measurements. : Twinning after both 001 and polysynthetic twinning after orthodomes at angles of 56° to 58° with the twinning lines of 001 occur, and occasionally divide the field into sextants of the same birefringence and all cut approximately normal to the acute bisectrix, the plane of the optic axes in the different sextants occupying different positions, as shown in the accom- panying sketches. (Figs. 1a, 10.) The polysynthetic lamellae are often extremely fine and resemble albite lamellae very closely. On a section almost precisely normal to the acute bisectrix the angle between the Wright—Three Contact Minerals. 549 plane of the optic axes and the fine twinning lamellae was measured at 57°°5 in sodium light. On this thick plate crossed dispersion was unusually clearly marked, the angle between the plane of the optic axes for red lithium light being about 57°-6, and for green thalinm light 57°-1. These measurements indi- cate a dispersion of the bisectrices ¢p:¢,: in the plane of sym- metry of about 0°15°. At the same time a slight dispersion of the optic axes was noticeable with 2Ep > 2Ky, The optic axial angle was measured on a number of different sections by the use of the double screw micrometer ocular and also of the universal stage, the average being 2V = 39°°5,+ 1°; whence 2E= 70°. Owing to the strong birefringence the interference figure is unusually well marked even in normal thin sections. The refractive dices were measured in sodium light on several different plates on an Abbe-Pulfrich total refractometer with reducing attachment, the different refractive index lines from the polished plates being clearly marked and easy to follow: eae —— 1-679) 2-002 y= © = WSY BNa = 1674 + -002 y — B = ‘005 a Na = 1°640 + :002 B—a = ‘034 From these values the calculated optical axial angle is 2V = 41° 12’, which agrees fairly well with the measured value. Optical character negative. The birefringence values were checked by direct measure- ments on plates in the thin section and closely accordant results obtained.* ; YG == a0 B — a = :036 In the thin section spurrite is recognized by its high bire- fringence, imperfect cleavage and small optic axial angle with negative optical character and in thick sections noticeable crossed dispersion. Still further evidence on the crystal system of spurrite was gathered from etch figures on the good cleavage face. Cleavage lakes were immersed for 10 seconds in cold 5 per cent hydrochloric acid and the etch figures of fig. 2 obtained. Many of these figures appear asymmetrical but the upper terminal endings are so variable and influenced by adjacent cleavage cracks to such an extent that the general symmetrical aspect of the figure with respect to a vertical plane of sym- metry may have thus been disturbed. It must be admitted, however, that the etch figures may be actually asymmetric, in * For these direct measurements of birefringence in the thin section the writer is indebted to Mr. E. S. Larsen, Jr., of the Geophysical Laboratory. 550 which case spurrite is triclinic instead of monoclinic; the rela- tions of the optic properties for different wave-lengths are then Fig. 2.—Etch figures on good cleavage face of spurrite produced by 10 sec. Wright—Three Contact Minerals. Fie. 2. immersion in cold 5 per cent HC1. 99°85 Magnification 150 diameters. la 2 ‘4467 27°13 0001. ‘0038 -0007* 0004 0058 email 62°98 ‘0008 +2919 9°89 1°7909 100-00 * Calculated as Fe.Os;. . Spurrite, Velardefia, Durango, Mexico, E. T. Allen ones ty Molecular proportions of 1. 2, Theoretical percentage weight composition for the formula 2Ca,SiO,.CaCO,. Wright— Three Contact Minerals. 551 such as to simulate very closely crossed dispersion of the mono- clinic system. Before the blowpipe spurrite shows strong calcium flame reaction, loses its glassy lustre, becomes white and porcelain- like but does not fuse even in thin splinters. Spurrite effervesces readily with weak hydrochloric acid, dissolves completely and gelatinizes thoroughly. The chemical analysis was made on carefully selected material. The agreement of the analyzed material with the formula 2Ca,SiO,.CaCO, is remarkably close and in view of the purity of the material analyzed can leave no doubt that spurrite is a compound of the above formula. The specific gravity at 25° was determined with pyeno- meter, both in xylene and in water, and the following results obtained : Spec. gr. at 25° in xylene = 3:013 3014 (74 ce 66 6¢ 7 ooo water = 3-016 Average spec. gr. at 25° = 3°014 Spurrite occurs in the hand specimens either in pure, unal- tered state, except for minute inclusions of magnetite, or together with yellow garnet, calcite and gehlenite. Its weather- ing products consist chiefly of carbonates in microscopic aggre- gates, which appear first along cracks and cleavage planes in the altering mineral. Through the courtesy of Mr. E. S$. Shepherd of the Geo- physical Laboratory, several experiments were made to repro- duce spurrite artificially by heating ten per cent solutions of sodium chloride with pure Ca,SiO, and CaCO, in finely divided state and in different proportions in silver-lined steel bombs from 6 to 9 days at temperatures of 350° to 400°. Although minute, well-shaped crystals were obtained in many of the preparations with refractive indices a and y, practically identi- cal with those of spurrite, the symmetry was orthorhombic and therefore not that of spurrite. Synthetic experiments on this compound are still in progress. This mineral is named in honor of Mr. J. E. Spurr of New York, who collected the original material and who has done much to further existing knowledge of ore deposits and their accompanying minerals. Hillebrandite.* Hillebrandite, unlike spurrite and gehlenite, is distinctly a fibrous mineral and occurs in aggregates often as radial spheru- * Type specimens from the 8th level of the Terneras Mine, Velardefia, Durango, Mexico. 552 Wright-— Three Contact Minerals. lites, the individual fibers of which are difficult to separate satisfactorily, and rarely measures *5™™ in length. In the hand specimen, especially when examined with a lens, these fibers tend to produce a faint silky luster on the otherwise vit- reous to porcelain-like mass. Cleavage, so far as could be observed, prismatic. Brittle. Hardness between 5 and 6, about 5°5. Color, pure porcelain white, often with faint tinge of pale green. Translucent in small chips. Streak, white. Under the microscope the optic properties are those of ageregates of fibers, often in approximately parallel orienta- tion, rather than of a single fiber. As a result the optical data are not easy to determine with great accuracy, although certain features of the mineral are so characteristic that its determination as such is a relatively simple matter. The refractive indices y and a were measured in sodium light on an Abbé-Pulfrich total refractometer on a polished plate of the mineral. It has been found by experience that even in the case of such fine-grained masses as hillebrandite, the phenomena in sodium light on the refractometer are sutfiici- ently distinct, when reducing attachment is used, to permit a fairly accurate determination of the two limiting curves y and a, although in the flood of hght from the different grains, the medium refractive index line does not appear with sufficient distinctness to allow of its determination. On such a plate the refractive indices were found to be I-612 ee 7008 1605 + °005 y a lI Il The birefringence is medium to weak, but difficult to deter- mine directly because of interweaving of overlapping fibers. The extinction is parallel, the ellipsoidal axis (c) being invariably parallel with the fiber direction which at the same time is the cleavage direction. The optic axial angie is not very large, 2Ep being possibly between 60° to 80°, while the dispersion of the optic axes is unusually strong and gives rise to peculiar, abnormal blue interference colors resembling those in certain epidotes and characteristic of hillebrandite. The optic character is negative with 2H, > 2K o. The plane of the optic axes was found to vary, being in the one plate parallel with the fiber direction and in the next perpendicular to the same, an abnormal phenomenon which may be due in whole or in part to the disturbing influence of the interlacing fibers which tend to veil the optic phenomena and often most effectively. On a section normal to the acute bisectrix the plane of the optic axes was parallel with the cleavage and direction of elongation. Wright—Three Contact Minerals. 553 From these optical and crystallographic data, it appears that hillebrandite is orthorhombic with possibly c=c, a@=a and cleavage after 110 (2). Its most characteristic optic features are: refractive index about 1°61, birefringence weak to medium, 2E medium with very strong axial dispersion 2E, > 2E , which in parallel ight gives rise to abnormal blue interference tints which are readily recognized. Optical character, negative. The absence of crystallographic faces of any degree of per- fection precluded any ewe at etching which might have been made. The specific gravity at 25° was determined by pycnometer with water at 2-692, and also in xylene at 2°692. The check determination in xylene was made because the analysis of hillebrandite shows it to have been slightly hydrolized. In hydrochloric acid (1:1) hillebrandite separates some silica at once but enters otherwise into solution. Hillebrandite decomposes very slowly with cold water as the test by adding a few drops of phenolphthaline to the mixture indicates. *~ Before the blowpipe thin splinters of hillebrandite fuse down with difficulty to.a colorless glassy bead, at the same time giving a pone calcium flame reaction and glowing briskly. CHEMICAL ANALYSIS. 1 1a 2 STC SSS nen et 32°59 "5398 O74 “Lo Le aia oe WOES a eine "02 "00038 WE ee 23 0023 ae SO RE: 15 0009* pte) Fata Sil oc ce ‘Ol 0001 Mie een yen ‘O4 0010 pee a ate pe ip G 1°0296 58°81 a OS ee fae AS 03 0005 HRY ig gi DI once 05 0005 EM eer ee ee _ 9°36 5019 9°45 Pipa ie Meee th Ui none 9G Dee RSS a ee eee ee none 100°24 2°0769 100°00 1. Hillebrandite, Terneras Mine, Velardefia, Durango, Mexico, K. T. Allen analyst. 1a. Molecular proportions of 1. 2, Percentage weight composition of formula H,O.2 CaO, 8i0,,. * Calculated as Fe20s. Am. Jour. Sc1.—Fourts Series, Vou. XX VI, No. 156.—DrEcremBER, 1908. 39 554 Wright—Three Contact Minerals. This analysis agrees closely with the formula Ca,SiO,.H,O,* it being a little higher in silica and lower in lime, a condition which is evidently due to a slight leaching of the lime. The water of the analysis was determined by loss in weight, the figure given being the average of two determinations, 9°34 and 9°39. By absorption by calcium chloride plus the little obtained at 110° the result was 9:18. The first figures, how- ever, are more accurate. Hillebrandite occurs usually with few inclusions and even magnetite is rare. Occasionally small grains of carbonate, yellow garnet and wollastonite occur with it and also earthy material of a secondary nature. Veinlets of wollastonite traversed several of the hand specimens and in each case the direction of elongation of the wollastonite fibers was normal to the vein walls. Experiments to produce hillebrandite synthetically have thus far not proved successful. The above optical and chemical data show beyond question that hillebrandite is a true chemical compound of unique chemical composition. It is with great pleasure, therefore, that the writer suggests the above name as a token of appre- ciation of the fundamental researches of Dr. W. F. Hillebrand of the U. 8. Geological Survey in mineralogical chemistry. Geophysical Laboratory. Carnegie Institution of Washington, Washington, D. C. June, 1908. * This formula may also be written CaSiOs. Ca(OH). or simply H.O. 2CaO.SiOz. Drushel—Estimation of Potassium in Animal Fluids. 555 Art. LIL—Zhe Volumetric Estemation of Potassium in Animal Fluids; by W. A. Drusaet. [Contributions from the Kent Chemical Laboratory of Yale Univ.—excii. ] Tue distribution of potassium in plant and animal tissues has been studied by Macallum* and others. Macallum pre- cipitated potassium in place as potassium sodium cobalti-nitrite, an insoluble potassium salt which by its crystalline form and color is easily recognizable under the microscope. To study the function of potassium in the animal organism it is desirable to have a simple and rapid method for its estimation in the various tissues and fluids. A number of quantitative methods have been proposed which, however, are not wholly free from objections. M. Kretschy, m 1876, after having carefully studied the several indirect methods for potassium and sodium in the presence of each other, finally adopted a modification of the chlorplatinate method for potassium in the presence of rela- tively large amounts of sodium in physiological work. He worked with quantities of potassium ranging from 3 mgrm. to 120 mgrm., precipitating it as the chlorplatinate in the usual manner. The washed and dried precipitate was carefully ignited, the residue extracted with water and the extract evapor- ated to dryness. This residue was gently ignited and weighed as potassium chloride. A small amount of platinum usually passed through the filter, giving a result which was too high for potassium chloride. To avoid error on this account the weighed potassium chloride was dissolved in water, any residue of platinum filtered off on ashless paper, ignited and weighed, and the necessary correction made in the weight of the potas- sium chloride. Some years Jater Lehmann,t Bunge,{ Heintz,§ and Pribram and Gregor| used different modifications of the Fresenius chlor- platinate method for the estimation of potassium in urine. In the methods of Lehmann, Bunge, and Pribram and Gregor the sulphate radical was removed by an excess of barium hydrox- ide or barium chloride ;. subsequently the excess of barium was removed by ammonium carbonate and ammonium hydroxide, or in case of barium hydroxide by carbon dioxide. Lehmann evaporated the urine with ammonium sulphate before ashing the residue, while Bunge treated the urine directly with bar- inm hydroxide. Pribram and Gregor oxidized the organic matter by heating the urine, acidified with sulphuric acid, to * Jour. of Physiol., xxxii, 95(London). +Zeitschr. physiol. Chem., viii, 508. t Zeitschr. Biologie, ix, 139. § Pogg. Ann., lxvi, 133 || Zeitschr. anal. Chem., xxxviii, 409. 556 Drushel the boiling point and adding an excess of potassium free barium permanganate. Heintz treated 20°™* to 30™* of clear urine with chlorplatinic acid and a threefold volume of a 1:4 ether and absolute alcohol mixture. After standing 24 hours the precipitate was filtered off, washed with alcohol, dried, ignited and weighed. The residue was then extracted with hot water and again dried and weighed. The amount of potassium chloride was found by taking the difference of the weights. It has been repeatedly shown that appreciable amounts of the alkal salts are carried down by barium sulphate, which can not be completely removed by washing. This objection applies: to all of the methods in which the sulphate radical is removed by means. of a barium salt. The loss of alkalis is especially appreciable where a large amount of barium sulphate is formed in the presence of relatively small amounts of the alkali salts. In 1898 K. Gilbert devised a method for.the separation of potassium which does not require the previous removal of the sulphate radical. He treated a potassium salt solution free from mineral acids with an excess of sodium cobalti-nitrite acidi- fied with acetic acid. After standing from 12 to 20 hours the potassium was quantitatively separated out as potassium sodi- um cobaltinitrite which could be freely washed with cold water without an appreciable loss. Gilbert decomposed this precipitate by heating with dilute hydrochloric acid and esti- mated the potassium as the perchlorate or chlorplatinate. A few years later Autenrieth and Bernheim* used Gilbert’s method for separating potassium in urine, subsequently esti- mating the potassium as the perchlorate. They used 6° to 10°™* of concentrated sodium cobalti-nitrite to precipitate the potassium in 50%* of urine. At this dilution it had been shown by Gilbert that the potassium is quantitatively precipi- tated and that the precipitate is apparently of indefinite com- position. In 1900, however, Adie and Wood* found that with a sufficiently high concentration of the reagent and of the potassium salt solution a precipitate of definite composition, represented by the formula K,NaCo(NO,),.H,O, is obtained. They further found that by decomposing this precipitate with boiling dilute sodium hydroxide and titrating the nitrites with standard potassium permanganate the potassium may be esti- mated with a fair degree of accuracy. In a previous paper* from this laboratory it was shown that it is unnecessary, after adding the cobalti-nitrite reagent, to let the mixture stand from 12 to 20 hours, if it is evaporated nearly to dryness, also that the precipitate may be directly oxi- dized with potassium permanganate without previously decom- * Zeitschr. physiol. Chem., xxxvii, 39. + Jour. Chem. Soc., Ixxvii, 1076. t This Journal, xxiv, 453, 1907. Drushel— Estimation of Potassium in Animai Fluids. 557 posing it with boiling sodium hydroxide. Later it was found that a half saturated sodium chloride solution is preferable to cold water for washing the precipitate since it permits the use of a coarser asbestos felt in filtering without danger of loss. The method used in the work on animal fluids is as follows: A potassium salt solution was obtained free from mineral acids and ammonia salts and treated with a liberal excess of concen- trated sodium cobalti-nitrite in an evaporating dish. The mix- ture was evaporated to a pasty condition over the steam bath. After cooling the residue it was stirred up with enough cold water to dissolve the excess of sodium cobaltinitrite. The precipitate was permitted to settle a few minutes, then it was filtered on asbestos in a perforated crucible and washed with the sodium chloride solution until the filtrate came through colorless. The precipitate and felt were transferred by means of a spray of water and a stirring rod to a beaker containing a measured amount (being an excess) of standard decinormal or fifth normal potassium permanganate, diluted about ten times and heated nearly to boiling. The permanganate solution was kept hot and stirred to facilitate the solution and oxidation of the precipitate, the oxidation being completed by adding 5°™* to 10°™* of dilute sulphuric acid and stirring for a minute or two. The excess of permanganate was then bleached by a measured amount of standard decinormal oxalic acid, and the solution titrated to color with standard permanganate. In this process the cobalt is reduced to the bivalent condition and the nitrites oxidized to nitrates, from which by a simple calculation it is found that one cubic centimeter of strictly decinormal perman- ganate is equivalent to 0°000857 grm. of K,O. The modified Lindo-Gladding method was used as a control in the experimental work of this paper. The potassium was obtained as the sulphate in the presence of sodium sulphate, and possibly traces of calcium and magnesium sulphate. The’ solution of these sulphates was treated with an excess of chlor- platinic acid, evaporated nearly to dryness, and the precipitate washed free from the excess of chlorplatinic acid with 85 per cent alcohol. The precipitate was then washed three or four times with a 20 per cent solution of ammonium chloride satur- ated with potassium chlorplatinate, and finally again two or three times with 85 per cent alcohol. By this treatment the sodium sulphate and the small amounts of calcium sulphate and magnesium sulphate are completely removed. A. Potassium in urine. The following table gives approximately the amount of the constituents present in a day’s excretion of urine of an adult in normal health and on an ordinary mixed diet: 558 Drushel Estemation of Potassium in Animal Fluids. 1h hiteroae Ge aye at Us 2 to 4 grm. Urea vist Oh ae grm. NaCl et eos ie tO a Urie acid 78 0:7 CaO and MgO ..-..-- Or84,,4° Creatinine _...- Ib ee INGER ee 2 ee perenne ORT Hippuric acid 0°71 ifs a (phosphates) .-. 1:5 “ Other organic -- H,SO, (sulphates) (222.275 |S bodies -..2/)- 32a In this list of constituents ammonia and the organic bodies, especially urea, are the only ones which should interfere with the volumetric method as previously described. To remove these bodies without the loss of potassium is apparently the only new problem. In the experiments recorded in Table 1 aliquot portions of urine of 10 to 50 cubic centimeters each were measured with pipettes or a burette into small platinum evaporating dishes, and evaporated to dryness over the steam bath in a good draught hood. ‘The residues of the aliquots of the first speci- men were treated with 5°™° of concentrated nitric acid and again evaporated to dryness. ‘These residues were then moist- ened with concentrated sulphuric acid and ignited to a white ash beginning with a low flame and increasing the heat until the organic matter was burned off, and the ammonium sulphate and excess of sulphuric acid completely removed. In subse- quent experiments it was found more expeditious to treat the dried urine residue with 5° to 10°™ of a 9:1 nitric-sulphuric acid mixture, in a covered evaporating dish, removing the cover when the first violent oxidation is over, evaporating to dryness and igniting without the further addition of sulphuric acid. By this treatment the ignition of the residue from 50°™ of urine could be readily made in 30 minutes without loss of material. The residue thus prepared was treated with a little water and a few drops of acetic acid to dissolve the alkalis. Without filtering about 10™> of concentrated sodium cobalti- ‘nitrite were added and the mixture evaporated to a pasty condition. From this point the process was carried out as previously described. In the control experiments the phos- phoric acid was removed by a slight excess of calcium hydroxide, and the calcium by ammonium oxalate, before the ignition of the residue. The results obtained by the two methods from a number of different specimens of human urine are given in the following table. B. Potassium in circulatory fluids. An additional difficulty presents itself here in the presence of a large amount of protein material which can not be removed by coagulation and filtration without a considerable loss of * Taken from Hammerstein’s Physiological Chemistry, (Mandel’s Trans- lation), 5th ed.; p. 628. Drushel—Estimation of Potassium in Animal Fluids. 559 T Urine taken Sp. gr. cm? PC 1) 10 1018 (2) 10 oe (en ei ae (4) 10 at EE CL) 10 1°022 (2) 10 Lee (3) 10 eae (4) 10 eure ill (1) 10 1°023 (2) 10 Biss (3) 10 ceaee rs (4) 10 Cee NCL) 10 1°024 (2) 10 ee (3) 10 eee (4) 10 eee Urine taken Sp. gr cm? V (1) 10 1°025 (oy 110 a (3) 10 apa (4) 10 rape VI (1) 25 1°025 (2) 25 ass ys (3) 25 FEN VIL (1) 20 ~—*1°025 (2) 20 a (3) 20 ais ee VIII (1) 20 ~—«*i1:-024 (2) 20 nite a (3) 20 Sones (4) 20 ee EX.(Y) 25 1°018 (2) 50 aE (3) 50 a Ye (4) 25 ee ABLE I, A K found grm. 0°012 0°0122 0°0123 0°0131 0°0180 0°0178 0°0178 00171 0°0241 0°0242 0°0238 0°0231 0°0241 0°0248 0°0242 0°0243 K found germ. 0°0293 0°0292 0°0293 —0°0292 0°0740 0°0747 0°0740 0°0757 0°0752 0:0764 00663 0°0662 0°0663 0°0662 K found per cent. 0°12 0°12 0°12 0°13 Oris 0-17 Ort 7 0°17 0°24 0°24 0°23 0°23 0°24 0°24 0°24 0°24 0°0425 0°0839 0°08438 0°0424 Method Vol. 66 Grav. 6 K in 24 hrs. erm. 218 Pe 2°78 20k Method Vol. Grav. (44 Vol. Grav. Vol. Vol. Grav. Vol. Vol. 6¢ Grav. 66 Vol. 52 Grav. 53 Vol. Grav. 560 Drushel—Kstumation of Potassium im Animal F luids. potassium. This is particularly true of the blood, where most of the potassium is intimately associated with the protein of the corpuscles. It is necessary therefore to decompose the protein material by oxidation. The nitric-sulphuric acid mixture was s first used for oxidizing the dried blood residue, but was found to work less satisfae- torily than in the case of urine. In the ignition of the blood residue oxidized in this way there was apparently a greater tendency to spatter, probably due to the presence of the sul- phuric acid. When, however, nitric acid alone was used for the oxidation, there was a tendency for the residue to burn off explosively on ignition. The analytical results from the first specimen given in Table II were obtained by treating weighed portions of defibrinated blood with about 2°° of bromine in covered evaporating dishes, allowing them to remain in a warm place under the hood for about one hour. The excess of bromine was then removed over the steam bath, the residue evaporated to dryness, and ignited sufficiently to char the organic matter. The residue was thoroughly extracted with hot water and the extract evaporated off with a few drops of: sulphuric acid. The residue was then ignited to remove any ammonium salt which might have escaped the action of the bromine and any organic matter which might have passed through the filter. The second specimen was a half liter of clotted sheep’s blood, from which no homogeneous portions could be taken. The whole mass was, therefore, evaporated over a steam bath and oxidized with concentrated nitric acid, getting everything into solution except a little lipoid material. The solution was made up to the original volume, and aliquots of 25°", repre- senting 30 grm. of blood, were ‘pipetted off. These portions were evaporated to dryness and gently ignited, but not suffi- ciently to produce explosive decomposition. The residues were then moistened with concentrated sulphuric acid and _ ignited carefully to remove the organic matter and the ammon- ium salts. For the results of the second and third divisions of Table II weighed portions of serum and lymph were similarly oxidized with nitric acid and subsequently ignited with a little sulphuric acid. The potassium in all the residues thus obtained was estimated gravimetrically or volumetrically as previously described. The results obtained for potassium in circulating fluids are given in the following table. C. Potassium in milk. In addition to lactose and the inorganic salts, milk contains a large amount of protein, chiefly casein, and varying amounts of fat. The ease with which casein is precipitated suggested the possibility of making a complete separation of the inorganic salts and casein, but it was found that after thoroughly wash- Drushel—Estimation of Potassium in Animal Fluids. 561 Taste II. Amount K,0 found Nature of taken a ae fluid erm. erm. per cent Method ay Defibrinated 10°89 0°0227 0-21 Gray. (2) pig’s blood - 11°21 0°0228 0°20 Vol. (3) 2 ere 20°33 00391 0°19 a: “reels 10°16 00208 0°21 “ 5) 5 ees Saga 10°85 00211 0°20 ss "5 en an 11:03 0:0936 0-21 “ Il (1) Sheep’s blood 30-00 0°0174 0°058 Grav. “Se ee 30:00 00179 0:060 Vol. Seberete ee 230-00) < 00181. 0-060. ‘Grav. a en 30°00 0°0181 0°060 Vol. 5) ee SR aes eee 30°00 0°0180 0-060 oo III (1) Serum of dog’s blood _. 1011 0:0024 0°024 Vol. (2) a ip ee eee 10°04 0°0024 0°024 Gravy. 2). See 10:07, )°:0-0023- ¢ 0-084" Vol. IV (1) Dog’s lymph 10°28 0:0018 0.018 Grav. aoe so 10°01 00019 0-019 Vol. ee 16700 3 00020 0020 “ (4) ee Se 10°03 | 0°0019 0°019 c peers Cir? 10°12 00019 0°019 ee (6) Pea 28) Sk 10°32 0°0022 0°021 Grav. ing the precipitated casein it still contained appreciable amounts of potassium. It was found preferable to evaporate weighed portions of milk to dryness, oxidize with concentrated nitric acid, again evaporate to dryness, and ignite gently until most of the or ganic matter was burnt off, finishing the ignition after moistening the residue with concentrated sulphuric acid. In the residues thus obtained the potassium was estimated gravi- metrically or volumetrically as in the previous work. The results obtained for potassium in two specimens of cow’s milk are given in Table III. TasieE III. Milk taken K.O found = aes — grm. germ. per cent Method I (1) Derg 2 as & 25°8 0°0413 0°16 Vol. (2) agnae Peas 25°8 0°0432 Ook? = (3) 28 yey Sere 25°8 0°0428 0°17 Grav. 05 eee 51°6 0°0833 0°16 Vol. 1 3 ah Of (eee comes 25°% 0°0454 0°18 Be (2) a sae ENS ree 25-7 0°0457 0°18 ee (3) Fok, 3 ae pS ay 0°0451 0°18 Gravy. Summary. Where protein does not occur m animal fluids it was found most advantageous to oxidize the dried residue with a 9:1 562 Drushel—Kstimation of Potassium in Animal Fluids. nitric-sulphurie acid mixture. In the presence of protein oxidation by bromine, or by nitric acid alone, finishing the ignition in the latter case with a little concentrated sulphuric acid was found more satisfactory. For the volumetric estimation the ignited residue was treated with a few drops of acetic acid and a little water. To this solution an excess of sodium cobalti-nitrite was added and the mixture evaporated nearly to dryness. The residue was cooled, treated with cold water, filtered on asbestos and well washed with a half saturated sodium chloride solution. The precipitate was oxidized by an excess of hot standard potas- slum permanganate, the solution bleached by an excess of standard oxalic acid, and titrated to color with permanganate. For the gravimetric controls the calcium, phosphoric acid and iron (in the case of the blood) were removed either before. or after the oxidation and ignition. The residue of sulphates was dissolved.in a little hydrochloric acid and water, and the potassium estimated as the chlorplatinate, taking care to wash the precipitate with alcohol, Gladding’s reagent and again with alcohol. Conclusions. The necessity of allowing the cobalti-nitrite mixture to stand from 12 to 20 hours to complete the precipitation, suggested by Gilbert, Adie and Wood, and others, is avoided by evapo- rating the mixture nearly to dryness. The removai of the cobalt from the precipitate before oxidation with permanga- nate is unnecessary, since the cobalt is reduced and not reoxi- dized in the titration process, and since with proper dilution its color does not interfere with the end point. For small amounts of potassium fairly accurate results are obtained by using the permanganate factor calculated from Adie and Wood’s formula for potassium cobalti-nitrite. Sutton has suggested that more accurate results are secured by obtain- ing a factor empirically from a pure potassium salt. All the results from this and previous papers were obtained, however, by using the theoretical factor calculated from the formula of Adie and Wood, their analyses of potassium sodium cobalti- nitrite having been verified by the analysis of a carefully prepared salt. The chief sources of error in the method appear to be in the slight solubility of the potassium sodium cobalti-nitrite, being one part in 25,000 to 30,000 parts of water at room temperature, and its tendency to inelude traces of sodium cobalti-nitrite. These sources of error tend in opposite directions, resulting usually in a positive error, which by proper washing of the precipitate may be kept within fair limits. The method requires less time and labor than the chlorplat- inate method, and is applicable in the presence of substances which form no insoluble cobaltinitrites and which neither oxidize oxalic acid nor reduce potassium permanganate. Wm. F. Prouty—Meso-Silurian Deposits of Maryland. 563 Art. LII.—TZhe Meso-Silurian oe of Me by Wu. F. Proury.* TE present article discusses briefly the lithological and faunal characteristics of the deposits in the state of Maryland lying between the massive Tuscarora (White Medina) sand- stone below and the Salina formation above. These deposits are approximately eight hundred feet in thickness and repre- sent the lower and middle portion of the Meso-Silurian series as shown in western New York. The lower portion corre- sponds to the pre-Rochester Silurian, the Clinton of common usage, while the upper portion is the approximate equivalent of the Rochester, or the lower division of the Niagara group of common usage. It is moreover probable that these deposits are equivalent to the typical Clinton of Hall exposed at Clinton, Oneida County, N. Y. The latter, according to both E. O. Ulricht of the U. 8. Geological Survey and to C. A. Hartnagelt of the New York Survey, contains in its upper portion a fauna equiv- alent to the Rochester or Lower Niagara fauna of western New York. It is very unfortunate that the fauna of the typ- ical Clinton should have gone so long unstudied, thus allowing at the present a double meaning of the term Clinton, as seen below. The Meso-Silurian deposits of western New York and the interior may be subdivided as follows : ( 3. Guelph ( B. Niagara 2. Lockport _ (1. Rochester Meso-Silurian ¢ | [ A. Clinton while the Meso-Silurian deposits of eastern New York, Penn- sylvania, Maryland, etc., may be subdivided as follows: ( 2. Rochester of com- 2 let Soe | ep ite | mon usage Meso-Silurian } A. Typical Sianon ‘Ce oe [ usage I shall not enter into a discussion of the nomenclature, but shall use the term Clinton in the restricted sense to include only the portion below the Rochester. In the study of the Maryland Meso-Silurian deposits all the important exposures in the state have been visited and at each * Published by permission of the Director of the Maryland Geological Survey. + From letters of recent date. 564 Wm. FL. Prouty—Meso-Silurian Deposits of Maryland. place careful measured sections and fossil collections have been.made. With the exception of the Ostracods and Bryo- zoa, the fossil study has been practically completed. In the pursuance of his work the author has had access to the collec- tions of the National Museum, the New York Museum of Natural History, the New York State Collection at Albany, and several smaller and private collections. He is also greatly indebted to Prof. Schuchert of Yale University, Dr. E. O. Ulrich of the U. 8. Geological Survey, Dr. C. K. Swartz of Johns Hopkins University, and others for assistance. He is further indebted to the Director of the Maryland Geological Survey for the privilege of publishing this article. Before turning to the present discussion of the Meso-Silurian in Maryland, let us make a brief historical review of what has already been done in this area and what have been and are ' the views held concerning it. [fistorical Review.—Nearly fifty years ago Philip Tyson, the state agricultural chemist of Maryland, published* the first report of importance dealing with the geology of western Maryland. He touched but briefly upon the formations under. consideration, using terms, “the Clinton” for the pre-Rochester Meso-Silurian’ and ‘‘the Onondaga” for the remainder of the Meso-Silurian and the Cayngan. In 1874 Prof. James Hall of New York, who had worked in Maryland, especially in the Lower Helderberg, did not recognize the presence of Niagara. It appears that he later did, as he cites Sparifer crispus and Homeospira apriniformis from the “ Niagara” of Maryland. In 1893 a preliminary geological map + of the state showed the whole Meso-Silurian deposits under the name ‘“‘ Rockwood.”’ James D. Danat in 1895 recognized in Pennsylvania, immedi- ately north of Maryland, a commingling of Clinton and Niag- ara fossils in the upper Clinton beds, and states that there frequently occur some distance above the top Clinton iron-ore, a succession of thin limestones which in many places contain Niagara fossils. | The author of the Piedmont Folio§ discussed a part of the Maryland area, using the terms Rockwood-(Clinton or pre- Rochester Meso-Silurian) and Lewiston (later Meso-Silurian, Cayugan and Lower Helderberg).. There is here a preliminary discussion of the two formations as Rockwood and the lower part of the Lewiston. The first detailed lithological study, however, of the Meso- * Geological Map and Report, 1861. + Map to accompany ‘‘ Maryland: Its Resources, Industries, and Insti- tutions.” G. H. Williams and W. B. Clark. + Manual, 4th edition, 1895. § Geol. Survey Atlas, Folio No. 28, 1896. Wm. F. Prouty—Meso-Silurian Deposits of Maryland. 565 Silurian of Maryland was made in 1900 by C. C. O’Harra.* He used the terms Clinton and Niagara, correlating the forma- tions in Maryland with those in western New York. It has long been known by all geologists working in the Appalachian districts that there is a marked stratigraphic change as one passes westward over the protaxis of these moun- tains. For the explanation of this fact the theory of a pro- found fault was suggested by some but accepted by few. Sir William Logan of the Canadian Survey, in 1866, suggested the possible existence of a narrow basin of deposition west of the protaxis to account for the stratigraphic break. It has also long been recognized that there exists a fau- nal difference in some of the deposits west of the protaxis in the Appalachians and in those west of the Allegany Mountains in New York and the interior. The explanation of this fact was little dwelt upon until recently, when E. O. Ulrich and Charles Schuchert, after careful investigation and faunal study, put forward a most admirably devel- oped theory} to account for the apparent condition. They conclude from paleontological evidence that there existed a barrier of great length, which separated from the interior sea a long and narrow body of water lymg in the Appalachian region, the “Cumberland basin,’ in which sediments were deposited bearing fauna more closely related to the east Canada and European deposits than to those of the interior sea; that the Atlantic crossed over the barrier forming the eastern side of this basin in about Beekmantown time, but was restrained by the western barrier of this basin from mingling with the ‘interior sea, possibly from Clinton but certainly from Niagara time well into Oriskany time. The barrier forming the west- ern side of the Cumberland Basin extended, according to Ulrich and Schuchert, from the region of Cayuga Lake, N. Y., southward to west of Altoona, Pa., thence parallel to the trend of the Appalachian Mountains through central West Virginia into eastern Tennessee. In discussing the fauna of the Cum- berland Basin, Ulrich and Schuchert state that few species are identical with those of the interior sea, and that the earlier fauna recalls the Clinton, while it passes above into one which may be compared to the Niagara. Charles Schuchert, in a later publication,t speaking of the Maryland deposits which include the Clinton and Niagara, says that save Atrypa reticularis and Leptaena rhomboidalis all the species appear to be new. Further, he directs attention to the absence of such characteristic forms of the western or * The Geology of Allegany Co., Md.; Md. Geol. Survey Rept., pp. 57-164, 1900. + Paleozoic Seas and Barriers in Eastern North America. N. Y. State Mus. Bul. No. 52, 1902. ¢ Lower Devonic and Ontario Formations of Maryland, 1903. 566 Wm. F. Pro uty— Meso-Silurian Deposits of Maryland. Mississippian sea as Spirfer radiatus, Spirifer Niagarensis, Spirifer Crispus, Spirifer sulcatus, Pentamerus oblongus, Caryocrinus, and Hucalyptocrinus. Of these, Sperifer radi- atus and Spirifer crispus are now known to occur in the Cum- Sa NS SF Kies de Dotted areas are Meso-Silurian. Scale 1 in.=16 miles. Ma ERLAND Kay) CUMB Ww ty mM mf yt MY 444 nine. As would be expected from the berland Basin. - Mr. C. A. Hartna- gel of New York, in an article* in 1902, shows, however, that the western or Held- erbergian barrier was not continuously and completely effective from Clinton to Oris- kany time, since the mingling of Guelph and Cobleskill fauna evidence the crossing of the barrier during Cobleskill time. | Up to the present less than half a dozen species of fossils have been definitely de- scribed from the Mary- land, Clinton, and later Meso-Silurian. General Descrip- tion.— The outcrop of the Meso-Silurian rock in Maryland is limited entirely to the region of the Alle- ghany Ridges proper, in Washington and Allegany Counties. The area covered is about twenty-four square miles, of which the Clinton occupies fifteen and the Later Meso-Silurian about general Appalachian structure, the rocks under discussion flank anticlinal folds which * Preliminary Observations on the Cobleskill Rept. N. Y. State Paleont. for 1902, p. 1156. Limestone of New York. Wm. F. Prouty—Meso-Silurian Deposits of Maryland. 567 have a general trend N. 20°-30° E. There are six such folds exposing these strata in Maryland: three in Allegany County and three in Washington County (see map, fig. 1). The axes of the anticlines are formed by the very resistant “ Tuscarora”’ or ‘‘white Medina” sandstone. Nearly all the folds are un- symmetrical, giving a much narrower outcrop on the west than on the east side. GENERAL CHARACTER OF THE DEPOSITS. Clinton Formation. The rocks of this age in Maryland consist essentially of red- dish and olive to grayish and brown argillaceous shales which are slightly lighter in color and less fossiliferous toward the bottom. The exposed surfaces of this shale often show a deep searlet color. Thin sandstone bands occur at irregular inter- vals throughout nearly the whole formation and become more numerous toward the bottom, giving the formation the appear- ance of grading into the Tuscarora quartzite. These thin sand- stone bands were in general originally more calcareous than at present and are uniformly more fossiliferous than the shale in which they oceur. Toward the top of the formation limestone bands become numerous and replace the sandstone layers. They seldom exceed six inches in thickness, and in some localities are very fossilferous. Immediately overlymg the limestone- bearing shales throughout the region there occurs a quartzitic sandstone, of variable thickness, which in character resem- bles very closely the Tuscarora. This sandstone thickens markedly toward the east, increasing from ten feet in thick- ness near Cumberland to nearly seventy feet in some eastern exposures. In the top portion of this sandstone is found the so-called top Clinton iron-ore, usually not more than a foot in thickness and commonly of too lowa grade to work. In places, however, it is thicker and has been enriched so that in the past it has served as an ore and might at present locally be so called were it more accessible to the railroad. In this latter respect the top Clinton ore differs markedly from the so-called bottom Clinton ore which occurs throughout the Maryland area from one hundred and twenty to one hundred and sixty feet from the bottom of the formation and which, though it sometimes attains forty feet. in thickness, is not sufficiently high in iron to be valuable as an ore. This lower zvon sand- stone occurs in two beds separated by from six inches to six feet of olive shale. Both the bottom and top “ore” bodies contain numerous though very poorly preserved fossils. The olive shales and thin crystalline limestone bands imme- 568 Wm. F. Prouty—Meso-Silurian Deposits of Maryland. diately above the top Clinton iron-ore sandstone contain a greater percentage of Rochester fossils than of those of the pre-Rochester Meso-Silurian, and consequently it would appear that the lower limit of the Rochester should be drawn at the top of the heavy sandstone laver. The thickness of the Clin- ton does not vary far from five hundred and fifty feet. The Later Meso-Silurian (fochester). The rocks of the Later Meso-Silurian formation, as herein discussed, include those lying between the heavy band of sand- stone bearing the top Clinton iron-ore and the bed of disinte- grated yellow rock full of Leperditia which marks the base of the overlying Cayugan, and are composed for the most part of thin-bedded limestones with shale partings. For a short distance above the bottom and for a greater thickness near the top the shales increase and preponderate over the limestone. The lowest limestone layers are grayish blue in color and, together with the shales which immediately overlie the Clinton — sandstone layer, are very fossiliferous. The upper limestones are uniformly of a darker color and more compact, sometimes occurring in more or less lenticular beds with thin shale part- ings. In general the middle limestones and shales contain few fossils, with the:exception of Ostracods and a few Favosites and Orthoceras. Toward the top, however, the brachiopod life begins again to abound and some beds are very fossilif- erous. The upper shales are usually darker it color than those below and often become arenaceous, bearing thin sandstone lenses, while the top of the formation is often formed by a bed of sandstone of variable thickness which is often very ferru- ginous. These upper dark shales usually carry a great number of Ostracods and poorly preserved Bryozoa. Going toward the east, one generally finds that both the lower shales and the upper ferruginous sandstone layers inerease in thickness. The thickness of the formation is in general not far from three hundred feet. At Pinto, where the most accurate measurements were made, though they are not entirely satisfactory because of some faulting, these rocks show a thickness of two hundred and eighty-eight feet. General Lithological Relations. It is concluded from.a comparative study of the different exposures in the state that during the deposition of these sediments there existed a shore line not far to the east, and a gradual deepening of the waters toward the west as far as the section exposed at Pinto (see section, fig. 2and map, fig.1). The Wm. F. Prouty—Meso-Silurian Deposits of Maryland. 569 main facts supporting this conclusion are: the top Clinton sandstone increases gradually in thickness from seven feet at Pinto to some sixty-five or seventy feet in the eastern exposures ; the bottom Rochester shales increase in thickness and the bottom limestones consequently decrease in thickness in passing from west to east; the top Rochester sandstone, like the top Clinton sandstone, thickens from eight feet at Pinto to some fifty feet or more in the eastern exposures. If we assume then Hig: i 2 3 4 5 6 7 8 Rochester | 2 A 5 1. Pinto Section. 4. Cumberland Section. 7. Great Cacapon Section. 2. Cedar Cliff ‘‘ 5. Six-mile-house ‘‘ 8. Keefers Mt. ee 3. Rose Hill *‘ 6. Flintstone le that the Maryland Meso-Silurian deposits were laid down in a narrow basin of deposition, then Pinto must be east of the central axis of the basin or else the western shore line of this narrow sea must have been much lower than the eastern land was. Itis not improbable that both these conditions were true. Fossil Zones. In Clinton.— The Clinton formation, so far as studied, has not shown any well-marked continuous fossil zones. Portions of the strata are much more fossiliferous, however, than others, and the formation as a whole is naturally divisible into four parts: Am. Jour. Sc1.—FourtH Series, Vot. XXVI, No. 156.—DrEcrempBer, 1908. 40 570 FIG. 3; | | i I | | HI | HAN Hy f) | | i Hl ; iF | HATE il itil | | | | Ei HE | | SUNTAN nit A | | | | | | i I ij lI \\-| | | | ss i i | Tusearora Camarotoechia obtusiplicata Zone Camarotoechia sf.nov. Lone Nucleospira ae Zone Dalmanites limulurus Zone Upher Fossiligerous Division Middle Barren Division LowerFossiliterous Division Lower Barren Division Wn. F. Prouty—Meso-Silurian Deposits of Maryland. (1) a lower relatively barren division includ- ing the shales and sandstones up to with- in some thirty feet of the lower “iron ore,” which bears locally at its base a few plant remains, the ostracod Berichia lata and the brachiopod Anoplo- theca hemispherica ; (2) a lower fossilifer- ous division, ncluding the remaining shales below the bottom ‘iron ore,’ the ore bed and some fifty to eighty feet of the over- lying shales and sand- stones; (3) a middle barren division, in- cluding some two hun- dred feet of relatively barren shales and sand- stones; and (4) an upper fossiliferous division, including about one hundred feet of shales and in- terbedded sandstones, replaced by thin-bed- ded limestone toward the top, about which OCCUTrS tite. wre shightly fossiliferous top Clinton sandstone (see sect. fig. 3). Of the two fossil- iferous divisions, the lower is characterized by a large number of tentaculites and ostra- cods and two species of trilobites. The upper fossiliferous Wm. F. Prouty—Meso-Silurian Deposits of Maryland. 571 division, on the other hand, is characterized by its branchi- opod fauna. The forms noted in the two fossiliferous divisions are as follows: From the lower fossiliferous division: Buthotrephis gracilis var. intermedia Hall * Anoplotheca hemispherica (Sowerby) Stropheodonta corrugata (Conrad) + Tentaculites sp. Clidophorus sp. * Calymene blumenbachit * Calymene sp. nov. * Ostracods From upper fossiliferous division : t Atrypa reticularis (Linnaeus) * Leptaena rhomboidalis (Wilckens) Camarotoechia neglecta Hall * Chonetes cornutus (Hall) * Chonetes novascoticus (Hall) * Chonetes tenuistriatus Hall Spirifer radiatus (Sowerby) Bucanella trilobata (Conrad) Calymene clintoni (Vanuxem) In Upper Meso-Silurian.— While the Clinton fossils seem to have a rather extended range vertically and a more or less local distribution, making a close zonal study of this formation seemingly impracticable, the overlymg Meso-Silurian may be divided into five portions, three of which are well-marked faunal zones. Immediately ‘overlying the top Clinton sandstone and extending vertically sometimes 30 feet, though usually more restricted, is a very prolific faunule, marking a zone called from one of its most widespread and characteristic fossils, the Dalmanites limulurus zone. ‘This zone includes more than one-half of the species | described from the whole Meso-Silurian series of Maryland. At Cumberland, where this zone was most studied, the foilow- ing species were observed : Fauna of the Dalmanites limulurus Zone. Anoplotheca hemispherica (Sowerby) * Atrypa reticularis (Linnaeus) Camarotoechia neglecta Hall Dalmanella elegantula (Dalman) + Dalmanella elegantula var. nov. * Leptaena rhomboidalis (Wilckens) Reticularis bicostata var. nov. * Abundant. + Very abundant. 572 Wm. F. Prouty—Meso-Silurian Deposits of Maryland. * Rhipidomella subcirculus (Simpson) Rhipidomella hybrida (Sowerby) +t Rhynchonella (tennesseensis) ? Roemer Schuchertella supblana (Conrad) Schuchertella tenuis Hall Spirifer crispus (Hisinger) Stropheodonta corrugata (Conrad) Stropheodonta corrugata var. pleuristriata (Foerste) Stropheodonta sp. nov. Stropheodonta sp. nov. Ctenodonta sp. nov. Pterinea emacerata (Conrad) Bucanella trilobata (Conrad) Conularia sp. * Diaphorostoma niagarensis Hall Platycerus sp. nov. Platyceras niagarense Hall Trochoceras sp. +t Dalmanites limulurus (Green) * Homalonotus delphinocephalus (Green) Cornulites sp. nov. Immediately overlying the Dalmanites limulurus zone there is a fossiliferous horizon of some 8 to 15 feet in width, which contains great numbers of Wucleospira pisiformis and some- times a few Schuchertella sp. nov. This Wucleospira pisi- formis zone, which occurs in the calcareous shale and limestone layers, is overlain by some 150 feet of limestone and shales in which the fossils are fewer and more scattered. This latter division has given us the following species : Buthotrephis gracilis var. intermedia Wall * Favosites niagarensis Hall Jamarotoechia neglecta Hall * Homeospira evax var. nov. Lingula sp. nov. Orbiculoidea sp. nov. Clidophorus sp. nov. Hormatoma sp. nov. + Hormatoma sp. nov. Near the top of the formation and usually between 50 and 80 feet below the “disintegrated yellow rock” of the Salina (the Niagara-Salina line) there occurs a band some 30 feet in thickness which is here called the Camarotoechia sp. nov. zone. It has yielded the following species: + Camarotoechia sp. nov. +t Camarotoechia obtusiplicata Hall Camarotoechia neglecta Hall Lingula lamellata Hall Spirifer sp. nov. Clidophorus sp. nov. Cuneamra sp. nov. * Abundant. + Very abundant. Wm. FF. Prouty—Meso-Silurian Deposits of Maryland. 573 Above the preceding there is a zone two feet in thickness in which occur great: numbers of Camarotoechia obtusiplicata ‘mingled with C. sp. nov. and Cuneamra sp. nov. The uppermost 50 feet of the Meso-Silurian rocks are in general largely shale and sandstone. They contain very abun- dant Ostracods and poorly-preserved Bryozoa and, in addition, so far as studied, only two fossils : Homeospira evax var. nov. Ctenodonta sp. nov. Table Showing Number, General Relation and Affinities of Maryland Meso-Silurian Fossils. Number of forms studied: Number of species and varieties studied -.------- 56 Number of species previously described -...----- 32 INumiber of mew species J... 222-22 25 Lopes 20 Number of new Varieties 2.22.25... 35-4-Len be 4 Percentage of new forms Petcentacte ol new species). 2. 222222 bee el 35°7 Pereentace of new varieties: 22.4. 22222 22--55522- yet Percentage of new species and new varieties -- ---- 42°8 Occurrence elsewhere : . Number of species occurring in Arisaig ----------- 10 x z: vi Zz. % — - * aa. iin 3 9088 01298 5776