Phe to cm EN my ete ot ly det Ba ome Ge ha Pele IN ek Oe te ee -—* Ph aye se Se he RA A Be et hee ee Ree wee oie he eet 2 el Ok et ce aed aaa aes ee i te PA WT Ne hy SN adene mie tbe a ate me eh teeta dts he at Wl
eee meetin 64 eee fe + v= ee Oe - e 4 * ro ° * => = ‘© awe et ~ Stk Seek See eee RAHN tee eM aR Be oath © Tee we we me eh tel A eh RY Mei tg Bee & oe 5 oes (Rete heer ee cote te te mee | fene
he = ee ear ne ~ —_ eee ee le a wee ee i et i i Et el ee ee ee Ser ee er ee ae a ee ce, a ee err i i rk ok kk Tee er mate ee teed ee Mage
eC. ne 2 ne _—™ 6 a etme = rare eS ~ + -- ~< aa - Pe me ot et re @ a> \a aed - Pe ee ee ee oe Oe De en ee oe ee el ne eee ee Oe
“?« ee ee ee joe ee ok <9 ooh ee “me “t Ce en ee ee ee ere 0 es ee ee ee ee tw We ek A tre ee we eh eth treet Hatin Rte Bete 2 in ee ee ee Redd og eek ed Sy
A ee ee ee © 5 = Se tte nme ee ere tm cies 2 fe +e hor .. ee ee Ce ea “4s Nee ee ee a Se ee en ee ee oe arn Mere Wee ot Me ee
ee ee ee a a ns hee Aad en we be baw ae 5 eke 4 - ee , are a eet ke nde bea te ae Se met As Mele SI | ten
a ee ee ee ee a & <== - - - -m es ak et. San we a ae owe ~ en
= . eee ae fm Ment te a1 —“e Me ae * * te ee . oe ee ee eo a 2 be tewtete ete ad ee ee ad
ee Rama etait ee ee - ee 1 ~ - q <> emt - “ + aan +e mere =a = < twee % amet ee ke ee ——* eo. or ne a ee
ee m, oe oe hy eee & Saar we dm See em ee” Lt = Pars — 4 ee ee y tte ade eo eter eR eh Rat albtt beter tale me Meike Modan me teh pete wo mah ine tp eee ee 8
i en or She Bete ee ee oe —— a 4 m6 ee a Ae Ree Ne eth Be eR te ee abd BO le te | tte Rte teed S eRe bem wel ten ke ete bees GE a ee
oe Se ee aM ke Re Fem tN em ey . —* : Od 6 et ny me ee Re em yee oe, ee ee roe . Sere ne, Or Se ar See ae eee ter Ore ee ee en Ae te Ngee ek tee be tet tote ib ee Be
ee ae or en ee ar . ~ ial ae te —_ Cs a oe a é eo we Sams Oe ee a He mee <a « bee Be we Hd Ree R ae ede ew etns be ates adh Lait tn Bel ete
-seee e — -——- > bl = ~ = 7 4 eh a 2 oe aoa 7 add ~@-+ 4, as er oe ete ee time ee te et te we Me ee we, a rm 5 a,
ot dew! alee een Bedi te - - = cnt ere Be ee he ete = . a * ae ~ “a Lad eee © whee Pee ee ee ee ~ 0, he tei he %o thy: en, en
a ies -" ~ as oe te ae 7 Sen Ah anaan ~* Sa Rat ee ek bw 4 seo ee ee ee tile tee ee BLY io ees Pie fo i
6 tomes e ot ~ ote A =% ma ce a . . aie ~~ mee aw Od Or eee ee eee ee ree ‘ “
= be law - <tr ~ oatiaiied ~ ~_— a “* a on 1 ate eethe c . * 7 rd a- 2 the .
am ee a lot oe « « aa ons - - Oe Tella tele tae
-- . he ey oe a -_- a ile J 4 “me . 7 bead ie ie *, ™ ‘~ . ’ 7 * ~ Oe ha ~~ er I ‘ ° Pens 4 am My ‘. be in: be te
— nthe em i et ee oe “ ee 16 few oe - eke ees 6m . Heme Meee A i te ly ewe Meet It g
ad 7 < te = S -— eh lads “ ? . 2 es pew . ve ab ay . . = a oe Seed ~ "8 a de ete eee ee NM ote te EO lite
= —. Na than anata mt ha -* » eel ra x se hom, wee ke on = ~ a a mw bo et 8 ee ad ae Se ote seals! © ints nee Sok ae, te ace Taadne
- oe eS “oe 4 - i _—- os stake oo 2 at oe ’ Ae me . - . ~~ bee eh ee tee ~~ A in aie a We Yedeo- -
- Ate mee) tet! fester tet! Rete al a To oa er am ~ * > « ee ‘ = <n eS yet we hl he . oo eee ed 3
Tn ed ie ~ “hs > 2 Aste te Oe ete eo 4 _ 26 i het
- ed - owe ~ a ts eee bcd iu hn = © >. a i“ - a - . sew eae a * Aan ~s
--- - . a in 1S Ae ae - - ; ot he » * > * - 1s
=e ie ee = ee en . i © - « mt © Naat -*- . t _ + wee a = -- 83a waa * . ‘ bat me
*. - ne - - — - a _ ~~ a i. ‘ . swe : ane = 4: “ot = « + a ‘ e - sar - - = + -
- mite te oe -- owe = «a et med ae me - - — ~~. ‘4 , “ = 74 ~he ie ’ 7 . “ so) we een ee me 1s id
: - e: mpl eee -- 7 ’ a. 12s vem ome . tld
. a =< sf é . os eee ‘ a» ‘ 4 oe ° _ oe 7 +8 wl oh. ee ‘ .
cies ee = — % sees 5 * « ‘ oe oe, : BY osp omg Pe
mack ¥ i oe -waw ps . . . . we * ‘ . i Sone foe oo)
bs = 7 7 ‘ _ - “ - tee . “ . a ® @. . . -
-- ve ale =’ a ee s+ - a : > +> — - -* & . es . . ‘ js re eee a) 7 ns
- += = whofe ~ - . « - . i. ° "ero += 1 * '
" . - 7 de ‘tg - > . ‘ -
- - +* Se - : - ‘ = an
: 7 = 4, ae =~ a a 7 . ‘ oa * a so: 2
- sete - - * ~ ‘ iw 1 , ‘ . , . e+ =o : ‘ F te a - 405 rtm la Pent +m) apa ee
7 a s 7 - , e a - ° «* ‘ “* ‘a 7 . oe’ ¥ shh i a data Bh hee ares - 2
ey - eet - * Ve es ‘ .- ate » #2) Gwe en & bet Se Wet Haters | i 2 ue Ae ts Kot
7 sicon - os 2 + —- - y ‘ ' : = . me ase wy ty (obs Sale hg © Pans o 7 » he
= oe os ~ thn SO sete &, “ 7 ae , ‘ ‘ ele oat —* “.% Was | & ' A et ae
- “ = o + - m — ' - as ~ s) fete ae ov - te
<*'t 2, 95 « - <n Ri 7 . — 7 P = 7 7 a
- = . = - 4 * oe . « feos it 2 one Lad She
mm ie Caied a= 7 - “ ” , - . o wo « - «te ‘ 1 *”
* Siri ‘a 2 . suse 8 . ‘a 7 . — Pe ee ere ee
= + > 7 * ‘ " soe eoae — = wl & t Sue a a
* ~- - ba « ? s -* + +
- = 4 7 - : : * ‘ “ <P we le —_— i 2 - Ps a
.o8) - 7 y i o's é« % ike - 2 cae and aia: Ves
. 7 =» - ees - ‘ : on hk ’ we a a & " : 7”
7 ra a 2 ‘ - en oe :
c =" - = - - - * - « ae ' * wh - oT) +
- - 7 » * ite ee 7 -
os ~ J y e - 2 ee -—. let BS Stes we a%e
- 2 oe a 7 - + ” u *. a -
2 ‘ - - + : . . : = toh ve
7 ws - a - P
= 3 7 ate a 's he etree
- 7 P $m te P oy - a= Se, Sa “ee ite
7 + ed ote = - = ' “ * * - ‘ S * = - . = ce ry -
= - Ray? ® 7 . - ‘= =4 Ye = . -
- Sian - . 7 -
fi * ~. “ 2 “ 7 - < , ees . -
- -* . a ue rs —_— - Sree
—- ' 7 ie Ps & oa i 7 cary -— oy P) mu ae
ta ‘ - “ . im
> “ = é = A 7 a : 4 < f.
- bd oid ° S -. ‘* - : "= a6 . < eee . .
- € +
’ -
- 7 7
- 2 :
1 =
» a . ‘°
’
. -
- + \
: r i
- » baal - cs _
+e OR ee a My el eh) me * ~ *% Z : - : -
cae : Ce tied . ee en eo ey ee ee a - oe .
1 ee te ee man me ee a 7 mi = - ee
< . a « -i- o&a~ —o <i — ~-in @ _- -
= em en He ~ + + ner - = -
1”

fem, 75

THE

AMERICAN
JOURNAL OF SCIENCE.

Epitor: EDWARD S. DANA.

ASSOCIATE EDITORS

Proressors GEO. L. GOODALE, JOHN TROWBRIDGE,
W. G. FARLOW anp WM. M.-DAVIS, oF CAMBRIDGE,

ors A. &, VERRILL, HENRY S. WILLIAMS anp
if V. PIRSSON, oF New Haven,

ProFessor GEORGE F. BARKER, oF PHILADELPHIA,
ProFEssor JOSEPH S. AMES, or Battimore,
Mr. J. S. DILLER, or WASHINGTON.

FOURTH SERIES.

VOL. XIV—[WHOLE NUMBER, CLXIV.]

WITH NINE PLATES.

NEW HAVEN, CONNECTICUT.
1902.

bp eto ae

THE TUTTLE, MOREHOUSE & TAYLOR COMPANY,
é NEW HAVEN, CONN.

AACR ca a age
\y be
ob ; we
ANU ReU MEET O LY |

CONTENTS TO VOLUME XIV.

Number 79.
Ss s hae Page
Art I—Spectra Arising from the Dissociation of Water
Vapor, and the Presence of Dark Lines in these Spectra ;

Bee xowERIvGEe. (With Plate 1.) -__-._...--2--+-. 1
IJ.— Occurrence of Greenockite on Calcite from Joplin, Mis-
Seer. b. CoRNwaALh( 22220... .--------- ee kes

IiI.— Quantitative Study of Variation in the Fossil Brachio-
pod Platystrophia lynx ; by E. R. Cumrines and A. V.

femme ith Plates ITT)... ea DB
IV.—Studies of Eocene Mammalia in the Marsh Collection,
Peabody Museum; by J. L. Worrman .- -_.-------- -17
V.—New Exposures of Eruptive Dikes in Syracuse, N. Y.;
Ermer mmInm Lo ee lee ee See 24
Vi1—Petrography of Recently Discovered Dikes in Syra-
Reese by OC. H. SmMyv1H, Jr. = 2 ence 2 gu =, 6 AE
Vil.—Preliminary Note on Silver Chabazite and Silver
Analcite ; by G. STEIGER.---. es 32th 8 PIZREACEETS Gl
VIII.—Significance of Certain Cretaceous Outliers in the
Klamath Region, California ; by O. H. Hersuey- ---- 33
{X.—Action of Copper Sulphate upon Iron Meteorites; by
RIEERRIGU ON 622 he eee 38
X.—Petrographical Contribution to the Geology of the
Eastern Townships of Quebec; by J. A. DresseR_.--- 43
XI.—Action of Carbon Dioxide on the Borates of Barium ;
emeemancrerrig ee eel 49
XII.—Studies in the Cyperacee; by T. Horm. XVI -_.--. 57

SCIENTIFIC INTELLIGENCE.

Chemistry and Physics—Probable Source of the Heat of Chemical Combina-
tion, T. W. Ricwarps: Lithium Silicide, Motssan: Black Color of the
Rocks in the Cataracts of the Nile, LorterT and Hucouneng: Bismuth-
lead Sulphide-halides, Ducatrn, 64.—Atomic Weight of Uranium,
RicHARDS and MERIGOLD: Qualitative Chemistry, Prescotr and SuLLI-
VAN: Qualitativen Analyse, W. BorttcerR, 65.—Reflection Power of
Metals for Ultra-violet and Ultra-red Rays, E. Hacen and H. RvsBeEns,
66.—Spark Discharge from Metallic Poles in Water, Wiustne: Electrical
Effect of Light, P. Lenarp: Interference Tubes for Electrical Waves,
A. Becker: Law of Velocity which Cathode Rays suffer by Reflection, E.
GEHRCKE, 67.— Role of Self-Induction in discharges of Electricity through
Gases, M. B. Eainitis and M. A. DE Gramont: Theory of Optics, P.
DrupeE, 68.—Elementary Principles in Statistical Mechanics, J. W. Gipzs :
Beitrage zur chemischen Physiologie, F. HormetstTEr, 69.

Geology and Natural History—U.S. Geological Survey : Minéralogie de la
France, Lacroix: Size of Grain in Igneous Rocks, QuENEAU, 70.—
Eparchzean Interval, Lawson: Former Extent of the Newark System,
Hosss: Quaternary History of So. California, HersHry: Plissements et
Dislocations de l’Kcorce terrestre, NEGRis, 71.—Tungsten Mine at Trum-
bull, Conn., Hogpss: Eruption of Mt. Pelee, VeRRILL, 72.—Maldive and
Laccadive Archipelagoes, GARDINER: Actinians of Porto Rico, DUERDEN:
Ascidians of the Bermuda Islands, Van Name, 74.—Embryology of
Terebratulina septentrionalis, ConKLIN: Yellow Fever investigations at
Havana, GARGAS, ‘75. ;

Miscellaneous Scientific Intelligence—Smithsonian Institution : United States
Weather Bureau, A. J. Henry: Scientia: Ostwald’s Klassiker, 76.

1V CONTENTS.

Number: So:
Page
Art. XIII.—Terraces of the Westfield River, Mass.; by W.
M. Davis., (With Plate IV) .._...__... 2-2) ee

XIV.—Cretaceous Turtles, Toxochelys and Archelon, with a
Classification of the Marine Testudinata; by G. R.
WIELAND 406. -b5s-cecc2! -stebs)- ts 95

XV.—Magnetic Effect of Electric Displacement ; by J. B.
WHITEHEAD, JT....- -=------ sseenss$ 43.422 eee

XVI.-—Certain Relations of Plant Growth to Ionization of

the Soil ; by A, B. Plowmancied 22.2 ossSl) Se 129
XVII.—Demaegnetizing Effects of Electromagnetically Com-
pensated Alternating Currents; by Z. E. Crook --_.--- 133
XVIII.—Nepheline-Syenites and other Syenites near Port
Coldwell, Ontario ; by A. P. CoteMaNn._._...... 22330
XIX.—Double Ammonium Phosphates in Analysis; by M.
AUSTEN dee 2 20.182 3 hs OTe See 156

XX.—Electrical Resistance of Glass, Quartz, Mica, Ebonite
and. Gutta-percha; by O. N. Reop_-..-..2=2.-.4.-222ee

SCIENTIFIC INTELLIGENCE.

Miscellaneous Scientific Intelligence—American Association, 166.
Obituary—Dr. Caro Riva, 166.

CONTENTS. Vv

Number 81.

Page
Arr. XXI.—Relationship of some American and Old World
Birches; by M. L. Fernatp. (With Plates V-VI) --- 167

XXII.—Fertile Fronds of Crossotheca and Myriotheca, and
on the Spores of other Carboniferous Ferns, from Mazon
Creek, Illinois; by E. H. Setiarps. (With Plate VII.) 195

XXUL—Validity of Idiophyllum rotundifolium Lesquereux,
a Fossil Plant from the Coal Measures at Mazon Creek,
Pees PT. SELLARDS) 4. .-.---- 22422 5----- 203

XXIV.—Precipitation of Ammonium Vanadate by Ammo-
nium Chloride ; by F. A. Goocu and R. D. Girpert..- 205

XXV.—Additions to the Alunite-Jarosite Group of Minerals ;

by W. F. Hitiesranp and S. L. PENFIELD ----------- 211
XXVI.—Niagara Limestones of Hamilton County, Indiana :

SEES ee a 221
XXVII.—Velocity and the Structure of the Nucleus ; by C.

RY as he a a i ke 225

XX VIII.—Corundum and a Graphitic Essonite from Bark-
hamsted, Connecticut ; by B. K. Emerson .--.--------- 234

vi CONTENTS.

Number 82.
Page
Art. X XIX.—Experimental Investigation into the Existence

of Free Ions in Aqueous Solutions of Electrolytes ; by

1. Osun tee web ae nee ee - cise +)

XX X.—Solution of Problems in Crystallography by Means
of Graphical Methods, based upon Spherical and Plane
Trigonometry ;, by 8. L. PENEIELD-_._. ... 22.23

XX XI.—Estimation of Bromic Acid by the Direct Action of
Arsenious Acid; by F. A. Goocu and J. C. BLakE __-_ 285

XX XITI.—Solubilities of Some Carbon Compounds and Densi-
ties of their Solutions; by C. L. Spemymrs._-.2-:---.22 293

SCIENTIFIC INTELLIGENCE.

Chemistry and Physics—Radio-active Bismuth, MARCcKWALD: Thermochemi-
cal Constant, F. W. CLARKE, 303.—Heatless Condition of Matter, Brinx-
WoRTH and Martin, 304.—Elementary Chemistry, by F. W. CLarkE and
L. M. Dennis: Liquid Hydrogen and Helium. J. DEwar, 305.—Vapor-
pressures of Liquid Oxygen and of Liquid Hydrogen, M. W. TRAVERs,
G. SENTER and A. JAQuEROD, 311.

Geology and Mineralogy—Martinique, 312.—Report of the Geological Survey
of Louisiana, G. D. Harris, A. C. VeEatcH and J. Pacuesco, 314.—Alpen
im Eiszeitalter, A. Penck and E. Brickner, 315.—Bacubirito, the Great
Meteorite of Sinaloa, Mexico, H. A. Warp: Origin of Hskers, W. O.
Crossy: Ueber Hussakite (Xenotim) und einige andere seltene gesteins-
bildende Mineralien, H. ROsuzr, 316.

Miscellaneous Scientific Intelligence—International Catalogue of Scientific
Literature, 317.—British Association: Experiments in Aerodynamics,
S. P. Lanctey: Elementary Physical Geography, W. M. Davis: Inter-
national Quarterly, 318.

Obituary—PROFESSOR RUDOLF VIRCHOW.

CONTENTS. Vil

Number: 83:

Page

Art. XX XIII.—Observations on the Eruptions of 1902 of

La Soufriére, St. Vincent, and Mt. Pelée, Martinique ;
bye. Hoyny. (With Plate VIII)-_-..----..-- ee OL

XX XIV.—Reflection of Electric Waves at the Free End of
a Parallel Wire System; by H. A. Bumsrrap.-_-.----- 359

XXXV.—Upper Permian in Western Texas; by G. H.
Pema ts tS Oe a eet oS er Aa re ee 363

XXXVI.—Reduction of Vanadic Acid by the Action of
Hydrochloric Acid ; by F. A. Goocn and L. B. Stookry 369

PME CeEGInY OWE. 2.2.02. 20. boo eee fo ee eck eo S77

SCIENTIFIC INTELLIGENCE.

Chemistry and Physics—Combination of Hydrogen and Oxygen, BAKER:
Arsenic in the Animal Organism, BERTRAND, 383.—Atomic Weight of
Radium, Curie: Double and Triple Thiocyanites, WELLS, 384.—Deter-
mination of Thallium in the Thallous State, THomas: Chemisches
Praktikum, A. Wo_trrum: Vacuum Thermo Element, P. LEBEDEW: Elec-
tromotive Force of Ozone, A. BRAND, 385,—Influence of Electrification of
the Air on Electric Sparks, E. LecHer: Reaction between Electrostatic
Tubes and Insulators, M. W. bE NicoLalEVE: Magnetic Detector for elec-
trical waves, Marconi: Radiations from Radioactive Substances, RUTHER-
FORD and Brooks, 386.—Induced Radicactivity in Air, J. J. THomson:
Deviable Rays of Radioactive Substances, RUTHERFORD and GRIER:
Removal of Negative Electricity from the air by falling drops of water, A.
ScumMavuss: Hlements of Experimental Phonetics, E. W. Scriprure, 387.

Geology and Mineralogy — United States Geological Survey, 389.—Iowa
Geological Survey, S. Catvin, 391.—Evolution of the northern part of the
Lowlands of Southeastern Missouri, C. F. Marsut: Geology of the
Potomac Group of the Middle Atlantic Slope, W. B. CuarKk and A.
BrBBins, 392.—Pleistocene Geology of Western New York, H. L. Fatr-
CHILD: Michigan Geological Survey, A. C. Lane: Vertebrates of the
Mid-Cretaceous of the Northwest Territory, H. F. OsBorn and L. M.
LAMBE: Queneau on Size of Grain in Igneous Rocks, A. C. LANE, 395.—Les
Roches alcalines caractérisant la Province petrographique k’ Ampasindava,
A. Lacroix: Ueber mariupolit, ein extremes Glied der Elaeolithsyenite,
J. Mornozewicz, 396.—Dahamit, ein neues Ganggestein aus der Gefolg-
schaft des Alkaligranit, A. Petikan: Mineral Resources of South Dakota,
C. C. O’Harra and J. E. Topp: Occurrence of Native Arsenic, N. N.
Evans: Inverness Earthquake of September 18th, 1901, and the Carlisle
Earthquake of July, 1901, C. Davison, 397.

Miscellaneous Scientific Intelligence—Beitrige zur chemischen Physiologie
und Pathologie, F. Hormrerster: Les Dirigibles, M. H. ANpR&: Introduc-
tion to Physical Geography, G. K. Grupert and A. P. Bricuam, 398.

Vill CONTENTS.

Number 84.

Art. XX XVII.—Instance of the Action of the Ice-sheet upon
Slender Projecting Rock Masses; by W. H. Hosss.
(With Plate PXyCr ls PL oP eee ee 399

XXX VITI.—New Species Melanochalcite and Keweenawite.
With Notes on some other known species; by G. A.

KOENIG \46 ci. coon 5 SS) oe
XX XIX.—Studies in the Cyperacex; by T. Homm __-...-- 417
XL.—Chemical Composition of Dumortierite; by W. E.
Horp. J. 2.i56 2.2222 ee oe 426
XLI.—Characteristics of Kau; by J. 8. Emerson._--.----- 431
XLII.—Titrimetric Estimation of Nitric Acid; by I. K.
PHELPS 2.202222 S. lel ee ee
XLII.—Clays of the Boston Basin; by R. M. Brown -.--- 445
XLIV.—New Form of Calcite-Sand Crystal ; by E. H. Bar-
Bourn and C. A. Wishme _22-@22. 215 3.2 2 451

SCIENTIFIC INTELLIGENCH.

Chemistry and Physics — Transformation of Carbon into Diamond, A.
Lupwic: Iodometric Titration of Thiocyanic Acid, THIEL, 455.—Iodine
Pentafluoride, Moissan : Separation of Manganese from Magnesium, Zine,
and Aluminum, DieTRIcH and HasseL: Plasticity and Adhesiveness of
Glass at Ordinary Temperatures, Piccarp, 456.—General Principles of
Physical Science, A. A. Nores: Lehrbuch der Allgemeinen Chemie, W.
OsTWALD : Spectra of Hydrogen and Reversed Lines in the Spectra of
Gases, J. TROWBRIDGE, 407.—New Holtz Machine, H. WOMMELSDORF :
Electrical Conductibility of Metals and their Vapors, R. J. Strutr: Ioniza-
tion of Nuclei Produced by Violent Agitation of Dilute Solutions, C.
Barus, 509.—Handbuch der Spectroscopie, H. Kayser, 460.

Geology and Natural History—United States Geological Survey : Geological
Survey of Kansas; Special Report on Mineral Waters, E. H. 8. BAILEy :
Quantitative Chemico-mineralogical Classification and Nomenclature of
Igneous Rocks, W. Cross, J. P. Ippines, L. P. Prrsson and H. 8S. WasuH-
INGTON, 461.—Petrography and Geology of the Monzoni region in Tyrol,
ROMBERG : Gesteinskunde fiir Techniker, Bergingenieure, etc., F. RINNE,
463.—Terlingua Quicksilver Deposits, Brewster County, Texas: Notes on
New Minerals, 464.— Gold in Meteorites, Liversipen, 466. — Triassic
Ichthyopterygia from California and Nevada, J. C. Merriam, 467.—
Zoological Results based on material from New Britain, New Guinea,
Loyalty Islands and elsewhere, A. WILLEY, 468.

Miscellaneous Scientific Intelligence—National Academy of Sciences, 468.—
American Association for the Advancement of Science : Smithsonian Insti-
tution, Annual Report for 1901: United States National Museum : British
Museum ; Catalogue of the Collection of Birds’ Eggs : United States Coast
and Geodetic Survey, 469.— United States Naval Observatory : Bureau of
American Ethnology: Ostwald’s Klassiker der Exakten Wissenschaften :
Hodgkins Gold Medal Awarded, 470.

Obituary--DR. OGDEN NicHOLAS Roop: Dr. RoBertT C. KEDZIE: Dr. ROBERT
RUBENSON : JOHN HALL GLADSTONE, 470.

F .
.
/
5
;
‘
3
=
va
>
my
' if
Sk
us
,
}
, H

Am. Jour. Sci., Vol. XiV, 1902. Plate |.

Fig. 1

Fig. 2

Fig. 3

Fig. 4

THE

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES.]

Art. I.—On Spectra Arising from the Dissociation of Water
Vapor, and the Presence of Dark Lines in these Spectra ;
by JoHN TRowsBRIDGE. (With Plate I.)

In passing from the study of the light emitted by gases
under the effect of electrical discharges of comparatively small
quantity to the investigation of the light produced by dis-
charges of great quantity, a new field of research is entered.
In previous papers on the spectra of hydrogen, I have stated
my convictions of the importance of the role played by
water vapor in spectrum tubes. The results of further study
emphasize these convictions. With powerful discharges in
hydrogen, oxygen, and rarified air, I always obtain the same
spectrum, which I regard as that arising from the dissociation
of rarified water vapor. The line spectrum, moreover, is
accompanied by a faint continuous spectrum on which are dark
lines, which indicate a selective reversibility in the silver salt,
which is of great significance, it seems to me, in the applica-
tions of photography to astrophysics.

It has long been recognized that spectrum analysis is an
extremely delicate method of recognizing the presence of a gas
or the vapor of a metal under the excitation of heat ; and when
the improvements in photography enabled us to obtain per-
manent records of the spectra of gases, it was supposed that we
had a means of escaping from the fallacies of eye observations
which arose necessarily from personal idiosyncracies. If the

hotographic plate were a perfect instrument for recording the
infinite number of vibrations which light can communicate to
atoms of matter, we should certainly feel that we had made a
great advance in physical science. When we reflect, however,
on the belief that emulsions containing silver salts are capable
of responding to all wave lengths of light in the portion of

Am. Journ. Sci1.—FourtTH SeErizs, Vout. XIV, No. 79.—JuLy, 1902.
1

2 J. Trowbridge—Spectra from the

the spectrum considered most actinic, even when the light ex-
ceeds a certain intensity, we are conscious that we cannot rely
upon an infinite range of photochemical action; and I shall
show in this paper the existence of a selective reversibility
produced on the photographic plate by powerful discharges of
electricity through capillary tubes.

Realizing the importance of studying the behavior of gases
under different forms of excitation, I have collected in the
rooms devoted in the laboratory to spectrum analysis three
forms of apparatus: an induction coil actuated by a very eff-
cient liquid break, giving a spark of thirty inches in air; a
step-up transformer excited by an alternating current, pro-
ducing with glass condensers of about °3 microfarad, dis-
charges of an inch in length, of great body; and a storage
battery of twenty thousand cells. A plant of this nature I
conceive to be necessary in the present state of spectrum anal-
ysis, for molecular motions excited in rarified gases vary greatly
with the kind of electrical discharge.. In the application of
photography to spectrum analysis, the investigator is immedi-
ately confronted with the necessity of submitting the gas toa
comparatively long electrical stimulus in order to obtain a
negative. Even with a concave grating of short focus several
discharges are necessary with a narrow slit. Hach discharge is
capable of modifying the condition of the gas. This fact is
well recognized by taking successive photographs of the light
emitted by a spectrum tube under different strengths of eur-
rent upon the same photographic plate. A simple form of
plate holder enables this to be done. One obtains a striking
example of the instability of the spectrum tube filled with
apparently dry hydrogen when one subjects it first to very
powerful discharges from a glass condenser of °6 microfarad,
charged by a storage battery of twenty thousand cells, with
practically no self-induction in the circuit; and follows this
excitation by the use of an alternating discharge of much less
quantity. The powerful discharge gives what I term the water
vapor spectrum; and after a certain number of these dis-
charges one obtains with the alternating current discharges the
spectra of argon. This results, I suppose, from the oxidization
of traces of air in the tube. The-hydrogen has disappeared
momentarily; possibly the dissociation of the water vapor
resulted in concealing its presence.

The period of the condenser discharges which I have em-
ployed varied from one five hundred thousandth of a second
to one millionth. The practically instantaneous current varied
from five thousand amperes to ten thousand. The revolving
mirror method showed that the pilot spark was mainly effect-
ive, and that the subsidiary oscillations were feeble. The

Dissociation of Water Vapor. 3

spectrum tube speedily became milk white from the sodium
set free from the glass.

Lord Rayleigh* has shown how to demonstrate the presence
of argon from very small quantities of air. My method is sub-
stantially his, except that I employ very powerful discharges
which set free a sufficient amount of sodium vapor from the
glass, and the oxygen is supplied from the dissociation of the
water vapor, which is always in evidence when such powerful
discharges are employed. The production of argon under
these circumstances is a striking proof that I am dealing in
this investigation with the spectrum of the dissociation of
water vapor. From the same tube one can, by modifying the
strength and character of the electrical discharges, obtain what
is termed the four-line spectrum of hydrogen, the spectrum of
sodium, the spectrum of argon, and the spectrum of the dis-
sociation of water vapor. Doubtless one could also recognize
the spectrum of helium; I am not yet sufficiently familiar
with it.

In the course of a study of the water vapor spectrum one is
naturally led to photograph the spectrum of the electric spark
under water. It is possible to obtain powerful discharges of
any snitable Jength under distilled water by enclosing the spark
terminals in glass tubes, allowing only a small portion of the
platinum terminals to project from the ends of the tubes. If
the terminals are immersed more than an inch under water,
the resulting explosion is apt to break the glass-containing
vessel. The light of these discharges under distilled water is
white and extraordinarily brilliant to the eye. When it is ex-
amined by the spectroscope one sees a continuous spectrum ;
and one obtains a continuous spectrum even by photography
in the most actinic portion of the spectrum. On bringing the
spark terminals to the surface of the water, leaving them barely
immersed, one immediately obtains the so called four-line spec-
trum of hydrogen. ‘To what is due the continuous spectrum
under water? Does it result from the production of the water
vapor spectrum under great pressure? ‘That there is great
pressure is shown by the sudden explosion, which is sufficient
to blow the small tamping of water out of both ends of the
containing tube. If the surface of the water is covered with
a thin film of oil, this oi] is immediately disseminated through
the water, a milky white emulsion, which remains for days.

When we turn to powerful discharges through Geissler or
Plucker tubes filled with hydrogen which has been dried with
great care, we also obtain a continuous spectrum on which are
superposed the bright lines of the water vapor spectrum,
together with certain reversed lines, giving on the positive

* Phil. Mag. (6), vol. i, 1901, p. 103.

t J. Trowbridge—Spectra from the

dark lines instead of bright lines. Moreover, what are ap-
parently the strongest bright lines of the water vapor spec-
trum are not reversed. There is aselective reversibility which
arises with high temperatures, and is made evident by the faint
background of the continuous spectrum. ;

This fact seems to me to be of great importance in the appli-
cation of photography to the study of celestial phenomena, for
reversal of spectrum lines do not necessarily indicate reversing
layers of cooler gas, and may be a photochemical action of the
silver plate.

One immediately thinks in this connection of the phenome-
non of dark lightning or the Clayden effect, and of the sug-
gestive experiments of Professor Nipher.* Spectrum analysis,
however, reveals a selective reversibility, which shows that the
subject of photochemistry must be carefully studied before we
can interpret properly the records of photography in spectrum
analysis.

There are doubtless many states of vibration even in the
actinic portion of the spectrum which are not recorded by the
silver salts, for a selective reversing action may obliterate or
prevent a permanent record.

I have obtained this reversing action with different emul-
sions on glass and also on celluloid films. The strongest
reversals are at wave lengths 4227, 3930, 3965. There is also
a reversed band between wave lengths 4315 approximately
and 4285, and a faint reversal at wave length 3953. Rever-
sals are often seen on the negatives which disappear in the
fixing bath.+

In this investigation ten thousand storage cells were em-
ployed to charge a glass condenser of ‘60 microfarads. The
charge was sent through Geissler tubes having an internal bore
of one millimeter; the length of the capillary of this diameter
was from eight to ten centimeters. The tubes were filled with
apparently dry hydrogen at a pressure of approximately -1™™.

I have reached a limit in submitting gases in glass tubes to
powerful electric discharges, and am now turning my atten-
tion to the subject of quartz tubes in the hope of obtaining
these in a suitable form for containing rarified gases.

In a previous paper,t I have expressed my conviction that
the four-line spectrum observed in the protuberances of the
sun is an evidence of the dissociation of water vapor in the
sun’s atmosphere, and an evidence therefore of the presence

* Transactions Academy of St. Louis, vol. x, No. 6, On certain properties
of light struck photographic plates.

+ In this connection the following paper is significant: Disappearance of
images on photographic plates. William J. S. Lockyer, Nature, Jan. 17,

+ This Journal, vol. x, Sept. 1900.

Dissociation of Water Vapor. 5

of oxygen. In the spectrum one sees two reversed bands
which coincide with the great H. H. lines in the solar spece-
trum; one also sees a reversed line at wave length 4227, closely
coincident with a strong solar line; one has also a reversed
band coinciding with a region of reversals in the sun between
wave lengths 4315-4285. These regions in the sun are doubt-
less composite photographs of many reversals of different ele-
ments. I believe, however, that the basis of the reversals is
the presence of dissociated water vapor.

The nomenclature, too, of the stars in relation to their spec-
tra I believe should be changed to dissociation spectra; an
excess of water vapor produces what is termed the hydrogen
type, and the dissociation of this water vapor in the presence
of other gases, nitrogen for instance, together with metallic
vapors, may account for other types.

The intense light due to the dissociation of water vapor
may, in some cases, mask the fainter light of the metallic lines
in the stars which show only gaseous spectra, especially when
we consider the varying distances of the stars. J have em-
ployed electrodes of platinum, copper, silver, aluminum, iron,
and aluminum, and no trace of the lines of their vapor can be
perceived in the photograph of the gaseous dissociation spectra ;
moreover, when sodinm is present in the spectrum tubes,
although it fills the tube with a brilliant yellow light at low
voltage and strong currents, and exhibits the sodium lines very
strongly, no trace is seen of it, with very powerful discharges.
The tubes then show only the brilliant white light due to the
dissociation of water vapor: a light which is the nearest ap-
proach to sunlight which I have been able to produce. Its
actinic effect is greater than that of any of the metals, such as
magnesium, aluminum or zine. Possibly the varying amount
of water vapor may be a factor in the variability of certain
stars; and one is led to conjecture whether the light of the
sun’s atmosphere is not due to an electrical dissociation due to
discharges of very high period.

The selective reversibility of the silver salts seems to me of
momentous importance in the subject of astrophysics; for we
have reversible actions on the photographic plate which are
not due to the radiations of a gas passing through colder layers
of the gas. In other words, we have actions recorded which
are photochemical and are not in the heavens. It may be that
certain reversals observed in the spectrum of Nova Persei may
be of this nature. This selective reversibility serves to reveal
certain lines which escape observation. Thus we see that
there is a gaseous line at wave length 4227, and a line not
shown on the reproductions but clearly seen on the negatives

between the great H. H. lines, having a wave length of 3953.

6 WS. Trowbridge—Spectra from the Dissociation, ete.

The silver salt, therefore, does not respond to all rates of
vibration ; or if it does respond, the molecule is unstable and
there is no resultant reaction which is evidenced by a photo-
graphic image. There may be spectra at very high instan-
taneous temperatures which we cannot phctograph. It seems
reasonable to suppose that the silver molecule is limited in its
rates of vibration and that the photographic plate as well as
the human eye is a limited instrument of research. :

On Plate I, A represents the solar spectrum in the neighbor-
hood of the great H. H. lines. B represents the gaseous spec-
tra. The photographs were taken with a Rowland concave
grating, and are not enlarged or touched in any way. Unfor-
tunately the reproductions do not give many of the reversals,
Figs. 1 and 2, B are spectra of air and oxygen taken with a
comparatively low voltage and amperage at a pressure of about
one millimeter. Fig. 3 is a spectra of hydrogen under the
same conditions. Fig. 4 is the spectrum arising from the dis-
sociation of water vapor with very powerful discharges. Fig. 5
shows a selective reversal line at approximately 4227. On the
negative the two bright gaseous lines, which in fig. 4 closely
coincide with the great H. H. lines, are seen to be reversed
and appear as dark lines. This reversal is shown in fig. 7.
B of fig. 7 closely resembles in general features the solar spec-
trum when the latter is photographed with a broad slit, so as
to obliterate details and give broad characteristics.

My conclusions are as follows:

Dissociation of water vapor takes place in the atmosphere of
the sun; oxygen, therefore, must be present. From a carefal
study of my negatives, I regard the evidence for the presence
of water vapor as conclusive as that accepted for the presence
of sodinm in the sun.

The dissociation of water vapor, under the effect of power-
ful electric discharges in the presence of small amounts of
atmospheric air, results in the production of argon even in
tubes presumably filled with dry hydrogen.

The spectrum arising from the dissociation of water vapor
contains dark lines as well as bright lines. These dark lines
are due to a selective reversibility of the silver salt employed
in the photographic emulsions, and are not due to the reversing
action of a cooler layer of gas. The great brilliancy of the
dissociation spectrum of water vapor, which obseures the
spectra of metallic vapors: and the presence of dark lines due
to photochemical reversals, make one cautious in accepting
photographic evidence in regard to states of development of
stars.

Jefferson Physical Laboratory,
Harvard University,

Cornwall— Greenockite on Calcite from Joplin, Mo. 7

Art. Il.— Occurrence of Greenockite on Calcite from Joplin,
Missouri ; by H. B. CoRNWALL. |

A YELLOW coating of greenockite on the sphalerite of the
Joplin zine district is no uncommon occurrence, and having
received, through the kindness of Messrs. Geo. L. English &
Co., of New York, specimens of calcite from this locality
which showed a very distinct greenockite coating, the writer
made tests upon the material, to determine what proportion of
cadmium sulphide the coating might contain.

The greenockite occurs as a bright yellow, dust-like coating,
which can be easily rubbed off with the finger. Beneath the
greenockite coating, and often extending beyond it, there is,
for the most part, a brown or yellowish brown, more adherent
deposit, in a thin layer which can be easily scaled off with the
knife-point. This seems to be sphalerite.

The greenockite was removed by very gentle scraping with
the knife and subjected to the tests below. Although in
places it is very distinct, the coating is also very light, and
from an area nearly two centimeters square not more than
0-0014 grm. of the purest material could be obtained. Even
this portion was found to be mixed with zinc compounds from
the scale-like deposit below it.

Gently roasted on platinum foil over a Bunsen flame, the
powder on cooling was brown in some parts, and nearly white
(distinctly yellow while hot) in others, indicating cadmium and
zine oxides.* The roasted powder, mixed with an abundance
of charcoal dust and heated to moderate redness in a closed
glass tube, gave a brownish sublimate (cadmium oxide), and
on stronger heating a less volatile sublimate, yellow while hot,
white on cooling (zine oxide).

On roasting 0:0014 grm. in a small platinum dish, dissolving
the residue in sulphuric acid, evaporating and gently igniting,
a residue was obtained which weighed 0°002 grm. Theory
requires from pure cadmium sulphide 0:00202 grm. of cadmium
sulphate; but as a mixture of sphalerite and calcite similarly
treated might give the same figures, this result does not
indicate with any certainty the proportion of cadmium in this
lot of material.

A second lot, also weighing 0°:0014 grm., was dissolved in
nitric acid, evaporated to dryness with sulphuric acid, and
heated to expel all free acid, the residue treated with a single
drop of hydrochloric acid, sp. gr. 1:20, and dissolved in four

* When not too strongly heated at first, the yellow coating assumes a

transient dark brown or reddish-brown color, and becomes yellow again on
cooling. !

8 Cornwall—Greenockite on Calcite from Joplin, Mo.

cubic centimeters of water. The whole was then saturated with
hydrogen sulphide gas, giving very quickly a yellow precipitate,
and this, after letting the gas act long enough, was filtered off,
dissolved in strong hydrochloric acid, evaporated with sulphurie
acid, ignited and weighed as sulphate. By this means the
coating tested was found to contain 23-07 per cent of cadmium,
or 29°66 of cadmium sulphide. It should be said that the
coating yields sulphur dioxide when roasted.

By comparing the liquid holding in suspension this cad-
mium sulphide with another in which a known amount of cad-
mium sulphide had been precipitated, it had already been
estimated that the yellow coating on the calcite contained not
less than 15 per cent of cadmium.

Finally, although the Joplin sphalerite on which greenockite
occurs will give a distinct reaction for cadmium when roasted
and reduced with charcoal in the closed tube, in the manner
already described, yet it was found that no precipitate could
be obtained from it by hydrogen sulphide, even after thirty
minutes action, the treatment being quantitatively, and in
every other respect, exactly the same as described for the
yellow coating on the calcite. A mixture corresponding to a
sphalerite containing 5 per cent of cadmium yielded a precipi-
tate of cadmium sulphide, under identical conditions, in a few
minutes. Dana (System of Mineralogy, 6th ed.) states that
the amount of cadmium present in any sphalerite thus far ana-
lyzed is less than 5 per cent.

It therefore appears certain that this yellow coating on the
calcite, like that on the Joplin sphalerite, is greenockite.

John C. Green School of Science,
Princeton University.

Cumings and Mauck— Variation in the Fossil, ete. 9

Art. IIIL.—A Quantitative serials of Variation in the Hossil
Brachiopod Platystrophia lynz;* by EpGar R. Cumines
and ABRAM V. Maucr. (With Tinine II and III.)

DuRinG the year 1900-01 the authors made extensive collec-
tions of the brachiopod Platystrophia lynx from the Upper
Ordovician rocks at Vevay, Indiana (Switzerland Co.).

Inasmuch as this species is extremely variable and at the
same time exceptionally abundant and well preserved, a quan-
titative study of the variations presented at once suggested
itself. Such an investigation is the more warranted because of
the different opinions current as to the taxonomic importance
of several of the forms under which Platystrophia presents
itself.

The specimens used in this investigation were collected from
a zone which at Vevay extends from 240 ft. to 360 ft. above
the level of the Ohio River. Part of the material was so col-
lected that the precise layer is known from which a given
specimen came; the object being to determine as accurately as
possible the factor of range.t The majority of the specimens
used are from the upper 50 or 60 feet, and the upper 20 feet
contains by far the most as well as the best of the material.
Examples could not be obtained in sufficient abundance from
the base of the Platystrophia zone to make a quantitative
treatment possible; but we were able to ascertain that in the
lower part of its range at Vevay, and in general throughout
the Ohio Valley, Platystrophia presents less variation than
at higher horizons ; and that in the lower beds only the small
pauci-plicate form is present.§

Beside the material collected layer by layer, a large collec-
tion was made from the upper beds wherever exposures could
be found in the vicinity of Vevay.

The data taken for study are: Ratio of width to length of
shell (equals shell index); ratio of depth to breadth of sinus

* Presented before Section E at the Denver meeting of the American Asso-
ciation for the Advancement of Science, August, 1901.

+ For synomony and bibliography of Platystrophia see Davidson, Silurian
Brachiopoda, 1871, p. 268; Hall and Clarke, Pal. N. Y., vol. viii, pt. i,
pp. 200, 201; Schuchert, Am. Foss. Brachiopoda, Bull. U. S. G. S., No. 87,
pp. 308-310.

¢ The Vevay section is published in the Am. Geol., vol. xxviii, Dec., 1901,
pp. 361-381. By reference to page 373 of that paper it will be seen that the
lowest specimens of Platystrophia (associated with Dekayia ulrichi) are small
and of a type similar to var. laticosta or var. dentata Meek (crassa James).
The form lynx does not come in till within 30 or 40 feet of the top of the sec-
tion.

§ Specimens have been examined from these lower zones at Cincinnati,
Ohio, Lawrenceburg, Aurora, Rising Sun, Vevay and numerous points on
Laughery Creek in Indiana.

10 Cumings and Mauck— Variation in the

(equals sinus- index); number of plications on ventral valve ;
number of plications on dorsal valve; number of plications in
sinus; number of plications on fold. 4

The width of the shell is in every case the greatest width,

whether this occurs at the hinge line or farther forward. It

was obtained by means of adjustable spring calipers. The
width of the sinus was obtained by spreading the points of

spring dividers between the anterior lateral angles of the sinus.

The depth of the sinus was obtained by spreading the points
of the dividers from one anterior lateral angle of the sinus to
the anterior extremity of the first adjacent furrow in the bot-
tom of the sinus. All measurements were read off on a milli-
meter scale and are correct to within 0°25™™.

The number of shells used varies for the several characters
indicated above, on account of imperfect material. Only
entire shells were used for measurements; but the number of
plications, especially the number in the sinus, may frequently
be determined with accuracy on very poorly preserved speci-
mens. The number of shells used in determining any one
sharacter is called a group.

Group I, Shell Index.

Width divided by length. Number of shells used, 679.
Range of variation from 1:0 to 1:8. Modet at 1°3, with fre-

quency of 300-4.{ Variation here is moderate in amount, and —

in the direction of greater width than length.

Group IL, Sinus Index.

Width of sinus divided by depth of sinus. Number of shells
used, 664. Range of variation from 0-9 to 3:0. Mode at 1-7
with a frequency of 112-2. Variation great in amount, and in
the direction of a shallow sinus.

Group ITI, Number of Plications on Ventral Valve.

Number of shells used, 1173. Range of variation from 10
to 28. Mode at 17 with a frequency of 176°3. Variation

large in amount, and in the direction’of the greater number of

plications.

* Prof. Shaler has given (Foss. Brach., Ohio Valley, p. 48) a series of meas-
urements of a limited number (20) of specimens of Platystrophia. His series
includes also forms from Richmond, Indiana, and other American localities.
In a forthcoming paper I shall take up the discussion of material from all
the provinces both American and European, where Platystrophia is known
to oceur.—E. R. C.

+ The term mode means the highest point of the curve, i. e., the class with
the greatest frequency.

t Forthe purposes of comparison all frequencies were Teduced tofreqtencies ~

per thousand.

0968-0 |9TET| | = | | | | | | | | | | | | I | F | CP | SSI | 988 | 0g | A dnorp
6802-0 |STFT a: Je! 0) i OFC re0t 881 Te) AT dno
SUIONHNOMU YT ALATOSay

{ . ns. Sera Se eee cole elle |68-0 99-8 |e9-28 [P9-eeTles-eoul6c-2e |e6-L1) A dnorg
IPL |PO-F |TL-68 |1S-FLTIGT-GeL/16-98148-FL| AT dnory
a aes lear el Neen Pagers, all: 8 hi 9 ¢ i 8 S I SOSSVTO
‘GNVSNOH YT, UAd SHIONDTAOAU YT F
GLIe-% |ISh ‘a ae & g E Geleoa nor a | 14v | 09 | 06 | 7 | of | @ | ge | i ee I | | MIIT dnoay
C6G8-0 [GATT LG es 20s On Tee hog) | Gens enh) air | kom] tet | eet 19 he ke ole ee | TT dno
‘SHIONGNOUUY ALATOSAYy
1-8/18-% |eP-F |eh-P [99-9 |0G-ST/SL-LT/LF-GE|F9-L6|18-FOT|G0-EET|GC-661|GL-611/98-OTLIG1-86 leP-ce |6g-F% l08-EL |Te-2 w{TT Inoay
G8- OL-T |OL-T |LT-G |6G-8 |06-LT|FS- 18/68-0G|80-ST1/6P-STL6G-CET9F-9LT/06-OST|TP-LOLILP-eP |T¢-48 196-4 |9¢-F |e9-0 | IT dnory
TOCi Ieee Ce Caro ete aoe eres Ole | Bi Yh.) Or: | op | Fr: | ere are anon Sassy)
‘ANVSOOHY, Ud SHIONTOAOAU
090¢-T [p99 | I [ | Lc |e | 0 | ¥ | 9 | 88 | | & | 09 | 69 | Pe | 4 Op) Fen| een cae | 9 | g Il dnosy
LOLD-T_|6L9 Ab ae | Ge See 23 GE LO PPL | POEs) - TAL eg li I dnotp
Q
ears ‘SHIONT OOM, GLA TOSTY
a B
° re ey
SB | «8 (09. 1|0G-1/10-€/F0-21/L¢-61|@1-08|29- 1e/GT-68|@e-4C/6L-GL'TS- 81/9806 |16-G0T/6-C11/26-90T|8L-0L |F8-09 [0%-1¢ |eT-¢8 [$0.6 [80-6 | &-F TL dnory
om | 85 | 6-6 08-OF |PS-1¢ |29-86 |90-818/SF-008|¢9-998|96-G¢ |LF-T ] dnoip
Po 0 GieOmieeelemcsindcel Gcdleterl tra CC IeenicOe Ore el (kek Ot | OT | pL ete let | TT 10-1 | 60 SO8SR1O
a3 | 2 $113 eee
He & ‘GNVSQOHY], UAd SHIONTAOAUT
le) (7)

12 Cumings and Mauck— Variation in the

Group Illa, Plications on Dorsal Valve.
Number of shells used, 451. Range of variation from 11 to
29. Mode at 18 with frequency of 199:°55. Variation corre-
lated with that of Group III.

Group IV, Plications in Sinus.
Number of shells used, 1412. Range of variation from 1 to 7.
Mode at 3 with frequency of 729-1. Variation in the direc-
tion of the larger number of plications.

Group V, FPlications on Fold.

Number of shells used, 1116. Range of variation from 2 to
8. Mode at 4, with a frequency of 793°8. Variation precisely
correlated with that of Group LV.*

The complete data are charted on page 11. The method of ;

constructing the curves of fig. 1 is the usual one of laying off
the classes along the axis of abscissas and the frequencies along
corresponding ordinates.

The lack of correspondence between Groups III and Ila,
which should be exactly correlated, is due to the relatively
small number of specimens used in Group Illa. It forcibly
emphasizes the necessity of using as large series of specimens
as possible.

All the curves are skewed. None of them show more than
one mode. They all rise abruptly and fall off less abruptly.
This shows a certain correlation of the different variants.

An attempt was made to determine the relation of the num-
ber of plications in the sinus to the number on the entire
valve. Hallsays: ‘*The prevailing number of lateral plica-
tions is seven on each side of the sinus or lobe; and so long as
the mesial plications remain three and four, so long the lateral
ones are seven. As soon as there is even an appearance of a
departure from this number on the mesial lobe and sinus, and
where the rudiment of an additional plait is visible, we then
find the lateral plaits to be nine or ten.” t

Our material does not bear out this statement. ‘To test the
point it was assumed that there is a precise correlation between
number of plications in the sinus and number of plications

on the whole valve. The area of the freyuency polygon for

any given number of plications in the sinus should then be
equal to the sum of the areas of the frequency polygons of a
definite number of plications on the valve. For example, the
frequency polygon of three to four plications in the sinus has
an area equal to that of the sum of the areas of the frequency

* While there is necessarily one more plication on the fold than in the
sinus, the fact that many specimens, owing to state of preservation, can be
used for the determination of one group that are not available for the other,

makes the two groups supplement each other.
+ Hall, Pal. N. Y., vol. i, 1847, p. 134.

———— ee

a

Fossil Brachiopod Platystrophia lynx. ae Es:

polygons of seventeen, eighteen, nineteen, and twenty plica-
tions on the ventral valve. Therefore any shell with four
plications in the sinus should have between seventeen and
twenty (both numbers inclusive) plications on the valve. It
should have no more and no fewer.

An inspection of our material does not show this to be the

ye tv

b eve OV IIla III

Fieg 1. Curves of Groups I to V. On the axis of ordinates, each small
division represents four individuals. On the axis of abscissas, each fifth (as
from a to b) represents one class.

ease. In a series of 153 specimens with four plications in the
sinus, the range is from 16 to 26 plications on the ventral
valve. The average is 18°8 plications. The specimens with
two plications in the sinus have from 11 to 22 plications on the
valve. Those with one plication in the sinus have from 10 to
15 on the valve. The correlation is approximate rather than
precise. ,

II

14 Cumings and Mauck— Variation in the

An inspection of Plates II and III shows the small size of
the extreme forms, and conversely the large size and robust
growth of those forms near the modes. The largest specimens
will not be found exactly at the mode in any of the characters
for which curves are plotted, but in classes to the right of the
mode. That is, while it is true, for example, that by far the
majority of specimens have three plications in the sinus, the
largest specimens are more likely to be found in classes having
more than three plications in the sinus. They will never be
found in classes having less than three plications. When, how-
ever, the class falls some distance to the right of the mode it
will again be found to contain small specimens. Smallness of
size is, therefore, an accompaniment of the general extinction
that prevails more and more, away from the mode. In the
case of the number of plications on the valve, extinction does
not become severe till the 20th is reached; and in the classes
between 17 and 20 will be found the largest individuals.

In regard to the validity of the several species or varieties
into which the dynz group of forms has been divided, namely,
Platystrophia lynx, P. laticosta, P. dentata Meek (= P.
crassa James), there is absolutely no character or combination
of characters that can be relied upon to separate any large col-
lection into distinct species. To a limited extent the above
forms differ in range; although the authors have frequently
seen all three represented on a single slab of limestone. It is
well, however, to distinguish for stratigraphic purposes such
varieties as laticosta, and dentata Meek.*

The following is a description of the modal form of Platy-
strophia lynx: shell three-tenths wider than long; greatest
width about half-way from the hinge line to the front of the
shell; width of sinus seven-tenths greater than its depth;
seventeen angular plications on the ventral valve, three of
which are in the sinus; eighteen on the dorsal valve, four of
which are on the fold.

* As stated in a preceding footnote, the other forms, acutilirata Conrad,
biforata Schl. and other foreign forms of Platystrophia are not considered in
this paper. The form identified by Miller, and doubtfully by Meek, as
dentata Pander is the costata of the latter author. Pander has figured
dentata and costata (= chama Eichwald) side by side on pl. xi of his Beitrage ~
zur Geog. Russ. Dentata has two plications in the sinus and three on the
fold, while costata has one in the sinus and two on the fold. The name
chama Eichwald cannot be retained for this variety, since the Spirifer cos-
tatus of Sowerby is a true Spirifer, and there is no other prior use of the
name costata among the Orthide. See de Verneuil Geol. de la Russie, p. 140,
and Sowerby, Tr. Geol. Soc. Lond., 2d ser., v, pl. lv, fig. 5-7. On the vari-
eties of Platystrophia see especially Meek, Pal. Ohio, i, 1878, pp. 112-121;
Hall, Pal. N. Y., i, 1847, pp. 133-1384; Winchell and Schuchert, Minn. Geol.
Surv., iii, 1893, pp. 454-457; Davidson, British Silurian Brachiopoda, 1871,
p. 268 et seq. ; Schuchert, Bull. 87, U.S. G.S., p. 308; Sardeson, Am. Geol.,
xix, p. 109.

‘OSVO TOVS UL SuUOTZVOTTC JO TOG MANU 0} AoJor SONY oY} MOlOd SLOquINNT “POF
WO SUOTZVOTT ‘OAISNPOUL Ge 07 LZ ‘SSI ‘“SNUIS UL SUOTZVOTTd ‘OATSHPOUL OY 07 OY “SSL ‘“OATVA [RIZUOA UO SUOTZVOTTA ‘OATSHPOUL GT O} T ‘SST

nal

Tan)
4 L 9 G p @ g V 9
C( wm wow qa WY

8 ee ite 08 6e 8¢ Le 9% rd

> OG ; > ie) o AiG ) NO

= G p @ g i 8g Le 92

n>

S

s

va)

& WZ WW)

Ss

_ *

Ry 7 eg ra 12 0% 61 ST Ni

=

& GB i GB 8 Ie 08 61 81

nS ~

SO

: VW Q

~

RQ

>

% 9T Cy a SI ra iA Or 6
it 91 CT val eT ral ia OT
8 L 9 G v g z I

‘Tl Bavig

yove Ul XOpUl OY} JO ONT[VA Yj SAIS selnsy oy} MOloq s1equInNy

= QD
D

~sS

"~

va)

(a)

Ss oe
D

< &.3
‘=

i) ;
ES

Ss

oe

S re
NK

1. ee
2

S

<< @Y
BS

= 91
a

~ hel
~:

=

=

SS)

16

0-€

ND

GG

!

WD

&G

V1

AW)

6-6

al

Yes)
=

8-6

0-6

‘TIL Sivitd

LG

QD

"XOPUL SNUIS ‘OATSN[IUL TE OF OT ‘SST

9-6

61

0-1

"OsBo
‘XOPUL [JOYS ‘OAISN[OUL G OF T ‘SOLA

LS a
9% von Coe OG
dail On
ST Fe
6-0 8.7
OT 6
Tal 0-T
z ji

Wortman—Studies of Eocene Mammalia, ete. a7

Art. IV. — Studies of Eocene Mammalia in the Marsh
Collection, Peabody Museum; by J. L. WortTMAN.

[Continued from vol. xiii, p. 448.]

Sinopa minor sp. nov.

A SMALLER species is represented in the Marsh collection by
numerous remains, all in a somewhat fragmentary condition.
Of these, I select as
the type a_consider-
able portion of a man-
dibular ramus of the
right side, contain-
ing the three pos-
terior premolars and
the first and second
molars in place, fig- eee
ure .96. Thereisalso FIGURE 96.—Right lower jaw of Sinopa minor
the anterior part of Wortman; side view; natural size. (Type.)

the opposite ramus, which contains the root of the canine and
those of the anterior premolars. With this, I associate as a
cotype a fragment of a right upper jaw bearing the fourth
premolar and first molar, figure 97. The
characters of this species, as exhibited by
the type and cotype, are as follows: The
size is much smaller than either S. rapaz or
S. gracilis ; the canine is of the usual pro-
portions; the small single-rooted first pre-
molar is implanted some distance behind the _ FicuRe 97.—Frag-
: : 2 ment of aright upper
canine, and is separated from it and the jaw of Sinopa minor
second premolar by considerable diastemata; Wortman; crown
the second premolar is two-rooted and sepa- View; natural size.
rated from the third by a short interval, (Cot?)
differing in this respect from both S. rapax and S. gracilis ;
the third and fourth premolars have the usual form ; the first
molar is relatively small, with low trigon and a large basin-
shaped heel; the second molar is larger, with high trigon and
large heel; the internal cusps are but little reduced, differing
in this respect from S. gracilis and agreeing with S. rapaw ;
the last molar is not preserved in the type; the fourth superior
premolar is like that of S. gracilis, except that the internal
cusp is more conical; the postero-external blade exhibits
faint traces of a bilobed condition, which is not seen in SV.
gracilis ; the two main external cusps of the first molar are
well separated ; there is a more distinct posterior intermediate,
and an internal cingulum, all of which are different from S.
gracilis.

Am. Jour. Sci.—FourtsH Serizs, Vou. XIV, No. 79.—Juny, 1902.
2

96

18 Wortman—Studies of Eocene Mammalia in the

There are three or four individuals that agree perfectly with
the type and cotype, but there are others that have a heavier
build and are perceptibly larger. I cannot detect any further
differences in the materials at hand, however, which would
warrant referring the latter to another species. It may be
that more complete specimens will necessitate their separation.
Additional characters derived from these specimens are:
Second lower molar equal to, or slightly larger than, third;
last superior molar without postero-external cusp, and second
superior molar, as well as first, with distinct internal cingula.

The measurements of the type are as follows:

Length of inferior dental series from base of canine,

exclusive of ‘lastmolar. 2 0007. oo 4. ee ee 42° mm
Length. of ‘pnemolars-.22 222 280. ee | Be eee 30°
Length of first _and second molars .._-.--. 5-222. fae 12°
Length of entire inferior dental series (estimated) ..-- - -- 48°5
Antero-posterior diameter of first molar._....-...---2-. ©
Antero-posterior diameter of second molar _---_.-------- os
Antero-posterior diameter of third molar (not type) ----- 64

Measurements of Cotype.

Length of fourth premolar and first molar __.......----- 18°
Antero-posterior diameter of fourth premolar_....-..-.. ©
Antero-posterior diameter of first molar.._...--.------- di
Transverse diameter of fourth premolar. ...-.---------- 5°
Aransyerse diameter of first molar: /2_550-.. 40‘) 25a 4°6

The type specimen was found by Professor Marsh, at Griz-
zly Buttes, and the cotype, by Mr. Kinney, at the same locality.
Other specimens are from Church Buttes, and one is doubtfully
from Henry’s Fork.

Sinopa major sp. nov.

The last species of this group to be considered is the largest
yet found in the Bridger horizon. It is apparently unde-
scribed, and the above name is proposed for it. Two reason-
ably complete mandibular rami, figures 98, 99, one of which
contains nearly all the teeth in place, together with three or
four other fragmentary pieces carrying various teeth, repre-
sent it in the collection. The type specimen has the three
molars and the third and fourth premolars in place in the jaw,
together with the roots and alveoli of the remaining premolars.
The chief characters of the species, as exhibited by this speci-
men, may be stated as follows: It is much larger than S. edaz,
S. gracilis, or S. minor ; the second premolar is separated but
very slightly, if at all, from the third; the first is small, single-
rooted, and placed about equidistant from the canine and
second premolar; the second molar is slightly larger than the

Marsh Collection, Peabody Museum. 19

third; the heels of the molars are large and basin-shaped, that
of the last bemg smaller than the other two; the shear is
rather transverse, and the internal cusp is less reduced than in
S. gracilis ; the jaw is much more massive and deeper, and
the distance from the posterior base of the canine to the third
premolar is proportionally shorter than in the other species.
This indicates a rather short-muzzled type. The specimens

98

FIGURE 98.—Right lower jaw of Sinopa major Wortman; side view; nat-
ural size. (Type.)

99

FIGuRE 99.—Left lower jaw of Sinopa major Wortman ; side view; three-
fourths natural size; restored from three specimens.

suggest an animal intermediate in size between a red fox and a
prairie wolf.
The principal measurements are:

Length of inferior dental series from posterior base of canine 70°™™

SemermetmpmeBBOL AES 2-2 2 2 ee bee eee eek 44°
Peer IMMeE AIS SS 558 ies Joes e+ = 4+ = --- 26°
Antero-posterior diameter of first molar___----.--------- 8:
Antero-posterior diameter of second molar.--.----------- 10°
Antero-posterior diameter of third molar ._--.-----.------ -F
Peo: jaw at sceond molar ____-...2_-_-.-...------- 20°
Meaieos jaw at. second premolar__.-.....-_.----------- 16°

The type specimen was found by Sam Smith at Church
Buttes.

20 Wortman—Studies of Eocene Mammalia in the

Discussion.—Two other species of this genus, aculeata and

insectivoru, were described by Cope from the Bridger, but
Matthew has shown* that the types are unrecognizable, and
he suggests that the names be dropped.

That the Bridger species should be more advanced than
those of the Wind River and Wasatch is quite natural. At
least three of them thus far known do not show any very great
advance in structure over their older representatives, but still
enough to separate them specifically. The remaining one, S.
pas exhibits very decided progress in the direction of the

igher specialization of the family. This is seen in the disap-
pearance of the posterior external cusp of the third superior
molar, the more decidedly connate condition of the two main
external cusps of the first and second superior molars, the
development of a more sectorial structure of the fourth pre-
molar, reduction in size of the heels of the lower molars, and
reduction of the internal cusps of the trigon, with a more
longitudinal shear. The internal cusps of the upper molars
are also reduced.

There appears to be very good evidence of the specific
descent of this species. Thus, Sinopa hians of the Wasatch,
as far as its structure is known, answers with tolerable accu-
racy to the ancestral requirements. Sinopa Whitie of the

intervening Wind River horizon appears to be intermediate in |

every feature of its osteology, as far as known, with the single
exception that the second lower premolar is not spaced. If
this character were variable in the Wasatch species, 1t may
well prove to be the annectent form.

That Hyenodon and very probably Pterodon, also, were
derived from this group, there appears to be very little doubt.
The fundamental similarity in the structure of the skull, teeth,
vertebrae, pelvic girdle, limb bones, and carpus, cannot be acci-
dental parallelism, but on the contrary affords very strong pre-
sumptive proof of genetic aftinity. The most specialized
species of the genus, Sinopa agilis, however, cannot be in the
line of direct descent because of the relatively small skull and
the sharp, compressed, and slightly fissured, bony claws ; whereas
in Hyenodon, at least, the skull is large in proportion to the
skeleton, and the claws are flat and deeply fissured.

It should be noted here that nearly every specimen of the
genus Sinopa represented in the Marsh collection comes from
the lower part of the Bridger beds, and that not more than
one or two have been found in the upper part of the horizon.
Their complete absence from the Uinta, as far as we now
know, is, moreover, fairly satisfactory proof of their extinction

*Bull. Amer. Mus. Nat. Hist., Jan., 1901, p. 24.

’

Marsh Collection, Peabody Museum. 21

or migration from this country at the close of the Bridger.

For this reason, I am persuaded that the evolution of the Oli-
gocene types took place probably in Asia or the north, during
later Eocene time, and that their subsequent appearance in this
country during the Oligocene was due wholly to migration. It
is highly probable that some generalized Wasatch species, such
as S. opisthotoma of Matthew, was the ancestral type from
which they were derived.

Further Observations upon the Marsupial or Metatherian
Relationship of the Creodonts.

In the foregoing descriptions of the various characters of
the Creodonts, frequent reference has been made to their rela-
tions with the Marsupials. Objections will doubtless be raised
to the use of this term, since some writers restrict the term
“ Marsupial” to the living representatives of the group, which
is perhaps, strictly speaking, correct. Huxley proposed the
term Metatheria for a hypothetical group, which was meant to
include both the modern Marsupials and the immediate Impla-
eental forerunners of the Eutheria. The all-important dis-
tinction of such a group would consist in the implacental
method of its reproduction—a character which would at once
separate it from the Eutheria, in which an allantoic placenta
is formed. In like manner, the absence of a distinct coracoid,
the union of the odontoid process with the body of the axis,
the lack of oviparous habits, and the more highly developed
reproductive system would distinguish it from the Prototheria.

Unfortunately there are no known characters of the skele-
ton in this group which are constant associates or infallible
correlatives of the implacental mode of reproduction, and since,
among the fossils, we are compelled to depend solely upon
osteological evidence, our judgments must of necessity rest
very largely on analogy. When I have spoken of “ Marsupial
characters” and “ Marsupial relationship,’ I have had con-
stantly in mind the Implacental Metatheria of Huxley, prefer-
ring the use of the term “ Marsupial,” because the large number
of primitive characters exhibited in the living members of this
group undoubtedly furnish us the safest ouide tor a proper
interpretation of similar characters in the fossils. By taking
the more primitive members of the existing Marsupials as the
basis of our comparisons, | am convinced that we shall be able
to arrive at a very much clearer understanding of what the
ancestors of the Creodonta were like, than by any other method
of study. That they were der ivatives or offshoots of any pre-
existing group of Placentals or Eutherians is exceedingly
unlikely, and the strongest evidence of this fact is that they
are practically as low in the seale of organization as any known

22 Wortman—Studies of Hocene Mammalia in the

Eutheria. Moreover, evidence of the existence of such a group
remains to be discovered. On the contrary, all the facts point
very strongly to their origin, along with the Carnassidentia,
from Implacental or Marsupial Metatherians. It is likewise
conceivable that from this same general substratum the other
Eutherian orders arose.

Some day, when our knowledge of these matters is vastly
more extensive and accurate than at present, it will perhaps be
necessary to abandon this horizontal system of classification,
implied by the use of such terms as Prototheria, Metatheria,
and Eutheria, and substitute for it the phyletic or linear sys-
tem—the only one expressive of the genetic relationship which
we seek to discover. It will then be determined, without
much doubt, that these several ordinal groups have recogniz-
able chains of ancestry, penetrating not only well down into
the Metatherian substratum, but almost to the very bottom
or beginnings of Mammalian existence. We shall then give
names to these lines of descent rather than to successive stages
of their development. But until this Utopia in the subject is
realized, we must content ourselves with cruder methods, more
in keeping with our ignorance.

Summary.

Having now completed a study of all the Eocene Carnivora
in the Marsh collection, I append herewith a brief summary
of the more important discoveries, opinions, discussions, ete.,
embodied in the foregoing paper. They are the following:
The general organization of the order and the relationship of
its more primitive members to the Metatherian Marsupials are
discussed ; the order is divided into three suborders, Creodonta,
Carnassidentia, and Pinnipedia, and their relations are con-
sidered ; of the Eocene Canide, the type of the genus Vul-
pavus is figured, a new species added, much of its osteology
described, and the progressive modification of the family con-
sidered; a new related genus, Veovulpavus, is proposed ; addi-
tional characters of Uintacyon, together with its position in
the group, are given; Prodaphenus is considered to be the
forerunner of the Amphicyon series, and four main lines of
canine descent are pointed out; the Viverravide are defined,
the type species of Viverravus is figured, and another species
added; a new genus of this family, Oddectes, is proposed and
a large part of its osteology described ; the relations of the
Viverravide to the living Viverride are examined, .and the
position is taken that the descent of the modern civets is prob-
ably traceable to this source; the type species of the three
genera T’riacodon, Ziphacodon, and Harpalodon, are figured,
and the opinion expressed that they are not valid genera; the

Marsh Collection, Peabody Museum. 23

type of “Alurothervum is figured and described, a new species
added, and its ancestral relationship to the Felidz discussed
and reaffirmed ; the organization of the Creodonta and their
relations to the Metatherian Marsupials are considered, a new
classification is proposed, and the families are redefined; the
Viverravide and Palzeonictidz are removed to the Carnassi-
dentia ; the family Mesonychide is considered, the genera of
the Mesonychine are defined, a new genus, Harpagolestes, is
proposed and defined, and the osteology of Dromocyon vorax
given in full, with numerous illustrations; the small pelvic
outlet may have indicated extreme helplessness of the young at
the time of birth, like the Marsupials, and a possible cause for
the extinction of the line; the evolution of the phylum is con-
sidered, and the progressive modifications of the teeth and
limbs are pointed out; the origin and homologies of the mam-
malian tritubercular molar are fully discussed, and dissent is
expressed from the theory of Migration or Rotation of Osborn ;
the family Oxyznide is considered, the arrangement of the
genera discussed, and two subfamilies, Oxyzeninee and Limno-
-eyoninee, are proposed and defined; the genera of the Oxye-
hinge are defined, the type specimen of Patriofelis ferowx is
fieured and described in detail, and the probable habits of the
species are rediscussed at length; the genera of the Limno-
cyonine are defined, one new species is added, and the synon-
ymy of the others given; many new points are added to a
knowledge of this group, its possible Insectivorous relation-
ship. is pointed out, and one species, at least, is thought to have
been aguatic in habit; the family Hyzenodontide is redefined,
and the Bridger species of Stnopa, of which two are added,
are described in detail; much of the osteology of Sinopa agilis,
with illustrations of skull, fore foot, and limb bones, is given ;
the probable origin of the Hyzenodonts and Pterodonts is con-
sidered, some further observations on the Marsupial or Meta-
therian relationship of the Creodonts are made, and their
probable origin is considered.

24 P. EF. Schneider— Eruptive Dikes in Syracuse.

Art. V .—New Exposures of Eruptive Dikes in Syracuse,
NYS by Pe rite Sci nmin: |

Igneous rocks in the horizontally stratified Paleozoic beds
- of Central New York are too rare to pass unrecorded; and
when, recently, excavations in Syracuse for the Butternut street
trunk sewer disclosed another of these occurrences in a new
locality and at so great a depth that ordinary excavations had
not reached it because of the thickness of the overlying drift,
it became important that some permanent and available record
_ should be made of the same.

The eruptive rock was first noticed April 16, 1901, some
three days after it was first penetrated. At this point, a short
distance beyond the place where the sewer crosses Highland
st., the eruptive rock occurred in the bottom of the trench and
was dug into only to a depth of some two feet. It was over-
laid by nearly three feet of decomposed peridotite, which had
been entirely changed to a soft greenish-yellow earth. As the
excavation progressed to the east of Highland st., the work-
men penetrated deeper into the rock, which for some distance
presented a slightly stratified appearance suggesting a sheet
branching from the dike proper, which subsequent excavations
proved to be the case.

The dike itself was first encountered 126 feet east of the
center of Highland st., and was so hard and firm as to be
removed with great difficulty. The width of the dike is 36
feet and it comes up to within ten feet of the natural surface.
From its location in the trench, which was five feet in width,
the strike of the dike appeared te be N..5° EB Wier ae
scarcely any sheet to the east of the dike, but to the westward
it extended over three hundred feet.

The rock in the main dike, with the exception of the upper
two feet, is perfectly hard and firm. It is of a dark green
color, some of it being almost black, and contains an abund-
ance of apparently jet black crystals. The upper ‘portion,
immediately beneath the drift, had changed to a soft greenish-
yellow earth, in some places to a yellowish earth. The fact
that the lower portion of. the drift contained much of the
serpentinous earth mixed with it would suggest that a con-
siderable area was covered with eruptive matter. The typical
rock contained few inclusions as compared with that at De-
Witt, N. Y., or even that in the Syracuse dikes at Green st.
The softer rock of the sheet, however, contained many of
them. No prominent fossils were found in any of these
inclusions, whereas in the rock at De Witt they were very

P. F. Schneider—Eruptive Dikes in Syracuse. 25

abundant. No traces of the enclosing walls of the dike could
be found. Sheet material banked it on the west, and heavy
Pleistocene clays with quicksand beyond them formed the
eastern border. At Green st. the enclosing walls are perfectly
shown, and it was hoped that further excavations would open
up other exposures showing the contact phenomena, but this
did not oceur. ‘The other excavations, where the strike of the
dike would seem to indicate its existence, were all on higher
ground with heavy mantles of drift which even this deep
sewer did not penetrate.

_ One other opening occurred in a return sewer on Highland
st. The excavations passed for 180 feet through rock which
was thought at that time might be merely sheet material from
the dike, but I am now convinced that it was another parallel
dike. This rock, while almost as difficult to excavate as that
of the first dike, decayed very rapidly after a few days expo-
sure. It also had more or less of a massive wedged appearance
in the trench like the first dike, and quite unlike the banded
appearance of the sheet, and furthermore contained many
inclusions. It also contained numerous small red crystals,
the “rubies,” which the neighboring school children collected
in abundance. None of these peculiar forms were found in
the De Witt, or in any of the Syracuse dikes at Green st.
Some of them appeared to be perfectly crystallized garnets,
but so rapidly did this rock break up, especially when dry,
that they usually fractured soon after being exposed. A
few crystals of greenish color were also obtained from this
same rock, but none of either kind were noticed in the hard,
firm variety. All of these facts would seem to indicate a
second dike more or less parallel to the first and less than 250
feet away. Through this second dike the excavations must
have passed very nearly longitudinally, while the sewer proper
erossed the main dike at nearly right angles. As the return
sewer stopped when connection had been made with the sewer
proper, and as the excavations up to this point did not pass
through to the farther side of the dike, no facts as to its width
can be given. The proximity of these dikes to those at Green
‘st., which are less than a mile away, suggests some under-
ground connection, and inasmuch as their general direction is
the same they may be merely a continuation of those dikes.
The intervening space has frequently been trenched, and at
such times the excavations have been carefully watched for
evidence of the dikes without revealing any trace of them.

Syracuse, N. Y.

26 Smyth, Jr.—Petrography of Dikes in Syracuse, N. Y.

Art. Vil.—Petrography of Recently Discovered Dikes in
Syracuse, V. ¥.; with Note on the Presence of Melilite
in the Green Street Dike; by C. H. SmytuH, Jr.

THE freshest specimens of the dike rock from Butternut st.,
Syracuse, N. Y., are nearly black, but in the average material the
color is gray, with a more or less decided greenish tinge. In tex-
ture, the rock is distinctly porphyritic, with phenocrysts ranging
up to 15™™ or more in diameter, scattered somewhat unevenly
through a moderately fine and even-grained groundmass.
Occasional veinlets and small pockets of calcite appear, but
these are exceptional, and the rock as a whole, aside from the
“ sheet ” material mentioned below, is quite homogeneous. Of
the specimens sent to the writer for examination, many appear
to the unaided eye quite fresh and unchanged from their
original condition; but they are very deceptive, as in most
eases alteration has progressed to a considerable degree. Sev-
eral sections cut from sound specimens of good color and with
lustrous dark phenocrysts, under the microscope show little
but a mass of alteration products, which, however, retain the
original texture of the rock very well. In consequence of this
alteration the following description of the petrography of the
dikes is based upon only a few sections of fresh material,
together with such data as are afforded by the altered speci-
mens. But the petrography is comparatively simple, and the
data at hand are sufficient to show a substantial agreement
with the other dike rocks of the vicinity.

Sections of the least altered material show that the pheno-
crysts are olivine, as indicated megascopically. Crystal outline is
generally lacking or at most quite imperfect, and the irregular
grains show the customary alteration to serpentine along cracks |
and around the margin. In some grains this alteration is but
slight, while in others the serpentine has wholly replaced the
olivine. Between the two extremes there is complete grada-
tion. Associated with the olivine there is a very little colorless
monoclinic pyroxene. It occurs in irregular grains, and shows
the usual high extinction angles. Though in very small
amount, this pyroxene is a fairly constant constituent of the
rock.

Biotite of a pale brown tint appears in large irregular plates
and shreds of small size. The former are primary, but the
appearance of the shreds and their relations to the other min-
erals suggest that they may be, in part, secondary. Perofskite
is abundant, though varying considerably in quantity in dif-
ferent sections. It appears in minute crystals of sharp outline.
In some cases, high powers show these crystals to be octahedra, —

Smyth, Jr.—Petrography of Dikes in Syracuse, N.Y. 27

and this is probably, as would be expected, the form of all of
them. They are honey yellow, and decidedly translucent, so
much so indeed as to show this property clearly with low
powers. In this respect, the perofskite differsfrom that of the
De Witt dike, in which Prof. Kemp* perceived the translucent
character only with high powers. It would be difficult to find
better microscopic specimens of perofskite than are shown in
the rock here described.

Magnetite is another primary mineral, occurring in the usual
small crystals and irregular grains. But it is far less abundant
than is often the case in rocks of this general type. In some
sections there is ‘a small quantity of a granular, isotropic min-
eral with high index of refraction. While this may be garnet,
no exact determination of the species could be made. In
addition to the foregoing minerals there is present, even in the
freshest specimens, a considerable amount of undeterminable
alteration products mingled with the serpentine, carbonates,
secondary magnetite, mica shreds and the minor primary
constitnents to make up the groundmass. While it is possible
that there may have been glass in the unchanged rock, none
is now recognizable.

Much of the foregoing description of the freshest parts of
the dike applies equally well to the more altered parts. In
these, the olivine has been almost wholly changed to serpentine,
which is greenish or yellowish and shows the characteristic net
texture to a marked degree. The double refraction of the
serpentine is very low, and in the thinnest parts of sections,
and often elsewhere as well, it is sensibly isotropic. Different
specimens show great diversity in the amount of secondary
magnetite in the serpentine. In some cases the former min-
eral is scantily sprinkled through the latter, while in other
cases the serpentine is rendered nearly opaque by the powdery
magnetite. The primary magnetite and perofskite are not
affected by alteration, the latter in particular retaining all of
its characteristic properties and showing in the sections of
thoronghly altered specimens perhaps even better than in the
fresher rock. All of the foregoing changes belong to the
process of alteration rather than weathering, and do not affect
the integrity of the dike as a whole, while modifying its com-
position, both mineralogical and chemical, to an unknown
depth. Weathering on the other hand, though superficial,
destroys the dike, as far as it goes, resulting in thorough
disintegration.

The rock first called a sheet, but now regarded as probably a
distinct dike, differs from that described above chiefly in con-
taining abundant inclusions of the wall rocks together with

* This Journal (3), xlix, p. 459.

28 = Smyth, Jr.—Petrography of Dikes in Syracuse, NV. Y.

occasional grains of garnet. The inclusions are so numerous as”
sometimes to equal or exceed the dike rock in quantity, and it
is probably due to their presence that the rock goes to pieces
very rapidly when exposed to the atmosphere. This may be
due in part to the chemical effects of the inclusions, but more
largely to mechanical influences, the homogeneity of the rock
being destroyed, while it is filled with fragments, many of
which disintegrate easily. The material is most unsatisfactory —
for study and affords very poor sections. The following
description is, for this reason, quite incomplete.

The garnet appears rather scantily in irregular, rounded
grains, sometimes reaching 4 or 5™™ in diameter. The color is
bright red, becoming very pale in thin sections. Around most
of the grains there is a very thin shell somewhat distinct from
the rest of the rock, but the nature of the material is such that
no good section of the shell could be obtained. Apparently
mica (perhaps with hornblende) is the chief constituent. In
one case the sliell consists of decomposition products, through
which are scattered great numbers of perofskite crystals.
Somewhat analogous to this is the occurrence of a mass some
3™" in diameter consisting largely of tiny crystals of perofskite,
the whole surrounded by a shell of biotite.

In every case seen by the writer, the garnet occurs in that
part of the rock which contains abundant inclusions, but never
in the inclusions themselves. This relation to the inclusions,
together with the scantiness and irregular distribution of the
garnet, suggests that rather than a normal product of the dike
magma like that of the Elliot Co., Ky. peridotite described by
Diller* and that of many European occurrences, it may result
from the fusion of parts of the wall rock in the molten dike,
somewhat as in the case of the sapphires of Yogo Gulch,
described by Pirsson.t However, conclusive evidence upon
this point is lacking. That the garnets are themselves inelu-
sions derived from the wall rocks is hardly possible. Were
this the case, garnets should be found in some of the inclusions
themselves, and even did this happen, it would be needful to
prove that they were primary, and not formed by the contact
action of the dike rock. Moreover, the only external source
of garnets would be in the underlying pre-Cambrian rocks,
and these are of such a nature that they would be most
unlikely to fuse completely away from the garnets, leaving the
latter free.

Embedded in one of the garnets was a bright green grain
about 0°5™™ in diameter. When broken and placed under the
microscope it shows one cleavage, with which the extinction
makes'a high angle, while the green color is hardly perceptible.

* Bull, 38; UNS Gs; + This Journal (4), iv, pp. 421-428.

Smyth, Ir.—Petrography of Dikes in Syracuse, N.Y. 29

The mineral is probably pyroxene, but the limited quantity
hardly admits of accurate determination.

Aside from the variations described, this rock differs in no
marked degree from the first dike rock described, except that
one section shows tiny prismatic crystals of pyroxene in the
groundmass. On account of its excessive alteration, sections
of the rock are largely made up of secondary products, but
these are essentially the same as those described above.

Of the two classes of inclusions, Paleozoic and pre-Cambrian,
the former are naturally more abundant. ‘The fragments are
often angular and sometimes show a faint zonal coloration
along the contacts. But there is no decided indication of new
mineral growth. ‘The pre-Cambrian fragments are usually
more rounded, the result of the attrition involved in their
upward journey of many hundred feet. They do not differ
materially from inclusions described in other dikes of the
region. |

In comparing the rocks of this locality with specimens from
DeWitt and from Green st., as might be expected, a strong
family resemblance is apparent. There is perhaps less mica
and rather more perofskite in the Butternut st. rocks, while
only one section shows the tiny crystals of pyroxene that are
so abundant in the groundmass of the DeWitt dike. But
these differences are much too slight to be of any moment.

Melilite in the Green St. Dike.—The marked resemblance,
both in petrography and geological relations, between the
Syracuse rocks and the alnoite of Manheim has led the writer
to expect, at any time, the discovery of melilite in the former.
In studying the DeWitt rock, Prof. Kemp* searched carefully
for this mineral but found none. The writer’s study of the
Butternut st. rock met with the same result; but in the exami-
nation of the rock of the original Green st. dike of Williams,t
melilite was found. This seems at first sight surprising, as,
doubtless, sections of the rock have been examined by many
petrographers ; but the fact that melilite has not been noted
before simply illustrates the elusive naturé of the mineral,
resulting from its tendency to alteration. In the case of the
original Manheim dike, the rock was first described by the
writer{ as a peridotite and its true nature was learned only
when fresher material was procured ;§ while in the largest
dike of that locality no melilite could be determined with cer-
tainty, although there can hardly be a doubt of its original
presence. |

Some years ago a section was made from a specimen of the
Green st. dike which very closely resembled the original

* Loe. cit. + This Journal (3), xxxiv, pp. 1387-145.

t Ibid. (8), xliii, pp. 322-327. § Ibid., xlvi, pp. 104-107.

|| Bull. Geol. Soc. Am., ix, pp. 257-268.

380 Smyth, Jr.—Petrography of Dikes in Syracuse, NV. Y¥.

Manheim rock and, moreover, seemed fairly fresh. This was
searched, with greatest care, for melilite, but to no purpose.
In the present study, three sections of the rock have been -
examined, one from an old specimen and two from new material.
In the first, a few very small individuals of melilite were found,
so inconspicuous that their discovery can be regarded as little
more than accidental. But in the other two sections, the
melilite. is present in some quantity. In habit and general
appearance it is very like the Manheim melilite, though less
abundant. Like the latter, it often shows abnormally high
double refraction, possibly the result of alteration, and, as a
rule, is optically positive, although the negative variety is also
present.

From the foregoing, it is evident that in rocks as thoroughly
serpentinized as are those of the Syracuse dikes, melilite may be
nearly or quite obliterated; and it is extremely probable that
perfectly fresh specimens would show melilite in all of these
dikes. Perhaps an exception should be made in the case of
the DeWitt dike, but even here there would seem to be sufii-
cient alteration to mask the melilite if originally present in
small quantities. Apparently, then, the petrographic affinities
of this group of dikes is with the melilite rocks rather than
with the peridotites, and this is interesting in bringing them
into close relationship with the Manheim dikes, and in deter-
mining another occurrence of a rare variety of rock. Of
course, incomparably the most important contribution to our
knowledge of the Syracuse rocks was the establishment by
Williams* of their igneous nature, but all new data in regard
to the locality are worthy of record.

In a note appended to his description of the De Witt dike,
Prof. Kempt givesa brief resumé of the occurrences of igneous
rocks in Central New Y ork, including two instances of bowlders.
Another case of the latter kind may be mentioned here.

In the Hamilton College collection are two specimens
labelled “ Hornblende Boulder, Syracuse, N. Y.” The rock
is coarsely porphyritic, with very little groundmass, and so
peculiar in aspect that the writer had a section made from it
some years since. Under the microscope this shows idio-
morphic pyroxene of pale violet color, with a somewhat smaller
amount of olivine. The individuals of both range up to
10-12"" in diameter, and the two minerals make up perhaps
90 per cent of the rock, being held together by a scanty ground-
mass of minute laths of plagioclase. While it is quite possible
that the rock is of Canadian origin, still it may well belong
with the other bowlders referred to, in a group of basic erup-
tives breaking through the New York Paleozoic rocks.

Hamilton College, Ciinton, N. Y.

* Op. cit. and Bull. Geol. Soc. Am., i, pp. 533-084.
+ Op. cit., p. 462.

Steiger—Silver Chabazite and Silver Analcite. 31

Arr. VII. — Preliminary Note on Silver Chabazite and
Silver Analcite; by GEORGE STEIGER.

In an investigation upon the constitution of certain silicates
an attempt was made to replace the alkaline metals by silver.
This has been done with chabazite and analcite, the only two
species upon which, so far, I have been able to test the reaction.

It will at once be seen that this work is analogous to that of
Heumann* and others in the preparation of silver ultramarine.
Heumann, by heating blue ultramarine in a sealed tube to 120°
for fifteen hours, obtained a silver salt which was very nearly
pure, and which upon fusion with various chlorides and iodides
yielded corresponding compounds of barium, zine, manganese,
etc., and also compounds of some organic radicals.

The first experiment which I attempted was to heat ammo-
nium chabazitet in a sealed tube with five times its weight of
silver nitrate for four hours at 250° C., at which temperature
the silver nitrate easily fused. After leaching with water, a
determination of silver was made in thedried residue, 25:06
per cent Ag,O being found.

Another portion was boiled for several hours in an open
dish with a 10 per cent solution of silver nitrate; the residue
from this gave 17:20 per cent Ag,O. These results show that
ammonium was at least in part replaced by silver. The
experiments upon chabazite were only preliminary, and are
to be carried farther.

More complete work was done in the case of analcite, and
three portions of it were treated as follows:

“A.” Natural analcite heated in an open tube with dry
silver nitrate for four hours to 400° C.

“B.” Natural analcite heated in a sealed tube with dry
silver nitrate for four hours at 250°C.

“C.’ Ammonium analcite,t heated in a sealed tube with
dry silver nitrate four hours ‘at 250° ©.

All were leached with water, and washed until the filtrates
gave no test for silver; the residue was then dried on the
water bath. The product in each case was a white powder not
differing in appearance from the original material.

The analyses of the different portions are given below,
together with the composition of the theoretical compound
Ag,O-Al,0,48i0,-2H,O, which is found in column “ D.”

* Liebig’s Annalen, cxcix, 253 (1879).

+ This: Journal, vol. xiii, p. 27. t This Journal, vol. ix, p. 117.

32 Steiger—Silver Chabazite and Silver Analeite.

A. B. eh D.
Leach water
Na,O (cale. to analcite)-. 13°13 12°57 - 60
Dried residue
S10, ici wipe k Pea bate aban inne: WM 41°31 40°08 42°69 aE! has ts)
Al,O, SEP ik, MEMEO Ce Pie Seah 16°44 . 16°29 18°22 1b:72
‘CANE ea Sinead MRI et® _ 87:45 3691 . 32°01 38°03
NAMP SP en ea ees "85 8] “68° "> a
BE 8) On 4-29 5°86. 6:08 5°90
NMR St cy 2) re a eee Se: Bape | ‘69 NO a
MO Gr Aves oie eek es none none none

100°34 99°95 100°37 100°00

It seems from the foregoing experiments that at least some
of the natural silicates, which have been supposed to be very
refractory bodies, are easily attacked and replacements effected
by simple operations.

Work is now in progress on other minerals, along these same
lines, and it is also proposed to study the action upon them of
thallinm and lead nitrates and other salts.

Chemical Laboratory, United States Geological Survey,
Washington, D. C.

O. H, Hershey—Cretaceous Outliers in California. 33

Art. VIII—The Significance of Certain Cretaceous Outliers
in the Klamath Region, California ; by OscAR H. HERSHEY.

CRETACEOUS remnants occur in the southern part of Trinity
County, California, in five very limited areas. They are dis-
tinguished from the Neocene deposits by their better lithifica-
tion, by the gravel being almost exclusively of quartz and
displaying the marine type of rounding, by their occupying
structural basins and not valleys of erosion, and by their being
identical lithologically with the Cretaceous deposits in the
neighboring portion of the Sacramento Valley.

The principal Cretaceous outlier is in the valleys of Indian,
Redding and Brown’s Creeks, a few miles northwest of Bully
Choop peak. It is about three square miles in area, and elon-
gated in a direction west-southwest. It consists of several
hundred feet of fine, well-stratified, yellowish sandstone and a
little of the olive-colored shale so characteristic of the Upper
Cretaceous in the Sacramento Valley. It rests unconformably
on the Abrams mica schist.

The next area is at the junction of the North Fork of the
East Fork of Hay Fork with the main East Fork Creek. It
occupies less than a square mile in area, but dips distinctly
toward the center of the small basin. The lower bed is a
coarse, mixed breccia and conglomerate. The angular frag-
ments were derived from the underlying Paleozoie slates and
cherts. This bed differs from the usual basal conglomerate of
the Cretaceous in the Sacramento Valley, but its identification
is made certain by its being overlaid by the olive-colored
shales, whose characters are tinmistakable.

The third area occurs on the divide between Hay Fork and
Salt Creeks, at the head of Dobbin Gulch. There is about a
thousand feet of coarse conglomerates with some green sand-
stone, yet the deposit is only about one square mile in extent.
On all sides it dips distinctly toward the center of the basin at
angles between 20° and 30°. It rests on Paleozoic slates and
cherts, the Clear Creek volcanic series and the Bragdon slate,
the latter of very late Jurassic age. The pebbles, cobbles and
occasional bowlders which compose the rock are, however,
chiefly of white, yellow, brown, pink, blue and black quartz,
derived from the rocks of the Klamath region bunt at some dis-
tance from their present position.

The fourth remnant occurs at the junction between Rattle-
snake and Post Creeks and is about one and one-half square
miles in area. It rests on the Clear Creek volcanic series, and

Am. Jour. Scl.—FourtH Series, Vou. XIV, No. 79.—JuLy, 1902.
3

34 O. H. Hershey—Cretaceous Outliers in California.

consists of several hundred feet in thickness of a fine pebbly
conglomerate, overlaid by a coarser conglomerate. The peb-
bles in the first are of quartz, have a remarkable uniformity
in size and are perfectly rounded. They weather out and
bestrew the surface as would a shower of marbles. They
unmistakably indicate marine action.

Another supposed Cretaceous remnant is found at the -

Sweepstakes mine, a few miles southwest of Weaverville, but
it has not been studied in detail and its extent is not known.
It consists of a hard conglomerate of fine pebbles resting on
the edges of the upturned Paleozoic slates and cherts.

The first four areas are Horsetown in age. The differences
which are observed between them in going southwestward are
exactly imitated by the basal beds of the Horsetown from the
village of Horsetown along the foot of the mountains to Har-
rison Gulch and Good’s Pass. At Horsetown the basal beds
are shale and fine sandstone, as in the Indian Creek area. To
make the identification complete, both Diller and Anderson
have found the Horsetown fauna on Indian and Redding
Creeks. Westward from Ono, the basal bed of the Horse-
town is a heavy conglomerate like that of the Dobbin Gulch
area; and near Harrison Gulch, there occurs in the Cretaceous
the same peculiar pebble conglomerate as that of the Post
Creek area.

Undoubtedly the sea of the Horsetown epoch transgressed on
the southeastern border of the Klamath region and deposited
probably several thousand feet in thickness of conglomerates,
sandstones and shales over what is now a rough mountain area.
At the time of the submergence it must have been compara-
tively even, but the heavy conglomerates tell of elevated and
perhaps mountainous land in the interior of the Klamath
region.

This country, that became covered by the Horsetown sedi-
ments, had previously suffered profound denudation. At the
close of the Jurassic period, the stratified and already largely
metamorphosed formations of Mesozoic, Paleozoic and perhaps
earlier age, were intruded by three great series of batholiths,
the first of peridotite, the second of gabbro and the last of

granodiorite and allied plutonic rocks. These intrusions were:

accompanied by orographic activity of the greatest magnitude,
for the whole area was thrown into a series of geosynelines,
with many minor folds and faults. At the close of the dis-
turbance the surface must have been very mountainous—a
mountain system, in fact, made up of quite different and more
massive ranges than those of to-day. Erosion slowly reduced
these mountains. Most of their destruction was accomplished
during the early Cretaceous time. Several thousand feet at

O. H. Hershey— Cretaceous Outliers in California. 35

least of Mesozoic slates and meta-andesites were removed and
the rivers cut deeply into the Paleozoic slates beneath. The
denudation reached even to the plutonic masses and laid bare
the coarse-grained gabbro and granodiorite which must have
solidified at a considerable depth beneath the surface. As
products of this erosion and of that accomplished in the Sierra
Nevadas at the same time, we have in the Coast Range region
many thousands of feet of sandstone and shale, the first chiefly
Franciscan and the latter Knoxville in age. Judging from
the work performed, this “early Cretaceous” period must
have been a very long one, perhaps equal to all the time since.
Base-leveling in the Klamath region continued through the
Horsetown and Chico epochs. Northeastward from Horse-
town the Chico formation laps past the Horsetown and from
Clear Creek to far beyond the Sacramento River it is in con-
tact with the metamorphic rocks as shown by Diller and
others. The basal. bed of the Chico is a fine sandstone, with
no conglomerate whatever present in this area. Apparently
the southeastern border of the Klamath region had been pretty
thoroughly base-leveled before the opening of the Chico epoch,
but the northeastern border seems to have still remained hilly,
since the base of the Chico in Shasta Valley and the Siskiyou
range is a heavy conglomerate. It is probable that even to
the close of the Cretaceous, the central portion of the Klamath
area remained elevated, but base-leveling had been affected
around its borders, and a very slight depression was sufficient
to cause the Upper Cretaceous strata to lap over on to it.
Subsequently to the Chico epoch there occurred another
profound orographic disturbance of the Klamath region. It
was characteristically different from that which closed the
Jurassic period. Much of the area and perhaps all seems to
have been deformed into a series of deep, elliptical basins.
The four principal Cretaceous remnants in Trinity County
mark the deepest portions of four of these structural basins,
each of which was elongated in a direction west-southwest,
adjoined each other and virtually constituted a syncline sepa-
rated from the Sacramento Valley by a rather prominent anti-
cline. Perhaps my meaning may be clearer if I state that the
four Cretaceous remnants are separated from each other by
tracts of metamorphic rock which rise about 2000 feet above
the deepest portions of the basins, but that the ridge of meta-
morphic rocks which separates them from the Cretaceous in
the Sacramento Valley is much higher.
The average altitude of the bottom of the Cretaceous in the
center of each basin is about 2500 feet. The ridge which
separates them from the Sacramento Valley rises from an alti-
tude of about 4000 feet at Harrison Gulch to over 7000 feet

36 C0. H. Hershey— Cretaceous Outliers in California.

in Bully Choop peak. On the south side of this range the
Cretaceous strata dip away at an angle about the same as the
general slope of the mountain, and Cretaceous remnants occur
at a considerable height on its flank. Indeed, Cretaceous peb-
bles and the gold derived from the basal conglomerate are
widely scattered over the slope and indicate that Cretaceous
strata in place have but lately disappeared from it. Near
Harrison Gulch and at Good’s Pass, where the ridge is lowest,
the Cretaceous strata reach to the very summit and curve over
its top. A careful reconstruction of the plane of the base of
the Cretaceous upon data furnished by the basins on the north
of the ridge and the Cretaceous border on its southern slope
shows that the summit and slopes of this range must corre-
spond roughly with the Cretaceous base-level upon which,
after submergence, the Horsetown was deposited. It is the
Cretaceous base-level brought to light by erosion. This base-
level was deformed near the close of the Cretaceous to just
about the same extent as is the present surface. The north-
east-southwest range of which Bully Choop is the principal
peak is, in a certain sense, a strnetural range although its
present topographic prominence is due entirely to erosion. It
almost exactly coincides in position and form with a post-Chico
range. The most remarkable thing about it is that this post-
Chico mountain range was thrown up almost at right angles to
the strike of the metamorphic formations forming its core and,
hence, to the post-Jurassic mountain system. If the post-
Chico Bully Choop range rose rapidly enough so that its sum-
mit was not materially reduced by erosion during the move-
ment, at its completion it was a ridge with a base on the
average 12 miles wide and rising in its highest peak over 7,000
feet above the lowlands on the south and 4500 feet above the
deepest basin on the north. It became the main drainage
divide and has continued so to the present.

We do not know whether the same sharp deformation of the
surface in post-Chico time persisted over the entire Klamath
region, but there are evidences that it did. Wherever the
Cretaceous strata are found on the borders of the Klamath
region, they are tilted at a considerable angle. On the north-
eastern flank of the mountains, in Shasta Valley, in the Sis- —
kiyou Mountains and in the Rogue River Valley in Oregon,
the Cretaceous strata rest upon the metamorphic formations |
with their original contact, but are tilted toward the northeast
at a high angle, in a few places nearly vertical. There isa
general rise of the Klamath region toward the northeast and
the highest portion of the upland occurs on the immediate
border, just beyond which the surface abruptly falls away to

C. H. Hershey—Cretaceous Outliers in California. 37

the deep broad basin on the floor of which Mt. Shasta and the
Cascade range have been built up. The phenomena are sim-
ilar to those which characterize the Sierra Nevada region with
its short steep slope on the east and.a long gentle slope on the
west, and suggest that the northeastern face of the Klamath
Mountains marks the line of a great fault; but the relation
between the steeply inclined Cretaceous strata and the older
rocks shows that the structure is that of a sharp monocline
rather than a fault.

The Sierra Costa Mountains rise abruptly above a supposed
Cretaceous remnant near Weaverville and on the northeast
they are terminated by the sharp monocline above-mentioned.
They must have been one of the abnormally elevated portions
of the post-Chico mountain system.

Berkeley, California.

38 O. C. Farrington—Action of Copper

Art. IX. — The Action of Copper Sulphate upon Iron
Meteorites ; by O. C. FARRINGTON.

In 1852 Wohler* announced the discovery that some iron
meteorites, irrespective of their nickel content or structure, are
passive to neutral solutions of copper sulphate, i. e., they do
not reduce copper from such a solution. This passivity he
stated could be overcome by introducing a piece of ordinary
iron into the solution or by adding a few drops of acid.

No data were given as to temperature, length of time allowed
for deposition or the strength of the solution. The following
meteorites were reported by him to be passive: Bemdego,
Bohumilitz, Braunau, Cape of Good Hope, Green County
(undoubtedly Babb’s Mill), Obernkirchen, the Pallas Iron,
Red River (probably Cross Timbers), Schwetz, Toluca, and
the terrestrial iron of Greenland. A number of other mete-
orites which need not here be enumerated he reported at the
same time as “active,” i. e., they reduced copper from the
copper sulphate solution ; while those of a third class were
designated as “intermediate,” i. e., they were nearly passive, but
after remaining some length of time in the solution brought
about a slow reduction of copper. That the passivity was a
property of the entire mass of the meteorite seemed to be
proved by the fact that when the surface of the meteorite had
been made active and the coating of copper was removed by
filing, it would again behave passive.

Since the publication of Wohler’s article several other
investigators of iron meteorites have tested their behavior
toward copper sulphate with the result usually of finding them
active, but the following, as listed by Cohen,t have been
reported passive: Toluca by Krantz, Knoxville by Smith,
Misteca by Bergemann, Octibbeha County by Taylor, Ohaba
by Bukeiégen (this is a stone), Charcas by Meunier and Lexington
Co. by Shepard. Two of these, however, were reported by other
observers to be active, viz: Toluca by Pugh, Jordan and
Taylor, and Charcas by Daubrée. The observers later than

Wohler also failed to give details regarding the conditions ©

of their experiments. Wicke and Wohler state that the

Obernkirchen meteorite, after remaining a day, reduced no

copper from its salts.t Smith states that Knoxville is passive

with reference to the action of sulphate of copper, but when

immersed in a solution of the latter and allowed to remain

several hours the copper deposits itself in spots on the surface
* Poggendorff’s Annalen 1852, vol. Ixxxv, pp. 448-9.

+ Meteoritenkunde, Heft 1, p. 68.
¢t Ann. d. Chem. u. Pharm., 1864, vol. cxxix,, p. 123.

EE a a

Sulphate upon Iron Meteorites. 39

of the iron.* Meunier states that the section of Charcas was
washed in alcohol and ether to remove grease and upon it a
drop of solution of sulphate of copper was placed. At the end
of some hours the drop had evaporated, leaving crystals of the
salt but no copper.t

Viewed simply as a matter of reasoning, it would seem quite
impossible, with our present knowledge, to account for a
passive behavior of meteorites independent of their nickel con-
tent. As is probably generally known, nickel is less active to
copper sulphate than iron, and it might be expected that the
activity of alloys of iron and nickel would decrease in propor-
tion as the percentage of nickel increased. But that some
individuals of the same alloy (for iron meteorites are practically
alloys of iron and nickel) should be entirely passive while
others were normally active would be quite extraordinary.
Wohler’s conclusion was that all iron meteorites were probably
in their original state passive, but that throngh local terrestrial
influences and the passage of time, some had since their arrival
upon the earth become active. Meunier expressed the opinion
that a peculiar molecular structure may be indicated by the
passivity.t In view of the lack of details regarding the experi-
ments upon which Wohler based his conelusions and the results
he obtained, it seemed to me desirable to reinvestigate the
matter to some extent.

Of the meteorites listed above as passive, there are to be
found in the collection of the Field Columbian Museum,
specimens weighing from 10 to 1100 grams each, of the fol-
lowing: Babb’s Mill, Bemdego, Braunau, Bohumilitz, Cape of
Good Hope, Chareas, Cross Timbers, Knoxville, Lexington
Co., Misteca, Pallas Iron, Toluca and the native iron of Green-
land. These therefore were available for investigation. In
order to test their activity each section was in turn immersed
in a solution of 200 grams of Baker & Adamson’s C. P. copper
sulphate to a liter of distilled water. The temperature of this
solution was about that of the room, or 18° C. In this test
every meteorite gave an active reaction. <A perceptible depo-
sition of copper took place upon each section within one to
four minutes after its immersion. The only exception to this
general rule was Knoxville, which in some trials deposited in
four minutes but in others required twelve or fifteen minutes.
In the same solution and at the same temperature about thirty
minutes was required before deposition took place upon a
piece of ordinary cube nickel the surface of which had been
rendered bright by filing. A specimen of josephinite, how-

* This Journal (2). xix, 155.

+ Ann. Chim. et Phys., 4th ser., vol. xvii, p. 92.
{ Loc. cit.

40 O. C. Farrington—Action of Copper

ever, gave no deposit at all, though left twenty-four hours in |

the solution. With the exception of Knoxville and josepninite
the period required for deposition seemed to correspond quite
closely to the percentage of nickel. Thus ordinary iron would,
-as it is well known to do, take on a coating of copper almost
instantly upon immersion in the solution. For the native iron
of Greenland, which contains 2 or 3 per cent nickel, about
thirty seconds were required. Irons like Chareas, Braunau,
Bemdego and Lexington Co., which contain 4 to 6 per cent of
nickel, required from one to two minutes, and Cape of Good Hope
and Babb’s. Mill, which have about 15 per cent of nickel, three
to four minutes. The somewhat eccentric Knoxville is a mete-
orite with a high percentage of nickel and josephinite contains
72 per cent nickel. Incidentally, therefore, the experiments
indicated that immersion of an iron-nickel alloy in a solution
of copper sulphate will afford a rough test of the percentage
of nickel contained in the alloy, and, since a meteorite must
contain nickel, a quick means of distinguishing meteoric from
terrestrial iron.

To the experiments made in the above manner as compared
with those of Wohler, the objection may be made that Wohler
described his solution as a “ Lésung von neutralem schwefel-
sauren Kupferoxyd,’ while a solution of copper sulphate has
normally an acid reaction. In order to remove doubt on this
point a solution was prepared according to the method given
by Blair.* Such a solution is one of 200 grams to the liter,
but to it sodium hydroxide is added until a slight precipitate
appears and the latter is removed by filtration. The solution
still has an acid reaction, but the possible presence of free acid
is avoided.

The sections immersed in this solution gave, however, as far
as could be observed, results exactly similar to those obtained
with the simple solution. As no mention is made by Wohler
or other investigators of treatment to reduce the acidity of
their solutions, it is highly probable that only the simple solu-
tion was used. Employment of solutions of different strengths
was likewise without effect on the results obtained. Tests

were made with a solution of 100 grams to the liter and some

intermediate strengths, but the fact of the deposition of the

copper and the period required for its deposition seemed —

unchanged.

‘Different temperatures were however of considerable influ-
ence in changing the time required for deposition. Thus the
cube nickel, which required thirty minutes at 18° C., reduced
copper in five minutes when the solution was heated to 63° C.

* The Chemical Analysis of Iron, 8d ed., p. 63.

Sulphate upon Iron Meteorites. 41

The specimen of josephinite which reduced no copper at 18° C.
gave a deposit in twenty minutes at 63° C. At temperatures
below 18° C. the time required for the deposit is greatly
increased. Thus, Cafion Diablo, which reduces copper in one
minute at 18° C., required sixty minutes at 10° C. and at 0° C.
was kept twelve hours without producing a deposit. Toluca,
which acquires a deposit in two minutes at 18° C., required |
eighty minutes at 10° C. and at 6° C. was kept for four hours
_ without receiving a deposit. It was not tried longer at this
temperature.

In these tests it was found necessary to immerse the sections
in water of the temperature whose effect it was desired to
study, for some minutes before transferring to the copper sul-
phate solution. Otherwise the heat of the section would cause
deposition at an earlier period. Also a fresh section which
had not been exposed to copper sulphate must be used, as
traces of a previous deposit of copper would hasten deposition.

The periods named as sufficient for deposition are in all
cases somewhat approximate and variable, since the exact
moment at which deposition begins can be determined only
roughly by the eye. The general results are however suft-
ciently accurate for the purpose. They show nothing to sup-
port Wohler’s belief that certain meteorites are inherently
passive, but on the contrary point to the conclusion that mete-
orites act like any other iron-nickel alloy.

Daubree considers the deposition of copper due to the fact
that the different alloys in a meteorite form electric couples,*
but it hardly seems necessary to appeal to this force when
deposition takes place so quickly upon a piece of iron not
made up of different alloys.

It is hardly essential to the purposes of this article to try to
account for the results reported by Wohler and other obsery-
ers, but one or two suggestions may be made in this connec-
tion. It is possible that some of-the surfaces tested by these
observers may have been rendered passive by treatment with
nitric acid, This was quite probably the case with Misteca as
treated by Bergemann, for he reports that upon freshly dis-
solved (geldst) pieces he could get no deposit, while pieces
from the outer surface of the meteorite without luster and of
dark gray color were quickly coated.t Wohler, however,
expressly states that the sections which he tested had not come
in contact with nitric acid.

Again a slight film of grease or oxide may have prevented
deposition of copper. A substance so active to copper sul-
phate even as iron will be found to have become passive,

* Comptes Rendus, 1867, vol. Ixiv, p. 685.
+ Pogg. Ann., 1857, vol. c, p. 246.

42 O. C. Farrington—Action of Copper Sulphate, ete.

if heated a moment before immersion. This is probably on
account of the film of oxide formed. The sections of Cape
of Good Hope and Knoxville, which had an active behavior
as treated ordinarily, I found to be thoroughly passive when
reimmersed without rubbing after an interval of two weeks.
No perceptible film of oxide could be detected, but on rubbing
the surface with emery cloth deposition took place as before.
There are some indications again that the molecular condition
of the substance and especially of its surface may determine its
degree of activity. For instance, a cube of nickel which was
perfectly passive when treated in the ordinary way, no deposit
being formed on it after standing 24 hours in the sulphate
solution, although fresh surfaces were frequently exposed by
abrasion, was made active by being heated to redness, cooled
and again made bright by abrasion. A deposit of copper then

formed within twenty minutes and the cube has shown active

behavior since. The activity was the same both when the
cube was cooled slowly and quickly. The specimen of josephi-
nite previously mentioned as being passive at ordinary tem-
peratures, was given a similar treatment, i. e., heated to
redness, cooled and rubbed. It also became active under this
treatment, reducing copper in a solution at the temperature of
18° C. at some time within a period of twelve hours, the exact
time not having been noted. It was noticeable in this case, as
in many others, that deposition began and was most active
about capillary craeks and fissures. This would indicate that
capillary tension was influential in effecting the reduction of
the copper, and it is reasonable to suppose that such would be
the case, since the molecules are probably brought into closer
contact in such situations. It is evident that several conditions
may affect the degree of activity of a meteorite or any other
iron-nickel alloy, but I have found nothing to accord with
Wohler’s observation that passive meteorites which had been
made active by contact with iron, after the reduced copper
was filed off became again passive. While several phases of
the subject invite further investigation, enough evidence seems
to be obtained to warrant classing all meteorites as active to
copper sulphate and dropping the designation of passive
meteorites.

I am indebted to my associate, Mr. H. W. Nichols, for a
number of useful suggestions during the progress of the
investigation.

Field Columbian Museum, Chicago, II1.

ee ee ee ee ee SS oe

Dresser—Contribution to the Geology of Quebec. 48

Arr. X.—A Petrographical Contribution to the Geology of the
Eastern Townships of the Province of Quebec; by JOHN
A. DRESSER.

THAT portion of the Province of Quebec which les south of
the St. Lawrence River comprises two physiographically dis-
tinct regions, viz.: The flat country of the St. Lawrence valley,
and that part of the Appalachian mountain system which
belongs to this province. The part of the latter which occu-
pies the Gaspé peninsula is known as the Shickshock mountains,

S
<
3

{ N
a S /

VY ae
Y

)> *
ey Lennoxvile. 8
\

ix]

Paleozorw
Scale. =
° to milés

\

A Part of the Eastern Townships of the Province of Quebec.

and that between the vicinity of Quebec city and the United
States boundary line as the Notre Dame Hills. The district
lying within the Notre Dame hills is commonly designated as
the “ Eastern Townships,” the geological structure of which
has furnished the theme of imuch well-known discussion
during the past forty years. This has been chiefly connected
with the question of the Quebec Group.*

* “Geology of Canada,” 1863, pp. 225-297, by Sir W. E. Logan.

**The Quebec Group in Geology,” Transactions of the Royal Society of
Canada, vol. I, 1882, by A. R. C. Selwyn.

““Notes on the Microscopic Structure of Some Rocks of the Quebec
Group.” Ann. Rept. Geological Survey of Canada, for 1880-1-2, by F. W.
Adams.

‘*“The Quebec Group.” Appendix A to Harrington’s Life of Sir W. HE.
Logan, by Sir J. W. Dawson, 1883.

Reports of the Geological Survey of Canada for the years 1886, J; 1887-8,
K ; 1894, J, by R. W. Ells.

44. Dresser—Contribution to the Geology of Quebec.

As originally defined by Logan and Billings, the Quebec ~

Group embraced all the rocks of the Eastern Townships that
are essential to the present investigation, and all were then
regarded as of sedimentary origin. Subsequently, however,
Hunt, on stratigraphical grounds, and Selwyn on stratigraphi-
cal and lithological evidences, distinguished certain older
measures, which were referred by the latter to the early Cam-
brian and pre-Cambrian ages. The amplification of the views
has been fully carried out by Ells in the reports of the geo-
logical survey of Canada for the years 1886 and 1894.

The only lithological changes embodied in these re-
ports are in the recognition of the eruptive origin of
the serpentines and the diorites, diabases, porphyrites
and granites generally associated with them. These had
been previously regarded as metamorphosed sediments.
The silicates of magnesia were correlated with its carbonates,
where dolomite occurred in the vicinity of the serpentine belt,
and even in 1886, Dr. Selwyn, who had been the first to
recognize the igneous origin of the serpentines, in a footnote
appended to Dr. Ell’s report, maintains that the hornblende
granites are probably products of metamorphism in situ and
not true intrusives through the serpentine, as the latter writer
correctly considers them to be.

In respect to age, the serpentines and certain of the clasties
were referred to the early Cambrian and the other eruptions to
middle or late Silurian time, while three belts of supposed
sedimentary rocks, running approximately parallel to the north-
easterly trend of the Appalachians were classed as pre-Cam-
brian. One of these appears for only a relatively short dis-
tance along the boundary line between the province of
Quebec and the State of Maine. The second crosses the St.
Francis river between the city of Sherbrooke and the vil-
lage of Lennoxville. This may be known as the Ascot or
Stoke Mountain belt, while the third, which crosses the St.
Francis river a little way north of the town of Richmond,
twenty-five miles northwest of the second, is generally designated
as the Sutton Mountain belt. The structure of these bands,
especially of the last mentioned, was long a crucial point in

the Quebec group controversy, they being interpreted as syn-.

clines by the earlier investigators, and as anticlines by the
latter.

PETROGRAPHY.—Recent petrographical investigations by
the writer have, however, shown that both the second and
third of these pre-Cambrian belts consist largely, and in places
entirely, of altered volcanicrocks. These are so highly
altered and consequently so much disguised, that they have

:

Dresser— Contribution to the Geology of Quebec. 45

been hitherto mistaken for sedimentaries, and have been
accordingly treated on that assumption in all the stratigraphi-
eal discussions regarding them.

(a) Inthe Stoke Mountain belt the principal rocks exam-
ined were taken from the township of Ascot, on the west
side of the St. Francis river, between Sherbrooke and Lennox-
ville. Others were taken from various points to the south-
eastward on the upper Belvidere road and from the vicinity of
the Suffield, Sherbrooke and Clark copper mines. The rocks,
which are generally fine in texture, are of various shades of
green and gray in color, and many are in advanced stages of
metamorphism. In the thin section, however, the microstructure
remains sufficiently distinct to clearly establish the igneous
origin of practically all of these rocks, and even to identify
the specific characters of several on the merely preliminary
examination of them that has yet been made.

Quarte-porphyry forms the hanging wall of the Silver Star
mine at Suffield. It is of light gray in color, and from the
prominence given the quartz phenocrysts by the bleaching of
the base of the rock on weathered surfaces, it has been com-
monly regarded as a species of sandstone.

By the aid of the microscope the structure is found to be
that typical of an effusive rock. A very finely crystalline
base holds phenocrysts of quartz, which show in basal sections
an uniaxial cross and positive sign ; and also of feldspar,
which extinguished in several cases with its principal axes
parallel to the planes of the crossed nicols and hence is ortho-
clase. A few feldspars are polysyntheticallv twinned and
accordingly are plagioclase. Their extinction angles are
small. Small rod-like individuals of colorless mica arranged
in lines are presumably of secondary origin. They often
occur within, or in association with, irregular areas of a
rhombohedral carbonate, apparently dolomite.

Granite-porphyry occurs near Lennoxville on the line of the
Canadian Pacific Railway. It differs from the quartz-porphyry
chiefly in the more advanced character of its crystallization,
both quartz and feldspar being distinguishable in the ground-
mass. A little chlorite, pyrite, brown iron oxide, and color-
less mica are also present. An incipient cataclastic structure
is beautifully shown in both of these rocks. The quartz-pheno-
crysts are sometimes reduced to mosaics of quartz grains, or
at others are crossed by lines of crushed material, while the
erystals on either side of the lines remain unchanged or show
undulatory extinction. Being the largest single masses and
of the most brittle material, the quartzes thus appear in
every case to be the first constituents to show the effects of
purely dynamic metamorphism. A much altered rock form-

46 Dresser—Contribution to the Geology of Quebec.

ing the hanging wall of the Clark mine was probably closely
allied to the quartz-porphyry in its original composition. It
rather closely resembles a specimen of “sheared felsite ” from
the Gettysburg Railway, south of Clermont House, Monterey,
Pennsylvania, seen in the petrographical collection of McGill
University.*

A greenish gray fine-textured massive rock is of large

extent, especially in the southern part of this belt. In the
thin section quartz is found in broken phenocrysts and also in

smaller grains, presumably primary, in the rather fine holo-

crystalline groundmass. Under high power (x220) feldspar
appears in the groundmass in small lath-shaped individuals
which extinguish at low angles with the principal axes. Epi
dote and chlorite are abundant representatives of primary
bisilicates. The rock would have originally been abont of the
character of a quartz porphyrite.

A large part of the central and southern portions of this
belt in the township of Ascot is occupied by a highly foliated
rock of green color and massive appearance. Under the
microscope a little feldspar is found in an aggregate of color-
less hornblende, chlorite, epidote, dolomite and sericite, all of
which are secondary constituents.

This rock agrees essentially with the sheared greenstone
from Jack’s Mountain tunnel, near Monterey, Pennsylvania, as
seen in the McGill University collection.

A similar but more dolomitized rock occurs in various parts of
the belt, one occurrence of which is-mapped as igneous on the

map to accompany the Annual Report of the Geological Sur-

vey of Canada for 1886.

These rocks, though probably of several different ages of
eruption, are generally much metamorphosed. But cutting them
thereare dikes of camptonite and olivine diabase, which are
quite undisturbed in position and comparatively little altered
in mineralogical composition. They also cut the Lower Tren-
ton strata along the edges of the belt, while these overlie
the other igneous rocks of the region. Whence it appears
that the Stoke Mountain, or Ascot belt, has been the scene

of volcanic activity at various periods through a long range

of time, from pre-Cambrian to post-Trenton. <A brief notice
of the lithological character of those rocks has been recently
given by the writer in a communication upon the copper-
bearing rocks of the area,t and their general agreement in
extent with the pre-Cambrian of the Geological Survey Map
of 1886 pointed out.

* Bulletin U. S. Geological Survey, No. 136, ‘‘ Ancient Volcanic Rocks

of South Mountain, Pennsylvania,” by F. Bascom.
+ Trans. Can. Min. Inst., Montreal, March, 1902.

a a, a a a

Dresser— Contribution to the Geology of Quebec. A7

(b) The igneous origin of a portion of the pre-Cambrian
area, which forms the Sutton Mountain belt, was pointed out
by the writer in a note to the Ottawa Waturalist of January,
1901. Since that time the extent of the igneous portion has
been more fully examined, especially to the northeast of the
St. Francis river, and its structural relations better ascertained.
Jt is found to comprise the greater part of the pre-Cambrian
of this area, as shown on the latest geological survey map of
the region, that of 1886.

As far as examined, for a distance of some forty miles on
either side of the St. Francis river the rock is an altered
greenstone, very commonly amygdaloidal. In the microscopic
section a little primary plagioclase sometimes remains, but in
many sections the whole field consists of a secondary aggre-
gate of chlorite, epidote, iron ore and leucoxene. The amyg-
dules usually consist of quartz and zeolitic minerals. The
rock is much foliated and has been generally described as a
chloritic slate. Its resemblance to the basic rocks of the Ascot
area is very close. It, too, is important as a copper-bearing
rock.

In geographical position these volcanic belts form a link in
the more westerly of the two chains of ancient volcanics des-
ignated by the late Prof. G. H. Williams,* while in their
lithologie characters they agree with certain of the South
Mountain rocks of Pennsylvania in all essential respects.+
The Quebec rocks are, however, largely made up of the acid
types, some of which agree essentially with rocks from Ascot.
Most of the specimens are in less advanced stages of recrys-
tallization than the rocks above described.

The basic phases from the Stoke Mountain area as well as those
which comprise the Sutton Mountain belt are closely analogous
to the greenstones from the vicinity of Monterey.

STRUCTURE.—As already stated, these rocks were formerly
regarded as presenting synclinal structures, while the later
investigators have believed them to form anticlines, the former
view making them later in age than the adjacent rocks, the
latter, earlier. With respect to relative age, the latter opinion
is doubtless the correct one. The lowest sedimentary rock,
dolomite, on the south side of Sutton Mountain belt in
the township of Melbourne, for instance, contains fragments
of the adjacent trap, and large areas of mica _ schists

* The Journal of Geology, Jan.—Feb., 1894.

+ A collection of rocks from this locality, which is in the petrographical
laboratory of McGill University, has been examined for comparison
through the courtesy of Prof. F. D. Adams, to whom my best thanks are
also offered for. valuable aid and advice in several matters connected with
this subject.

48 Dresser—Contribution to the Geology of Quebec.

in the vicinity are made up of its débris. Other evidences
also point to the later age of the sedimentary rocks, which
generally belong to the Trenton formation, and none to
the reverse relation. The change of view in regard to the
origin of these rocks, from sedimentary to igneous, does not,
therefore, necessarily alter the recent determinations either of
their relative or of their actual geological age. It would,
however, be well that the evidence on which the latter deter-
mination was made should be carefully reviewed in the light
of the present knowledge of their volcanic character.

On the other hand, the elimination of these volcanics leaves
the stratigraphical structure of the sedimentary rocks a much
simpler one than has hitherto been supposed.

A detailed study of the Sutton belt and adjacent forma-
tions in the vicinity of the St. Francis river shows the sedi-
mentary rocks of the Quebec Group to occupy a trough be-
tween the volcanic ridge on the north and the serpentine belt
at the south. Both these have apparently been covered by
sediments which are now in part removed by denudation.

A summary of the results of this study, therefore, shows:

1. That at least the greater part of the pre-Cambrian or
crystalline belts of the eastern townships of Quebec is of
igneous, not sedimentary origin, as has been hitherto sup-

osed.
: 2. That these rocks are allied to the voleanics of South
Mountain, Pennsylvania, especially to the basic types, and
indicate the continuance of this class of rocks throughout the
Appalachians, as was suggested by Williams.

3. That the sediments of the region, which probably all
belong to the Quebec. Group, were deposited between and
upon the preéxisting ridges of igneous material, which are
now being uncovered by denudation, while the intervening
valleys still remain deeply filled.

The pre-Cambrian area near the international boundary line,
before referred to, and also the extension of the Sutton belt
into the State of Vermont, would furnish subjects of investi-
gation of especial interest in connection with the areas herein
described.

Richmond, Quebec, Canada.

Jones— Action of Carbon Dioxide on the Borates, etc. 49

Art. XI.— The Action of Carbon Dioxide on the. Borates
of Barium ; by Louis CLEVELAND JONES.

[Contributions from the Kent Chemical Laboratory of Yale University—CVIII.]

In 1888 Morse and Burton described “A Method for the
Separation and Determination of Boric Acid.”* I published
an article in 1898 entitled “The Action of Carbon Dioxide
on Soluble Borates,’’+ an account of certain results not in accord
with those of Morse and Burton. More recentlyt Morse and
Horn record at some length “ preliminary experiments ” made
in an effort to substantiate the original analytical results of
Morse and Burton.

I have waited for some time for the promised resumption of
these experiments, and lest further silence be misunderstood,
will show that the action of carbon dioxide and boric acid, in
the presence of the one base—barium hydrate—both in the
dry state and in solution, is that which I pointed out in my
original paper, and, further, that these later experiments of
Morse and Horn, correctly interpreted, completely substan-
tiate this position.

The question is concerning the action of carbon dioxide gas
upon a mixture containing boric acid and an excess of barium
hydrate, viz: what part of the total barium is converted into
carbonate and, upon evaporation and ignition, how much, if
any, of this barium carbonate is reconverted into borate ?
Indirectly, and stated differently, it is : Can boric acid, when
once separated, be brought into definite and weighable condi-
tion by the method recommended by Morse and Burton?
(Loe. cit.) Despite the title of the contribution of Morse and
Burton, Morse and Horn now regard the separation of the
boric acid “as the most, and in fact the only, important feature
of their contribution.” -

The following is the entire description of that part of the
process relating to the determination of the boric acid as
originally given by Morse and Burton: “ The quantity of the
dilute barium hydroxide solution” (about 25°) “which is
equivalent to 25° of the standard sulphuric acid ” (1/10) “ is
run into the flask, and the apparatus attached to a filter-pump.”
This is done preparatory to receiving the alcoholic solution of
boric acid—a little over 75°. “Finally, the excess of the
barium hydroxide is precipitated by passing into it a current
of carbon dioxide: the contents of the flask are transferred to
* Am. Chem. Jour., March, 1888, vol. x, pp. 154-158.

+ This Journal, vol. v, pp. 442-446, June, 1898.
¢ Am. Chem. Jour., August, 1900, vol. xxiv.

Am. Jour. Sci.—FourtH SERIES, Vou. XIV, No. 79.—Jcuy, 1902.
4

50 Jones— Action of Carbon Dioxide

a platinum dish, evaporated to dryness, and heated to a con-
stant weight over a triple burner.”*

In my original article I made three claims: ist, that the
metaborate of barium may be decomposed by gaseous carbon
dioxide; 2d, that the boric acid thus “ liberated ” may be lost
by volatilization; 3d, that the metaborate of barium and the
carbonate may interact at high temperatures with the loss of
carbon dioxide. These facts preclude the use of the method
described by Morse and Burton for the determination of boric
acid. The more recent experiments of Morse and Horn not
only substantiate these facts but disclose a fourth objection to
the use of this method for its original purpose, viz: the diffi-
cult dehydration of the metaborate of barium, concerning a
mixture of which with barium carbonate, Morse and Burton
originally said that it ‘ being neither markedly hygroscopic nor
capable of absorbing carbon dioxide, can be brought to a con-
stant weight more readil y than one containing oxide of calcium
or magnesium.”

The decomposition a barium metaborate.—That barium
metaborate may be decomposed by carbon dioxide in aqueous
solution, is proved by my original experiments; and now is
corroborated by those of Morse and Horn. These authors,
however, urge that, in alcoholic solutions—78 per cent—the
boric acid would be in the form of metaborate of barium
only and this should be unattacked to the extent of its insolu-
bility. To test this point directly, carbon dioxide was passed
into equivalent parts of barium hydrate and boric acid in 78 per
cent alcoholic solution. The precipitate gave abundant evi-
dence of the presence of carbonate, showing decomposition of
the metaborate. The extent of this decomposition was deter-
mined quantitatively by determining the carbon dioxide in
another similar precipitate. Barium hydrate and an excess of
boric acid in a mixture consisting of 78 per cent alcohol, were
treated in the cold for three hours with carbon dioxide. The
precipitate was filtered, washed, and analyzed. for carbon
dioxide with the following result :

Ba(OH), taken, 0°7883 grm.; B,O, taken, 0°3576 grm.;
excess B,O,, 0°0357 grm.; CO, found, 0°2042 grm.; COQ, in
total barium carbonate, 0- 2024 erm.

Evidently the metaborate of barium in aleohol was completely
decomposed by carbon dioxide and the boric acid lzberated.

Another experiment was made as above, except that equiva-
lents were taken and the treatment with carbon dioxide was |
discontinued after thirty minutes. After stopping the flow of
carbon dioxide, the mixture was boiled for three minutes, to
remove any excess of gas. The precipitate was filtered, washed

* The italics, as in other quotations to follow, are mine.—L. C. J.

on the Borates of Barium. 51

with 150° of hot water and analyzed for carbon dioxide. The
result was as follows:

Ba(OH), taken, 0-4379 grm.; B,O, taken, 0°1788 grm. ; alco-
hol, 76 per cent; CO, found, 0°1135 grm. ; CO, in total barium
carbonate, 0°1124 orm.

Upon evaporating to small volume the filtrate from this
experiment, boric acid in abundance crystallized from the soiu-
tion. Under these conditions, then, well within the limits of
the recommendations of the Morse and Burton process, it is
clear that barium metaborate is completely converted into
carbonate with consequent /zberation of boric acid, and that
alcoholic solutions in no way prevent the action.

It might be argued that, in an evaporating mixture contain-
ing barium carbonate and boric acid in 78 per cent solution,
the reaction would be reversed after the volatilization of the
alcohol and the solutions became more concentrated. Indeed,
Morse and Horn use this explanation (p. 110). To test this
hypothesis, equivalent parts of barium hydrate and borie acid
in alcohol were treated with carbon dioxide for two hours, the
mixture evaporated to dryness and the carbon dioxide remain-
ing in the dry residue determined. The following results
were obtained :

Ba(OH), taken, 0°4554 grm.; B,O, taken, 0°1860 grm. ;
aleohol, 82 percent; CO, found, 071018 grm.; CO, in total
barium carbonate, 0°1169 grm. This shows that even with the
very strong alcoholic solution, which was used by mistake,
about 90 per cent of the metaborate was converted into carbo-
nate and remained as carbonate after bringing to dryness.

To demonstrate, in another way, the presence of boric acid
after the treatment recommended by Morse and Burton, I
have used the methyl alcohol flame-test. Equivalents of
barium hydrate and boric acid (I have used the word equiva-
lents repeatedly, meaning quantities to form exactly the most
important salt under discussion, viz: metaborate) were evapo-
rated to dryness, methyl alcohol added and burned without
getting the slightest indication of free boric acid. Again,
equivalents of barium hydrate and boric acid were treated
in aleoholic solution with carbon dioxide, the ethyl alcohol
and water removed by evaporation, and methyl alcohol applied
and. burned with unmistakable indications of free boric acid—
burning with a solid green flame.

Since free boric acid does not decompose barium carbonate
upon evaporation, the question arises as to what is the tem-
perature required to bring about this decomposition. The
following experiment throws some light upon this point : finely
divided anhydrous borie acid (071925 grm.) and barium car-
bonate (0°5177 grm.) were mixed in a platinum crucible,

52 Jones— Action of Carbon Dioxide

moistened with water and allowed to stand sixty hours. The
contents of the crucible were then dried on a steam bath and
afterwards heated for thirty minutes in an air bath at 350° C.
The carbon dioxide remaining in the mixture was determined
and found to be 0:0844 grm. Total CO, in BaCO, taken,
0°1154 grm.; CO, liberated by B,O,, 0°0310 grm. ; CO, dis-
placeable with formation of Ba,OB,O, by B,O, taken, 0°1210
erm. Therefore, 74°38 per cent of the boric acid taken, if
still unvolatilized, remains uncombined—at 350° C.

From the foregoing experiments, then, it is obvious that
carbon dioxide decomposes the metaborate of barium in either
aqueous or alcoholic solutions, and a portion, at least, of this
boric acid liberated not only remains uncombined after evapo-
rating to dryness but even upon heating to a high temperature.

Loss of boric acid by volatilization.—It is scarcely conceiv-
able that a mixture of boric acid and barium carbonate in water
should not lose an appreciable amount of bori¢ acid on evapo-
ration. Morse and Horn have found a detectable loss of boric
acid from treating aqueous mixtures of barium hydrate and
boric acid with carbon dioxide and evaporating. The carbon
dioxide was stopped shortly after the alkaline coloration with
phenolphthalein was destroyed. The quantity found by them
is given, and varies from 0°0001 grm. to 0°00016 grm.; in one
instance exactly 0:00012 on 0:0633 grm. of total borie acid

( 0: 3 0°02538
ee 13-77 | + 8448 = 0-0633 grm.

present:

The decrease in weight should be even more noticeable in the
presence of alcohol, though Morse and Horn did not detect
more than 0°0040 grm. B,O, [ (1/14200)x 50 = 0:0040] and
this only when the concentration of alcohol arose above 92 per
cent. Amounts below one part in 14,200 they were not able
to discover.

The actual amount of boric acid volatilized on evaporating
such mixtures to dryness is, of course, small, but certainly
appreciable. Furthermore, under the analytical conditions
described by Morse and Burton, the greater loss of boric acid
doubtless occurs with the escape of water of combination
between the temperature 100° C. and that at which the boric
acid completely replaces its equivalent of carbon dioxide in
barium carbonate. This temperature [ have shown by experi-
ment to be at least above 350° C. On the other hand, as was
shown in my original paper, under extreme conditions, by
the use of methyl alcohol and a continuous current of carbon
dioxide, it is possible to volatilize all the boric acid from a
mixture of barium hydrate and barium metaborate.

The interaction at high temperatures of the carbonates and
borates of barium.—\l have shown that the treatment recom-

a!

|

on the Borates of Barium. 53

mended by Morse and Burton may result in a mixture con-
taining barium carbonate and barium metaborate, or barium
carbonate, barium metaborate and boric acid, or, indeed, barium
carbonate and free boric acid. Of these compounds, both the
barinm metaborate and borie acid retain water considerably
beyond 100° C.

Morse and Burton assumed that they had, after evaporation,
a mixture containing only barium carbonate and barium
metaborate. Morse and Horn explain it as follows: “It may
be urged that even if Morse and Burton had the metaborate in
insoluble condition, and it had been sensibly attacked during
the treatment with carbon dioxide, nevertheless, in the subse-
quent attempt to heat the dried residue to constant weight,
the metaborate must have attacked the carbonate. It has
since then come to light that some caution must be exercised
in this part of the manipulation. In the original description
of the method, it was simply stated that the residue was
heated to constant weight-over a triple burner. The practice
then, and on the few occasions when the process has since
been used, was to hold the burner in the hand and rapidly
play the flame over the platinum dish in order to secure as
uniform a temperature of the contents as possible. The object
in using a triple burner was not to obtain a very high tempera-
ture, which is not necessary, but, rather, to employ a flame
large enough to keep the whole dish hot. No difficulty was
experienced, when the heating was conducted in this manner,
in quickly obtaining constant and correct weights. Neverthe-
less, it must be admitted, the authors of the method did not at
that time suppose that the temperature at which the meta salt
will attack the carbonate is so low as it has since been found
to be. The temperatures employed by Jones were evidently
Jar above those to which Morse and Burton heated their mix-
tures of metaborate and carbonate.” |

The use of a triple burner to secure gentle and uniform
heating is unusual. Indeed, Morse and Horn use it in experi-
ment XIV fora quite different purpose, i. e., to secure a tem-
perature which decomposed the carbonate and borate mixture.
present. That the temperatures employed by me in my
original experiments were not “far above those” used by
Morse and Burton, can be answered best by the words of Morse
and Horn: “ At a full red heat the meta salt attacks the car-
bonate with expulsion of carbon dioxide” (page 130). ‘The
temperature at which a mixture of the metaborate and the car-
bonate of barium is stable and which is still sufficiently high
to insure the dehydration of the former, appears to be just
under a red heat” (page 130). Morse and Horn do not, how-
ever, produce any evidence to show that the meta salt is "dehy-

54 , Jones—Action of Carbon Dioxide

drated below red heat. In fact, their experimental evidence
points the other way: “The experiment of heating the mate-
rial, after treatment with carbon dioxide, until the weight fell
below that calculated for a mixture of metaborate and ear-
bonate, and then of exposing it in the air until the weight
became very nearly constant, and finally, of removing the
slight excess of weight by placing the material over calcium
chloride or sulphuric acid, was many times repeated and
always with a satisfactory degree of success except in two
cases. In these it was suspected that, while some portions of
the material had been heated high enough to decompose the
carbonate, other portions had not reached the temperature
required for the complete dehydration of the metaborate ; for
it was observed that the weights gained in the air on these
occasions far exceeded the calculated deficits.” Evidently, the
temperatures at which decomposition of the carbonate and
dehydration of the metaborate of barium takes place, cannot
be far apart when both processes are going on in the same
crucible. .

I quote further (page 133): “In general, it was found diffi-
cult, by heating in our bath, to bring the too great initial
weight of the material down to that calculated for a mixture
of metaborate and carbonate, though in one case (No. XXX)
this was accomplished by heating in the bath for three hours
over three burners. The temperature on this occasion, how-
ever, rose so high that the thermometer was removed.” The
thermometer used registered as high as 550° C.

The conditions described above can only be accounted for
in two ways, viz: (1) The metaborate has been decomposed
by carbonate dioxide, though this treatment was stopped five
minutes after the disappearance of alkalinity with phenolph-
thalein—and the barium carbonate has not at this temperature
been decomposed ; or (2) the metaborate still contains water.

A combination of both these explanations, doubtless, more
nearly expresses the truth. Morse and Horn, however (page
134), seem to prefer the latter explanation. “ This,” they say
(i.e., the behavior of the ignited mixture in experiment XXX),
“led us to suspect that the temperature at which complete
dehydration occurs and that at which the metaborate begins to
attack the carbonate cannot be very far apart.”

The experiments made by Morse and Horn in an effort to
show how these mixtures can be “ quickly” brought to “ con-
stant and correct weights” are interesting, but not conclusive.
They proposed (page 119) “to determine also the effect of
exposing these mixtures, after heating, in an atmosphere con-
taining carbon dioxide” with the object of finding “the tem-
perature at which it was presumed a mixture of the metaborate

on the Borates of Barium. 55

and the carbonate of barium is stable, or rather the temper-
ature at which it becomes stable.” All these experiments,
and those in which the mixtures were exposed to the air and
dried in a dessicator, prove nothing; in fact, Morse and Horn
only draw inferences from them (page 134): “‘ From this it
appears probable that the more basic borate which is formed
at high temperatures, is decomposed at ordinary temperatures
by the carbon dioxide of the air and reconverted into the
borate and carbonate. A similar absorption of carbon dioxide
takes place when material whose weight has been reduced
below the normal amount, is reheated in the bath at temper-
atures under 500° C. From this it is tzferred that normal
weights could be quickly obtained by first heating to a high
temperature and then at a lower one, but we have not yet
tried the experiment.”

Those final and satisfactory weights which are sometimes
obtained may be due to the presence of an excess of carbon
dioxide or water, or both, in an amount sufficient to replace
any boric acid volatilized in evaporation and especially with
the water of hydration. That Morse and Horn did have
_ present in these experiments at the beginning of evaporation
some free boric acid—or, if it is preferred, a more acid borate—
and a corresponding increase in the amount of carbonate, is
certain, because in all these experiments of which the results
are recorded, carbon dioxide was passed for not less than
fifteen minutes, or when phenolphthalein was used as an indi-
cator, usually for five minutes after the disappearance of the
alkaline reaction.

In my original paper I suggested the use of phenolphthalein
to prevent too great an excess of carbon dioxide and corre-
sponding decomposition of the metaborate. In a recent experi-
ment I have tested the action of a metaborate in alcoholic
solution upon this indicator, and find that exact equivalents of
barium hydrate and boric acid in 78 per cent alcohol give a
decided alkaline reaction, and that this alkaline coloration dis-
appears only when about 36 per cent more boric acid is added.
It is clear that phenolphthalein gives indication, not when the
excess of barium hydrate is converted into carbonate, but at a
point considerably beyond. Morse and Horn, then, were
working with barium carbonate and an acid borate, and in
many cases, most certainly, with barinm carbonate and free
boric acid.

If by accident—and I see no other way of doing it—the
action of carbon dioxide is stopped just when the excess of
barinm hydrate has been converted into carbonate, still noth-
ing is gained; for the experiments of Morse and Horn—if

56 JSones—Action of Carbon Dioxide on the Borates, ete.

they show anything—make it clear that even these salts can-
not be weighed with safety.

These experiments leave nothing more to be done to prove
the impossibility of bringing to definite and constant weights
the mixtures resulting from the method described by Morse
and Burton. Furthermore, in the promised forthcoming con-
tinuation of this study by Morse and Horn, I see no possibility
of any other deduction.

In conclusion, then, I reiterate:—First. Carbon dioxide
decomposes the metaborate of barium in either aqueous or
alcoholic solutions.

Second. The boric acid liberated may in part escape during
evaporation to dryness and especially before reaching that
temperature at which it has completely recombined with the
barium, replacing carbon dioxide.

Third. A mixture, containing barium carbonate and hydrated
boric acid, or barium carbonate with the hydrated metaborate
and boric acid, or even barium carbonate and hydrated metabo-
rate, cannot with safety be brought to definite composition for
weighing.

T. Holm—Studies tn the Cyperacee. 57

Art. XII.—Studies in the Cyperacee; by THro. Hom.
XVI. Carices (C. genuine) physocephale and leucocephale.
(With figures in the text.)

THE old section Psyllophore has generally been abandoned
in later years as being too artificial, because it comprises both
Vigneew and Carices genuine, and several authors have striven
to classify these monostachyous species in sections among
higher developed types as “forme hebetatw” of these. But
it is often a most difficult task to undertake such arrangements,
inasmuch as there are not a few monostachyous species that
seem to stand as isolated types so unlike the others, that in
several cases one feels obliged to consider them as species with
no immediate affinity to any section of the higher developed
types.—This is, for instance, the case with the remarkable
Carex Fraseri Audrews, with C. Breweri Boott and the puz-
zling C. physodes Bieb., which considered by themselves
exhibit characters that are sufficient to make them represent
groups of their own, in other words, “ monotypic sections.”
Nevertheless, the only suggestion that has been made, so far,
in regard to the classification of these three species was to refer
them all together with a fourth species, C. Angelmannii Bailey,
to one section: Physocephale, alluding to the inflated utricle.

A critical examination of these species has, however, con-
vinced us that they do not demonstrate any very close aflinity
to each other, hence that they ought to be regarded as consti-
tuting several sections instead of one; moreover C. physodes
is truly’a member of the Vzgnew. Having made a special
study of monostachyous species for the purpose of connecting
them with the higher types, C. /raserz has actually been one
of those whose possible affinity has aroused our curiosity more
than most of the others. So far, we have been unable to trace
any affinity to the other sections, as defined by Drejer,* near
enough to unite it with any of these. And a consideration of
its supposed allies, enumerated above, has not thrown any light
upon the subject, but rather induced us to make a segregation,
Let us examine each type of this section “ Physocephale” by
itself, and first of all C. Frasert. In regard to the peculiari-
ties of this species it is not necessary to present any long dis-
cussion, since we have already described the species in this
Journal,t but we may recall a few of the facts to which we
have called attention, viz: The monopodial ramification of
the rhizome with its single assimilating leaf; that this leaf has
no sheath or ligule differentiated, no midrib and no bulliform

* Drejer S., Symbole Caricologice. Opus posthumum ab academia scient.

Danica editum, 1844, p. 9.
P Vol, ii, 1897; p. 121.

58 T. Holm—Studies in the Cyperacec.

cells, which is the more peculiar when we consider the unusual
breadth of the leaf-blade in contrast to other broad-leaved
species; furthermore that the stem above ground, the culm,
is flat and hollow. These are features of such importance that
the species may be well separated from all the other Carices as
occupying a section of its own among the Carices genuine.
Its snow-white inflorescence and broad, deep-green leat gives to
the plant a very singular aspect, yet the minor structure of the
staminate and pistillate flowers is that of a true Carex.

The utricle is small for the size of the plant and the texture
is very thin on account of the very delicate structure of the
epidermis and the poorly developed mesophyll, besides that the
very weakly developed mestome-bundles are only supported by
a few strata of thin-walled stereome. None of these characters
are mentioned by Professor Bailey, who established the section
Physocephale. *

In passing to describe the other species, formerly supposed
to represent allies, C. Brewert Boott also shows, to some
extent, a peculiar habit, though only in respect to the structure
of utriculus. Boottt+ describes the utricle in this way: “ peri-
gyniis ovalibus amplis inflatis tenuissimis, leviter nervatis glab-
ris fulvis rostellatis, ore albido oblique secto,’ and he com-
pares the species, as far as concerns the large, inflated perigy-
nium, with C. Mertensiz and C. Banksii ; the plant, however,
shows no affinity to any of these two species. The utricle of
C. Brewert is, thus, very large and inflated, altogether very
unlike that of C. Fraseri, and we might state furthermore
that the anatomical structure is, also, very different. Only the
outer epidermis persists; it is very thin-walled and constitutes
the only tissue between the two mestome-bundles, which are
surrounded by a few layers of mesophyll, while no stereome
was observed in our material. The leaves of C. Breweri are
very narrow, almost filiform, with a shallow groove on the
upper surface (fig. 5), and glabrous with exception of the mar-
gins, where short, prickle-like projections are developed in a
few rows. The epidermis shows the outer cell-walls very much
thickened, and the stomata, which are here restricted to the
dorsal face of the blade, are slightly sunk below the outer cell-

wall of epidermis; no bulliform cells are developed. The

mesophyll consists of about six layers of palisades, arranged
vertically on the leaf-face, partly surrounding the mestome-
bundles and located close to the dorsal face of the blade. But
the greater portion of the leaf is occupied by a colorless tissue,
which soon breaks down and forms lacunes of quite considera-
ble width. The stereome is thick-walled and accompanies the

* A preliminary synopsis of North American Carices, 1886, p. 152.
+ Ill. genus Carex, vol. iv, p. 142.

:

T. Holm —Studies in the Cyperacee. 59

larger mestome-bundles as-hypodermal groups on the leptome-
side of these, while it is much less developed on the hadrome-
side; the smaller mestome-bundles have mostly only a few
stereome-cells on the leptome-side; an isolated group of ster-
eome is located in the leaf-margin. In regard to the mestome-
bundles, these are all arranged in one band, large and oval in
alternation with small, almost orbicular; they are surrounded
by a thin-walled, green parenchyma-sheath, besides by a mes-
tome-sheath of which only the inner eell-walls are thickened ;
no thick-walled mestome-parenchyma was observed.

Carex Breweri exhibits a leaf-structure, which corresponds
in most respects with that of the two similarly narrow-leaved
Carices from the dry rocks in the alpine and arctic regions: C.
nardina Fr. and C. elynoides nob. besides with that of Alyna
Bellardi Koch. When compared with the narrow-leaved spe-
cies, which occur in wet places, for instance: C. pulicaris L.,
gynocrates W omsk}., dioica L., paralela Leest. and exilis Dew.,
we noticed a prominent difference in the leaf-structure, depend-
ing upon the non-differentiation of the large colorless tissue,
besides that the cross-section is more or less triangular in
contrast to the almost hemicylindric outline, observed in C.
Breweri.

In regard to the culm, this is cylindric, furrowed and gla-
brous; the internal structure corresponds with that of the leaf
and the central portion is occupied by a large, thin-walled pith,
separated from the narrow layers of mesophyll by wide lacunes.
There is, thus, a very important difference from a morphologi-
eal and anatomical viewpoint between C. Hraserzand C. Brewerr,
and no evidence of affinity, close enough to consider them as
members of the same section is, so far, apparent.

The third species, which by Professor Bailey is enumerated
as one of his Physocephale, is C. Engelmannii Bail., very
briefly described in his preliminary synopsis (I. ¢.). This inter-
esting species was first collected by Dr. Geo. Engelmann, in
the upper Clear Creek region of Colorado, at an elevation of
12,000 ft., and it appears to be very rare, as it has only been
found once since by Dr. P. A. Rydberg, who rediscovered it on
the high mountains near Silverplume, Colorado, in the same
region from where it was first recorded. Since the species is,
thus, very little known, we have thought it worth while to
present an illustration of it (fig. 1), showing one of the speci-
mens, which Engelmann collected, and which was kindly loaned
to the writer by Professor William Trelease. Furthermore we
append a diagnosis, since, as stated above, the plant has hitherto
only been briefly described, and since some of its more salient
characters have been overlooked.

60 T. Holm—Studies in the Cyperacea.

Fic.1. Carex Engelmannii, natural size.

Fic. 2. Scale of pistillate flower of same, magnified.

Fic. 3. Pistil of same; R = the rhacheola, magnified.

Fie. 4. Longitudinal section of utriculus with the pistil and rhacheola of

same, magnified.

Fic.5. Transverse section of leaf of Carex Breweri ; V.F. = ventral face ;
magnified.

Fic. 6. Transverse section of same; Ep. = epidermis of the ventral face ;
C. P. = part of the colorless tissue ; magnified.

Fic. 7. Inflorescence of Carex physodes ; copied from Boott’s plate ; two-
thirds natural size.

T. Holm—Studies in the Cyperacee. 61

Carex Engelmannii: Rhizome stoloniferous, densely matted :
culms numerous until 15™ in height, erect, cylindric, sulculate,
glabrous: leaves a little shorter than the culms, often curved,
filiform, scabrous above, their sheaths light brown, persisting,
not fibrillose: spike one, andypgynons, globular, 1°" in length:
scale of staminate flower lanceolate, reddish brown with pale
midrib: scale of pistillate flower (fig. 2), broader, acute, red-
dish brown with pale, excurrent midrib and membranaceous
margins: utriculus (fig. 4) membranaceous, glabrous, tapering
into a distinct bidentate-beak, at maturity not surpassing the
scale, two-nerved, the nerves obsolete: caryopsis stipitate,
obovate: rhacheola (R. in fig. 3) quite long: stigmas three.

The presence of the very distinct rhacheola was observed by
Engelmann, who had made some sketches in lead-pencil repre-
senting the utricle and the pistil, which accompanied the
specimens. In a pressed and dried state the species somewhat
resembles C. Breweri, but when prepared in diluted alcohol
the utricle shows a marked difference by being much narrower,
much less inflated and especially by the beak being bidentate;
but the anatomical structure of this organ, the utricle, agrees
otherwise with that of C. Breweri.

In considering the leaf of C. Engelmann, this is somewhat
broader and distinctly conduplicate, with a larger ventral face;
the outer cell-wall of epidermis is heavily thickened, and
roundish papille abound on the upper surface; no bulliform
cells are developed, and there are only two or three strata of
colorless tissue on the upper face in the middle of the blade,
separating epidermis from the green parenchyma. The meso-
phyll consists of very irregular palisades with large intercellular
spaces and shows besides several lacunes of considerable width.
The stereome and the mestome bundles show the same develop-
ment as observed in C. Breweri.

The culm is very thin and furrowed, and exhibits a very
open internal structure. The cortical parenchyma is traversed
by wide lacunes separated from the pith only by a single or
two layers of green parenchyma, and the pith itself is very
thin-walled and broken down in the center.

The fourth member of Professor Bailey’s Physocephalw is
Carex physodes Bieb., but it is hardly necessary to mention
this species any further since it being a “ Vignea” can not
possibly be placed in this section of “Carices genuine.” We
have copied one of the inflorescences, figured by Boott (fig. 7),
and it is at once noticed that the utricles exhibit a very anoma-
lous structure, described by Treviranus in Flora Rossica as
follows: “ Utriculus maturescens magnitudine et forma fruc-
tuum Corni mascule et Taxi baccate, angulis destitutus,

62 T. Holm—Studies in the Cyperacee.

multisuleus, glaberrimus, colore foliorum V2tzs aridorum, rostro,
quod ante maturitatem bifidum est, nunc obsoleto. Species
nulli ceterarnm affinis.’” Boott, in illustrating the species,
reprints Bieberstein’s diagnosis and declares himself unable to
decide its affinity unless it be near C. baccans Nees, which,
however, is one of the “Jndicw.” It isa very rare plant, but
has been recorded again by Meinshausen* from sandy soil on
the Southern plains between Volga and the Caspian sea, and
this author enumerates it among the Homostachye “ spiculis
homogeneis androgynis.”

But neither Boott or Meinshausen have expressed any doubt
as to the normal condition of this singular plant. However,
Treviranust himself, but several years after the publication of
Ledebour’s Flora Rossica, calls attention to the possibility of
the utricles being abnormally developed, and the abnormality
being evidently due to the puncture of some insect, as we are
used to observe in other species of Carex, e. g. C. vulpina, C.
paniculata and very commonly in C@. disticha. And there are
actually some points in Bieberstein’s diagnosis which seem to
confirm Treviranus’ suspicion, namely, the structure of the
anthers, “ polline emisso, ad basin usque in lobos duos dehis-
centes, ut tune filamentum apice biantheriferum appareat,”
and the structure of the utricle, ‘“dentibus oris duobus conni-
ventibus exilibus, in maturo vix superstitibus,” besides the
comparison of the utricle with the fruit of Cornus or Taxus.

The section Physocephale as established by Prof. Bailey is
thus, to say the least, an assemblage of incongruities and can
not be maintained. Carex physodes, whether it be abnormal
or not, must be placed among the “ Vignew” on account of its
inflorescence and two stigmas. Carex Frasert can not by any
- means be regarded as an ally of Carex Breweri on account of
its morphological and anatomical peculiarities, and we suggest
to place it in a section of its own, to which we designate the
name “Leucocephale,” characterized as follows: Spike one,
androgynous, snow-white; perigynium thin, glabrous, elliptie
with the beak very short, obliquely cut; leaf broad and flat
without sheath, hgule or midrib; culm flattened; stigmas 3.
Only species Carew Frasert Andrews.

Carex Engelmann appears, on the other hand, well refera-
ble to Drejer’s Physocarpe as a “forma hebetata,’ and of
which the diagnosis of the section is defined as follows:
“Spicee mascule feminezeque plures, breviter pedunculate.
Bractes foliaceze, evaginate. Perigynia membranacea, nervata,
glabra, valde inflata, rostrata ore distincte bifido vel bicuspi-

* Cyperaceen der Flora Russlands. (Acta hort. Petrop. vol. xviii, 1901.) -
+ Ad Caricographiam Rossicam, Moskow, 1863, p. 4.

T. Holm—Studies in the Cyperacee. 63

dato. Stigmata terna, stylus plerumque intra perigynium
flexuosus, caryopsis trigona vel triquetra.”

In several respects the utricle of O. Engelmannit reminds of
that of C. pulla Good. and its nearest allies, while we cannot
agree with Drejer, that such types as C. ‘microglochin, C.
paucifora and C. pulicaris should be looked upon as pri-
mordial types of this same section. C. pulicaris, for instance,
differs greatly from any of these by the structure of its utricle
alone, and we suggest the affinity of C. microglochin and C.
paueiflora rather to be sought among the Echinostachye Dre}. :
“perigynia vix inflata, demum plus minus squarrosa ” both
of which characters are readily recognized in C. paucifiora
and its allies. As regards C. Breweri, this may “ad interim ”
be retained as the only member of the Physocephale, and this
section appears to be very near the Physocarpe, evidently
parallel with this; but at present we are unable to refer it to
any of the already established sections in which higher devel-
oped types are represented.

Brookland, D. C., April, 1902.

64 Scientific Intelligence.

SCIENTIFIG ANGEL LET GE Nae

I. CHEMISTRY AND PHYsICcS.

1. Probable Source of the Heat of Chemical Combination, and
a New Atomic Hypothesis.—Professor T. W. Ricuarps has
brought forward the very suggestive generalization that the con-
traction exhibited during chemical combination is in many cases
approximately proportional to the heat evolved; hence he be-
lieves that the chief source of heat of chemical combination is
the work performed in compressing the material. He gives an
explanation upon the same basis of the mechanism of the heat of
adsorption, adhesion and change of allotropic form. He is led,
from these considerations, to the interesting hypothesis of com-
pressible atoms, for the details of which, with ingenious applica-
tions, reference must be made to the original article-— Proc.
Amer. Acad., xxxvii, 399, and Zeitschr. physikal. Chem., x\, 597.

H. L. W.

2. Lithium Silicide—By heating silicon with metallic lithium
in a nickel boat in an exhausted tube, Moissan has found that the
two elements combine and that the excess of lithium used may be
distilled off between 400 and 500°, leaving a bright indigo-blue
product in the form of brilliant crystals. The compound has the
composition represented by the formula Si,Li, and corresponds to
the spontaneously-inflammable hydride of silicon, Si,H,, which
was recently described by Moissan and Smiles. The lithium
silicide is very hygroscopic, is violently decomposed by water,
acids and the halogens, and, as would be expected, has very
powerful reducing properties. Thus, when heated with the
oxides of iron, chromium, manganese and calcium it gives the
metals or alloys, but it does not decompose alumina.— Comptes
Rendus, cxxxiv, 1083. H. L. W.

3. The Black Color of the Rocks in the Cataracts of the Nile.
—Lortet and Hugounrng state that the rocks in the Nile rapids
at Wady-Halfa and at Assouan, which are covered at the time of
high water, are very smoothly polished and present a deep black
color when visible at low water. This is not due to the colors of |
the rocks themselves, which are eruptive rocks and sandstones of
various colors, but to a thin black coating. When this surface
is treated with strong hydrochloric acid a brown liquid is obtained
which loses its color upon heating, chlorine is given off and the
liquid gives reactions for manganese. This shows that the black
coating is due to manganese dioxide. The rocks contain small
quantities of manganese not in the form of dioxide, but the lat-
ter is produced upon their polished surfaces.— Comptes Rendus,
Cxxxiv, 1091. H.-L... W:

4. Bismuth-lead Sulphide-halides.—Three compounds, made
by fusing lead halides with bismuth sulphide, have been described
by Ducarrr. They form very small acicular crystals in the

Chemistry and Physics. 65

cavities of the mass which has been fused. The formulas given
are
PbS - Bi,S, - 2 BiSCl,
PbS - Bi,S, - 2BiSBr, and
PbS - Bi,S, - 2BiSI.

No evidence is given to show that the halogens are combined
with bismuth rather than lead, hence it is not clear why the
above formulas have been adopted instead of simpler ones which
are perhaps more plausible on account of their less complicated
form; for instance, PbCl, - 2Bi,S,— Comptes Rendus, cxxxiv,
1061. Ee La WY.

5. The Atomic Weight of Uranium.—RicHarps and MErtI-
GOLD have made a new investigation upon the atomic weight of
uranium by analyzing uranium bromide, UBr,. The method
differs essentially from any that has been used hitherto for the
purpose, and apparently the results are more reliable than those
of previous investigators. ‘The results show that this atomic
weight is about a unit lower than the value that has been pre-
viously accepted, the average of several closely agreeing deter-
minations being 238°53.— Proc. Amer. Acad., xxxvil, No. 14.

H. L. W.

6. First Book of Qualitative Chemistry ; by Prescorr and:
Suttivan. Eleventh Edition. 8vo, pp. 148. New York, 1902
(D. Van Nostrand Company, price $1.50). — This text-book,
which has been in use since 1879, has now been entirely re-written
in order to present the subject in the light of the modern theories
of solution and of mass-action. Some attention is paid also to
the classification of the Periodic System.. The first 37 pages of
the book are devoted to an introduction, chiefly of theoretical
character, including concise presentations of the subjects of
electrolytic dissociation, chemical equilibrium, solubility, etc. In
the practical part of the book only a moderate amount of use is
made of the ionic theory. The equations introduced are limited
to the more difficult ones, and these are given without regard to
ions. The book contains many satisfactory variations from the
time-honored methods that have come down through Fresenius
from Will and Rose, but it is surprising that the calcium sulphate
method for testing for strontium, which was abandoned by
Fresenius, is retained here, since the treatment of the nitrates
with amyl alcohol is a much more satisfactory method.

The course outlined in this book is a rather extensive one,
which is better adapted for beginners in college than for younger
students. Abbreviated tables for analysis are adopted, but they
are used in connection with many explanatory notes. Man
teachers would prefer to have these aids to thoughtless labora-
tory work omitted altogether. H. L. W.

7. Grundriss der qualitativen Analyse, von Dr. Wi1iH. Bort-
GER. 8vo0, pp. xxii, 249. Leipzig, 1902 (Wilhelm Engelmann).
—This text-book on qualitative analysis has been written from

Am. Jour. Sc1.—FourtH SERIES, VoL. XIV, No. 79.—Juty, 1902.
5

66 Scientific Intelligence.

the standpoint of the ionic theory, and comes from an assistant
in Professor Ostwald’s laboratory. It is evident that the work
has been very carefully carried out on the proposed plan, and it
will be useful to those who desire to employ the modern theories
of solutions in teaching the subject. Perhaps it is doubtful if it
is advisable to give as much prominence to the doctrine of ions
and the law of mass-action as is given in this book, particularly
with beginners in the study of chemical analysis, but the tendency
of teachers seems to be in this direction.
The equations in the ionic form, which are extensively used in
this book, appear more complicated and less easy to grasp than the
usual equations, on account of the separation of the ionic symbols
by the plus sign. For instance, the form of the usual equation,

2AgNO,+K,CO,—Ag,CO,+2KNO,,
is changed to
2Ag*+2NO0,'+CO, "4 9K =Ag CO, 42K: +2NO,’.

This last equation does not represent the actual conditions with
complete accuracy, because in moderately concentrated solutions
ionization is not complete, and moreover it is not as easy to see,
as in the first equation, what salts are mixed to produce the
reaction. It would seem preferable, if ions are to be marked, to
give equations some such form as the following one :

2Ao*NO,'+Na,"CO,"=—Ag,CO, + 2Na NO,’

A somewhat loose statement is noticed in the explanation of .
ionization (S. 13) where it is stated that “always an equal num-
ber (eine gleiche Zahl) of positive and negative ions is formed or
disappears.” Ofcourse these numbers vary according to valency.

The author has not used the results of the more recent investi-
gations in all cases. If he had carefully consulted Fresenius’s
last edition, he would probably have pointed out the difficulty of
detecting small quantities of silver chloride in the presence of
mercurous chloride, and would have given a method for over-
coming this trouble. He should have avoided the unreliable
method of testing for strontium, in the presence of much calcium,
with calcium sulphate solution. He should not have given the
formula NH,Hg,Cl, but NH,HgCl+Hg for the familiar black
substance produced by the action of ammonia upon mercurous
chloride. The formula SbCI,5CsCl should have been 2SbCI,-
8CsCl, as has been shown by at least four different investigators,
the first of whom published his results in 1882. HL, We

8. Reflection Power of Metals for Ulira-violet and Ultra-red
Rays.—E. Hacen and H. Rusens extend their previous work
on this subject, which extended from X= 450 to 700 pp. They
abandoned their previous photometric method and adopted a
heat method, consisting of the use of a linear thermal element
(H. Rubens, Zeitschr. f. Instrumentenk., 18, p. 65, 1898). This
heat method could be ‘used as far as 221 LE 5 for the authors

Chemistry and Physics. 67

discovered that an arc lamp gives carbon bands as far as this
wave length in the ultra-violet which are of sufficient heating
effect. Some of the results of the previous photometric work are
corrected.

The reflection power of silver falls off quickly below 450 py
and reaches a minimum of 4 per cent approximately at 320 py.
While silver for visible rays is the best reflecting of all metals, it
is the poorest between 250-350 wu. Platinum, iron and nickel
show approximately the same reflecting power. The reflecting
power of speculum metal was investigated and is of especial
interest in view of concave gratings. Magnalium shows an
exceptionally high reflecting power. The reflecting power in the
extreme ultra-red possesses far more reguiarity than in the visible
and ultra-violet regions.—Ann. der Physik, 5, 1902, pp. 1-21.

el JE

9. Spark Discharge from Metallic Poles in Water. — Dr.
Wiutsine of Potsdam believes that the peculiarities of the spec-
trum of Nove can be explained by pressure, and accordingly has
studied spectra produced by the electric spark in water where it
is presumable that great pressure results. He arrived at the con-
clusion that displacement of lines and double lines occurred
which were in every way similar to those observed in the spec-
trum of Nova Aurige. Sir Norman Lockyer has repeated Dr.
Wilsing’s results and comes to the conclusion that shifts are
obtained under water which are in opposite directions to those
noticed in the dark lines of the spectra of Nove and that the
shift in the star does not arise from pressure.—Vature, May 22,
1902.

10. “Hlectrical Effect of Light. — P. Lenarp contributes an
exhaustive paper on the effect of ultra-violet light in generating
cathode rays, and the bearing of this phenomena on radio-active
bodies. The velocity of the rays so excited is approximately
ivov Of the velocity of light. These rays show diffusiveness and
must be absorbed in great measure by gases. ‘The method of
measure was by means of loss of charge.—Ann. der Physik., 5,
1902, pp. 149-198. J. T.

11. Interference Tubes for Electrical Waves.—V arious attempts
have been made to employ Quincke’s resonating tubes for the
detection of electric waves. Aucusr BecxEr has made a num-
ber of measurements with such tubes and with a coherer. The
waves were propagated along the surface of the tubes and issued
into the air with scarcely diminished energy. Their length
within the tubes was 10™ and in free air 7°5°™. The tubes were
filled with various liquids and the dielectric constants determined.
A list of such constants is appended.—Ann. der Physik, 5, 1902,
pp. 22-62. J. 2:

12. Law of Velocity which Cathode Rays suffer by Reflection.
—If there is such a law, the reflected rays should be more power-
fully diverted by a magnet than the direct rays. E. GEHRCKE
finds that this happens. As an exciter of the rays he employed

68 Scientifie Intelligence.

a static machine which gave a constant current. Different reflect-
ing surfaces were inv estivated. The writer refers to the discov-
ery of Hertz that thin metal films are penetrated by the cathode
rays and draws the conclusion that the corpuscles can move
within a metal without losing their charges. If one makes the
further assumption, that the corpuscles also can lose in internal
reflection, and that the loss of velocity of each corpuscle increases
with the number of collisions with the particles of the reflectors,
one comes to the conclusion that a strongly absorbing body like
platinum will give forth fewer corpuscles than a weaker absorb-
ing body like magnesium. The author’ s results seem to confirm
this theory. — Ann. der Physik, 5 , 1902, pp. 81-93. J, 7.

13. Role of Self-Induction in aaseheh: ges of Electricity through
Gases.—Various observers have shown that certain spectral lines
disappear when suitable self-induction is employed in oscillatory
currents. M. B. Eeiniris studies the changes in heat effects pro-
duced in the spark terminals by varying self-induction. M. A.
DE GRAMONT believes that the proper use of self-induction may
lead in physical chemistry to a method of separating the spectra
of dissociation of mietalloids:— Comptes Rendus, May 5, 1902,
pp. 1043, 1048. JT

14, The Theory of Optics ; by Paut Drupz. Translated from
the German by C. Riborg Mann and Robert A. Millikan. Pp.
546+xxl, New York, 1902 (Longmans, Green, and Company).—
Very nearly a half of the volume is devoted to a compendious
treatment of geometrical and physical optics, written in a clear
but very condensed style. The rigid deduction and application
' of the Principle of Huyghens, founded upon the theoretical inves-
tigations of Kirchhoff and of Voigt, will offer the most novel
extensions to the familiar discussions of the current texts.
It is difficult to conceive how a more admirable review of
physical optics could be contained in so concise a form, and
although it can hardly be regarded as a sufficient guide to a
_ novice in this field, it is sure to give pleasure to every reader
who possesses the requisite elementary knowledge.

Section II, of 220 pages, develops the electromagnetic theory
of light and applies it to the explanations of the enormously
varied optical properties of bodies. It is impossible to indicate
in a brief review the surprising extent of the phenomena dis-
cussed with: highly satisfactory completeness, or to point out
the vast departure in style of treatment from the best known
text-books on light. The last chapter of the section, however, in
which the theory of Lorentz regarding the modifications which
motion of the bodies in question introduces in optical phenomena,
may fairly be cited as not the least interesting.

The final Section treats, in three chapters, of the energy of
radiation ; of the application of the second law of thermo-
dynamics to pure temperature radiation; and of incandescent
vapors and gases. These are broadly philosophical and illustrate
the essential unity of all departments of physics.

Chemistry and Physics. 69.

As a whole, the book must be regarded as one of the most
important recent contributions to scientific literature. The work
of the translators appears to have been exceptionally well done.

isi rE

15. Elementary Principles in Statistical Mechanics developed
with special reference to the Rational Foundation of Thermo-
dynamics; by J. Wittarp Gises, Professor of Mathematical
Physics in Yale University. Yale Bicentennial Publications.
Pp. xvili + 207. New York, 1902 (Charles Scribner’s Sons).—
In this last volume by Professor Gibbs we have, as it were, the
conclusion of his earlier papers published so many years ago on
the subject of the Equilibrium of Heterogeneous Substances.

It rarely happens in science that there is one particular branch
of it in which only one or two men can be considered as preémi-
nent. To all intents and purposes this is the present condition
in advanced study of thermodynamics. The scientific world, as
a whole, owes a debt of gratitude to Professor Gibbs for at last
giving in printed form such a clear and definite exposition of the
Principles of Statistical Mechanics, with special reference to
Thermodynamics. It is true that in this present volume the
problem is treated as one of pure mechanics and not of thermo-
dynamics ; nevertheless, one’s thoughts always turn to the latter
to appreciate the physical meaning of the various equations.

‘Fhe best description of the scope of the volume is contained in
the preface written by Professor Gibbs himself, from which the
following remarks are quoted more or less accurately. In the
first chapter there is the general problem of which may be called
the determination of the fundamental equation of statistical
mechanics and its integration. In the second chapter certain
principles proved in the first are applied to the “Theory of
Errors”; while in the third chapter the same principles are .
applied to the “Integration of the Differential Equations of
-Motion.” In the fourth and following chapters, making fifteen
in all in the book, attention is directed to the consideration of
statistical equilibrium of conservative systems, with special appli-
cations to Thermodynamic analogies, such as Entropy, Tempera-
ture, etc.

It cannot be denied that the present volume is extremely diffi-
cult reading, but one feels, as one masters it page by page, that
there is obtained a deeper insight into the ultimate truths of
Physics. It is not a book which a university student would in
general find profitable reading, but it is one which every such
student should add to his collection of choice volumes in order to
have it at his disposal for reference or for consultation concerning
many of the perplexing questions of Thermodynamics. J. s. A.

16. Bettrdge zur chemischen Physiologie und Pathologie, heraus-
gegeben von Fr. Hormetster. II, 5-6, 1902 (Vieweg u. Sohn).—
in an earlier number of the Beitrige, Lewin concluded that indol
and phenol may have their origin in metabolic processes in the
animal body independent of putrefactive changes in the intestine.
This view is now combatted by P. Mayer, as the result of an

70 Scientific Intelligence.

experimental investigation of the excretory products found in
phlorhizin diabetes. In the latter he finds no increase in the
output of phenol, indol or glycuronic acid. LanesTErn has con-
tinued his study of the end-products of peptic digestion on crys-
tallized egg-albumin. The compounds thus far definitely isolated
include: leucin, tyrosin, phenylalanin, glutamic and aspartic
acids, cystin, lysin, pentamethylendiamin, oxyphenylethylamin
and a nitrogenous carbohydrate. The differences between peptic
and tryptic digestion thus appear to be largely quantitative
rather than qualitative in kind ; although in peptic digestion the
biuret reaction persists. NEUBERG and BLUMENTHAL report the
formation of isovalerylaldehyde and acetone from gelatin. Bret-
FELD, in an investigation of the iron content of the liver, found
that the liver cells of women are, as a rule, decidedly poorer
in iron than those of men. These differences do not appear
until the age of 20 to 25. Macgnus—Lervy has published the
results of a comprehensive and interesting investigation of
acid formation during the autolysis of the liver, under aseptic
and antiseptic conditions. The non-volatile acids thus obtained
include two lactic acids and succinic acid ; the non-volatile acids
formed are: formic, acetic and butyric acids. The gases H, H,S
and CQO, also arise. The application to our knowledge of certain
pathological conditions is pointed out. L. B. M.

Il. Gerotocy AnD Natura History.

1. United States Geological Survey. Examination of Forest
Reserves and Adjacent Regions; by Hmnry Gannetr. 21 Ann.
Report, Pt. V, 711 pp. ; 143 pls.; maps in accompanying port-
folio. This report includes descriptions of seven reserves and
wooded areas in different parts of the United States.

— 2, Minéralogie de la France et de, ses Colonies. Description
- physique et chimique des Minéraux; Etude des Conditions géol-
ogiques de leurs Gisements; par A. Lacroix. Tome troisieme, —
1°* fascicule, pp. 1-400. Paris, 1901 (Librairie Polytechnique,
Ch. Beranger, Editeur).—After an interval of nearly four years,
the second volume of this most valuable work is now followed by
the third. It contains the Oxides, Hydroxides, Nitrates, Plom-
bates and Manganites. A fourth volume, which is promised soon,
will complete the whole. To the mineralogist, who is acquainted
with the admirable features of the preceding volumes, it is un-
necessary to say more than that the part now issued is carried
through on the same plan and with like success.

3. Size of Grain in Igneous Rocks in relation to the distance
From the Cooling Wall; by Aueustin L. QuENEAv. Contribu-
tions from Geol. Dept. Columbia University, vol. ix, No. 80, pp.
181-195.—The Palisades (Triassic Diabase) and the minette dike
of Franklin Furnace, N. J., were chosen as fields for the study
of the causes of variation in size of grain in igneous rock. The
method used is based on Lane’s work (Geol. Survey Mich., vol.
vi), and the mathematical treatment upon the general theory of
cooling presented by R. S. Woodward (Annals of Math., vol. iii).

Geology and Natural History. TL

4, The Eparchean Interval—a Criticism of the use of the
term Algonkian; by ANDREW C. Lawson. Bulletin Dept. Geol.,
Univ. of Calif., vol. iii, No. 3, pp. 51-62.—In common with many
geologists Professor Lawson believes that the term Algonkian
has been used without sufficient regard for the accepted rules of
nomenclature, and in such a way as to minimize the great interval
of erosion between the Archzan and the Animikie series of Lake
Superior. The time sequence of the important rock groups of
the Lake Superior province is believed to be as follows:

{ CamBrian (Upper division or Potsdam only)

Unconformity.
PALEOZOIC + | Keweenawan.
| ALGONKIAN - Unconformity.
/ Animikie—Penokee= Upper Marquette.

EPARCHAZAN INTERVAL.
¢ Hurontan= =Upper Keewatin=Lower Marquette, etc.

Unconformity.
: | LAURENTIAN, so called, granite gneisses, etc., (intrusive
ARCH ZAN } in the Ontarian) and the Carltonian anorthosites.
[ Keewatin=Lower Huronian—Crystalline
ONTARIAN J schists of south shore invaded by
L Unconformity. [ granite-gneisses

. Coutchiching.

5. Former Extent of the Newark System ; by W. H. Hopss.
Bull. Geol. Soc., vol. xiii, pp. 139-148, 5 tigs.—The prevalent
opinion of geologists inregard to the Newark beds of the Atlantic
border is that they were “deposited in “ local basins,” and never
extended far beyond their present boundaries. Professor Hobbs
argues that the marginal faults favor the “broad terrane”
hypothesis and that the distribution of the coarse conglomerates
is not opposed to this theory. For a number of other reasons
_ Professor Hobbs joins Russell in the belief that the Newark areas
of the entire eastern border of the United States are remnants of
one widespread formation.

6. Quaternary History of Southern California ; by Oscar H.
Hersuey. Dept. Geol., Univ. California, vol. iii, No. I, pp. 1-30.
1 pl.—In his studies of the California Quaternary, Mr. Hershey
has examined the evidence of earlier writers and has added much
that is new in regard to the sedimentation and crustal move-
ments during the Quaternary. A time scale divided into seven
epochs has been constructed and the character of the erosion and
deposition of each epoch is given.

7. Plissements et Dislocations de [écorce terrestre en Gréce ; by
Pa. Neeris. Pp. 210; 2 maps. (Athens: C. Beck; Paris: C.
Béranger.)—Earth movements, principally foldings ‘on a large
scale, have affected Greece and parts of Turkey since Jurassic
times; the first movement (the Olympic) was pre-Cretaceous, then
followed movements during the Cretaceous and the Eocene. The
last folding (the Tenarus), which runs from N. to S., began late
in the Pliocene and affected all of Greece. The Tenarus folds
are considered part of the world-wide crustal movements of late

72 Scientific Intelligence

Tertiary times. Outbursts of igneous rock—peridotite and tra-
chyte—accompany the disturbances.

8. The Old Tungsten Mine at Trumbull, Conn.; by Wm. H.
Hozss. U.S. Geol. Survey, 22d Ann. Rept., Pt. II, pp. 13-22.—
This report gives a short description of the geology of this min-
eral locality together with a detailed geological map of the
region. Thedeposit is of interest chiefly on account of the occur- g
rence of the rare tungsten minerals, scheelite and wolframite, the |
latter occurring as a pseudomorph of the former. Other charac-
teristic contact minerals are found associated with them. These
minerals are found at the contacts between a crystalline limestone
and two beds of hornblende-gneiss, which lie, one above and the
other below it. The hornblende-gneiss is of igneous origin hay-
ing the characteristics of a hornblende diorite, which, however,
becomes gneissoid and altered near its contacts with the lime-
stone. Considerable work has been done at the locality in the
effort to develop a paying tungsten mine, but with little success.

Woe |

9. Peculiar Character of the Kruption of Mt. Pelee, May 8th. .
—The writer, in a communication to the Connecticut Academy
of Sciences, "May 14, when only the first great eruption had |
occurred, maintained that the déstruction of St. Pierre was due ;
to the eruption of a great volume of exploding oxy-hydrogen .
gases, due to’ the decomposition of water by contact with the |
intensely heated lava deep within the voleano.* |

Subsequent investigations of the nature of the destruction |
wrought there, and the character of the later er uptions, have |
fully confirmed this view.

At 2500° C. 50 per cent of the steam will be converted into |
these explosive gases at ordinary pressures, but at the enormous
pressures within a volcano it would require a higher temperature
to effect this. But there is no doubt whatever that temperatures
above 3000° exist within many volcanoes.

It is, therefore, only necessary to suppose that the highly heated
lava, by bursting through the intervening rock, came into sudden
contact with a more or less extensive body of subterranean water,
or even with porous rocks saturated with water, in order to ex-
plain the sudden generation of explosive gases.

If the subterranean waters thus converted into gases and
steam were sea-water, the chlorine of the salt would also be dis-
sociated from the sodium at the same or even lower temperatures,
and this gas would also form an explosive mixture with a part of
the hydrogen. By their union hydrochloric-acid gas would be
produced, which is a highly irritating, poisonous and suffocating
gas, powerful enough, even when present in small quantities, to
destroy animal and vegetable life, as has happened in several of
the eruptions of Vesuvius and other volcanoes near the sea.
However, in the case of St. Pierre, nearly all the people were
probably killed instantaneously by the outburst of intensely hot

* See Science, May 16, 1902.

Geology and Natural Aistory. 73

flames, due to the oxy-hydrogen gas explosions, which must have
penetrated to the interior of all buildings, scorching and burning
everything within reach, while the force of the explosions,
analagous to that of dynamite, was sufficient to demolish the
strongest stone buildings, tear the branches from the trees and
rend their trunks, dismount large cannon and transport them
some distance, and produce scenes of indescribable ruin of all life
and property, as is well shown in recent photographs. Probably
there were several of these explosive outbursts, following in rapid
succession, though a single one might have destroyed all life in
St. Pierre.

The recent reports of Mr. George Kennan and of Professor
Heilprin, who seem to have been the first to visit the new crater
which destroyed St. Pierre, show that this crater is an oblique
pit on the side of the mountain, facing St. Pierre, and with
several tunnel-like openings in the perpendicular wall, pointing
directly toward the city. Therefore, when the great eruption
took place, the vast jets of explosive gases were ejected hori-
zontally toward St. Pierre, which was thus swept by a “tornado of
fire,” as if it had been in the focus of an enormous blowpipe flame.

The burning gases and the very hot steam, resulting from this
explosion, instantaneously spread out on all sides, far beyond the
focus of the explosion, causing a wide zone of scorching and
burning, but without the total destruction seen in St. Pierre
itself. These hot and flaming gases caused the sudden burning
of the upper works of the vessels and great loss of life, far
_ beyond the limits of the more violent explosive action.

The explosive gases and flames evidently reached St. eee
sooner than the cinders and stones thrown out by the same out-
burst of gases, but the oblique direction of the mouth of the
crater was such as to produce a regular bombardment with hot
missiles of this kind, which added to the destruction of the build-
ings and to the debris in the streets.

That the woodwork left on some of the buildings and the heaps
of broken timber and trees in the streets were not burned up,
was probably due to the heavy fall of rain that immediately fol-
lowed the eruption, caused by the condensation of the steam in
the upper regions of the atmosphere.

Professor R. T. Hill, who reached Morne Rouge and witnessed
a late eruption (May 26) of Mt. Pelee from there, saw horizontal
flashes of light in the midst of the vast column of steam and
cinders emitted. These lightning-like flashes he attributed, no
doubt correctly, to explosive gases. They were evidently due to
the oxy-hydrogen gases, at that time arcaily diluted with steam
in the column of cinders, ete.

Had the openings of ‘the new crater been directed vertically
upward, as is usually the case, the loss of life and property
would have been comparatively small, because the destructive
energy would have been expended in the upper regions of the air.

This eruption, though of small volume, is of “ereat scientific

74 Scientifie Intelligence.

interest because of its rare and exceptional character. In fact, it
is the only example of this highly explosive type of eruption that
has occurred in modern times under circumstances that permitted
scientific examination, or the testimony of intelligent observers.
Eruptions of this kind might well be designated as the ozxy-
hydrogen explosive type, the principal destruction being caused
by the explosion of vast volumes of these gases, while in the
much more common type of eruptions the explosions are caused
mainly by the escape of steam under very high pressures. Ex-
plosive eruptions of intermediate kinds, due to various mixtures
of steam and the explosive gases, also occurred here. The later
eruptions (as on May 26) from the summit crater of Mt. Pelee
were evidently of this mixed kind, in which steam predominated.
A. E. VERRILL.

10. The Kuuna and Geography of the Maldive and Laccadive
Arehipelagoes. Edited by J. Stantey Garpiner. Vol.1, Part II,
pp. 119-222, 7 plates, 16 figs.—The principal coral reefs of the
Indian ocean were made the special object of study by an expedi-
tion headed by J. Stanley Gardiner in 1899 and 1900. The pres-
ent paper is a continuation of a previous report (this Journal, vol.
Xlll, p. 8321) and contains the following papers on fauna :

Amphibia and Reptilia, by F. F. Laidlow; Lepidoptera, by
Ed. Meyrick ; Echiuroidea, by A. E. Shipley; Sipunculoidea, by
A. E. Shipley; Land and Freshwater Mollusca, by Edgar A.
Smith ; Pigments of Corals, by C. A. MacMunn; Marine Crusta-
ceans, by L. A. Borrodaile ; Chaetognatha, by Leonard Doncas- |
ter; Dragon Flies, by F. F. Laidlow.

Mr. Gardiner has continued his studies on the origin of the
coral island groups (pp. 146-184) and finds that the subsidence
theory is untenable for the great Maldive plateau. To quote the
author: ‘‘My conclusion then is that an almost flat plateau at a
depth of 160 fathoms was at one time formed, and that on this
the banks severally arose.” The details of the formation of the
group are given graphically on page 183. Part III of vol. 1 is
promised for October, 1902.

11. Report on the Actinians of Porto Rico; by J. EK. DUERDEN.
Bulletin U. 8. Fish Commission for 1900, pp. 323-374, 13 plates.
—This valuable paper is devoted largely to the anatomy and
histology of the species treated, of which there were only thirteen |
in the collection. A single new species (Bunadosoma spherulata)
is described. Two of the plates give general views of five
species. The other plates are devoted to sections, and are well
_ executed. As VE

12. The Ascidians of the Bermuda Islands ; by W. G. Van
Name. Trans. Conn. Acad. Sciences, vol. xi, pp. 325-411, 19
plates. Feb. 1902.—This is an excellent monographic work on
the Bermudian Ascidians, which had previously received but
little attention. The total number of species and marked varieties
is 46, representing 23 genera. Four new genera are described
and 21 new species, besides 11 subspecies or marked varieties.

Geology and Natural [listory. 15

Most of the new forms and all the new species are compound
ascidians, which are unusually numerous at the Bermudas. Most
of these were studied from living specimens, as well as from large
series preserved in different media. Nearly all the species are
illustrated by anatomical drawings, and most of the new ones by
reproductions of photographs, showing the general appearance.
Some of the photographs were made from living specimens, under
water. It is altogether the most important work hitherto pub-
lished on the ascidians belonging to the West Indian faunal area.
A. EB. V.
13. The Embryology of a Brachiopod (Terebratulina septen-
trionalis) ; by E. G. Conxity. Proc. Amer. Phil. Soc., xli, No.
168, pp. 41-76. 10 plates. 1902.—This i, in many respects, one
of the most complete accounts of the early stages of a Brachiopod
that has been published. The author adds a discussion of the
zoological affinities of the Brachiopods, concluding that from
embryological data they are nearest to Pharonis and the Polyzoa.
A. EV:
14, Results obtained in Havana from the destruction of the
Stegomyia fasciata infected by Yellow Fever; LI. The Propa-
gation of Yellow Fever; by Major W. A. Gareas, Medical
Corps, U.S. Army. Sanitary Dep., Havana, Ser. 4, 1902.—These
two papers are of great interest and importance as demonstrating
that yellow fever in Havana is transmitted by this particular
mosquito, and in no other way. The disease was fully controlled
simply by destroying these mosquitoes in various ways, and pre-
venting them from gaining access to fewer patients, by the liberal
use of screens. By these means, and without special disinfection
of rooms or clothing, the fever was reduced to a minimum after
March, 1901, when this method wascommenced. No cases what-
ever occurred during the four months from October to January
inclusive, which has not happened before in 150 years or more.
The average number of deaths from yellow fever from April 1st
to December Ist, since 1889, had been 410.54, but by the anti-
mosquito methods it was reduced to5,in 1901. Yet in 1900, with
the most careful and elaborate methods of ordinary disinfection,
very little impression was made on the yellow fever, for there
were 1244 cases and 310 deaths in 1900, but in 1901 there were
only 18 deaths, 12 of which occurred in January and February,
before the destruction of the mosquitoes was commenced. Yet
the conditions were in other respects very favorable for a bad epi-
demic in 1901, for about 40,000 non-immune emigrants had
arrived,—a larger number than ever before. In view of such
results there seems to be no doubt whatever that the true source
of the yellow fever infection has been demonstrated and also
that the disease can be easily and surely controlled in all cases, if
suitable care be used to destroy this pernicious mosquito. More-
over, the same efforts will simultaneously eradicate the malarial
fevers. A. E, V.

76 Scientific Intelligence.

III. MisckELLANEOUS SCIENTIFIC INTELLIGENCE.

1. Smithsonian Institution.—The following papers have been
recently issued from the Proceedings of the U. 8S. National
Museum : |

1269. The Mammals of the Andaman and Nicobar Islands ; by
Gerrit S. Miller. Pp. 751-795.

1271. Review of the Larks of the genus Otocoris ; by Harry
C. Oberholser. Pp. 801-884.

1272. New Decapod Crustaceans from the West Coast of
North America; by Mary J. Rathbun. Pp. 885-905.

1273. A newly found Meteorite from Admire, Kansas ; by
George P. Merrill. Pp. 907-913.

1274, Distribution of three new Birds from the Southern
United States ; by Edgar A. Mearns. Pp. 915-926.

Descriptive Catalogue of Meteorite collection in U.S. National
Museum to Jan. 1, 1902; by Wirt Tassin. Rept. 1900, pp. 671-
698. 3 pls.

2. United States Weather Bureau: Wind Velocity and Fluc-
tuations of Water Level on Lake Erie; by Atrrep J. Henry.
Dept. of Agriculture, Bulletin J. 22 pp., 13 plts.—The data col-
lected by the Weather Bureau for many years has -been tabu-
lated and furnishes the basis for an interesting study of the
*‘ seiche ” of Lake Erie.

3. Scientia.—The valuable series published in Paris by G.
Carré and C. Naud under the above name has received the fol-
lowing additions :
ees par F. M. Raoult. (Phys.-Mathematique, No. 18) pp. 106,

Franges d’interférence et leurs applications Métrologiques ; par J. Macé de
Lépinay. (Phys.-Mathématique, No. 14) pp. 101, 1902.

La Géométrie non Euclidienne ; par P. Barbarin. (Phys. -Mathématique,
No. 15) pp. 79, 1902.

Le Phénoméne de Kerr et les phénoménes Electro-optiques ; par Eugene
Néculeéa. (Phys.-Mathematique, No. 16) pp. 91, 1902.

Théorie de la Lune; par H. Andoyer. (Phys. -Mathématique, No. 17)
pp. 86, 1902.

Géométrographie ou Art des Constructions géométriques ; par Emile
Lemoine. (Phys.-Mathématique, No. 18) pp. 87, 1902.

4. Ostwald’s Klassiker der Kxakten Wissenschaften. Leipzig,
1901-1902 (Wilhelm Engelmann). The following are recent
additions to this valuable series:

Nr. 124. Abhandlungen zur Thermodynamik, chemischer Vorgiinge; von
H. Helmholtz. Pp. 83.

Nr. 125. Untersuchungen iiber den Salpeter und den salpetrigen Luft-
geist, das Brennen und das Athmen; von John Mayow. Pp. 56.

Nr. 126. Experimental-Untersuchungen itiber Electricitit (IX bis XI
Reihe) ; von Michael Faraday. (London, 1835.) Pp. 106.

Nr. 127. Die Auflésung der bestimmten Gleichungen; von Jean Baptiste
Joseph Baron Fourier. (Paris, 1831.) Pp. 263.

Nr. 128. Experimental-Untersuchungen tber Electricitét, XII und XIII
Reihe); von Michael Faraday. (London, 1838.) Pp. 1833.

Am. Jour. Sci., Vol. XIV, 1902.

Plate IV.

| Fieure 6.

| Sketch of the Terraces of Westfield river, in bird’s eye perspective

?)
looking northeast. Te ae

! a 5

| N Se : BOSTON,
| ; .

T CP | ,

e
WORCESTER,

SPRINGFIELD

ee

|
|
| De i *
| we?
|

THE

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES.]

Art. XII.—The Terraces of the Westfield River, Mass.; by
7 W.M. Davis. (With Plate LV.)

1. The several theories of river terraces.—The river terraces
that are so abundantly developed in the stratified drift of our
New England valleys, receive scanty explanation in the text-
books to which reference is ordinarily made for accounts of
such forms, and are, indeed, not exhaustively treated in essays
of a more advanced character. Their most significant feature
is an arrangement in the form of a flight of steps, of unequal
tread and rise, and usually of unlike sequence on the two sides
of avalley, but necessarily exhibiting a less cross-valley breadth
between the terraces at the bottom of two opposite flights than
between those at the top. It is generally agreed that each
terrace plain is the remnant of a flood plain, that was formed
during the process of valley-carving by the river that now
flows on the flood plain or “interval” between the lowest
terraces of the series; and that the terrace fronts or scarps
have been carved by the wandering river as it swung laterally
on its successive flood plains. The slope of the terrace plains
down the valley and the pattern of the terrace scarps in curves
concave towards the river, frequently uniting in cusps, give
convincing proof of these conclusions. It follows that our
terracing rivers habitually had a greater breadth of swinging
on the flood plains at high levels, when beginning the work of
sweeping the drift from their valleys, than at the low levels on
which they are now flowing. The special point that needs to
be accounted for is, therefore, the restriction of the belt over
which the river swings to a less and less breadth in passing
from the initial to the present stage of terrace development.

There are three theories which offer an explanation for this
restriction. The first and most popular postulates a decrease

Am, Jour. Sc1.—FourtTH SERIES, Vou. XIV, No. 80.—Ave@ust, 1902.
6

78 Davis—Terraces of the Westfield River, Mass.

in river volume during and after the uplift of the region by
which the erosion of the valley was prompted. The river is
by this theory supposed to have been so large when terracing
began that it needed a broad space on which to swing; now
that the river has diminished in volume it is relatively enfeebled
and is contented to swing over a narrower belt than formerly ;
so the later formed terraces do not undercut and destroy those
of earlier date.

The second theory postulates successive uplifts of the region.
The river, revived by each uplift, wears its channel beneath its
previous flood plain and, on reaching grade, begins to swing

laterally. It is then further postulated, sometimes tacitly, that

the later uplifts have succeeded each other at shorter and shorter
intervals, allowing less and less time for lateral swinging as the
valley was worn deeper and deeper.

2. Millers theory of river terraces.—The third theory, sug-
gested in explanation of terraces in Scotland by Hugh Miller,
the younger, in 1882,* recognizes slow regional uplift as the
cause of valley erosion and then calls attention to the increase
in the number of resistant obstacles—rock ledges, boulders,
till—that the degrading river will encounter as it swings later-
ally while eroding the valley floor to lower and lower levels,
and ascribes the decrease of interscarp breadth to this simple,
effectual and observable cause. :

As Miller’s theory has not, to my knowledge, been quoted in.

this country, except in a brief note of my own,t a brief exposi-
tion of its merits may be made. I have found it very generally
applicable to the terraced valleys of New England, and nowhere
more so than in the valley of the Westfield river, between the
eastern base of the Berkshire hills and the village of Westfield,
Mass., a distance of about five miles, where I have repeatedly
examined it. This district was the scene of an intercollegiate
excursion in the autumn of 1901, in which Yale, Amherst,
Williams, Wesleyan, Institute of Technology, and Harvard,
together with six secondary schools, were represented by

teachers and students to the number of forty-six; and it is not

too much to say that at the end of the day’s walk along the
north side of the valley no doubt remained as to the competence
of Miller’s theory to explain the occurrence and the pattern of
the terraces there seen. Decrease of volume and intermittent
uplift seemed to be of altogether secondary importance, if
indeed they had produced any recognizable effect.

The Westfield terraces.—The following pages give a brief
account of the Westfield terraces, beginning with those on the

* River terracing, its methods and their results, Proc. Roy. Phys. Soe.

Edinb., 1888, 263-305.
+ Bull. Geol. Soc. Amer., xii, 1900, 483-484.

:

Davis—Terraces of the Westfield River, Mass. 79

north side of the valley near Westfield railroad station, thence
going west abont two miles to a little settlement known as
Pochassic Street; and returning by the south side of the valley
to the village again. The general pattern of the terraces is
indicated in figure 6 (Plate IV), where the attempt is made to
show them in bird’s-eye view, as if looking northeast from a
height of two or three thousand feet above a point about a mile
south of Pochassic Street. The scale in the further part of the
view is somewhat smaller than in the foreground. The vertical
scale is significantly exaggerated. The Boston and Albany rail-
road is not so straight as it is here drawn through the valley ;
about a mile and a half of its length is shown. Defending
ledges are drawn in black. Roads are dotted.

3. The open Westfield plain.— Just north of Westfield
station, a good view is had from the terrace near Prospect hill
schoolhouse, A, figure 6, over a broad plain that the river has
excavated east of the village. The plain is limited on the
north by a single high terrace, b, rising at once from the
marshy abandoned river channels at its base to the level of
the highest drift plain of the district. On the opposite side
the flood plain of the Westfield is confluent with that of (West-
field) Little river (not shown in fig. 6), limited on the south by a
single high terrace, to-day undercut by Little river at two
places. The low plain thus gains the unusual breadth of a
mile or more for a distance of about two miles eastward from
our point of view. The two rivers here show no incompetence
whatever in the process of lateral swinging. Whatever loss
of volume they may have suffered—and some loss since the
disappearance of the ice sheet is highly probable—and how-
ever recent the last uplift of the region may have been, these
two streams are here sweeping over a broader valley floor
to-day than they have opened at any earlier time; for whatever
earlier flood plains they have formed at intermediate levels,
during the process of excavating their valleys, are now com-
pletely undercut and destroyed (except for a few low terraces)
by the opening of the present broad plain. The prime reason
for so striking an exhibition of competence on the part of the
streams is believed to be the absence of rock ledges in this
section of the valley. No ledges have here been discovered
by the degrading streams ; their high terrace scarps consist of
clays, sands and gravels (the latter near the top), very easily
eroded wherever the streams flow against their base, as may
be seen where Little river is now swinging against the terrace
scarp on the southern side of the plain. A secondary reason
for the competence of the river to swing so broadly at the
present level is probably to be found in the delay of further
valley-deepening and in the resulting detention of the streams

80 Davis—Terraces of the Westfield River, Mass.

close to their present grade by the trap sheet which their
united current has come upon in the notch in the trap ridge
about two miles east of Westfield. 7

4. Terrace diagrams.—The single high terrace scarp, here
so well developed and typified in figure 1, is of more general
occurrence in the terraced valleys of New England than is
generally supposed to be the case. It may indeed be taken as
typical of many terraced valleys where no ledges have been
discovered to prevent the free swinging of the rivers. A few
lower terraces may still be unconsumed here and there, as in
figure 2; this arrangement of terraces also bemg of common
occurrence where ledges are wanting. The ordinary diagram of
a terraced valley, such as is here reproduced in figure 3, is there-
fore misleading in implying that the stepping terraces have
been carved in drift alone, without relation to the rock beneath.
These many-stepped flights of terraces need for their preserva-
tion a number of ledges, as in figure 4; but this figure is very
faulty in implying that ledges occur all along the base of the
terrace scarps. Asa matter of fact, the ledges do not occupy
more than a small percentage of the terrace lengths, and far
from all lying on the line of a single cross-section, as figure 4
implies, they are frequently distributed somewhat irregularly
np and down the valley sides. The true relation of ledges
and terraces is better shown in a block diagram, such as figure
5; and even here the area of the ledges exposed in the terrace
cusps is exaggerated over that commonly observed.

5. The Westfield spur (Prospect hill).—If the terrace by
Westtield station is now crossed to the west, it is soon found
to be a spur of a high, but not of the highest plain, extending
forward (southward on the north side of the valley) from the
broad plain in the background. It is known as Prospect hill.
The apex of the spur, E, figure 6, reaches the river; its
breadth, east and west, is about a quarter of a mile. From its
western side the valley floor again widens, making a recess or
re-entrant in the high plain on the north; but here four sub-
ordinate terraces are found beneath a stronger scarp that rises
to the highest level; they combine in a charming landscape,
when seen from the western scarp of Prospect hill. As the
river is thus found to have been competent to widen its valley
both east and west of the terrace spur, some reason for the pre-
servation of the spur should be looked for; and it is discovered
clearly enough on descending the western scarp, which is found
to be defended at various points along its bank by sandstone
ledges, O’, 0’, 0’. The sandstones are, to be sure, relatively
friable; they are weak compared to the schists and gneisses of
the Berkshire hills to the west; so weak, indeed, that while
the Berkshires retain in the equable height of their uplands a

Davis—Terraces of the Westfield River, Mass. 81

—_"
il)

1)

Fic. 3

Z\\\

Fie. 5

82 Davis—Terraces of the Westfield River, Mass.

fair indication of the altitude to which the Cretaceous peneplain
of this region has been raised, the Triassic sandstones in this
part of the Westfield valley (part of the greater Connecticut
valley lowland) have been reduced by later Tertiary erosion to
a lowland of a second (or 7+1) generation. Yet compared to

the silts and gravels of the terraces, the sandstones are very ~

strong ; whenever the river has, in the process of sweeping the
drift from its valley, swung against a sandstone ledge, pre-
viously buried, further lateral swinging has been peremptorily
stopped and the terrace behind the ledge has been preserved.

It is evidently because of the abundant defending ledges
here that the Westfield spur has not been destroyed. The
stream has made a most determined effort to destroy the spur by
scouring out a hollow, C, at its back, sweeping around in so
great a curve that its normal eastward course was locally
turned back to the southwest; but the ledge, C’, on which the
stream was caught, could not be removed, and hence the spur
still stands there. The river seems to have been withdrawn
from the deep recess, C, by taking a short-cut across a more
axial part of the valley floor of that time, probably during a
flood. When it again swung northward towards the recess, at
a somewhat lower level than before, a lower part, D’, of the
same ledge was encountered a little farther forward than the
point, OC’; and a small lunate area, C, of the earlier sweep was
. therefore preserved back of the new terrace, D. On the third
return of the river into the same locality at a still lower level,
it was caught by a small ledge, D”, several hundred feet
farther forward, and again by larger ledges, OC”, C’”’, and
hence a good stretch of the flood plain between D’ and D”
was preserved in a terrace.

The Westfield river is at present making still another effort
to remove the spur, and it is now for at least the fourth time
stopped by a member of this group of ledges, for a strong reef
of sandstone is seen in the river bank at E in the. southern
corner of the spur. The spur may in the distant future be
somewhat sharpened by losing ground on its undefended south-
east corner, should the river chance to swing that way; but
such swinging will be for the present strongly resisted by the

buttresses of two bridges and by artificial embankment where

the Boston and Albany railroad follows the river; it will cost
less to restrain the river than to remove the bridges and the
tracks.

6. Wo local evidence of decrease of volume in the Westfield
river.—Evidently the preservation of the spur of Prospect hill
is not due to any incompetence or to any want of effort on the
part of the river to remove the sands and gravels. The river
has repeatedly and energetically attacked the spur; it has in

:
:
.

Davis— Terraces of the Westfield River, Mass. 83

the most competent fashion swept away all the drift that could
be reached. It was only upon encountering the stubborn
resistance of an ambushed and impregnable ledge that the
river first withdrew, and even then it withdrew only to renew
the attack by returning bravely toward the spur on a later
swing. Unfortunately for the reputation of the river, the
ambushed ledges have been found entrenched farther and
farther forward at every successive attack that has been made
upon them ; hence the river, losing ground at every advance,
has come to be looked upon as a weakened and shrinking
stream that voluntarily abandons its earlier enterprises and
accepts a narrower limit for its conquests to-day than when in
a youth of (imagined) greater vigor and aggression ; but this is
a most unjust interpretation of its behavior. The river is
making a most determined and heroic effort to carry out its
original plan of campaign. It is eminently successful in open-
ing the valley east of Westfield, and if it is defeated at the
spur, this is only because of the reinforcement of the uncon-
solidated drift by the invincible strength of the entrenched
ledges. The river may be accused of want of foresight in not
more carefully reconnoitering the ground that it originally
proposed to excavate, but it is notorious that rivers are heedless
of buried ledges, on which they often become inextricably
superposed. The river may be thought headstrong to return
to an attack, a forlorn hope, where defeat is inevitable. Heed-
less and headstrong it may be, but it does not deserve the
reproach of being looked upon as enfeebled. Even if its
volume is now less than formerly, the river is as: competent
to-day as it ever was to open a wide flood plain in drift, and it
does so wherever free opportunity is offered of carrying out its
original enterprise.

7. Lelation of ledges, terraces and river swings.—It should
be noted that the ledges of the Westtield spur have not in any
case determined the depth to which the river has cut its valley.
The ledges here were not encountered in the river bed but in
the river bank, and hence have controlled only the breadth of
lateral swinging at the point where they were by chance dis-
covered. The depth at which the ledges were encountered was
dependent simply on the amount of valley excavation that had
been accomplished by the graded river at the time that the
discovery was made. It is also important to note that while a
ledge thus encountered in the bank of a swinging river will
defend and preserve that part of any flood plain previously
formed at a higher level, above and back of the ledge, it is
not at all necessary that every former flood plain of the river
should be thus recorded. If the successive northward swings
of an east-flowing river have by chance less and less ampli-

84 Davis—Terraces of the Westfield River, Mass.

tude, successive terraces will remain, even without defense by
ledges. There can be no question that some of our terraces are
of this accidental kind. But the undefended lower terraces will
be undercut and destroyed if the river swings more strongly
north again. If the river again swings north until it strikes a
ledge, the uppermost terrace may alone be preserved, and that
only back of the defending ledge. It thus becomes evident
that in order to discover the number of times that a river has
swung across its valley, making laterally sloping flood plains at
lower and lower altitudes at every swing, we must not trust to
the chance preservation of flood plain remnants in terraces here
and there, but must seek a flight of terraces, systematically
grouped on a long sloping ledge, which may preserve a lateral
remnant of every flood plain that has been formed, as in figure
5. It is certainly a striking fact that the number of steps in a
flight of valley terraces always reaches a maximum in just such
situations. The preservation of numerous terraces of moderate
height on long sloping ledges—however few such ledges there
may be in the valley and however few terraces occur elsewhere
on the valley sides—goes far towards excluding the theory of
successive uplifts and pauses as a cause of terracing. It goes
far also towards supporting the theory that the wandering river
has been swinging from side to side across its valley, always
degrading its channel but always acting as a graded river,
during the whole period of terracing, whatever may have been
the cause that determined the excavation of the valley drift.
If uplift were the cause, the uplift must have been slow and
relatively uniform.

In illustration of this conclusion, we may return for a moment
to the broad basin east of Westfield. No terraces at intermedi-
ate levels are found here to prove that the river did repeatedly
swing laterally while degrading its valley floor in this part of its
course; yet there can be no reasonable doubt that the river really
did swing back and forth here, for the remnants of four flood
plains at intermediate altitudes are found in the terraces on the
west side of Prospect hill, only half a mile away. Farther up
the valley one may find a flight of nine defended terraces,
described below, whose subequal heights range from ten to
fifteen feet; thus proving that even the four terraces on the
Westfield spur preserve a very incomplete record of the river’s
activity. Evidently the maximum number of steps observed
in any terrace flight gives only the minimum number of swings
that the river may have made during the whole period of
degradation.

8. Ledges outcrop on the up-valley side of terrace spurs.—
In the terrace spur here called Prospect hill, as in many others,
it is noteworthy that the up-valley side of the defending ledges

Davis—Terraces of the Westfield River, Mass. 85

has been clean swept by a swinging curve of the attacking
river. This may be explained as a result of the normal progress
of a river meander down the valley, until it is stopped by
coming on a ledge, or abandoned by withdrawal of the current
to a short-cut or cut-off course. The fact of the down-valley
progress of a meander does not seem to have received much
attention from physiographers, judging by the silence of text-
books concerning it; but it must be a familiar matter to river
engineers, so conspicuously is it exhibited on such maps as those
prepared by the Mississippi River Commission. The cause of
the down-valley progress is evidently to be found in the con-
tinued displacement of the thread of fastest current to the
down-valley side of the channel on entering the tangent of
inflexion between two meander curves.

9. Brown's spur-—On the other hand, a terrace may trail
some distance down-valley from its defending ledge, unless the
stream should by any chance swing in again and sweep it
away. This chance has not happened in Prospect hill, the
spur thus far considered; but it has in Brown’s spur, F, half
a mile farther west. This spur is well defended by a large
sandstone ledge, at whose forward-reaching base the river is
now flowing in a vain effort to widen its valley. Four terraces
in the next up-valley reéntrant curve forward to the apex of
the spur, and all agree in the most unanimous manner to sweep
tangent to the slope of the defending ledge. The ledge is well
exposed in the cuts made by the passing road and railroad.
The scarp of the next higher terrace, G, is, however, pushed
back a quarter of a mile farther north ; evidently because when
it was made the river was swinging at a slightly higher level
than the summit of the defending ledges in the apex of Brown’s
spur. It seems undeniable, when one looks at these terraces
on the ground, that the river would have pushed back the
lower members of the series about as far as the higher member,
if its lateral swinging had not been stopped by the ledge.

The peculiar feature of this spur is, however, the close trim-
ming that it has suffered on its down-valley side. The river
has at least three times swung northward so near the eastern
side of the ledge as to narrow the spur into a sharp point.
The normal down-valley progress of the meanders cannot be
appealed to as a cause of this close trimming on the down-
valley side of the ledge; some special cause must be looked for
to direct the several northward swings of the river over so
nearly the same course. It seems probable that some constraint
has been exerted on the river further up its channel, whereby
it has been repeatedly guided to the down-valley side of the
ledge. A possible explanation of this peculiar feature will be
suggested on a later page. :

86 Davis—Terraces of the Westfield River, Mass.

10. Lhe flight of terraces by Pochassic Street.—The finest
fiight of terraces hereabouts is preserved a little east of a small
settlement, known as Pochassic Street, on the southeastern
slope of Pochassie hill, a drumlin, around whose base abundant
ledges have been discovered. The settlement and drumlin are
just to the left of the limit of figure 6. The highest terrace,
H, shows waterworn cobbles and pebbles on its plain; the
bouldery slope of Pochassie hill rises behind it. Terraces are
usually counted upward from the valley floor in the reverse
order from that of their production. It will be convenient
here to follow the natural order and count downward, beginning
with the plain and scarp of the highest terrace as number one.
Thickets of small trees and bushes obscure many details here,
and some of the terrace plains are inconveniently swampy near
their back border, perhaps because of ledges farther forward
by which the ground water is held up. Hence the correlation
ot some of the terraces in this locality is doubtful, as indicated
by the blanks left in the figure. A flight of at least nine steps,
H-M, may, however, be counted, all presenting characteristic
concave fronts in what may be called the Pochassic reéntrant,
all curving forward at their down-valley ends to defending
ledges, and all of similar height, roughly from eight to fifteen
feet. A later terrace occasionally undercuts an earlier one, so
that the two scarps are locally united in a slope of more than
the average height; such being the case with the fifth—sixth
scarp, along whose base lies a narrow country road in the mid-

dle of the reéntrant; a little farther east and west the scarp is.

divided into two (or more) parts by a narrow terrace that comes
forward at an intermediate level; thus what seem to be the fifth
and sixth swings of the river may be identified.

The uppermost terrace may be followed along its scarped
front through the second growth of bushes and trees past two
defended cusps, H and H’, beyond which it turns to the north-
east and at a distance of half a mile or so seems to run tangent
to another drumlin. The second terrace is not identified on
the first ledge, H, but appears on the second, H’; it seems to
fade away on the broad plain a quarter of a mile to the north-
east. The third terrace is caught on both ledges, H and H’;
it then runs eastward half a mile and is undercut at G by the
large reéntrant between Prospect hill (the Westfield spur) and
Brown’s spur. Shortly before it is cut away, a low terrace
turns off northeastward from it and seems to continue some
distance; hence what has just been called the third scarp might
be taken to represent the fourth northward swing of the river.

It is important to note that the sharp curvature and large
are of the reéntrants in the first and third scarps in connection
with the ledges at H and H’ are much more consistent with the

Davis—Terraces of the Westfield River, Mass. 87

behavior of a river similar in volume to that of the present
Westfield than with the behavior of a much larger river.

The fourth terrace in the Pochassic reéntrant is caught on
a ledge of loose-textured sandstone, J, that stands forward from
the first small reéntrant in the higher terraces. The ledge is
not directly exposed, but abundant angular fragments of pebbly
sandstone are found in the apex of the blunt cusp on the terrace
front. This terrace is believed to be the one that runs forward
in a long sweeping curve to the apex of Brown’s spur; but it
has not been followed all through the bushes and some details
of its form may not be shown in the figure.

The divergence between the eastward course of the third and
fourth terraces, G and F, is highly significant. The third scarp
was cut a quarter of a mile back of what is now the apex of
Brown’s spur, because at the time of the third northward swing
of the river, its channel had not been worn deep enough to
eatch upon the summit of the ledges in the spur. But at the
time of the fourth northward swing the river had eroded its
plain to a lower level, so that it was held by the topmost ledge.
The fourth terrace, therefore, could not be cut so far back as
the third; it makes a long sweep forward from the Pochassic
reéntrant to the apex of Brown’s spur and leaves a rather broad
plain between its scarp and that of the third terrace. It is per-
fectly evident that this arrangement of the two terraces was not
due to any decreasing strength on the part of the river, but to
the constraint imposed upon its wandering by the ledge at F.

11. Perry's spur.—Below the fourth terrace, J—F, come
several others, which run forward to the rounded front of
Perry’s spur, K, K’, where several blunt cusps are determined
by ledges of very friable sandstone that would hardly be seen
but for cuts made by the road and the railroad. This item is
of importance, for it shows that certain ledges which are strong
enough to defend a terrace are not always bold enough to keep
themselves in sight. After having done their duty in fend-
ing off the river, they strategically weather under cover and
thus ambush themselves again beneath a thin sheet of their
own waste mixed with creeping drift from the terrace they
have protected. It is possible that another example of this
kind may be found in the well defined cusp, D’”’, of a low
terrace in the reéntrant between Prospect hill and Brown’s
spur; it seems at first to be only the free intersection of two
curves, for there is no sign of a ledge at its base and there are
no angular fragments of sandstone by which the presence of an
ambushed ledge is sometimes revealed. A little digging or a
boring with a soil auger would suffice to determine this point.

The indefiniteness of some of the terraces in the Pochassic
reéntrant may be due to the discovery there of till underlying

88 Davis— Terraces of tre Westfield River, Mass.

the stratified drift. Tull is seldom well carved by a swinging
river; its texture is significantly firmer than that of stratified
sands and clays. A small stream coming out from the north
cuts little trenches in the terrace fronts and spreads its gravel
fans on their plains, thus further obscuring their forms.

The eighth and ninth members of the Pochassic flight are
well defined, but instead of conforming to the concave pattern
of the higher members of the series, they spring forward to a
defended cusp, M, nearly opposite the middle of the reéntrant.
Here as elsewhere it is perfectly evident that the failure of the
river to cut back these lower terraces is not due in the least to
any loss of its original strength, but to the increase of resistance
offered by the ledges. The river has now swung away from
the northward meander that it followed while carving the
lowest terrace, N’, to a correspondingly strong southward
meander, T’, which is now cutting a low terrace, T, on the
south side of the valley. A broad flood plain, N, has thus
been opened. It happens curiously enough that the down-
valley progress of this southward meander has just now brought
it to such a position that it is impinging against a large exposure
of the sandstones at M. The normal down-valley advance of
the meander, T’, will soon carry it past the ledge, unless the
eaving bank at T is protected. For the present, the obstruction
caused by the ledge in the normal flow of the river seems to
have produced a slight bend in the channel a little farther up-
stream. When the meander is sufficiently advanced the river
will impinge directly on the unprotected bank between M and
L, and consume it rapidly, leaving a sharp cusp at M. It is
probably in some such way that the sharpening of Brown’s
spur, referred to above, has been accomplished. There has
already been some undercutting of the low terrace east of M,
for the main valley road was swept away by the northward
swing of the river at La few years ago. The road has been
set back so as to cross Perry’s spur, K’, north of the railroad,
thus causing the desertion of several honses on the low plain,
R, in front of the spur. The railroad itself is threatened by
the river near L; the caving bank has been worn back danger-
ously near the track, and a quantity of coarse rock blocks has
been thrown in there to stop the caving. But for this resist-
ance the river would probably continue to swing northward
until it encountered a low member of the group of ledges in
the southwest base of Perry's spur, K’.

West of Pochassic Street another large reéntrant has been
swept out on the north side of its valley; its down-valley side
is well defended by ledges in the southwest base of Pochassic
hill.

Davis—Terraces of the Westfield River, Mass. 89

The flight of nine terraces in the Pochassic reéntrant, above
described, constitutes the best series in the valley, as far as I
have seen it. Closer study will probably increase the number
of steps. The subequal height of the scarps suggests that these
terraces record nearly every northward swing of the river in
their locality. A similar number of swings has probably
occurred elsewhere ; the record of them is incomplete or want-
ing only because of the absence of defending ledges. It may
therefore be concluded in general that it is only in localities
well provided with ledges that one may expect to see preserved
in terraces the lateral remnants of all the flood plains that were
formed by the swinging river during the excavation of its
valley ; and that the maximum number of steps in a terrace
flight is only the minimum number of lateral swings made by
the river. All of these terraces testify to the graded condition
of the degrading river at the time the terrace plains were made.
The nine chance samples of river condition thus preserved may
be fairly taken to show that the degrading river was meander-
ing and swinging at grade during the whole period of terracing
in this section of the valley. There is no indication that the
individual terraces depend in any way whatever on individual
uplifts.

12. Terraces south of the river.—The terraces on the south
side of the valley may now be considered. No ledges are seen
on this side of the river within the two-mile stretch included
in figure 6. The first ledges are found half a mile further
west; they defend the eastern or down- valley side of a stron
reéntrant a little west of the town farm. Correlated with the
prevailing absence of ledges is the absence of upper terraces.
The high plain is habitually bordered by a strong scarp of
forty or more feet, beneath which there may be several low
terraces; but in one case the low flood plain of the present
river enters a strong reéntrant, P, to the base of single scarp
by which the whole descent is made from the high plain.
Nowhere in this district is there a flight of low-scarped terraces
from the top to the bottom level on the southern side of the
valley. The reason for this contrast between the two sides of
the valley may be with confidence ascribed to the general south-
ward shifting of the belt within which the river swings, because
of the southward slope of the ledges on the north side of the
valley and the undefended condition of the terrace drift on the
south side of the valley. It is a sort of monoclinal shifting of
a river course. The several low terraces on the south might
seem at first to contradict this explanation. They may, how-
ever, be reasonably explained as being not yet swept away.
The river cannot attack the whole length of the southern side
of the valley at once; it will swing against the southern

90 Davis—Terraces of the Westfield River, Mass.

terraces only here and there, now and then; and hence the
destruction of the lower southern terraces by the southward
shifting of the belt of river-swinging can only be accomplished
progressively. Thus viewed, it is no wonder that some of the
lower southern terraces still remain. This corollary to the

explanation above suggested receives much support when

examination is made of the relation between the defended
cusps of the terraces on the north and the concave reéntrants
of the terraces on the south, to which we now proceed.

13. Correlation of terraces on the two sides of the valley.—
It will be readily understood that, wherever a terrace scarp
curves forward from a concave reéntrant to a defended cusp,
the river must have once flowed along the base of the scarp
and must have continued the line of the terrace curve past the
cusp toward the opposite side of the valley, there to recurve
toward the general valley axis. The farther forward the
defended cusp reaches toward the axis of the valley, the more
likely it is to direct the departing river strongly against the
terraces on the opposite side of the valley. Thus the curved
scarps, K’, K’”’, K’”’, of the northern terraces, suggest that the
river has formerly flowed on past the defending ledges of
Perry’s spur at these several levels and has thus entered the

reversed curves of the southern scarps, 8’, 8S’, 8S’. The levels

of the terraces concerned seem to correspond by pairs, but they
have not yet been accurately measured. Of the three northern
scarps, the one leading to K’ has the strongest curvature, for
the middle of its reéntrant is cut farther back than the middle
of the others. When the river flowed at the base of this searp,
its current must have departed from the defending ledge almost
transversely to the general eastward course of the valley. Oon-
sequently the corresponding southern scarp, 8’, makes a strong
reéntrant on the south’ side of the valley. The second scarp,
K’”’, of the northern group was more gently curved; the third
scarp, K’”, still more gently curved; and the same "relation is
seen in S$” and §’”.. It is possible that even the higher southern
reéntrant, S, is indirectly related to the ledges of the northern
valley-side about K; the general curvature of the river being
determined by the ledges and a southward swing of the curving
river there causing the excavation of the reéntrant, 8. It is
proposed to make a careful measurement of all these terraces in
order to test their correspondences.

Another example of this cross-valley relation is found far-
ther west. The curved scarp of the lowest northern terrace,
N’, that sweeps out to the defending ledges at M, is the north-
ern member of a double curve whose southern member made
the strong concave sweep, P, already mentioned, under the
high southern terrace plam. The ledges at M stand unusually

Se

ES

Davis—Terraces of the Westfield River, Mass. 91

far forward toward the valley axis; this southern sweep is
therefore the only one that has consumed all the southern ter-
races at intermediate and low levels. The abandoned channel
of the sweep is still swampy and the enclosing scarp is still
uneven with landslides, so lately has the river been withdrawn.
It is natural enough that the river should not have swung so
far south at higher levels, for the guiding ledge, M, is rather
low; it was not encountered until the river had cut down its
flood plain nearly to the present level.

Inasmuch as the southern sweeps, S’ and P, seem to have
been guided by the northern ledges, K’ and M, it is natural
to find the unconsumed remnant of the terrace between the
two sweeps in the form of a cusp, Q. This cusp is about
forty-five feet high. It is entirely undefended and might,
therefore, at first sight be classed with free cusps, and re-
garded as the consequence of a chance intersection of succes-
sive meander sweeps on the valley side. But the dependence
of the sweeps, P and S’, upon the defended spurs on the
northern side of the valley shows that the strong southern
cusp, Q, is not altogether accidental. It is in reality a natural,
although an indirect and temporary, product of the northern
ledges at M and K’. Unlike the defended northern spurs,
which are relatively permanent features, the indirectly de-
fended cusp, Q, will not endure. Its apex is already trun-
cated by a chance swing of the river against it; it will be
more and more consumed as such swings are continued and
repeated. It is safe only so long as the river flows on curves
determined by the ledges at M. and Kk’.

The truncated free cusp, W, on the south side of the valley,
is probably related to Brown’s spur, F, in much the same
way that the truncated spur, Q, is related to Perry’s spur, K’.

Another southern cusp, T, is the remnant of a 15-foot ter-
race projecting far into the valley between the southern re-
entrants, T’ and P. A strong scarp, T’’, with blunt salients
rises to the high plain back of T. The ‘sharp apex of the
cusp points directly to the ledges at M, yet it is entirely unde-
fended south of the river. It is probable that there was
something of up-valley carving on the eastern side of the
cusp;a relatively unusual process, for, as has been stated
above, river meanders normally progress down-valley. But
in this case, the down-valley progress of a northern meander
was stopped by the ledge at M. ” The ledge probably acted as
a sort of fulcrum as soon as the river impinged upon it; the
deeper the northern reéntrant, N’, was cut, the more nearly
the river must have turned square across the valley at M, and
the more it must have been turned against the down-valley side
of the spur, T. Something of the same kind probably occurred
when the reéntrants, 8S’, 8S’, S’’, were scoured out.

92 Davis—Terraces of the Westfield River, Mass.

14. Development of future meanders.—The river is eyvi-
dently tending to become more curved at its northward bend,
L, and its southward bend, Q. It was formerly less curved ;
and it was probably then that it ran into the strongly concave
reéntrant, V’. No immediately local cause is found for the
preservation of the low spur, V, for no defending ledge is to
be seen at its apex; but one may be there, buried in the flood
plain deposits ; the sandstone outcrops in the little island near

by and in the cusp F nearly opposite. It is possible that these

ledges held the river in the channel between them (on the
north side of the present island) while the river scoured out
the sharply concave reéntrant, V’; and that, as the river in-
creased its curvature at L and Q, it withdrew from the curve
of V’. Such aseries of changes would not be inconsistent
with what is known of the development of river meanders.
Their greatest dimensions are attained only where the curves
are well organized, and such organization requires time for
its accomplishment. A limit is set to the size of the curves,
less by an equilibrium between current and bank than by the
abandonment of the curves when short cuts and cut-offs are
made. The river course is thereby made nearly straight
again, after which a new series of curves is gradually estab-
lished. The Westfield river hereabouts is comparatively
straight to-day. Its course for several miles eastward from T
must be much less curved now than when the concave terrace
fronts were carved at various earlier dates. But a strong
curve is seen to-day at T’; the curves at L and Q are increas-
ing and their maximum curvature is not yet reached, and
hence it may be expected that another period of organized
meandering is approaching. The restraint of the ledges at M
will soon be avoided by the down-valley progress of the mean-
der T. The northward curving at L will be resisted by the
railroad. The southward curving at Q may be delayed by the
abundant fall of gravels from the truncated end of the spur;
and indeed there are already some indications that the river
may bend southward into, the low flood plain west of the spur.
A stronger northward turn toward F would thus be induced
and a stronger southward turn might then follow farther east-
ward. The latter item in this series of changes would be
made more probable if the swinging river would again pass
the ledges at F on a southeastward course, as it did when
carving the up-valley side of Brown’s spur.

All these details are relatively trifling, yet they have a value
in that they unite in showing the competence of ordinary pro-
cesses, appropriate to a meandering, swinging and slowly de-
grading river, to produce even the most minute forms of our
terraces. There is no demand for an ancient river of great

a

;

Davis—Terraces of the Westfield River, Mass. 93

volume or for repeated movements of uplift in the terrace
problem of the Westfield river.

15. Southern terraces at Westfield. — Nearer Westfield
the complication of the southern terraces increases somewhat,
and there is one member of the series at a higher level than
elsewhere on the south side of the valley in this district. It
may be therefore inferred that the river belt has hereabouts
been shifting northward, and this would be confirmed by the
high terrace of the opposite northern reéntrant. Yet no
ledges are found on the south side of the valley as a cause of
this shifting. The only explanation that I have thought of
for it is that Little river once entered the Westfield from the
south near the present site of Westfield village, and thus
slowly, pushed the Westfield river northward from the course
it had previously followed. Interference of one river with
another in this way has been suspected in the eastern basin,
beyond Westfield, and in several other localities in the Connec-
ticut valley.

16. Conclusions.—The most manifest conclusion to be
drawn from this study is the one already announced ; namely, that
Miller’s theory of defending ledges gives a better explanation
than any other for the terraces of our New England valleys.
It is not desired to imply by this that all our terraces are de-
fended, but that most of them are; and especially that all the
many-stepped flights of terraces owe their preservation to
defending ledges. Decrease of river volume and intermittent
uplifts do not seem to have had any significant part to play in
the restriction of the swinging rivers to narrower and _nar-
rower belts. Another conclusion is that the normal action of
a meandering and swinging river suffices to account for prac-
tically all the details of terrace form: and hence that terraces,
like other land forms, are susceptible of explanation, even
down to their most minute elements. Following this there is a
third conclusion, of interest to those who concern themselves
especially with the study of land forms ; namely, that in this
division of the subject as well as elsewhere, observation is
greatly aided by the discovery of a successful theory ; for
the essential facts are then quickly acquired by well dir ected
search. It is also apparent that here as elsewhere description
is greatly facilitated by explanation, for explanation enables the
student to bring the local example into proper relation to the
generalized type. There may seldom be necessity of giving
minute description of forms so small aud so ephemeral as
drift terraces; but when that necessity arises it will be met
better by characterizing terraces In terms explanatory of their
origin than by an attempt at absolute or empirical description ;

Am. Jour. Scl.—Fourts SERIES, Vou. XIV, No. 80 —Aveust, 1902.
7

94 Davis— Terraces of the Westfield River, Mass.

the defended and the free cusps, the high scarps without out-
cropping ledges, the flights of advancing terrace steps in asso-
ciation with groups of outcropping ledges, the correlations of
the terraces on the opposite sides of a valley, all these items
are best told by explaining them.

Finally it may be noted that even the geologist who is con- —
cerned only with the underlying rocks may well afford to give
some heed to the pattern of drift terraces; for he will be most
quickly guided to his desired outcrops if he examines the
points and the up-valley sides of the terrace cusps.

Cambridge, Mass., June, 1902.

\
\

r

G. R. Wieland—On Marine Turtles. 95

Art. XIV.—Wotes on the Cretaceous Turtles, Toxochelys and
Archelon, with a Classification of the Marine Testudinata ;
by G. R. Wreianp.

I. The Front Flipper of Toxochelys latiremis.

A ¥rNne specimen of Zoxochelys latiremis Cope* from the
Niobrara Cretaceous of Kansas (Accession No. 2491 of the
Yale coll.) includes a well preserved skull and lower jaw, six
cervical, and several dorsal vertebrae, anterior portions of the
carapace and plastron, and two nearly complete front flippers.
Although some of these bones are somewhat crushed, their
preservation is in all other respects exceedingly good. The

Figure 1. Toxochelys latiremis Cope. Dorsal view of the left front
flipper, x14.

“OD ag

aes:

a Ta Sy
i, Ce

a, b, c, e, g, respectively the head, radial and ulnar processes, ectepicon-
dylar groove, and entocondyle of the humerus; R, radius; U, ulna; In,
intermedium ; Ul, ulnare; P, pisiform ; C, centrale ; 1-0, first-fifth carpalia ;
I-/, first-fifth metacarpalia.

left of the front flippers is evidently the more complete, and
apparently only lacks the radiale, first carpale, and some distal
portions of the terminal phalanges of the third, fourth, and
fifth digits. The various elements, with the exception of
carpalia 4 and 5, were separated in collecting, or else had
originally been more or less scattered. N evertheless, the
arrangement of the parts, as shown in figure 1, is given with
considerable confidence as to the correctness of ‘the main facts.

Hitherto our knowledge of the front flipper of Yoxochelys
has been limited to the proximal half of the humerus and two
fingers as described by Case,t although the phalanges are men-

* K. D. Cope.—Vertebrata of the Cretaceous Formations of the West,
vol. ii, Rep. U. S. Geol. Sur. of the Territories. Washington, 1875.

+ E. C. Case.—University Geol. Sur. of Kansas, Paleontology, part iv,
Toxochelys. Topeka, 1898.

96 G. R. Wieland—On Marine Turtles.

tioned as “flattened” by Cope. Case figures as the first finger
what is here denoted as the second, but, as is further mentioned
below, there is a striking peculiarity in the strong resemblance
of the second metacarpal to the first, which would lead one to
err in this respect in the absence of the other digits.

Description. The humerus, rather broader in the fossil
than in life as the result of crushing, is, as I have already
pointed out on the basis of the proximal portion cnly, of
the thalassocd type characteristic of various of the older —
Chelonide.* That is, while the general outline closely ap-
proaches that seen in typical oceanic turtles, certain persisting
older characters are present, such in particular as the high
position of the rather prominent radial crest. The most closely
related humeral form is to he seen in Lytolomat and in
Neptunochelys tuberosa (Leidy, Cope) Wieland.*

The radius is somewhat longer than the ulna as in other
Chelonidan forms, and also be it noted as in Acichelys (Eury-
sternum) Wagleri. (See figure in Zittel’s Handbuch.)

The large tntermedium is also of the same general outline
as that of Aczchelys.t The arrangement of the carpals is,
however, mainly as in Chelone, etc., the first carpale not being
excluded from contact with the centrale, while the first meta-
carpal is set high, and the small and thin pisiform is attached
to the fifth carpale.

General Observations. The present is probably the first
fairly complete restoration of the fore flipper of an ancient
marine Chelonian which has been given, and as one might
well expect in the case of a Cretaceous turtle presenting vari-
ous primitive characters, no little light is shed upon the man-
ner in which the evolution of the Testudinate flipper has
proceeded. The most striking peculiarity is the flatness of
metacarpal II. This bone presents characters distinctly inter-
mediate to those of metacarpals land Ill. As flattening is
usually only present in the bones of the first finger, there is in
Toxochelys a very marked use made of this means of adapta-
tion for marine life. Zoxochelys has two well developed
claws, not noticeably differing in this respect from Lretmo-
chelys, but the short and robust phalanges of the first and
second fingers, all of which are exceedingly well preserved,
are decidedly more suggestive of those seen in land forms than
in the case of any other known distinctly marine turtle.

* G. R. Wieland.—Some Observations on Certain Well-marked Stages in
the Evolution of the Testudinate Humerus. This Journal, fourth series,
vol, ix, June 1900.

+ L. Dollo.—On the Humerus of Euclustes. Geol. Mag., vol. v, pp. 261-
267. London, 1888.

t Zittel—Handbuch der Paleontologie.

G. R. Wieland—On Marine Turtles. 97

With respect to the distal phalanges of the fourth and fifth
fingers, it should be stated that the present restoration is of
necessity more or less generalized. These phalanges being
more slender than those of the first and second fingers, have
suffered crushing, and having been isolated also, one cannot be
wholly sure of their exact order. Never theless, it is fully
evident that fingers 3-5 are long, and have the structure seen
in strictly marine turtles. In the appended table of measure-
ments only those elements are given which may be deterinined
with reasonable certainty, though the estimated lengths ot
fingers 3-5 are doubtless very near the truth.

At least we are enabled to deduce the principal variants in
the evolution of the Testudinate flipper from some generalized
type of foot like that of Chelydra. The humeral changes in
general form I have already outlined at some length. In the
case of the remaining elements it would of course be desirable
to consider an approximately phyletic series, but in the absence
of this, the general trend of change may best be made clear by
considering the percentage of leneth of the elements of the
flipper in Chelydra, T: owochelys, Lretmochelys and Dermo-
chelys. These are, together with Acichelys of the Jurassic,
and the known por tions of Archelon, as follows:

/ 5 e a : : : : A

Ss 4 Fo 3 mr

= 3 A 2 < Zo zs re ‘a

ae a =) = R an) > 1D Ay

Dermochelys, 100 43 ao A387) “160%, 209 7: 86 23

Eretmochelys, 100 53 44 49 Se) A258)... 105 44. 12

Archelon, 100 54 51 — — — 21

Toxochelys, 100 58 50 51 73 100 = 104 de AO ery eh
Chelydra, 100 52 53 50 72 73 a) 50 (small)

Acichelys
(=Eurysternum), 100 S57 51

i
oO
or
mse

63 66 d1 ti

Inspection of the above table shows:

1. Strongly marked radial and ulnar decrease in length.

2. Greater or less elongation of the radius as compared with
the ulna.

3. Nearly static length of the first finger in the Chelonidan
forms, with sharp increase in Dermochelys. |

4. Persistent increase in the length of fingers 2-4.

5, More or less variable tendency to increase in length of the
fifth finger.

6. Great pisiform increase, which began relatively early.

98 G. R. Wieland—On Marine Turtles.

Measurements of Toxochelys latiremis.

(All are from the same individual.)

The Skull. M.
Extreme length from beak to extremity of the supra-
occipital process, which is complete_-__-__-.----.----- 152
Length from beak to occipital condyle ___.._..-.-......) “14
Exiteme width -~.. 2... 8. See ee ee ee
Width between articular surface of quadrates ___-_--...- 078
Length from beak to anterior border of internal nares __-- 025
Extreme length of ramus of lower jaw.---.---.-------.. ‘ll
The Cervicals.
Leneth of ist cervical centrum (2. +=. ee 6 014
Length of 2d = wens etc .oh 2G Ge ee
Length of 3d Re PES \ hen So hoo yd Soe Se ae —
Length of 4th ‘“ Ase) Chasis oe oe Se Ce 033
Length of 5th “ Fi Nit De ee See eee —
Length of 6th “ ST ety oe et eae ae St oe
Length of 7th “ ett tek ee ee 031
Length of 8th “ Bore) Be Disk tees ae "03
Estimated total length of the eight cervicals _.-.-------- 226
Length of Ist dorsal centrum £2 /°_.. 22-22 22352 ee ee
The Humerus. M.
Greatest lenoth .__-.. = @. £222 85 135
Depression of radial crest beneath proximal extremity.-.- °02
Distance of ectepicondylar groove from distal anterior
perder 202 23. .005
Greatest length of the other elements of the front flipper :
Padi sk Pe ee ee ‘075 | Metacarpale, TIT __-___ S3ga8
pte eee 066 — $ TV ee 055
iupermediom 2. 22 018 | <7 YA ee 037
Diaare ©. 22-2. a Ae ee 017 }Phalanx, I-12." _) >=
eemuemie 222 So S50 Mas O15 60 Teo. (claw) 2-00 ee
Warpale) eo be sii). P83 ae sep DSA aE Ris | Everett "02
aire hae. Tesi 015, Ke top Eas Sek ee 019
77 (Sac. Dagar baths Repti 013 | “<* Ii-3 (claw). 2-025
"OS ee i Oe Ra eae 013) Phalanx TEi—t 2). ee ‘034
a Wig ett Fd 2 ‘O11. cI }i-9 '._-..) eee
Pipi fee oS OES EY 1. ee * +036
Metacarpale, 1 ....---. "023 | MN ole eee 031
ae 1, Sipe dai Saag ‘035
Total lene ch first, finger __._ 2. 2. eee ee "65
“- ” second finger Bebe ain = = pep ee Ec, ee
Estimated lensth, third finger... . coe 2s “13
3: fourth finger a no 2c ny eo are a
oe SST eiet Rewer. 62. | 2! ole 09

~~:

G. R. Wieland—On Marine Turtles. | 99

Il. The Front Flipper of Archelon ischyros.

In my first description of the gigantic turtle skeleton from
the Fort Pierre Cretaceous of South Dakota which constitutes
the type of Archelon ischyros,* I figured, in addition to other
skeletal parts, the humerus, radius, ulna, femur, tibia, and
fibula, and mentioned the fact that a number of carpals and
tarsals with several phalanges were also present. This was
practically the first contribution presenting the main features
in the limb organization of the [Protostegine]. Hitherto our
knowledge of the flippers of these great turtles had been lim-
ited solely to that given in Cope’s original description of the
first member of the group discovered, Protostega gigas.+ In
this form the humerus was described and figured together with
the radius. and ulna, although the latter were then supposed to
be “ metapodials”’ rather than bones of the forearm.

As there was at the time of my first publication an absolute
dearth of information concerning the carpal and tarsal structure
in marine turtles from the American Cretaceous, not one hav-
ing been described, the fear that | might make some serious
error prevented my publishing the restoration of the carpus
which I then made, the various elements having been found
only partly in position. |

Now, however, Professor Willistont has just described the
hind flipper of Protostega, and this enables me by exclusion
to determine with a reasonable assurance of correctness that the
elements [ originally assigned to the carpus in my study of the
closely related Archelon truly belong there. Iam hence able
to add some further facts concerning the skeletal organization
and systematic position of these highly interesting Testu-
dinates.

The partial restoration of the left fore flipper shown in
figure 2 is based on the radius and ulna, with what are consid-
ered to be all of the carpals but two, together with the first
and fifth metacarpals. Several phalanges are present, but as
finger proportions may vary markedly it is not deemed advis-
able to attempt a complete restoration. The value to be
attached to this preliminary restoration is provisional, as
follows :

(a) The radius and ulna are simply drawn in a generalized
position, but their orientation is based on that found in a sec-
ond specimen where these bones were in an approximately
normal position with respect to the humerus, only.

*G. R. Wieland.—A New Gigantic Cryptodire Testudinate from the Fort
Pierre Cretaceous of South Dakota. This Journal, fourth series, vol. ii,
December 1896.

+ Loe. cit.

¢S. W. Williston.—On the Hind Limb of Protostega. This Journal, fourth
series, vol. xiii, April, 1902.

100 G. R. Wieland—On Marine Turtles.

(b) Carpalia 3-5 were found in succession, and are doubt-
less correctly placed.

(c) There is little doubt but that the intermedium and
ulnare are correctly determined, but their precise orientation
is not so certain. Thus it may be that the intermedium should
be rotated in a vertical are of 90°. This, however, would not
greatly alter its general contact outline, as it is a very robust
and much rounded instead of flattened bone.

(d) Since metacarpal I and the pisiform are beautifully
preserved, both as to form and surface markings, the possible
remaining margin of error as to the identity of the several
parts is shght.

Figure 2. Archelon ischyros Wieland. Partial restoration of left front
flipper, xt

R, radius; U, ulna; P, pisiform ; I, carpale 1; V, carpale 5.

Description. Since the peculiarities of the Awmerus have
been quite fully dealt with in my article on the Evolution of
the Testudinate Humerus, this bone need not be further men-
tioned now, except to recall the fact that it resembles the
humerus of Dermochelys more closely than any other, although
differing in some very essential features.

The radius is only slightly longer than the ulna. Proxi--
mally it is triangular, and distally, rather elliptical in trans-
verse section. The most marked characteristic whatsoever is
the strong proximal bow, which recalls the lesser proximal
bow seen in the radius of Dermochelys. Otherwise this form
has a rounder, heavier head. The proximal articular face is
only slightly concave, its general outline being that of an isos-
celes triangle, with the base in contact with the inner face of
the ulna.

G. R. Wieland—On Marine Turtles. 101

The wlna@ is short and massive. The proximal articular sur-
face is slightly crescentic in general ontline, and somewhat
concave except for an oblique, low, saddle-shaped ridge which
divides this face into subequal areas, the larger facing towards
the radius. The distal articular surface is moderately convex
antero-posteriorly, and rather flat in the dorso-ventral direc-
tion. The bone has a distinct broad and shallow grooving on.
the proximal ventral, and on the distal anterior side, marking
the proximal and distal contact with the radius and producing
the effect of a marked twist corresponding to the high angle
between the general trend of the proximal and distal articular
faces. Asin the case of the humerus and radius, there is a
certain correspondence with Dermochelys, but the ulna of the
latter is more rounded. |

The intermedium is much rounded and very robust. The
ulnare is suboval in external outline, with the proximal edge
much the thicker. The pzszform is very large, of subcrescen-
tic outline, thick, and quite flat, but with a raised border on
both faces.

The carpalia and metacarpals present have more the appear-
ance seen in the Chelonine than in Dermochelys, in which
respect there is in fact a wide difference, the pronouncedly
marine appearance of the rounded subcylindrical and flat-ended
metacarpal and phalangal bones of the latter being quite
absent.

General Resemblances. Inso far as now known, the manus
and pes of Protostega and Archelon resemble those of Dermo-
chelys rather more than any other form. Briefly pointed out,
the more marked similarities of the manus are, the approxi-
mately equal length of the radius and ulna, the heavy proximal
bow of the radius, the earpal organization with the centrale
excluded from contact with carpale 1, and an enormous pisi-
form set high up near the ulna, and mainly on the ulnare.
The point of most importance and necessarily of the greatest
difficulty to settle with complete satisfaction is as to the
assumed contact of the centrale with carpale I.

Professor Williston has, I think, omitted one of the tarsalia
from his restoration of the hind limb of Protostega, so that
there is likewise an even closer general correspondence between
the hind limbs of these several forms than he suggests. The
diminution of the fifth finger of the hind flipper, as he shows,
to a single metatarsal is, I suppose, not to be regarded of as
much importance as any reorganization of the carpalia or
tarsalia (loc. cit.).

102 EDRO Fg Wasa on Marine Turtles.

Measurements of Archelon ischyros.
(From the type specimen.)
The Radius.

Meters.
Ui) Mite perenne iy AeA. peed De aMpm yar earne 35
Proximal diameters: 2... J. oo ke ele. . Ree
Least. diameters.of shaft-.o2-...2 002-1. .:.._.. "05 / meee
DDistaleeiametens:= 0 see e ee A kg ee "049 and ‘114
The Ulna.

Hength’_ os 242. SL ee Ss So ee ee "33
Greatest and least - pr ‘oximal diameters ......-.--- 093 and *14 |

< “diameters of shafti2: oo 22) O64 |

« Geo a eittr* Guistall: diamiebers<2 os. Thee eres ‘088 and 138 ;

Measurements of carpals and metacarpals, viewed from dorsal
surfaces :
Greatest length. Greatest width.

Entetmedinancte. ete) te °8 shet
Wibaater S50 dew Gast. Pe he “Th 08
Psi. 221 be Boe aie ae eee 14 12
Centrale. fo. is. ee Ye "048 ‘06
Canpalecg hast. 1.) Seer es Cooke 04 07
ce Tk SP Tas ict es kT Bete a ee

se sy I Aegean amet SAE a Ey tia pe "06 "056

ue EV: 2 lier OE yi ein aa ened ‘068 ‘068

ne i GCA, MAES MR acl eRe ‘06 072
Metacarpale 1S OS Pees ‘09 "10
imal ee hc Ra A Sle "156 °056

The extreme length of the accompanying humerus is °65 meters.

Ill. Zhe Cervicals of Toxochelys and Archelon.
(See Measurements, p. 98.)

The cervical formula has not hitherto been given for any of
the Cretaceous marine turtles, so far as the writer is aware,
our knowledge having been restricted to scattering or isolated
vertebre. I may hence give the formula for Toxochelys, and
will show that that of Avehelon may also be determined as
quite similar, smce we know the vertebrae most susceptible of
change, the last three.

In the Yale specimen of Zoxochelys (see Measurements),
the first, second, fourth, sixth, seventh, and eighth cervical
vertebree are present. The nature of the other two, the third
and the fifth, of course follows, as included arbitrarily in the
subjoined table, in which are given the formule for the
present forms and several others presenting interesting or im-

pue ‘A[101109s0d xeauoo A[qnop puv ‘ATLoLIeyUR eAvoUOD A[Quop=uRe,tAoIq-oO]~MoIg ‘ eAVOTIOD-a[(Nep=URITOD | xeAMoo-e]q(nop=ure4tsO

uveytho
-oy1AOTG

GE TOOT
-04th—)

bi)

9)

bp)

9)
uUBIpODOyILD,

xel duo

‘wafiuds
LAUT,

ae

)

uvayts00[0D

?
orddiydy
uveytsyy
xaduiog

‘snUDYLMaWwng
srMawoopog

IITA

)

Le)

””
?

“Pe

uBeyAD0][AD
uve4y1Ad
xe[dutog

‘(purry) sasuar
-Unaspbvpvyyy
shvoyoouwmhay

ITA

weet A01q,
-OgaAOLg

WeTTPOOT
-o[oorg

uveyt£01q
“OT@)
weayTADOTCOL

we94akO
?
UWBITA904AALD

xo[dutog

‘snucaydlijod
opnysa],

IA

uva4ztho
-O[OoIg

weoqaAOLG,
-O[ MOTT

uvoypthorq
Of)
wWRaytADOTHOHD

uraytd—a
)
URT[CI0}.LAL)

xotdurog

*punuaduas
paphjeyO

A

uReyLAo

-O[OOTg |

twRaytAdIq

AYR] |

uvdye[dopap

uvoytAd0[a~—)

uvaytdD

i]

WRI[HOI0yLAD,

xeTduton

“DPDIVQUL
shpayooujany

AI

|

|

ueeytho
-O[Md0Tg
|

uveytA01q,
-O[))|

1)

uvoayprho0]e09
uvoythp
2)
UBIToIOyI AEA

xo[dutog

*“ad0D1LN09
shjayoouldad

TE

uvoythoojaoQ | ureyt£o0p~ag
uvaythy uvoqiho
9?) )
uvIpoooytdg | uvije00,1h9
xo[dutog xo{durop
*s7Ua dn) ‘souliyose
ShpAYIOXOT, U0)AYIN
II E

?)

?

Pr)

.

“AT[VLOZVT Yoou oy} Surzoerjor
FO FIQVY oY} JO 4[NSet oyy sv ssey}qnop ‘sparq Jo osoy} sv podvys-e[pprs ATjouLsSIp sv oyinb ore spumauoopog JO s[eolAteo Y4XIS—paAlyy ey, *
‘shjoyouyanigy UL se wveytdorqAqeid soyovordde shjyayoowuag Jo } ‘ON
“eIZUed pezyeutuLte} A[ZVy Lo ‘xesuod-e[qnop ‘eAvouod-o[qnop oyut dofeasp AT[wngov you op ynq, *peorq aie s/izayooxoy, JO S[VOLATEO 4ST OUT,
; ‘eg, “d ‘6681 “Qsnony 10j [eUIMOL sTyy veg

9

9

?

V

"[ROTALOD

104 G. R. Wieland—On Marine Turtles.

portant comparisons. All taken together display the great
variations seen in particular in the last five Testudinate cervi-
cals. I may explain that I have introduced a modified nomen-
clature I proposed in this Journal for August, 1899, p. 163,
since the ordinary and incomplete terminology does not ade-
quately express the complicated forms seen in Testudinate
cervicals.

The cervicals of Zoxochelys are distinctly intermediate in
character between those of Chelydra and the Chelonide, being
most like the former. The ends of the sixth, seventh and
eighth broaden, but there are no distinctly biconcave, bicon-
vex, or flat terminations, these vertebra. still being proccelons
(ceelocyrtean). Their centra have strongly marked elongate
and thin or blade-like keels, quite similar to those of the sixth
and seventh, but not the eighth centrum, of Chelydra. The
total length falls far short of that seen in the Chelydride.

In Archelon, the sixth, seventh and eighth cervicals were
found in place in the type specimen, as well as several others.
All are characteristically proccelous (=ccelocyrtean), and none
have their anterior ends markedly broadened as in Zoxochelys.
They are also relatively much shorter than in Zoxochelys, and
very robust (loc. cit.). I have estimated the cervicals of the
type specimen of Arvchelon as having a length of -72™. The
cranial length must be about the same. But in Toxochelys
the total length of the cervicals is about one and a half times,

and in Chelydra twice that of the cranium. We cannot doubt:

that the fourth cervical was biconvex (Cyrtean), since it is so
in all known marine Testudinates. The vertebrae of Archelon
are on the whole rather more primitive than in any other
marine turtle, and it is certainly very interesting that there
should be a closer agreement with Zoxochelys than any other
form.

IV. Béaring of the Foregoing Data, and Classification.

A phylogenetic classification of the marine Testudinates will
still be held more or less difficult to deduce, according to the
view that is taken of the much debated descent of Dermochelys.
In weighing the evidence at present available, however, it
needs to be borne in mind that the wide distribution of the
turtles in latitude and time necessitates the consideration of
slighter differences than in the case of more variant forms.
While this must finally be a great advantage, it is a fact that
at present brings home to us with force our imperfect, but
happily rapidly increasing, knowledge of the- fossil record.
Again, there is a constant danger that in such a case one may
regard evolution as having taken a far simpler course than has

a
Z
a

:
‘
"

G. R. Wieland—On Marine Turtles. 105

really been the fact. It would appear that several hypotheses
yet require consideration, as follows:

(1) All the known marine Testudinates may be the de-
scendants of a single littoral species.

(2) Dermochelys on the one hand, and all the other marine
forms on the other hand, may have descended from two dif-
ferent littoral species of the same genus, or from different
genera.

(3) Dermochelys, Toxochelys, Protostega, and the living
Chelonide may represent the descendants of four genera of
the same, or closely allied families.

(4) Dermochelys is of ancient descent, and stands phyleti-
eally and morphologically opposed to all other Testudinates.

Doubtless the final truth will be found to lie somewhere be-
tween the first and last extremes. Certainly it is. difficult to
overthrow the conclusion as to the general fact of descent as
thus expressed and as defended with such signal ability by
Baur :*

“Dartiber aber ist kein Zweifel dass Dermochelys und
Psephophorous keine urspiinglichen Formen sind, sondern
dass sie von wahren ‘Thecophoren’® und zwar yon den
‘Pinnaten’ abstammen, um mich hier dieses Ausdrucks zu
bedienen.”

Dollo,t originally a strenuous opponent of this view, has
recently adopted it. Taking up the question in further detail,
he holds Dermochelys to be descended from a pelagic Theco-
phore with an extremely reduced carapace and plastron, but
the descendant of a littoral Thecophore with a fully developed
carapace, and a plastron without fontanelles. And this eminent
scientist has proposed the ingenious hypothesis that such a
Thecophore again acquired littoral habits, resulting in the
formation of a heavy mosaic carapace, which, with a second
resumption of pelagic habits, again began to disappear, and is
still in process of reduction. The persistence of the nuchal
is held to be due to its value as an attachment for the nuchal
ligaments. However involved such an_ evolutionary process
may appear, it is skilfully presented, and has much in its
favor.

On the other hand, Hay{ has presented at considerable
length facts favoring a very early origin of the Dermochelan
line.

* G. Baur.—Biologisches Centralblatt, Band ix, 1889, p. 191 (Erlangen).

+ L. Dollo.—Sur Vorigine de la Tortue Luth (Dermochelys coriacece). Ex-
aay Bull. Soc. roy. des Sciences Med. et Nat. de Bruxelles, Seance 4 fevrier,

tO. P. Hay.—On Protostega, the Systematic Position of Dermochelys, and

the Morphogeny of the Chelonian Carapace and Plastron. American Nat-
uralist (Boston), Dec. 1898.

106 G. R. Wieland—On Marine Turtles.

_ Presumably the evidence in favor of Baur’s view is increas-
ing. The writer so regards it. The fact that the cervicals of
Toxochelys and Archelon agree in general, and at the same
time differ most widely from the cervicals of Dermochelys and
of the Chelonine, is, however, rather unexpected. Did the
vertebrae of these Cretaceous forms tend to simplify, or has
there been a more or less remote homoplastic parallelism in
the course of complication in the case of the sixth, seventh
and eighth vertebree of the modern sea turtles, and such widely
different forms, for instance, as the Testudinide ?

The question at once arises, what is to be regarded as a very
primitive Testudinate cervical, and what was the form in the
species, genus, or group which made its way into the sea and
gave rise to the marine group? We may most reasonably
suggest as avery primitive cervical type, that of a turtle like
the Plenrodiran Arymnochelys, in which there is a well-nigh
complete agreement with the modern Crocodilia, the second,
and not the fourth, centrum being biconvex. And we assume
that some descendant of such a primitive type, with double
convexity moved back to the fourth cervical centrum, the
fifth-eighth centra remaining simply proccelous, stood in some
common ancestral relationship to the sea turtles and most other
existing Cryptodirans. It would at present, therefore, seem
that even since the sea turtles split off from their littoral an-
cestry there has been a certain parallelism in the secondary
cervical modifications undergone by them and the most nearly
related land forms. This may hence prove, once we know the
record more completely, to be another example of the fact
that a course of evolution and change once established in a
persistent group, may long continue, after the invasion of wholly
new environments. There is in biologic, as in physical evolu-
tion, inertia.

As to the carpus of Avchelon. It will certainly be very
interesting if my surmise that there is no union between ear-
pale 1 and the centrale should prove correct. This, although
to be regarded as a secondarily acquired character, would in-
deed go far toward narrowing the gap between the extreme
ends of the marine group. I may point out that the greatly
accentuated bow of the radius of Archelon would make it
probable, even in the absence of more direct evidence, that
there was present some marked change in the order of the
carpals. |

Systematic Position of Archelon. While the data given in
the preceding notes go far towards showing that Protostega
and Archelon present more osteological resemblances to Der-
mochelys than any other turtles whatsoever, living or extinct,
their structure is essentially that of the Chelonidae, of which

G. R. Wieland—On Marine Turtles. 107

they may best be regarded as a subfamily, the Protostegine.
Moreover, these Dermochelan resemblances are only what we
might well expect in Cretaceous turtles. There are likewise,
as we see, certain Chelydran resemblances in the general type
of skull, just as there are also Chelydran resemblances in
Ti oaochelys. Weare simply following convergent lines back
sufficiently far to somewhat accentuate general relationships.

Position of Toxochelys. The fore flipper and cervicals of
Toxochelys present some additional Chelydroid characters to
those of the cranium and lower jaw, as already pointed out by
Cope and Hay. TZoxochelys hence proves: to be one of the
most interesting of turtles. Like Pvotostega and Archelon, it
points with more or less distinctness toward a Chelydra-like
ancestry. Baur has said that this genus should be placed in a
distinct family, and in this has been followed by Hay and
Case. But I think that Zozochelys may more conveniently be
considered as representing a subfamily of the Chelonide, cer-
tainly if of common ancestry.

V. Provisional classification of the marine Testudinates.

CHELONIOIDEA (Baur).
(Superfamily of the Cryptodira.)

A parieto-squamosal arch; palatine foramen and free nasals
sometimes present (Desmatochelydinz); fourth cervical bicon-
vex, with the centra of the sixth, seventh, and eighth usually
greatly moditied.

I. Dermochelydide.

No descending parietal processes; no palatine foramen ;
other cranial and limb characters not remote from those of the
Chelonidz ; carapace represented by the nuchal only, and body
enveloped 1 in a leather y hide with an osteodermal mosaic ; no
claws. Genera: Dermochelys, Psephophorous, Eosphar gis.

‘Il. Chelonide.

Skull with descending processes of parietals, so far as
known ; palatine foramen sometimes present ; vomero-premax-
illar union often, but not constantly present ; a normal, though
often much reduced, carapace and plastron : nuchal with or
without process on under side ; claws, one or two.

1. Protostegine.—No free nasals; no palatine foramina ;
obturator foramen small and enclosed by ischiopubic contact
on median line, as in many land forms.

Genera: Protostega and Archelon ; Protosphargis? Pseu-

dosphargis ?

108 G. I. Wieland—On Marine Turtles.

2. ‘Toxochelydine.— No free nasals; palatine foramen,
pterygoids, and lower jaw distinctly Chelydra-like ; two strong
claws.

Genera: Toxochelys, Porthochelys, Cynocercus(?), all of
the Niobrara Cretaceous of Kansas, also Veptunochelys from
the Cretaceous of Mississippi.

3. Desmatochelydine.—Free nasals; distinct palatine fora-
mina (except Lhinochelys?). |

Genera: Desmatochelys, Rhinochelys, Atlantochelys.

4. Chelonine.—No free nasals; no palatine foramina ;
vomero-premaxillar union often, but not constantly present ;
obturator foramen presumably not enclosed as in Archelon in
any member of the group; claws, one or two.

Genera: Osteopygis, Allopleuron, Lytoloma, Argillochelys,

Eretmochelys, Chelone, Colpochelys, Thatassochelys.

* ; a9 * vis * 9

ftesume.—In the foregoing notes the following additions to
the osteology of the marine Testudinates have been made :—

The elements and organization of the front flipper of Zozo-
chlys

The-main elements of the wrist region of the front flipper
of reer ischyros. (These are described from the type of
the genus and species. The writer, in his first announcements f
the discovery of this gigantic turtle, figured and described the
accompanying humerus, with the radius and ulna, as well as the
femur, tibia, and fibula,—this being the first instance in which
all these limb bones were made known-in the case of any extinct.
eure ‘Testudinate.)

3. Important measurements for the codrdination of various
skeletal elements of Toxochelys and Archelon.

4. The deduction (with the exception of the secondary rear-
rangements of the carpals) of the principal lines of change in
the evolution of the fore flipper from the foot of some primitive
swamp, or littoral, Chelydra-like turtle.

5. The cervical organization of Yoxochelys and Archelon,
which is compared with that. of living turtles.

6. A classification of the marine turtles.—This has chiefly been
made possible by the description during the last few years of
large portions of the skeleton of Protostega and of Toxochelys,
and especially by the discovery of Desmatochelys, Archelon, and
Porthochelys.

Yale Museum, New Haven, Conn.
April, 1902.

Se ee ee ee

Whitehead— Magnetic Lifect of Electric Displacement. 109

Art. XV.—The Magnetic Lifect of Electric Displacement ;
by Jonny B. Wuirrnran, JR.

HistroricAL REVIEW.

In the development of his theory of the electromagnetic
field, Maxwell assumes that the phenomenon of polarization in a
dielectric consists of an actual propagation or displacement of
charge in the direction of the polarization ; that in the case of
the charging of a condenser, for instance, this act of displace-
ment is equivalent to a current at any instant equal to the
rate of change of the surface charge on one of the plates, i. e.
ae 4s q being the current density, K the specific inductive
capacity and F the difference in potential per unit length, and
that therefore the current is continuous throughout the circuit.
He assumes further that the displacement current has the same
magnetic effect as would be produced by a conduction current

of density g= eek so that in the case of the condenser the

4 dt’
magnetic effect in the neighborhood, incident upon any change
of charge, would be due to the combined influence of the cur-
rent in the charging wires and the displacement current in the
dielectric.* ‘It appears, therefore, that at the same time that
a quantity Q of electricity is being transferred along the wire
by the electromotive force from B towards A, so as to cross
every section of the wire, the same quantity of electricity
crosses every section of the dielectric from A towards B by
reason of the electric displacement.”’ (Art. 60.)......

“The variations of the electric displacement evidently con-
Ememievelectric currents.” -(Art.60.)......

“One of the chief peculiarities of this treatise is the doctrine
which it asserts, that the true electric current C, that on which
the electromagnetic phenomena depend, is not the same thing
as K, the current of conduction, but that the time variation of
D, the electric displacement, must be taken into account in
estimating the total movement of electricity.” (Art. 610.)

The remarkable consequences of Maxwell’s theory seem to
justify his assumptions beyond all question. There need only
be mentioned the finite velocity of propagation of electric and
magnetic actions, this velocity being the same as that of light;
the electromagnetic theory of light, which accounts for the
results of experiment, practically without exception; the

* Maxwell: Electricity and Magnetism, Art. 60, 75, 76, 111, 328-334, 608,
783, 791.

Am, JOUR. a SERIES, VoL. XIV, No. 80.—Avaust, 1902.

110 Whitehead—Magnetic Effect of Electric Displacement.

experiments of Hertz on electric waves showing them to be
identical in behavior with light waves. The direct magnetic
effect of the displacement current, however, has never been
satisfactorily observed, if at all; published accounts of work in ~
this direction seem to be limited to those of Roéntgen, S. P.
Thompson, Nicolaieff and Blondlot. .

Léntgen.—The earliest attempt, apparently, was that of
Rontgen in 1885.* He rotated a rubber disc between two
stationary glass plates all in horizontal planes. The upper
plate was coated with tin-foil which was grounded; the lower
plate had on it a ring of tin-foil which was split along a diam-
eter and the two halves oppositely charged. As the dise
rotated there was a change in polarization at those portions of
it passing over the opening in the tin-foil, and the resulting
displacement current was in opposite directions at the two
ends of this opening. Over the whole system and as close as
possible to the upper plate, he suspended an astatic needle
whose direction was along the line of the opening in the tin
foil, and the line of suspension the continuation of that of the
axis of the rotating disc. The length of the lower needle
brought its ends to the center of the width of the tin-foil rings ;
any magnetic effect of the displacement current would thus
tend to deflect the astatic system. On commutating the
charges on the rings the deflection read by a mirror and scale
was never over 1°5™", and the needle was subject to an oscilla-
tion of that amount due to disturbing influences. The ob-
server, being ignorant of the direction of commutation, was
supposed to take into account the motion already possessed by
the needle in giving the direction of the resulting impulse.
Rontgen states that after 1000 observations he acquired such
practice as to be able to observe the proper direction nearly
every time.

The results of this work can hardly be considered conelu-
sive. That little importance is to be credited to it is evidenced
by the small notice it has attracted. The genuineness of the
observed deflection is particularly to be questioned in view of
the fact that the poles of the needle were not in proper posi-
tion to experience the full magnetic force of the displacement
current. Each pole was entirely above the plane of the sur-
face of the dielectric, and it may well be questioned what is
the distribution of the magnetic field beyond the terminating
surface of an ‘open current.”

Thompson.—In 1889 S. P. Thompsont wound an iron ring
with many turns of fine wire and imbedded it in a block of
paraftin having on two sides parallel to the plane of the ring

* Rep. der Physik, No. 21, 1885, p. 521.
+ Proc. Roy. Soc., xlv, p. 892, 1889.

Whitehead— Magnetic Effect of Electric Displacement. 111

metal coatings which were connected to an induction coil.
The winding on the ring being connected to a telephone
receiver, sounds were obtained on charge and discharge of the
coil. Very slight conductivity in the paraftin would have
caused this.

Nicolaieff—N icolaieff in 1895* suspended a ring of paraffin
between the two poles of an electromagnet, the plane of the
ring being vertical and making an angle of 45° with the axis of
the poles. The E.M.F. induced in the ring when the magnet
is excited by an alternating current, sets up a displacement
current which, by its magnetic reaction on the exciting field,
creates a couple tending to make the ring turn and set itself
perpendicular to the field. He excited the magnet first with
direct current, getting a deflection due to the magnetic qualities
of the ring, then with alternating current of the same intensity,
causing a different deflection, the difference being due to the
displacement current. The record of observations is not given
in the paper, the author simply stating that the deflection was
greater when the magnetic field was alternating than when
direct ; and was greater the greater the frequency. For 12
eycles per sec. the increase was 9% over the constant magnetic
field, for 15 cycles 12%; in this case also conduction would
cause the observed effect.

Blondlot.— Recently M. R. Blondlott+ looked for an electric
displacement in a mass of air moving in a magnetic field. He
forced a blast of air through a rectangular passage, two sides
of which were pole faces of a magnet, the other two a pair of
condenser plates. The plates were connected by a wire which
was broken at the instant the blast was at its maximum, so that
the plates should be left charged if there was any electric dis-
placement in the mass of air. He calculated the expected
effect and calibrated his electrometer to read it; the results of
a number of experiments were invariably negative. Making
use of this experimental fact, he then proves theoretically the
absence of any electromagnetic action of a magnetic field on a
displacement current.

PrReEsENT RESEARCH.
General Theory.

The series of experiments herein described were undertaken
in view of the continued uncertainty of the results of all effort
to observe the magnetic effect of electric displacement, and in
view of the fact that the method employed had certain advan-
tages over those used heretofore.

* Jour. Phys., vol. iv, 1895, pp. 245-254.
+ Jour. Phys., Jan. 1902, p. 8.

112 Whitehead— Magnetic Effect of Electric Displacement.

In principle, the method is to subject a piece of dielectric to
an alternating electric field and also to an alternating magnetic
field, the directions of the two being at right angles in space ;
to adjust the phases of the two to give the maximum effect of

the displacement current reaction against the magnetic field,
and to look for motion of the dielectric in a direction perpen-
dicular to the plane including the directions of the electric and
magnetic fields. In this form the idea was suggested to the

Whitehead— Magnetic Hiffect of Electric Displacement. 118

writer by Professor Rowland shortly before his death. Some
twelve years before he made the same suggestion to Dr. Louis
Duncan, who constructed a crude form of apparatus, but

cis)

abandoned the work on account of the difficulty of obtaining
a uniform field. :

In each of four modifications of the form of the apparatus
embodying the above principle, a block of dielectric was hung

114. Whitehead—Magnetic Effect of EHlectrie Displacement.

rigidly at each end of a tight beam which was suspended hori-
zontally on a quartz fiber attached at its center. The phase of
either the electric or the magnetic field on one block being 180°
from that on the other, the other field having the same phase
for each block, the reaction of the displacement current on the
magnetic field would be opposite at the two ends of the beam,
causing a couple to act on the fiber suspension.

Apparatus.

First form.—The first form of the apparatus is shown in fig.
1, and fig. 2 gives a central section. AA are the rectangular
blocks of dielectric, each attached to the beam D, which was
constructed of bamboo or glass; B BBB are brass electrodes
each turned so that its surface is a part of a circular cylinder
with its axis coincident with the line of suspension of the beam ;
this was thought to give the best approximation to a uniform
electric field between the electrodes. Each pair of electrodes
was connected to the terminals of a transformer giving 8800
volts at 133 cycles per sec. CC are circular coils of wire, one
surrounding each pair of electrodes, the planes of the turns
being horizontal; they are supplied with alternating current
from the same generator which excites the electrodes. The
magnetic field is seen to be vertical, the electric field horizontal,
so the resulting deflection of the beam should be out of the
plane of the paper at one end and in at the other. The whole
rested on a wooden base, and each half of the apparatus was
enclosed in a brass case, there being between the two only a
small connecting trough in which the beam could swing, thus
making the interior as small as possible and so minimizing the
disturbance due to air currents. A close fitting brass cylinder
inside the coil, a ring on top, and tin foil on the floor, all con-
nected to the case, which was grounded, protected the dielectric
from any electric field due to the coil; all possible secondary
circuits were split, the openings being closed with hard rubber
or fibre. The damping vane shown in the sketch was a strip
of mica immersed in water, attached to the deflecting beam by
means of a fine glass rod. The quartz fiber used was approxi-
mately 102™ (40) long in all cases; it was enclosed in a
glass tube fitted with a convenient torsion head permitting the
adjustments of torsion and length independently; the fibre
carried a small copper hook at its lower end which fitted into

a corresponding hook sealed to the glass beam. Deflections -
Pp Ss =

of the beam were observed by means of the movements on a
ground glass scale of the image of an incandescent lamp fila-
ment reflected from a small mirror on the center of the beam ;
during the later experiments the lamp was replaced by a Nernst

j
j
4

Whitehead— Magnetic Effect of Electric Displacement. 115

filament, which is beautifully adapted to this purpose. The
distance of the scale from the mirror was 140. The connec-
tions are shown in fig. 3. The generator, G, was a Westing-
house “smooth-body,’ giving a pure sine wave, 1000 volts,
133 cycles; the transformers, T, T, T, were all Westinghouse
make. The transformer T,, giving the high voltage, was built
privately ; it was of closed magnetic type and had a ratio of 1
to 80. Bya suitable arrangement of switches 110, 200, or 400
volts could be put on the coils. The two magnetizing coils
were connected in multiple and so that they worked together ;
i. e., at any instant the fields were equal, parallel, and in the
same direction. The electrodes were also in multiple, but
worked in opposite directions, i. e., at any instant the fields

were equal, parallel, but in opposite directions. In each circuit
there was a reversing switch, RS.

Calculation of Lffect Hxpected.—F ollowing is a calculation
of the magnitude of the effect in the apparatus as constructed
and experimented upon.

Each pair of electrodes were 1°9™ apart; the dielectrics were
blocks of rock salt, glass or paraffin, 1°" x 1™ x °63™ and they
were hung with their square faces parallel to the electrodes, so
that one-third the length of a line of force was within the
dielectric ; thus V2-8800 being the maximum E.M.F. between
the electrodes the proportion of it active on the dielectric is
given by:

vy = V2 - 8800
ae ee are)
1G
Haiintensity of EALY. hoes

4/2. 8800

(QK +1) x63 (maximum value in volts).

je =

116 =Whitehead—Magnetic fect of Electric Displacement.

The density of the displacement current is ¢= pak = since ~

4 dt
the E.M.F. is alternating we have:
Pe Sin. a
ie da.
and 7=G-im cosa 5,
also ee = angular velocity = 27N = 7. 133

> dt
Ta nK’ 4/2. 8800°" Qaedpamans
Ty ae (OK 41) 63" SS

cos a in C.G.S. electrostatic

units.
Ka 2. S800 Qa . 133. 10° Me
= 4a ° (2K +1)X63' (8X10)? X10= cOS a, amperes
Keg i
OS A ollie
K

Peek ns amperes, effective, per unit cross section.
Each of the coils for the magnetic field contained 1200 turns
of No. 18 B. &.S. magnet wire; H, the intensity of the
magnetic field at the center of these coils, as calculated with a
neglect of the influence of the ends, was 200 7,7 being the
current in amperes; as measured by the electromotive force
induced in an exploring coil of known area and number of
turns, H was 166 7, (area of exploring coil 25°™’, 400 turns ;
observed E.M.F., with 2-4 amps. im coil, 33°5 volts, 133
cycles). The resistance of each coil was about 9 ohms, and
the calculated value of the coefficient of self-induction was

L = -109 henry; at 114 volts, 133 cycles, each took 1:3 amperes ~

which corresponds to a value 104 for L. If @ be the differ-
ence in phase between the E.M.F. and current, i. e., the angle
of lag, we have:
_ amNL _ 6°28 x 188 X “104
‘gpa: ae 9

Now consider the phase relations between the displacement
current and the alternating magnetic field from the sketch of
the connections. Since the difference in phase between the
primary and secondary E.M.F’s of a transformer is approxi-
mately 180°, the E.M.F. on the electrodes having passed
through two transformations has returned to coincidence of
phase with the generator E.M.IF’. and may be indicated by
sina. The resulting displacement current has the phase cos a
as shown above, i.e., differs by 90° from the generator E.M.F.
The E.M.F. impressed on the coils has been transformed

tan 6 = 9'3, whence 6 = 83° 52’

\

Whitehead—Magnetic Effect of Electric Displacement. 117

once, suffering retardation of 180°; owing to the self-induction
of the coil the resulting current lags approximately 90°, and
so, therefore, does the resulting magnetic field. Consequently,
the displacement current and magnetic field are approximately
180° apart, that is, in the proper relation to give the electro-
magnetic reaction. Actually, since the angle of lag in the
coil is 83° 52’ instead of 90°, the effect as calculated for coin-
cidence must be multiplied by -9942 = cos 6° 8’.
The current in the dielectric is thus acted on by a force:

| lq K 10°3. 1661-2
a= G3. A { .°9942.
10 Oe) bor 10 :
= - fae dynes
ment 10 2

when 1:2 amperes flows in each coil, 7 being the thickness of
the dielectric, or the length of the displacement current acted
on by the field, and H the intensity of the magnetic field.
This force is applied in opposite directions at the two ends of
the suspended beam, which was 17°78™ long, giving thus a
couple of
K 2°28
2K+1° 10%

The angle of twist of a thread of length 7, radius 7, coefficient
of rigidity n, acted on by a couple w is:
2 tee
o——~

rm on

dyne-centimeters.

The length of the quartz fiber was 101-6; its radius was
estimated by comparison under a microscope with the spaces
of a grating ruled on glass, the grating space of which was
known, and did not differ greatly from -0006™. The value of
m being taken at 310" for quartz,* we have

K ZSCLOEG X S38 K
Ss = 4 4 ia abet poe
2K +1 ° «~x:0006**10°**3x10 2-3

Taking the value of K for rock salt as 5-8, we have ®=-173,
which, since the distance from the mirror to scale was 140,
represents a deflection of 2 1400 x‘173=485™", or more, since
we have taken the tangent as equal to the are.

The caleulation above assumes that the electric and magnetic
fields remain constant through the range of movement of the
dielectrics. It was of course foreseen that the electric field
would be most intense at the center of the space between the
electrodes and so would tend to hold the dielectric within that
region ; nevertheless it was thought that a couple of the magni-

. 37 radians.

* Threlfall and Boys.

118 Whitehead—Magnetic Effect of Electric Displacement.

tude above indicated would to a certain extent overcome this
tendency. ‘The test for the true effect seemed to lie in a
reversal of the sense of the deflection when the phase of either
the electric or the magnetic fields was changed by 180°.

Lesults—When the electric field alone was applied, the
beam, as was expected, found a stable zero position near the
center of the electrodes. With a fiber 20™ long this position
could be upset by 10 to 20™™ on the scale, by a very small twist
of the torsion head. This test was not so satisfactory when the
long fiber was used, owing, apparently, to the large imertia of
the syspended system, and to the arising of other disturbances
within the time necessary for the torsion of the fiber to make
itself felt. The zero point was always quite steady, having
practically no oscillation, but was not always the same, varying
within two centimeters. This indicated that near the center
there was a region of fairly uniform electric field, and also that,
owing to the mass and damping of the suspended system, the
torsion of the fiber could not be depended on to give the zero
position. The mass of the beam and dielectrics was from 2 to
3 grams, varying with the nature of the dielectric ; of this the
greater part was in the dielectrics. Fibers of the size here
required broke if the mass were 3 grams or more.

The method of procedure was to allow the beam to come
to rest under the influence of the electric field alone, the long
time required for the oscillations about the central position to
die out being another evidence that the field there was fairly
uniform. Then the magnetic field was put on and the result-
ing deflection noted. Then either the electric or magnetic field
was reversed, the deflection noted, and so on. The deflection
was generally small and quite slow im all cases, requiring con-
siderable time to become steady; in most of the observations
only the sense of the deflection was noted. Proceeding in this
way, several hundred observations were taken ; they were gen-
erally negative in result, that is, the deflection did not change
in either direction or amount when either field was reversed.
At times, however, there were indications of the effect looked
for, as is shown by the following description :

The electric field and then the magnetic field being put on,
as described, with the values and conditions as given above,
there resulted an uncertain deflection of about 1™™, which did
not reverse with a reversal of the magnetic field. The circuits
were then arranged so that 200 or 400 volts might be impressed
upon the coils. With 200 volts a deflection of from 5 to 10™
was obtained, which, however, did not reverse with a reversal
of the magnetic field; this deflection was towards a position
which the beam tended to take when under the influence of
the magnetic field alone. With paraftin as the dielectric the

——— a Se

Wihitehead—Magnetic Lifect of Llectric Displacement. 119

proper reversal was obtained to the extent of 4 or 5™™ on two
occasions, one with a reversal of the electric field, the other
with a reversal of the magnetic field. In both instances the
deflections were from a zero position constant within 1™™, and
appeared as a slow motion immediately on closing the switch.
The above mentioned disturbing effect of the magnetic field
was always evident through all the experiments, but in this
form of the apparatus in several instances there were distinct
initial deflections in opposite sense to this tendency.

When the switches remained closed long enough, the above
disturbing tendency seemed to predominate regardless of the
direction of the field. The question at once arose as to whether
it was a magnetic or an electric effect. Experiments with bits
of iron filing on the dielectrics, with hard rubber known to be
slightly magnetic, with rock salt which gave no trace of mag-
netic impurities, and with both direct and alternating currents,
indicated that a part of this disturbing influence was due to the
difference in the intensity of the magnetic field at the center
and close to the side of the coil. For this reason only rock
salt, glass and paraffin were used, as they appeared quite free
from magnetic impurities; with them the disturbance was less
than with sulphur and hard rubber, though it was still the
predominating influence. Further experiments with rock salt
several times gave evidences of the proper reversals to the
extent of 2 or 3"; and with glass on one occasion the expected
reversals of deflection were noticeable for six switch reversals ;
these deflections were only 2 to 3™™. Very long waits between
readings were necessary owing to the difficulty of obtaining a
steady zero; this was probably due to air currents set up within
the case caused by the heating of the coils when on the 200
volt circuit ; the six readings mentioned required one afternoon.
On the next day the proper reversals of deflection were observed
for four reversals each of magnetic and electric fields; they
‘were from 2 to 5™".

It must be stated, however, that these proper deflections
were taken from among a mass of attempts which resulted
sometimes in no deflection at all and at others in deflections in
the wrong sense. While it was sometimes possible to account
for these by a steady shifting of the zero position, I still ques-
tioned whether I had obtained the effect, and looked to the
improvement of the apparatus.

Second and Third Form of Apparatus.—In the second form
the air space was reduced and the electric screening made bet-
ter by enclosing only the electrodes and suspended system
within the case, the coils being outside. The case was entirely
of brass, the posts of the electrodes being set in insulating
bushings. To prevent secondary currents in the cylindrical

120 Whitehead—Magnetic Effect of Electric Displacement.

cases, they were each split down an element of the cylinder;
these splits were covered and were on the front of the case,
i.e., diametral planes through them cut the beam, in its central
position, at right angles. The conditions were not improved
by the apparatus in this form; if anything, the disturbance
due to the magnetic field alone was worse, and it always
caused a movement of the dielectric towards the side of the
casing to which it happened to be nearer. Still thinking it a
magnetic effect due to the more intense field near the coils, a
single large coil which would completely surround the entire
case and have its center in the line of the fiber, was wound.
The field due to this coil would be uniform ar ound any circle
concentric with the coil; the coil was 7:62™ high, had an
internal diameter of 26-6e" and contained 360 turns of No. 15
wire. The conditions were not improved by the use of this
coil, and it was still impossible to look for the true effect.
This seemed to indicate that the disturbance was not magnetic,
nor could it be due to electrostatic influence of the coil since
the shielding was practically perfect. It was thought then
that the electromotive force induced in the cylindrical metal
case must set up a field in the slit down its side sufficient to
cause the disturbances. To test this the slits were closed with
solder, and new ones cut directly back of each of the outer
electrodes; the disturbance disappeared almost entirely. Tests
were then made with the single large coil and also with the two
single coils; the constants for the large coil introduced no con-
siderable difference in the calculated effect; the conditions in
each of these series of tests were quite good; the electric field
alone caused a good steady zero position ; on closing the switch
of the magnetic field there was usually no deflection at all,
only a slight oscillation; sometimes there was a deflection of
1™™ or 2™™, which, however, did not reverse on reversal of field
and was pr robably, a survival of the old disturbance.

Fourth Form of Apparatus.—As has been pointed out, the
design of the electrodes in the experiments above described
caused a position of stable equilibrium for the dielectrics, and

it was thought that if the electric as well as the magnetic field

were uniform and constant for all positions of the beam, the
conditions would be greatly improved and practically as favor-
able as the method permits. To this end the final form of the
apparatus was constructed: The large coil described above was
used for the magnetic field; as shown im fig. 4, the electrodes
were two complete rings of brass, set concentric with each
other and with the coil. The surfaces were carefully turned
and polished, and each ring was split in one place to prevent its
becoming a short circuited secondary circuit. The true cir-
cular form of the ring was maintained by bridging the split

— —_.

Whitehead— Magnetic Effect of Electric Displacement. 121

with a strip of hard rubber; the splits in the two rings were
in the same diametral plane. The rings were set on small, hard

rubber chairs, and carefully levelled and spaced so as to be as
nearly as possible the same radial distance apart at all points ;
this distance was 1°9™ (3’’). In the earlier experiments this

122 Whitehead—Magnetic Lffect of Electrie Displacement.

distance was varied, and 1°9™ was found to be the minimum
practicable separation, anything less resulting in the dielectrics
receiving a charge which would draw it against the surface of
one of the rings; for any distance between the electrodes this
always happened in damp weather. The coil was carefully
sheathed with tin-foil which was connected to earth. The
suspension hook on the beam was lengthened so as to reach
into a glass tube shown in the sketch at M and the mirror
placed on it there. By moving the source of light and scale
in a circle about the apparatus, it was thus possible to take
observations in any position of the beam. The walls of the
glass tube scattered the light somewhat, but the intensity of
the Nernst filament was so great as to always give a well de-
fined line on the scale. The whole deflecting apparatus was
suitably enelosed so as to be free from disturbance by air
drafts.

Calculation of Expected Lffect—With 8800 volts on the
electrodes and with dielectrics of glass, paraffin and rock salt,
the order of magnitude of the displacement current is not
different from that already calculated. The angle of lag of
the current in the coil, that is, the phase of the magnetic field,
and also the intensity of the field, had also to be determined.
The angle of lag was obtained by measuring the current taken
by the coil at a known voltage and frequency. At 400 volts
and 133 cycles this current was 10 amperes, the resistance
being 4 ohms:

iO = als whence L = 05 henry ;
/16+835 L?
‘also the angle of lag, 6, is given by
tan d= moe = 10°45, or 6 =)84" 289

The intensity of the magnetic field at various positions within
the coil was determined by sending a current through it and
measuring the E.M.F. in an exploring coil of known area and
number of turns. At 200 volts the current in the coil was 5°3
amperes. ‘The exploring coil consisted of 400 turns of No. 33
wire, and had an effective area of about 25:8. Moving the
exploring coil along a radius by equal steps, the following read-
ings give the curve, shown in fig. 5, of variation of the inten-
sity of the magnetic field with the distance from the coil;
ordinates are values of the intensity and abscisse distances
from the center, the coil having a radius of 13°6™. The de-
flecting beam was 22°2°" long, so that the dielectrics swing at
a radius of 11:1, and at this distance the field, with 5°3 am-
peres in the coil, had an intensity of approximately 220, or 42

ee a

OE

Whitehead— Magnetic Effect of Electric Displacement. 128

per ampere, thus of the same order of magnitude as in the cal-
culation for the first form of the apparatus. The suspension
fiber was also the same, and so, therefore, the order of magni-
tude of the calculated deflection.

Position Volts

Center 7°8
1 8°]
2 8°6
3 9°9
4 (near coil) Pt

fesults.— With this apparatus, when the electric field alone
was put on, the beam after a sufficient time always found a

ee

H per ampere.

re G 8 is 12 14
2 * edge

center

position of equilibrium, in some places stable and at others
unstable. The former seemed to be in a diameter which passed
through the hard rubber supporting chairs of the rings, doubt-
less owing to a slight concentration into the chairs of the field
of force; a slight departure from this line was sufficient to cause
the beam to wander to some other position, and the indications
seemed to certify a quite uniform electric field. On putiing
on the electric and then the magnetic fields, (5°3 amperes in
coil) there was a deflection of from 2° to 3°™, which, however,
did not reverse or show any regular difference in amount
when either field was reversed ; it was found that the magnetic

124 Whitehead—Magnetic Lffect of F lectric Displacement.

field when on alone gave a deflection in the same direction.
Thus the magnetic field when on alone also had an effect on the
position of the beam. With the beam at rest, when this field
was made a slow drift away was generally seen, the beam com-
ing to rest in some other position. Repeated efforts failed to
locate single positions that it seemed to prefer, though in cer-
tain neighborhoods the direction of the deflection seemed the
same. In some positions there was no deflection, though these
were never the zero positions for the electric field alone; the
problem then seemed to bring the two zero positions together.

Since the disturbance due to the magnetic field did not ap-
pear to be the influence of the splits in the ring as before, for
it seemed independent of the distance of the dielectrics from
them, and since the intensity of the magnetic field was uni-
form, it therefore seemed that direct electric influence of the
coil or leading-in wires must be the cause, though the tin-foil
screen seemed as perfect as possible. The attempt was then
made to bring one of the positions in which the beam seemed
unaffected by the magnetic field alone, into coincidence with a
zero position for the electric field alone, by turning the coil
into various positions about its center. This process was most
tedious, involving as it did long waits for the beam to come to
rest, and at its best it seemed only a hit-or-miss method. The
condition sought, however, seemed so desirable, that repeated
efforts were made to bring it about. One of them succeeded 3.
and while the result obtained must be discredited in view of
the failure of all attempts to repeat the proper conditions, and
of other negative attempts, it is nevertheless thought that the
following quotation from a note taken at the time should be
inserted here: “this condition was obtained March 18th; the
beam was allowed to come to rest under the electric field; this
was then cut off and 400 volts put on the magnetic coil, which
caused no deflection of the beam. On putting on both fields
there was a slow deflection to the right of between 1™ and 2,
about one minute being required. The magnetic field was
then taken off and the beam allowed to move under the electric
field alone; it resumed the former zero after one or two min-
utes. The magnetic field was then put on reversed; there
resulted a much stronger deflection to the left which was
allowed to continue several minutes, resulting in a deflection of
5™, Repetition could not be obtained, due to the fact that
after the large deflection the beam did not return to the origi-
nal zero position. (The heating of the apparatus due to the
large current in the coil (10 amperes) when applied for any
length of time frequently upset the equilibrium positions.)
The conditions, however, of the above observations were so
favorable that my belief in the presence of the effect was

Whitehead— Magnetic Effect of Flectric Displacement. 125

much strengthened after the many recent failures to detect any
evidence of it.”

Failing to repeat the condition as above described, a light
strip of wood was placed along a diameter on top of the inner
ring. This strip took up a sufficient charge to give the beam
the position of delicate equilibrium when the beam was paral-
lel to it, 1. e., directly over it; it could be rotated to any posi-
tion on the ring without opening the case, by means of a device
not shown in the sketch; the rod of the damping vane passed
through a hole at its center. A number of experiments were
made with the following order; the beam was allowed to find
a stationary position under the influence of the magnetic field
alone. ‘The interior being visible through the glass cover of the
casing, the wooden strip was turned so as to be parallel to and
directly under the beam, with the idea that the zero positions
for the two fields would be brought into coincidence. It was
found that in most cases any movement of the strip would
upset the position of equilibrium due to the magnetic field.
Frequently, however, the condition of coimcidence was ap-
proached, and in all these instances when the effect was looked
for the results were either spurious or negative. That is, if
there was a deflection, as sometimes to the amount of 1™™ or
9™™, neither its direction or amount changed with a reversal of
field; or often there was no deflection at all for either phase
of -the field.

Differences in deflection for the two phases of either field
were now looked for; the wooden strip was discarded and the
beam allowed to find a stationary position under the action of
both fields. That is, the electric field was put on and a steady
state reached ; then the magnetic field, resulting in a deflection
of usually between one and two centimeters to a new position
of equilibrium. The magnetic field was then quickly reversed
and a change in the position of the beam looked for. Ina
series of observations in which the conditions seemed perfect,
no change at all could be noted when the field was reversed,
there being either no deflection at all, or less than 1™ without
regularity. In these experiments the coil only carried 2°6
amperes, giving a calculated effect one-half of that estimated
above; 5°3 amperes, i. e., 200 volts on the coil, during the time
required for the beam to come to rest, heated the interior to an
extent sufficient, by reason of air currents or expansion of the
ring electrodes, to cause erratic oscillations of the beam about
its zero position.

While it was thought that at the comparatively low frequency
of the alternating circuit, 133 cycles per second, a possible lag
of the displacement current, owing to molecular friction or other
cause, could hardly make itself felt, it nevertheless seemed

Am. JouR. ee Pg Surges, Vou. XIV, No. 80.—Avueust, 1902.

126 Whitehead—Magnetic Effect of Electric Displacement.

worth while to shift the phase of either the electric or the
magnetic field by a quarter of a period. For this purpose a
two-phase machine was used, one phase being put on the pri-
mary of the high tension transformer, the other directly on the
magnetizing coil. This separates the displacement current and
magnetic field by one-quarter of a period, if there is a lag of
one-quarter of a period in the coil due to its self-induction,
as was approximately the case in the experiments already
described. The machine operated at 38 cycles, and 85 volts
per phase; the resulting displacement current is thus

88 85

133 > 1107)
of the value calculated for the first arrangement; at the lower
frequency and voltage, however, the coil took 6°2 amperes, so

Mee

5G 1-17 times that in the first

that the magnetic field is 5

arrangement, so that the calculated magnitude of deflection is
still appreciable ; the angle of lag is given by

Q2nNL 237X + 05
Bai Sf, esis

The cireuits were also arranged so that both electric and mag-
netic fields might be put on the same phase or one on each ;
this permitted, with the use of the reversing switches, a rapid
change of phase of either approximately 90° or 180° between
the displacement current and the magnetic field. In a series
of observations with this arrangement there appeared to be no
difference among the results of the several combinations of
phase. The beam was allowed to come to rest under the
electric field alone; on putting on the magnetic field there
was the usual deflection, but there was no evidence of any
change either in direction or amount of this deflection when
the phase relations were altered at random.

tan 6 = = 3°05, .. 0= 72°.

Summary and Conclusions.

The mass of the evidence of this research is against the pres-
ence of the magnetic effect of electric displacement in an amount
given by Maxwell’s expression. The single positive result
obtained with the final form of the apparatus must be ques-
tioned, for although the favorable conditions under which that
result was observed could not be obtained a second time, it is
thought that at times they were approached sufficiently to give
traces of any effect as great as that noticed. The final form of
the apparatus was undoubtedly that most likely to give results ;
in it there appear to be only two influences which might pre-

Whitehead—Magnetic Effect of Electric Displacement. 127

vent the effect sought. The first is a variation in the intensity
of the electric field from pomt to point and the tendency of
the dielectric to remain in the region of greatest intensity ; and
the second, the presence of the damping vane. It is thought
that neither influence was great enough to mask a couple of
the value of that sought; as has been stated, the rings were
carefully turned and adjusted in place so that their surfaces
_ were the same distance apart at all points. Indications of any
choice of position.by the beam in the electric field were of the |
slightest, the stationary positions being very easily upset, and
often changing by one or two centimeters between readings.
The distribution of the electric field due to the coil, though
uncertain, could not have departed greatly from symmetry
about the center, and since the disturbance due to it manifested
itself in deflections rarely exceeding 1°”, it does not seem possi-
ble that the variations of intensity were great enough to account
for the absence of visible effect when the displacement couple
acting on the beam was reversed. In the earlier experiments
several series of observations were made without the damping
vane; no difference in the behavior or in the nature of the
results was observed except the increase in time necessary for
the beam to become stationary. Observations could not be
made with the final form of the apparatus if the damper were
taken off; the electric field was so nearly uniform that the
beam would not retain a stationary position definite enough to
be taken as a zero point.

The following arrangement, independently conceived, but
afterwards discovered to be an improved modification of the
apparatus used by S. P. Thompson, offers a possible means of
detecting a magnetic effect of the displacement current. An
Annulus 1:9 thick, built up in laminations of sheet iron rings
3°18™ inside and 7°6™ outside diameter, was wound with a coil
of several hundred turns of fine wire, which was connected with
a telephone receiver. In the opening in this magnetic circuit
was placed a cylinder of dielectric 2°54™ in diameter and 2°54™
in length. Flat electrodes were brought very close to the ends
of the cylinder, but were not allowed to touch it. When now
an alternating E.M.F. is applied to the electrodes the resulting
displacement current. should set up an alternating field in the
surrounding magnetic circuit, and so induce a current in the
telephone receiver. The space between the electrode and
dielectric prevents in great measure a possible conduction cur-
rent due to conductivity of the material of the dielectric. With
8800 volts at 133 cycles on the electrodes, neglecting the air
gaps, and with parafiin (K=2) as the dielectric, the displace-
ment current density is:

128 Whitehead—Magnetic Hiffect of Hlectric Displacement.

1D 2472) S800 Ors. 183°) 10°
oF ge GTA S 10") 105 ie

and the total amount is
6°45 g = 3°34 X 10° amperes.

The experiment was tried with paraffin and with hard rubber
and no sound could be detected in either case. The question
whether the telephone receiver would respond to such a small
current was of course still open, nor could it be completely
answered. The following approximation was, however, made:
a conductor was placed through the opening of the iron annulus
and included in a circuit with an ammeter, a resistance, and a
source of alternating electromotive force in low values; the
lowest division on the ammeter scale was ‘02 amperes, and the
voltage was adjusted to give this reading, the telephone receiver
giving a good note. Resistance was then inserted by steps up
to about 100 ohms (inductive), apparently bringing the ammeter
needle to zero; sound in the telephone receiver was still audi-
ble, though faint. The experiment was performed very roughly
and hurriedly, and is susceptible of considerable improvement.

The writer’s thanks are extended to Professor Ames for his
interest and kindness in furnishing every facility needed in the
course of the investigation.

The mechanical work was done by Mr. Charles Childs; for
his skill and interest at all times the writer wishes to extend
his appreciation. —

It is proposed to carry this work further in both forms of
apparatus.

Physical Laboratory,

Johns Hopkins University,
May, 1902.

Plowman—Ffelations of Plant Growth to Lonization. 129

Art. XVI.— Certain Relations of Plant Growth to Loniza-
tion of the Soil; by Amon B. PLowMAN.

In a series of experiments now in progress at the Harvard
Botanie Garden, there have been observed some interesting
phenomena in the relations of plants to electricity. The
experiments have been of a widely varied nature, dealing with
both static and kinetic charges, through a range of potential
from 0°5 to 500 volts. Either platinum or high-grade carbon
—usually the latter—was used for electrodes, and special care
was taken to insure normal conditions of temperature, light,
and moisture.

The experiments were of course checked off by controls,
always using at least two different sets of controls in order to
avoid accidental errors. In a well-lighted greenhouse it is not
a difficult matter to locate the different plants in such a way as
to give fairly uniform light to all. For soil-cnltures, pots of
the same size were used, filled with equal amounts of care-
fully prepared soil, and supplied with measured quantities
of water.

The regulation of temperature presented serious difficulties
whenever a considerable amount of current was used, for the
resistance of the soil is very great and much heat is evolved in
forcing the current through it. Thus a 500 volt circuit,
joined through a body of soil 1025™ in cross-section and
40™ long between electrodes, gave a current of 0°2 ampere, and
in one hour raised the temperature of the soil from 16° to 44°C.

However, with a slight excess of moisture and a current not
exceeding -05 ampere, it has been possible to prevent a rise of
temperature of more than 38° above the normal; and in most
eases the temperature has been kept within a range of 1°
above that of the control. In the matter of temperature-
regulation the water-culture method possesses many advan-
tages. Hither by a slow circulation of the electrolyte, or by a
more rapid circulation of water about the vessel containing the
electrolyte, it is easy to maintain a constant temperature even
when using currents of 2 amperes. In this way it is possible
to introduce a very considerable electrical factor without
seriously disturbing the other relations of the plants.

Among other facts recorded in the course of these experi-
ments is the following: Seeds placed near the anode are
always killed by current amounting to -003 ampere or more, if
continued for twenty hours or longer, while seeds placed near
the cathode have in most cases been but little affected, and
under some conditions have been apparently stimulated by
such currents.

130 Plowman—Lelations of Plant Growth to Ionization.

When the seeds were germinated in water this difference was
most pronounced when a relatively heavy current was passed
through the water for only a short time, in which case the
seeds near the anode were killed, while those near the cathode
were not apparently injured. If the current was allowed to
flow for twenty hours or longer the ill effects were produced
at all points between the electrodes, even when the current
amounted to only ‘003 ampere, at a voltage of 2 or more.
When the current was passed through ordinary sandy soil in
which the seeds were planted, the same results were obtained
at the anode, but a much longer time was required to extend
the injurious effects to the region of the cathode. Indeed, so
long as the current does not exceed about -08 amp. there is a
considerable increase in the rate of growth of the seedlings
near the cathode.

In explanation of these phenomena the following provisional
theory is offered.

Whenever two points in any electrolyte are electrically
charged to different potentials, the movement of the free ions
in the solution* is given definite direction, and, if the differ-
ence of potential is sufficient, further dissociation of the
electrolyte follows. The anions with their negative electrons
move toward the anode, and the cations with their positive
charges pass to the cathode. Since the movements of ions in
solutions are relatively slow,t it is reasonable to suppose that
in the region of the anodes there would be a slight excess of
positive ions, due to the rapid neutralization of the negative
ions by the positively charged electrode. In like manner the
cathode removes the positive ions in its immediate vicinity, and
is consequently surrounded by a slight excess of negative ions.
The more slowly the ions move through the electrolyte the
more marked wiil be the difference of conditions about the
two electrodes.

How will seedlings respond to these differences of condi-
tions? Itis clearly indicated by the results of our experiments
that vegetable protoplasm is paralyzed and quickly killed by
the conditions existing about the anode, while within certain
fairly broad limits it is stimulated by the conditions about the
cathode. While the dissociation of the atoms and electrical
separation of the ions must bring about slight differences of a
purely chemical nature in the region of the electrodes, yet

* Ostwald and Nernst. Ztschr. phys. Chem. iii, 120, 1889.

+ Kohlrausch (Wied. Ann. 1. 403, 1893) has shown that, in an electro-

lyte of unit potential gradient, the velocity in centimeters per second for
certain ions is as follows :—

CATIONS | ANIONS.
H_ ‘00320 NH; °00066 | OH -00182 if ‘00069
Na ‘00045 Ag “00057 | Cl -00069 Ni ‘00064

Li ‘00036 K ‘00066

|

Plowman—felations of Plant Growth to Ionization. 131

these mere chemical differences can hardly account for the
effects upon the plants, even near the electrodes, and certainly
not for certain effects at points mid-way between the elec-
trodes. This conclusion is borne out by the following facts :—

1. When seeds are germinated in distilled water through
which a weak current is forced, the O ions are in excess in
that part of the solution where stimulation occurs, and the H
ions are in excess where the plants are killed. But it may be
shown that small quantities of hydrogen are not markedly
harmful to plants. And the quantity of oxygen set free in
distilled water by a weak electric current is certainly no greater
than that normally present in ordinary: tap-water, such as was
used for the controls; yet germination is often more rapid
near the cathode than in the case of the control.

Apparently, the effects are produced by the electrical

charges of the ions, rather than by any mere chemical activity
of the atoms.
_ 2. When seeds are placed in solutions of various acids,
bases, or salts, of a degree of concentration far below the
“ killing point,” they will germinate quite as well as in ordi-
nary distilled water. But when a current of electricity of
sufficient strength to propel the ions is passed through the
solution, that part about the anode becomes destructive to
plant life.

From these and other facts we conclude that negative
charges stimulate, and positive charges paralyze, the embryonic
protoplasm of these plants. This is strikingly in accord with
the conclusions reached by Matthews* in his experiments on
the nature of nerve stimulation, in which he shows that the
sciatic nerve of the frog is stimulated by negative ions, and
rendered less irritable by positive ions.

In support of the theory here advanced the following facts
may be mentioned :—

1. When a flower-pot containing several lupines of about
four weeks growth is charged to relatively high potential
(500v.) with positive electricity, the plants cease to grow,
gradually lose their turgidity and finally die. On the other
hand, when a negative charge is used these effects are not pro-
duced, but the plants are actually stimulated.

It is evident that, in the first case, the negative ions in the
soil are drawn to the positive terminal, while the positive ions
are driven to the plant. In the second case, the positive ions
are drawn to the negative electrode, and negative ions are
driven to the plants.

2. When seedlings are grown in an aqueous culture-medium
through which a weak current of electricity is flowing, the

* Science, vol. xv, No. 378, 1902.

132 Plowman—feelations of Plant Growth to Lonization.

root-tips turn toward the anode. An attempt has been made
to show an analogy between this movement and a similar
turning of root-tips “up stream” against a water current.
However, the analogy is of little value, owing to the fact that
in the electrolyte there is a streaming of ions in doth direc-
tions. Hence, the question is not so much a matter of the
direction of the electric current as it is of the difference in
effects of the two streams of ions.

Upon examination of the conditions which prevail about
roots growing in an active electrolyte, we find that the side of
the root toward the anode is being bombarded by a stream of
positive ions moving toward the negative electrode, while the
side toward the cathode is exposed to the stream of negative
ions on their way to the anode. Consequently the side of the
root toward the cathode is stimulated, and that toward the
anode is retarded, in its growth, and of necessity the root-tip
curves toward the anode.

Seedlings grown in ordinary soil show this curvature even
more strikingly. The main axis of the plant is frequently
curved almost 90° just below the surface ofthe soil. The
curvature is toward the anode and away from the cathode,
whether these be in circuit or isolated.

3. Normally, the plant body is electro-positive to the soil
in which it grows. The potential difference appears to be a
function of the physiological activity of the plant. The
positive charge of the plant attracts the negative ions of the
soil to its roots. Thus it seems that negative electrons are
being constantly discharged to the plant as a natural condition
of its life-activity. Any circumstance which would facilitate
this electrical interchange we should naturally expect to be
beneficial to the plant, while the reverse condition would be
detrimental. .

-It must be borne in mind that the phenomena with which
we have been dealing are conditioned not only by temperature,
light, aeration and moisture, but also by the nature of the
electric current used, the solution-tension of the electrodes, the
osmotic pressure of the electrolyte, the degree of dissociation
of the electrolyte, the valence of the ions, the physical state of
the ions, the chemical relations of the ions to metabolic
processes in the plant, besides certain peculiarities of the living
protoplasm with which we are working, and which should
doubtless be taken into account in connection with each of the
conditions named above.

Evidently, any theory which may be advanced in explanation :

of the various phenomena of plant-growth in the electrical
field, must stand the most exacting tests of physical chemistry,
in order to be worthy of serious consideration.

Harvard University.

Crook—Electromagnetic Alternating Currents. 188

Art. XVIl.—Demagnetizing Liffects of Llectromagnetically
Compensated Alternating Currents; by ZENO E. Crook.

Tue effect on the hysteresis cycle of passing an alternating
current longitudinally through an iron wire, was given in a
paper published in 1891, by Dr. G. Fienzi and Professor G.
Gerosia.* Ata later date, Ignas Klemenicic published a paper
on the relation of circular to longitudinal magnetism in iron
and steel wires. He madea very brief mention of the effect
due to alternating current. In the first named paper the
writers found a very evident reduction of the area of the
hysteresis cycle, and that a current of three amperes per sq. mm.
was sufficient to totally destroy the hysteresis. When in this
condition, the ascending and descending magnetization curves
coincide and the effect of the magnetizing force approaches the
- theoretical condition expressed by Frohlich’s formula

aH
1+8H

in which aand # are constants for a given specimen of iron, and

|

a is the saturation value of I. The writers thought the effect

observed was due only to the demagnetizing power of the rap-
idly oscillating circular magnetic field. Their reason for this
was that there was no perceptible jarring or vibration in the
iron.

An attempt will be made in this paper to go still further into
the study of the effects of alternating current on the magnetic
properties of iron and steel. The main object of the study
will be to discover if there is an effect due to the current inde-
pendent of that producéd by the circular magnetism.

The first question to be solved was how to eliminate the
electromagnetic action, so that the effect of the current alone
could be studied. This was done, approximately, by making
use of the principle discovered by Ampere, that when two
adjacent currents in parallel conductors flow in opposite direc-
tions, the external field of force of the one tends to destroy
that of the other. In order to be effective in destroying the
internal as well as the external magnetic actions of the current,
it was necessary to build up the iron from very thin laminze,
and send the current in opposite oo in each alternate
layer.

*Rendiconti de Instituto Lombardo, vol. xxiv, fasc. x, April, 1891; Elec-

trician, April 5, 1891; Ewing, Magnetic Induction in Iron and Other Metals,
page 319; Wiedemann’s Annalen, vol. lvi, page 530.

134 Crook—Electromagnetic Alternating Currents.

Two methods of studying the hysteresis were adopted, so
that the results of the one might be a check on those of the
other.

First Method.

The first method was a modification of the apparatus used in
the Hopkinson’s bar and yoke method. The bar of iron, fig. 1,

Tig. t.

was made up of eighteen strips of sheet iron, each 40™ long,
1-9" wide and -045°" thick. These were joined at the ends,
and insulated from each other by strips of shellacked paper so
that a current would flow through each successive strip in the
opposite direction. Allowing a small amount for the effect
due to insulation between the iron strips, it was estimated that
the internal magnetic action of a current flowing through the
iron in the manner described above, was approximately one-

Bie. 2:

2 »,
a

+ T aes

Crook—Electromagnetic Alternating Currents. 135

eighth of that due to a current of the same current density
j flowing directly through the iron strips in multiple.
j The yoke, fig. 2, used was one designed by Prof. D. B.
| Brace and already described.*
The advantage of this apparatus over the old form of yoke
is that at all times during the hysteresis test the total induction
_ --_— ean _be measured without changing the susceptibility of the
7 magnetic circuit, and the only change in the induction is that
caused by a change of current in the | magnetizing solenoid.

Diagram I.

rere — =

25 Sane Eeees
Esssseseea

iat a ee auc ected Sees aerae | eee

i eee sea ence ute) a i 8 Hisaee
Ee,
aul _.

ieseeeeeteetece ae
Hate

peuevistasessaviasa
gtaseszaass

BETT6 INSTITUTE OF TECHNOLOGY,

oe

FEE
SEzazongssesusverey ensbsteens eres

The alternating current used for demagnetization was

a obtained from a 50-volt circuit, transformed from a-1000 volt,

b 40 h. p. Westinghouse dynamo. The frequency of the alter-
nating current was 133 periods per second.

The three hysteresis cycles given on Diagram I were all taken

= the same evening and under practically the same conditions.

=. * This Journal, vol. xi, p. 365, May, 1901.

oe ~~

136 = Crook—Llectromagnetic Alternating Currents.

The curves were taken in the order in which they occur on the
plate and were plotted from the readings given in Table I.
Curve 1 is the normal hysteresis cycle.
Curve 2 is a cycle taken with an alternating current density
in the iron to 10 amperes per sq. mm.

TABLE I.
PB:
i. Cycle 1. Cycle 2. Cycle 3.
0) 174 120 1443
lis 846 600 1815
6°3 QAAT 2010 S252
9°45 4329 4AQ45 4890
12°6 5400 _ 0415 © 5706
Lv 6279
18°9 6765 6750 6939
25°2 7500 7440 TOs
31°5 75930 7686 7740
VAS ab 7594 7604 7689
18°9 fpey el 12:75 7320
12°6 6675 6630 6600
6°3 5730 5655 5331
0 3285 3128 2232
6°3 — 2418 — 2550 — 3525
126 —5750 —5750 — 6600
18°9 —6999 —6747 —7011
pA es — 7677 — 7446 — 7560
31°d — 7983 — 7800 — 79206
BH<2 —7794 — 7650 — 7776
18°9 — 7560 — 7360 — 7440
12°6 —7215 — 6086 —6981
6°3 —6204 - —5765 - — 5607
0) — 4050 — 3666 — 2265
6°3 1863 2100 3249

A close inspection of the curves will show that the reduction
of the area of the cycle is directly proportional to the reduction
of the width and also to the reduction of the residual mag-
netism remaining in the iron when the magnetizing force is
reduced to zero. This being the case, the problem can be
considerably simplified by studying the effects of the alterna-
ting current on the residual magnetism from a constant mag-
netizing force, assuming that the effect on the hysteresis cycle
is directly proportional to this effect on the residual mag-
netism.

The curves in Diagram II were taken to show the effect on
the residual magnetism of varying current density in the iron
strips in series. They were plotted from the readings given
in Table IJ. Ordinates represent total induction and abscissae

_—

Crook—Electromagnetic Alternating Currents. 137

represent current density per sq. mm. Curve 1 shows the
effect due to a constant current; curve 2, that due to an alter-
nating current.

Diagrams IT and ITI.

Curve 1 (Diagram II1) shows the reduction of the residual
magnetism due to an alternating current of varying density
passing through the iron strips in series.

138 Crook—Llectromagnetic Alternating Currents.

TABLE II.
Current density Total residual induction.
per sq. mm. For curve 1. For curve 2.
0) 745 745
2 740 712
3 743 684
5 743 665
7°95 740 630
9 710 580
10 695 350
15 580 360
19°2 520 290

Curve 2 (Diagram IIT) shows the corresponding effect due
to an alternating current passing through the iron strips in

multiple.
Readings for these curves are found in Table III.

TABLE III.
Current density Total residual induction.
per sq. mm. For curve 1. For curve 2.
‘0 (22 122
728 719 Bite alae
"56 714 708
84 708 689
1°25 700 670
1°68 690 654
2°2 682 631
2:5 675 618
Second Method.

The second method of studying the alternating current
effect on hysteresis was carried out in the same way as the
first, the only difference being in the apparatus used.

Instead of a number of narrow strips of iron connected in
series, the iron specimen was made from two thin sheets of
iron, separated by shellacked paper and rolled into a compact
tube. The iron sheets were each 56™ long, 50™ wide and
018 thick. Along the edges of the plates corresponding to
the upper end of the tube were soldered “leading-in” con-
ductors, short distances apart. This was to insure a uniform
current density in all parts of the plate. The edges of the
plates at the other end of the tube were bared and soldered
together, so that a current entering the upper edge of one
plate would flow down, across the soldered junction, back on
the other plate and out by the leading-in conductors soldered

to its edge. 3

Crook— Electromagnetic Alternating Currents. 189

The tube was 50™ long and had an inside diameter of 2°5™.
Inside of the tube was a core of wood with a copper wire
running along its cylindrical axis. This system is practically
equivalent to sixteen concentric iron tubes ‘018™ thick, with
their ends so connected that a current of equal current density
flows through each alternate one in the opposite direction or
_ through all of them in multiple, according as the experimenter
desires. The magnetizing coil was of the same length as the
iron cylinder and wound with 300 turns No. 16 copper wire.
The test coil used for measuring the magnetic flux was wound
on a wooden spool made to slip over the magnetizing coil and
sufficiently large and heavy to fall freely and uniformly when
liberated. The cylinder was mounted inside the magnetizing
coil in a vertical position, about five feet from the floor, and
so arranged that the test coil would fall from the center of the
cylinder to about two feet from its lower end. In this way it
was possible to measure the total magnetic flux in the iron.
The other apparatus was the same as that used in the first
method.

It was found that the amount of residual magnetism remain-
ing in the cylinder after applying a strong magnetizing force,
was very small, and all attempts to show the demagnetizing
effects on the hysteresis cycle were unsatisfactory. However,
the reduction of the residual magnetism by the alternating
current was quite satisfactorily shown and the results are as
given below.

The residual magnetism’ left in the iron was stronger when
an alternating current was present during magnetization.
Because of this, the alternating current was thrown off each
time the magnetizing force was applied, and readings were
taken to find the residual induction before again applying the
alternating current. These all gave almost constant results.
The readings were all reduced to the mean value. Readings
taken while the alternating current was flowing and those
taken after it was thrown off were nearly always the same.
However, in curve 2 the values of the residual magnetism
after the alternating current was thrown off were different and
are represented by the dotted portion of the curve.

The alternating current was applied in four different ways
and the effects studied. First, it was passed through the iron
plates in series, so as to get the effect of the magnetically
compensated current. Second, it was passed through the
plates in multiple so as to get the combined effect of the cur-
rent and circular magnetism. Third, it was passed through
the iron plates in multiple and back through the copper rod.
Fourth, it was passed directly through the copper rod. The

140 Crook—Electromagnetic Alternating Currents.

curves, plotted from the values given in Table IV, are given
on Diagram IV and are numbered to correspond to the order
in which they are here mentioned.

Ordinates represent total induction, and abscissae, current
densities per sq. mm. 3

Diagram IV.

MABSACHUSETTS INSTITUTE OF TECHNOLOGY,

Curves similar to those given on Diagram IV were taken
with the frequency of the alternating current reduced one half.
They are given in Table V, and plotted on Diagram V.

i oe ~~ ve

Crook—Electromagnetice Alternating Currents. 141

The effect due to a direct current of varying current density
is shown in Diagram VI. The direct current was applied by
suddenly closing the circuit through the iron. A slow build:
ing up of the current had no effect on the residual magnetism.

Diagram V.

This shows that the effect was due to the sudden impulse of
the current through the copper rod or through the iron.
Readings for these curves are given in Table 6.

AM. Jour. Sc1.—Fourts SeRies, Vou. XIV, No. 80.—Aveust, 1902.
10

142

Current
density

per sq. mm.

O°

ae

~T Or
Or

ie)

H He CO ND DO DS eS et pe
& wo

Or
ie)

Current
density
in iron.

0)
25
9)

, ~yJ Or
Or

Or

Hm HR OO OO DD Nw ee
eb .
—

on
S

Current
densi ty
in iron.

O .
“25
5
“75

af e

1°5

The hysteresis cycles given on Diagram I shows that the
alternating current as magnetically compensated has an effect

Crook—Electromagnetic Alternating Currents.

TABLE IV.
Iron and
Iron in rod in Tron in
multiple. series. Rod alone. series.
F054 1051 1045 1055
962 8628 861 ee
856 660 702 1003
760 492 533 ge
728 286 340 oe
705 Ane oes ao 970
681 “ee Lee hae
re is. 80. 163 aie oe
Crise Pieces Ra S23 935
meee Bees Deere 903
Ee 8s une oat NS ries 870
TABLE V.
Total residual induction.
‘Gua Sos eee | —
Tron in Iron and rod Tron in
multiple. in series. Rod alone. series.
1150 1146 1150 1150
1090 Saba mee pee
1004 1012 1010 TI25
865 and 855 900 908 1095
725 and 695 a ike Pert: ee |
665 and 610 hares eee age ay
ay an =yiaeh 658 686 1045
560 and 465 esos SEOe gues
fe om Lik, ® 400 460 975
fiat aS ed See 945
te neds 179 230 seta
ae se 45 132 870
TABLE VI.
Total residual induction.
(aa SS ay
Tron in Iron and rod Iron in
multiple. in series. Rod alone. series.
1000 1001 1000 1004
996 tpt se rots sees thes
990 940 920 1000
980 Eee ersten Ee aes
ete 875 850 1003
rere s 820 792 10083

on the magnetic induction in iron.

Crook—Electromagnetice Alternating Currents. 148

The eurves on Diagram II show that for low current densi-
ties, the reduction of the residual magnetism in iron, or the
reduction of the area of the hysteresis cycle, is proportioned to

Diagram VI.

HE Wf fF jaa Sneeneeece
So)

eiiuusesaaiaasz

Sssesserscsesszevseseess

Suan eases PeBeaGEBeaCuuaueaes

HH Ppp rity ry
BSE tE teeta

Serene wetiens SRSRERRaS See
aN GEES eRe Se

See Sirti smite

fecal SSeS ei

He ieee Henini

SES LeneeEeel HoH

Sree acetal eras bd nt ertae? ee
SSSSCGSSU GEESE Se rT Ener snes SEaceeSESse’
Seduce ewe detain teen ects

BEE

aisssseesercesseraas

cngaee siti

33.amSg tare 20s, USSR RS eee Sel
Hert SBu8BRawe hee
ror

HIER IHTET Heed iessparttesspossiostioast

as
MASBACHUGETTS INSTITUTE OF TECHNOLOGY,

the strength of the current. Above 7°5 amperes per sq. mm.
this relation does not hold, but heating effects and the effect

144 Crook—EKlectromagnetic Alternating Currents.

of the residual magnetism in the yoke come in to affect the
induction measurements. If the values of curve 1 are sub-
tracted from those of curve 2, the result is nearly a continua-
tion of the straight line which gives the effect at low current
densities.

The similar curves on Diagrams IV and V taken by the
second method, and at different frequencies, show that the
effects of the alternating current on the residual magnetism in
iron are different for different frequencies, that for the lower
frequency being the greater.

Curves 3, Diagrams IV and V, show that the total reducing
effect of the current is approximately proportional to the
strength of the current.

Curve 3, Diagrams IV and V, show the reducing effect of the
external circular magnetic field from a current flowing through
the central copper rod. Curve 4 shows the effect of passing a
current down the central rod and back on the iron.

The force at a point in a metal tube due to a longitudinal
current in the tube is expressed by the formula

SS ere (r°—r,”)

in which F is the force, w the current density per unit area of
the cross section, 7 the distance of the point from the axis of
the tube, and 7, the inside diameter. From this it is seen that
at the inner surface of the tube, where r=7,, the force is zero,

: : ee
while at the outer surface, where 7=7,, the force is a The
. 2 e e
average force in the tube due to a current flowing longitudi-

e . e 7 e e
nally along it is proportional to — where 7, is the mean radius
Tv
3

of all points in the tube.

The magnetic force at a point outside a linear conductor is
expressed by

| 24

F = —
‘:

where 2 is the total current and 7 is the distance of the point
from the axis of the conductor. It is, therefore, evident that
the average force in the iron due to a current flowing along its

e e e e e 2 e e
cylindrical axis is proportional to = ithe magnetic force in
ais

the iron when the system was connected for taking curves 4,
Diagrams IV and V, was therefore one-half what it was when it
was connected for curve 3, and if the reduction of the residual
magnetism was due to the circular magnetic flux alone, we

Crook—Electromagnetic Alternating Currents. 145

should expect the values of any point on curve 4 to be one-
half that of a similar point on curve 3. Any added effect
caused by the presence of the electric flow in the iron would
tend to make curve 3 approach curve 4, and in fact it is found
that the latter falls below the former.

If the apparatus used in the second method was as effectual
in reducing the circular magnetic effects as was expected,
eurve 1, Diagram IV, can be taken as giving an approximate
measure of the reduction of the residual magnetism by an alter-
nating current flux, with a frequency of 133 alternations per
second; and curve 1, Diagram V, a similar measure when the
frequency is at 66 alternations per second. The effect of the
magnetically compensated current as given by these curves
is about one-fourth of the total effect of the current and
circular magnetic effect together.

Other Methods.

The following experiments were made to study the effects
of electric oscillations on magnetized plates (fig. 3 A) placed
in a magnetizing solenoid of very thin sheet iron (trunk cover-
ing) cut into strips 7 inches X3/4 inch and separated by sheets
of mica made to project past the edges of the iron an eighth of
an inch. This system was connected to a Holtz machine and

Fig. 3.

the spark gap adjusted so as to get a rapidly oscillating dis-
charge. The magnet, after being magnetized to a maximum
amount of residual magnetism, was placed carefully in the
field of a sensitive magnetometer. This magnetometer was
made of an astatic magnetic system suspended by a silk fiber
in a glass tube. The magnet was so placed that the mag-
netometer was deflected a few degrees from the meridian.
The system was carefully tested to prove the effect produced
was not due to conditions external to the iron. Any disturb-
ance in the magnetic field of force thus produced could be

146 Crook—FLlectromagnetic Alternating Currents.

readily detected with a telescope and scale. It was found that
an oscillation of the static charge in the “ magnetic condenser ”
produced a demagnetizing effect which could be readily
detected with the magnetometer. The effects were lacking
when a thin foil condenser was substituted for the iron one.
The residual magnetism of the laminated electro-magnet
was also found to be affected by the oscillating electric charges,
when connected in series, fig. 3 B, with a condenser to the
terminals of the static machine. The amount of this effect
was approximately equal to the effect when the magnet was
used as a condenser. |

Fig. 4.

Another test, fig. 4, was made by placing a pile of these
japanned iron sheets between terminals of tinfoil which were
insulated from the iron by mica. This proved to be just as
effective in producing results as the other method. The
frequency of oscillation was an indeterminate factor and did
not seem to affect the results, as the demagnetizing effect was
apparently produced during the first oscillations. A single
spark between the terminals of the machine, in most cases,
proved to be as effective as a continuation of the oscillatory
discharges.

Nore.—This work was done at the University of Nebraska under the imme-
diate direction of Dr. D. B. Brace, and much of the success of the investiga-

tion was due to suggestions offered by Dr. Brace and by Professor B. E.
Moore.

Physical Laboratory, University of Nebraska,
Lincoln, Nebr.

Coleman—Nepheline and other Syenites in Ontario. 147

Art. XVILI.—Wepheline and other Syenites near Port Cold-
well, Ontario; by A. P. COLEMAN.

A FEW years ago some very interesting dikes of a rock con-
taining analcite named “heronite” or analcite tinguaite were
described by the writer from the north shore of Lake Superior
between Heron Bay and Peninsula on the Canadian Pacific
railway;* and the opinion was expressed that nepheline
rocks should be found connected with them somewhere in the
region. Dr. Adams also has suggested the same idea, basing
his belief on some rock specimens from the region of Peninsula
in the Geological Survey collection. They are augite syenites
of an unusual kind often associated with nepheline : syenites.t

In connection with an excursion to the iron range of the
Slate Islands an opportunity was taken to examine the railway
and shore near Port Coldwell, and it was intended to visit Pic
Island a few miles off shore, where Professor Pirsson and others
have suggested that nepheline rocks would probably be found,
but unfortunately no suitable boat could be got at that little
harbor, and this had to be given up.

No syenites of any kind were found between Heron Bay
and a point three miles east of Peninsula, where augite syenite
had been obtained a few years before; but west of Peninsula,
more than half way to Port Coldwell, considerable stretches of
nepheline syenite were discovered. So far as the study of the
specimens has gone one can say that a great area of syenites
and associated rocks rich in alkalies, derivatives of a magma
differentiated into a whole series of related species, like those
so elaborately described by Broégger in the Christiania region
of Norway, occurs in this region.

The first rock of the group, going west, is the dark. augite
syenite which commences three miles east of Peninsula, and

with some interruptions of red syenite and more basic rocks

extends to a long trestle at mile 818, a distance of nine or ten
miles, with an unknown width. From the trestle west to a
cutting beyond Peninsula, the prevalent rock is a gray or
purplish gray nepheline syenite, having an extent of about four
miles. It is probable that detailed mapping of this little

_ explored region would show large areas of this syenitic group

of rocks, including Pic Island, and it is hoped that in the
future these interesting eruptives may be studied more at
length.

The only previous references to this group of syenites are to

* Bureau Mines, Ontario, 1899, pp. ee ae and 1900, pp. 186-191.
+ Jour. Geol., vol. viii, No. 4, PP. 322

148 Coleman—Nepheline and other Syenites in Ontario.

be found in the reports of the Geological Survey of Canada,*
where rocks containing red and white feldspar, some grains of
orange-red elaeolite, and a few zircons, are said to occur in
Pic Island and the mainland to the north; and of the Bureau
of Mines,t where the occurrence of augite syenite and other
associated rocks is referred to, though the nepheline rocks
were overlooked. Acknowledgments must be made to Pro-
fessor L. V. Pirsson and Dr. H. 8. Washington for having been
good enough to send chips and larger specimens of various
nepheline syenites and related rocks from other localities
which have proved most useful for comparison; and to Profes-
sor Pirsson for suggestions as to rock relationships.

Nepheline Syenites.

The syenites and associated rocks are very well exposed
between miles 818 and 822 in the numerous rock cuts and
cilffs where the railway winds along the rugged shore of Lake
Superior, so that an almost continuous section is presented.
The first nepheline syenite observed is just east of the long
trestle at Red Sucker lake, where it forms irregular dikes and
larger masses in gabbro, which appears to be the older rock
of the two; and similar relations are found at the rock cut
west of the trestle, though red syenite interrupts it at the third
cutting. Beyond this, toward the west, hills of nepheline
syenite rise 200 or 300 feet above the lake and continue with
few interruptions to Port Coldwell station and the next cutting
beyond it, the last point where it was observed being a little
beyond mile 822. The second cutting west of the station is in
red syenite. The rock varies from almost compact to very
coarse-grained kinds having crystals an inch or more long;
and in color from pale to dark gray, sometimes running into
purplish tones or having brillant red spots. The black horn-
blende and augite crystals stand out sharply giving a fresh look
to the rock, which unfortunately is not borne out in thin sec-
tions. In some specimens the hornblende crystals are long
slender prisms, but in others they are short and _ stout.
The different textures are often mixed intimately, fine-grained
parts enclosing coarser grained ones or the opposite; and
large or small blocks of the gabbro mentioned above are
enclosed in the nepheline syenite. Dikes of a fine-grained pur- .
plish gray rock, sometimes with the look of an amygdaloid, eut
the syenite; and last of all, there are sharply defined dikes of
black diabase.

In general appearance the nepheline syenites are very dif-
ferent from those of Eastern Ontario, never showing the

* 1846-7 ; also 1863, p. 80. +1897, p. 147.

Coleman—Nepheline and other Syenites in Ontario. 149

eneissoid structure so common there, nor having dikes of peg-
matitic rock consisting of large individuals of nepheline and
muscovite. Nor are they like specimens from Barkevig,
. Norway, nor Litchfield, Maine, but some of them have much
the appearance of specimens sent by Professor Pirsson from
Highwood Mts., Montana, and Red Hill, Moultonborough,
New Hampshire, having a tendency to flat plate-like forms of
the feldspars, and long prisms of the darker minerals. These
would apparently be classed by Professor Brégger as foyaites,
though unlike a specimen of foyaite from Lougenthal, Norway,
sent me by Dr. Washington.

In the considerable number of specimens collected near Port
Coldwell four fairly well marked types may be distinguished,
so far as megascopic structure is concerned :

1. Medium to coarse-grained gray rocks having a dioritic
appearance with light and dark minerals in about equal
amounts and the grains isodiametric.

2. Medium-grained, reddish, purplish or violet-gray rocks,
with about twice as much of the lighter colored minerals as of
the darker ones, and with a tendency to plate-like or elongated
forms in the minerals.

3. Violet-gray, fine-grained rocks with porphyritic feld-
spars and other minerals.

4, Narrow veins of coarse-grained rocks, often mottled
with red, gray and black.

There are, however, many intermediate varieties between
the four here mentioned, illustrating the usual variability of

. the nepheline rocks.

1. The first variety occurs as fresh looking material about
two miles east of Port Coldwell, and was supposed to be diorite
when collected. The white minerals are nepheline, orthoclase,
and a less amount of finely striated plagioclase having a very
small extinction angle, probably oligoclase ; all badly weathered
and turbid, the nepheline sometimes changed into a brown-
ish substance having aggregate polarization, perhaps a zeolite.
The dark minerals include hornblende in fairly well formed
crystals having a pleochroism of dark green, brownish green and
brown, and an extinction angle of 23° ; and also augite in about
equal ‘amounts, sometimes enclosed in the hornblende. The
augite has a slight pleochroism, sea-green, gray-green and
brownish green, but its extinction angle is normal and the
nearly rectangular cleavages in cross sections show that it is
really a pyroxene. The only accessory minerals noticed are
magnetite and apatite, both in considerable quantities.

A very similar but duller rock occurs in the first railway cut
west of Port Coldwell, with the difference, as seen under the
microscope, that the hornblende is dark brown and the augite

150 Coleman— Nepheline and other Syenites in Ontario.

grayer and not pleochroic. One or two large masses of
magnetite and serpentine probably represent olivines com-
pletely decomposed.

2. The gray or purplish gray variety, with relatively small .
amounts of the dark ingredients, contains all the minerals
mentioned as belonging to No. 1, with the exception of the
probable olivine; but the ferro-magnesian minerals are, of
course, less in amount, and occasionally a little brown biotite
occurs, in addition to the hornblende and augite. The red
color of spots in the rock is due to infiltration of iron oxide in
portions of nepheline completely changed to zeolites, and the
usual reddish or purplish tone of the rock is due to the
general diffusion of the same oxide. As distinguished from
the previous variety this one is leucocratic. In some examples
the minerals have plate-like or long prismatic forms with a
suggestion of the trachytic structure. In one section the augite
is almost entirely replaced by hornblende, often dark brown in
the middle and green at the edge with very deep colors but
not strongly pleochroic, perhaps barkevitic in character.

3. The porphyritic varieties of the nepheline syenite occur
partly a mile or two east of Port Coldwell and partly to the
south of the station near the harbor. Specimens from the
former locality are dark bluish gray, fine-grained, with por-
phyritic feldspar, nepheline (rarely) and hornblende crystals.

One thin section from mile 819 shows very small crystals of
nepheline having the prism and basal planes, enclosed in ortho-
clase and possibly oligoclase, as well as in hornblende, the
latter mineral forming sieve-like structures, the holes being .
filled with lighter minerals, a good example of poecilitic inter-
growths. The other minerals are augite, magnetite and apatite.
A second specimen shows less of the poecilitic intergrowths
but contains one or two long porphyritic prisms of nepheline.

Porphyritic examples from south of the station have a
purplish gray ground in which bluish erystals of feldspar and
black crystals of biotite are embedded. The groundmass does
not differ much from the former rock, but the numerous
phenocrysts are orthoclase, oligoclose and brown biotite, having
strong dichroism.

4, The fourth variety forms narrow pegmatitic veins in the
other varieties, and consists of the same minerals but of larger
dimensions, sometimes more than an inch in length, though
never rivaling the giant nepheline pegmatites of Hastern
Ontario, as described by Dr. Adams, with crystals more than
a foot long. The nepheline in the Port Coldwell specimens is
often changed to a turbid orange-red material, mentioned by
Sir William Logan as elaeolite,* the feldspars (orthoelase) are

* Geol. Can., 1868, p. 81.

Coleman—Nepheline and other Syenites in Ontario. 151

pale gray, and the hornblende prisms black, making a very
showy rock. In spite of the striking differences in appearance
of the varieties mentioned above, the range of minerals found
in the thin sections examined is not great, much less, for
instance, than in the nepheline rocks of Dungannon and York
branch in Eastern Ontario,* and none of the rarer minerals
have been found by myself, though zircon is mentioned in the
1863 report. The absence of muscovite, scapolite, sodalite,
and of the usual microcline and microperthite is peculiar ;
though in some cases weathering has gone so far as perhaps to
obscure the structures of the feldspars.

Augite Syenites.

The other important group of rocks in the region includes
the augite syenites, which occur in two well marked varieties,
one dark brownish gray to black in color, coarse-grained and
with more or less of a plate-like character in the feldspars ; the
other red or reddish gray, finer grained and usually granitic
in texture. The first variety is much the more extensive of
the two and will be described first.

In the dark variety, which is no doubt the trap that Logan
reports from the region, the feldspars are the prominent
ingredient, forming broad plates or narrow shining strips,
often carlsbad twins, attracting the eye in the sun; while the
relatively small amounts of ferro-magnesian minerals escape
notice. While dark brownish gray to black is the prevalent
color, there are phases of a dull brown or a dull red; and
weathered glaciated surfaces may even be white by the bleach-
ing of the feldspar, when the augite and magnetite show as
angular or black filling material between the feldspar crystals,
which tend to be idiomorphice.

The syenite is always coarse-grained, the crystals averaging
about a quarter of an inch in length, and also in breadth when
seen broadside, but often only a tenth of an inch in cross sec-
tion. There are coarse pegmatitic veins in the finer grained
rock having individuals of feldspar an inch or two in diameter,
and often fairly well built out in occasional cavities.

As this rock has been quarried by the railway for bridge
construction, etc., it is easy to get fresh material.

Thin sections consist of feldspar in more or less idiomorphic
forms with augite wedged in between, resembling, so far as one
ean tell from a description, Brégger’s laurvikite.t The
feldspars show no twin striations but have partly the appear-
ance of microperthite and partly of microcline. They are

* See Bureau Mines, 1899, Corundum and Other Minerals, p. 205, etc.; and

Corundiferous Nepheline Syenite, p. 250, etc.
+ Zeitschrift fir Kryst. u. Min., Band 16, 1890, pp. 29, 30.

152 Coleman—Nepheline and other Syenites in Ontario.

fairly fresh and in some directions have a handsome bluish
shimmer. No nepheline nor sodalite nor quartz was observed.
The dark minerals are chiefly augite with brown interior and
dark green exterior, but some dark green and brown horn-
blende, and some magnetite occur also, as well as apatite.

An analysis of this rock made by Mr. A. H. A. Robinson of
the Chemical Department of the School of Practical Science,
Toronto University, gives the following results in column I:

ip II. TUL,

s) LO pee eseie i 2 58°81 58°88 °980
AOL Ge a 13°37 20°30 131
FeQ.0.6) ene 3°63 024
ReQ 2. See ee soy is ee "097
MnO 2 oe ad eee 0°20 ae |
Wee 22 acc vslnis eee 0°51 0-79 012
Oe Oe VRE Lite yey 3°89 3°03 069
NG AD NOs oT Sb eee 4°96 5°73 080

HO Aikc2 ee Ee 5°49 4°50 056
HOt D008 eee 0°29 1:01
H,O above 100° C. ..-. 0°75
MSOs cb sok Ne ok e te O75
iG) ai. 2 vides. Pegientel 0°31 0°54

Total 2 -isses opeet AVOOSOG 100°99

Specific gravity at 17°°5 C., 2°75.

For comparison an analysis of Norwegian laurvikite from
Byskoven, Laurvik, by A. Merian is given in column IIL.* The
two analyses agree fairly closely except for the relative pro-
portions of alumina and of ferrous iron oxide, which differ
greatly. The analysis shows the rock consists largely of micro-
cline and microperthite, the small amount of magnesia and
large amount of ferrous oxide that the augite must consist
mostly of the hedenbergite molecule. The amount of alumina,
though smaller than in the laurvikite, is that which is required
by the soda and potash to turn them into feldspars and is
therefore correct, as shown in the molecular ratios given in
column III. .

Associated with the dark augite syenites with plate-like feld-
spars are numerous other varieties in much smaller amounts,
some merely having the red color of the ordinary syenite, due
to diffused hematite particles, but with the same ingredients
and the same shape of the feldspars; others differing more
widely in appearance and composition, but all more weathered
and less satisfactory for study with the microscope. It will be
sufficient to refer to the kinds having granitic structures,

* Thid., p. 30.

Coleman—Nepheline and other Syenites in Ontario. 158

grains with equal diameters. These are on the average finer
grained than the Laurvikitic syenite, and may be divided into
leucocratic red syenites with comparatively little of the ferro-
magnesian minerals; and melanocratic varieties containing
more than half dark minerals. It must be admitted that the
term leucocratic syenites is not happy for the less basic varie-
ties, since they are strong red and not white or pale colored.
Several specimens from east of Port Coldwell are in reality
quartz syenite consisting of feldspar pegmatitically intergrown
with quartz, and small quantities of hornblende, augite, mag-
netite and apatite. The feldspars, which tend to be por-
phyritic, so far as their weathered condition permits one to
decide, are orthoclase, microcline and oligoclase. These rocks
seem to have the same composition as Brégger’s nordmarkites,
though no mention is made of pegmatitic intergrowths in the
latter rocks.* Aegirite has not been recognised in the rocks
from Port Coldwell, another point of distinction.

The melanocratic syenites, consisting to the extent of at least
half of dark colored minerals, are dark gray rocks, usually with
a red tinge, not very coarse-grained, with about equal diam-
eters to the grains. The light colored minerals are ortho-
clase, some plagioclase and occasionally nepheline; the dark
ones hornblende, pale blue-green augite and brown biotite in
not far from equal amounts: while magnetite and apatite are
always present, the latter often as numerous large prisms.
They appear by analogy to be related to the essexites.

Plagioclase Rocks.

Three kinds of plagioclase rocks accompany the syenites of the
Port Coldwell region, coarse-grained gabbro-like rocks older
than the syenite and penetrated or carried off as blocks by the
nepheline syenite; fine-grained gray-brown rocks occurring as
dikes without well defined edges, in the nepheline syenite ; and
green-black dikes of diabase or diabase porphyrite, which are
latest of all,

The gabbro is a speckled gray, coarse-textured rock showing
plates of mica and often a few porphyritic plates of plagioclase
megascopically. Thin sections are made up of half or less
than half of a plagioclase having the extinction angle of
andesine or sometimes labradorite; of pale bluish green
augite, often idiomorphic, and brown biotite in about equal
amounts, while brown hornblende and olivine are in smaller
quantities. Magnetite and apatite in thick prisms are the chief
accessories; and serpentine, chlorite and iron oxides occur
as secondary products. In one section the biotite surrounding

*Ibid., p. 5d.

154 Coleman—Nepheline and other Syenites in Ontario.

a weathered olivine crystal is modified so that the parts near-
est the olivine are more strongly dichroic than the rest, bright
green and orange-brown in the two directions.

The fine-grained dikes of brownish plagioclase rock are not
very sharply defined as a rule, perhaps because they were
erupted before the mass as a whole had completely cooled
down. In many cases these rocks are specked with white or
pale flesh-colored spots as if amygdaloidal, and they often con-
tain what seem to be fragments of an older, fine-grained reddish
rock.

The general mass of these rocks consists of some greatly
weathered, lath-shaped plagioclase, partly with a radiating
arrangement, enclosing biotite, augite, and magnetite in larger
amounts. In this groundmass are often well-formed crystals
of augite, sometimes in groups, pale green or brown, somewhat.
dichroic and with a zonal structure; and of dark brown horn-
blende. The lighter patches, suggesting amygdaloids, are com-
posed chiefly of plagioclase, often with well-shaped prisms
projecting inwards, the center being of some transparent
mineral having low double refraction, perhaps a zeolite.
There are a few prisms with parallel extinction, probably
nepheline, though so badly weathered as to leave their char-
acter uncertain. Without an analysis it would be difficult to
place this rock with certainty, so for the present it will be left
unnamed.

The dikes of dark gray or black diabase and diabase porphy-
rite have been little studied. The only one of which thin
sections have been made is an olivine diabase with compara-
tively little angite, often in slender fibers or prisms having the
usual extinction angle, but sometimes as broader portions
between feldspar laths. The magnetite, too, has elongated rod-
like forms, and when the numerous needles of apatite are
included, there is evident a tendency to elongation in almost
all the constituents of the rock. The large well-formed crys-
tals of olivine, still fairly fresh, are, however, an exception to
the rule just mentioned. These quite fresh rocks are probably
of Keweenawan age like most of the diabase dikes on the north
shore of Lake Superior, while the other eruptives described
appear to be older, though not so old as the Huronian schists
_which they penetrate.

It is believed that with the possible exception of the gabbros,
which may be older than the syenites, and the diabases which
are distinctly younger, all the rocks which have been referred
to belong to the same magma and represent phases of differ-
entiation. The dikes of heronite or analcite tinguaite, though
found several miles to the east of the nearest syenite, are to be
looked on, no doubt, as split off from the large mass described.

Coleman—Nepheline and other Syenites in Ontario. 155

The older nepheline or elaeolite rocks and their associates

ean no longer be considered rare. In the province of Ontario

they are known to occur very widely spread in Dungannon
and adjoining townships, where they were first noticed by Dr.
Adams, and where they have been followed up for many miles
by Professor Miller because of their connection with the
corundum-bearing band of the Laurentian. The series of
eruptives described in this paper form another large mass of
nepheline syenites and related rocks, though of a very different
type; and the malignites described by Dr. Lawson from
Poohbah Lake, west of Lake Superior, make a third, each with
ite own peculiarities differing markedly from the others. The
nepheline rocks of Montreal make another Canadian locality,
though on a small scale, and with their alndite dikes, as
described by Dr. Adams, present still another type; the whole
illustrating strikingly the great variability of this group of
pintonic and dike rocks as contrasted with most others.

To refer to the areas described in the United States by
Pirsson, Washington, Osann and others, would lead too far;
and a mere list of the localities in Europe, India, South
America, etc., would require considerable space.

In coneluding this notice of the Port Coldwell and Penin-
sula syenitie rocks, it should be mentioned that a number of
them are handsome ornamental stones, as proved by polished
specimens prepared by the Bureau of Mines for the Buffalo
Exposition, where they attracted considerable attention. The
dark gray augite syenite with its gleams of blue from the
feldspars is a particularly fine stone, resembling the famous
Norwegian syenite though on the whole finer in grain. As
it can be obtained close to the Canadian Pacific railway and
beside an excellent harbor on Lake Superior, in quarries afford-
ing blocks of almost any required dimensions, it should prove
of importance in the future.

Toronto University, Toronto, Ontario, Canada.

156 Austin—Double Ammonium Phosphates in Analysis.

Art. XIX.—The Double Ammonium Phosphates in Analy-
sis; by Marta AUSTIN.
[Contributions from the Kent Chemical Laboratory of Yale University—CIX.]

THE function of ammonium salts in the formation of double
ammonium phosphates for the purposes of analysis has been
the subject of a good deal of study. Manganese, which tends
to fall as the tribasic phosphate, Mn,P,O,,* appears in the
presence of a sufficiency of ammonium salt in the form of the
ammonium manganese phosphate (NH,)MnPO,, which gives
the pyrophosphate, Mn,P,O,, on ignition. Magnesium,t on
the other hand, tends to pass under the influence of ammonium
salt toward the condition of the diammonium magnesium phos-
phate which yields the metaphosphate on ignition. Further,
in the cases of zine and cadmiumtf it has been my experience
that the presence of a considerable amount of ammonium salt
is essential to the precipitation of the ideal double ammonium
phosphates of both these elements, although in the case of cad-
mium, too large an amount of the ammonium salt prevents
complete precipitation and apparently occasions the formation
of a phosphate too rich in ammonia. The solvent effect of the
reagents in the case of mercury phosphatet is very great, ©
yielding a new salt the composition of which has not been
investigated in recent years. Beryllium§ is only partly con-
verted to that ammonium beryllium phosphate which yields
the pyrophosphate on ignition, even in the presence of large
amounts of ammonia salt. Neither an ammonium salt nor
ammonia in solution converts the phosphates of barinm, stron-
tium, and calcium to the double ammonium phosphates. Of
these three elements barium alone falls in the form of the
hydrogen barium orthophosphate. In certain recent articles
upon the precipitation of the double ammonium phosphates
some of the facts mentioned above have been called in question.

In the determination of zinc and manganese Dakin|| has pro-
posed to substitute ammonium phosphate as the precipitant in
place of microcosmic salt (hydrogen ammonium sodium phos-
_ phate) in presence of a considerable amount of ammonium
chloride, and to wash with a one per cent solution of the pre-
cipitant followed by alcohol. Dakin’s analytical results, taken
without scrutiny, would seem to show that the precipitate pro-

* Gooch and Austin, this Journal, vi, 233.

+ Gooch and Austin, this Journal, vii, 187. Neubauer, Zeitschr. anorg.
Chem., ii, 45-50; iv, 251-266; x, 60-65; Zeitschr. angew. Chem., 1896,
435-440 ; Jour. Am. Chem. Soc., xvi, 289.

t Austin, this Journal, viii, 206.

S$ Roessler, Zeitschr. anal. Chem., xvii, 148.

| Chem, News, Ixxxii, 101; Ixxxiii, 37.

Austin— Double Ammonium Phosphates in Analysis. 157

duced by ammonium phosphate (in only moderate excess) and
washed with a one per cent solution of the precipitant and alco-
hol has the ideal constitution of the double phosphate and
leaves upon ignition the pyrophosphate.

From the standpoint of theory it is difficult to see why the
ammonium phosphate should be more effective in producing
the normal ammonium double salt than is hydrogen ammonium
sodium phosphate in amount sufficient to supply the ammo-
nium equivalent, or than a mixture of the equivalent amount
of any soluble phosphate in association with énough ammonium
salt to supply the same amount of the ammonium radical or
ion. Whether the conversion of the tribasic phosphate of the

type R a, O, to the double ammonium phosphate of the type

(NH )RPO, be viewed either as ‘dependent upon the inter-
action according to mass of reagents assembled, or from the
point of view of the ion theory, the action of a solution of
ammonium phosphate should not differ much from that of
equivalent amounts of other ammonium salt and soluble phos-
phate. My own experience with the double ammonium zinc
phosphate confirms this view.t In the following table are
reproduced from my former paper the data of certain experi-
ments bearing upon this point.

Zn2P20;,
corre- Error in Error Zine Time
sponding terms interms left of
to ZnSQOx,. of of in the stand-
Taken. Found. Zn.P.0;. Zinc. filtrate. (NH,);PO,. NH,Cl. ing.
erm. grim. grm. germ. erm, erm. grm. hours.
A
(1) 0°6355 0°6206 0°0149— 0°0060— __ trace 3°13 x. pe 14
(2) 0°63855 0°6254 0:0101— 00040-— trace 3°13 bg es 16
(3) 0°6355 0°6800 0°0055— 0:0022— __ trace 3°13 es ae 16
B
HNaNH,PO,
.4H.0
erm.

(4) 0°68355 0°6271 0:0084— 0:0034— ~— trace 4°47 0°5 1
(5) 0'6355 0°6256 0°0099— 0:0040— none 4°47 0°5 20
C

(9) 0°6855 0°6335 0:0020— 0:0008— none 4°47 10 16
(10) 0°6855 0°6381 0°-0026+ 0:0010+ none 4:47 20 4
(11) 0°6355 0°6379 0°:0024+ 0:0009+ none 4:47 20 2
(12) 0°6855 0°6886 0°0031+ 0:0012+ none 4:47 20 4
(18) 0°6355 0°63893 0:00388+ 0:0014+ none 4°47 20 $
(14) 0°6367 0°6355 0-0012+ 0°0005+ none 4°47 30 16

*Gooch and Austin, this Journal, vi, 233; Boettger, Ber. Chem. Gesell.,

Xxxiii, 1019.

+ This Journal, viii, 210.

Am. Jour. Sci.—FourtH SERIES, Vou. XIV,

ll

No. 80.—Avu6usT, 1902.

158 Austin— Double Ammonium Phosphates in Analysis.

The results of the experiments of Section B, in which the
precipitation was made by microcosmic salt in presence of a
small amount of ammonium chloride, show an average consti-
tution of the precipitate similar to that of the precipitate
obtained in A by an equivalent amount of ammonium phos-
phate, but the results of both series, A and B, vary considerably
from the ideal much more closely approximated in the experi-
ments of O, in which precipitation was made in the presence
of a large excess of ammonium chloride. With these results
those of Dakin are not in accord. Upon examination, how-
ever, Dakin’s procedure appears to be open to criticism.

In the first place, it appears that the asbestos employed by
Dakin was of the hydrous (serpentine) variety, which disinte-
grates when heated and is readily acted on by many reagents.
Though previously treated with hydrochloric acid, it was, by
Dakin’s account, perceptibly soluble in a solution of ammonium
phosphate. After drying at 100°-105° the crucible and felt
lost some decimilligrams on ignition. It was on this account
that Dakin weighed precipitate and filter, after drying at 100°—
105°, when the double phosphate was to be estimated as such,
or after ignition when the zinc pyrophosphate was weighed,
and then, dissolving the precipitate in nitric acid, reweighed
the crucible and asbestos, and by difference from the former
weight obtained the apparent weight of the precipitate. The
inevitable loss through disintegration or solubility of the filter
in the nitric acid used to dissolve the precipitate in Dakin’s
procedure, must raise the corresponding apparent weight of
the precipitate itself. Any series of results based upon the ©
use of such material must of necessity be imperfect to the
extent to which the ignited filter disintegrates or dissolves
under the action of the nitric acid used to dissolve the precipi-
tate.

Anhydrous (amphibole) asbestos, the material of which I
made use, is insoluble, under the conditions of analytical work,
in ordinary reagents including even the strong acids, as has
been abundantly shown.* It is likewise completely insoluble
in ammonium phosphate under the conditions of the work
described, as I have found by experiment. Herein lies one
cause of difference between Dakin’s results and mine.

A second source of error in Dakin’s procedure is found
in the fact that precipitates of the double ammonium phos-
phates are washed in a one per cent solution of the reagent,
ammonium phosphate followed by “redistilled alcohol.” Dakin
calls attention to the fact that absolute alcohol must not be

* Gooch, Am. Chem. Jour. i, 317. Mar, this Journal, xii, 288; xliii, 521.
Browning, this Journal, xliv, 399. Phinney, this Journal, xlv, 468.

Austin—Double Ammonium Phosphates in Analysis. 159

used, but makes no mention of the exact strength of the
alcohol employed in his work.

In order to test the insolubility of ammonium phosphate in
alcoholic washing [ have made various experiments, the details
of some of which are given in the following account. Portions
of a solution of manganous chloride, standardized as the sul-
phate according to the method* described in a former paper,
were carefully measured from a burette, ammonium chloride
and microcosmic salt were added in the proportions formerly
recommended, and precipitation of the ammonium manganese
phosphate was completed exactly as directed. After cooling,
the precipitates were collected on asbestos under pressure in a
perforated platinum crucible. In the experiments the results
of which are given under A of the table, the precipitate was
in each case washed with distilled water made faintly ammoni-
acal, while in those recorded under B, C, and D of the table
the precipitate was washed with two hundred cubic centi-
meters of a one per cent ammonium phosphate solution, and
rinsed with forty cubic centimeters of alcohol of different
strengths,—sixty, eighty and eighty-eight per cent,—applied
in successive portions.

Mn.P.0, Washed with Washed with Washed with
correspond- Washed with 1%H(NH,)2PO, 1%H(NH,)2PO, 1%H(NH,),PO,
ing to ammoniacal rinsed with rinsed with rinsed with
MnCl, water. 60% alcohol. 80% alcohol. 88% alcohol.

taken. A B C D
germ. erm. erm. erm. erm.
0°3020 0°3042 0°3050 0°3058 0°3066
0°3020 0°3041 0°3074 0°3086

From these experiments it appears that the procedure
involving the washing of a precipitate of ammonium manga-
nese phosphate by a one per cent solution of ammonium phos-
phate, draining with the pump, and then completing the
washing with alcohol, results in a contamination of the pre-
cipitate proportionate to the strength of the alcohol used.
Even the thin felt of asbestos alone when washed with
ammonium phosphate followed by alcohol retains amounts of

Weight of asbestos
felt after washing
ten times with a
1Z solution of am- Weight of asbestos felt after washing

monium phosphate with a 1% solution of ammonium
Weight of and rinsing with phosphate and rinsing with alcohol
asbestos felt. distilled water. of the strength given.
60% 80% 88%
germ. grm. erm. erm. erm.

0°0681 0°0681 0°0682 0°0684 0°0686
* This Journal, v, 209.

160 <Austin—Double Ammonium Phosphates in Analysis.

that salt proportionate to the strength of the alcohol employed.
This is shown in the accompanying statement.

So it appears that in washing with ammonium phosphate
and alcohol, according to Dakin’s procedure, the precipitate
of the double ammonium phosphate must be loaded with
foreign material to a degree dependent upon the conditions.
This error is in the same direction as that introduced by the
solubility of the serpentine in nitric acid. With these two
sources of error active, the most remarkable thing about
Dakin’s results is that they happen to agree so closely with
mine, the solubility of the serpentine asbestos and the insolu-
bility of the ammonium phosphate in alcohol together counter-
balancing the deficiency due to imperfect constitution of the
double ammonium phosphate.

As to the experience of Miller and Page* upon the ammo-
nium cadmium phosphate, in consequence of which they
concluded not only that my method for the estimation of
cadmium is unsatisfactory, but that “asbestos filters should
be avoided on account of the solvent action of either ammo-
nium phosphate or nitric acid,” it is sufficient to direct atten-
tion to the fact that their analytical examination of this process
is wholly worthless by reason of their use of serpentine asbestos.

With suitable anhydrous asbestos at disposal, the estimation
of magnesium, manganese, cadmium and zine precipitated as
the double ammonium phosphates, according to the directions
and with the precautions laid down in the former papers from
this laboratory to which reference has been made, is perfectly
practicable. Concerning the manganese and zine phosphates,
in particular, it may be reiterated that with a sufficiency of
ammonium salt, preferably the chloride, and with an excess of
the precipitant present, the precipitation is complete and the
phosphate formed of nearly ideal constitution. In the pre-
cipitation of the ammonium cadmium phosphate the propor-
tions of the reagents must be regulated with care. The use of
ammonium phosphate followed by rinsing in alcohol is not
only wholly unnecessary but always altogether undesirable.

* School of Mines Quarterly, xxii, 391.

Rood— Electrical resistance of Glass, Quartz, etc. 161

ArT. XX.—On the Electrical Resistance of Glass, Quartz,
Mica, Ebonite and Gutta-percha ; by Oapren N. Roop,
Professor of Physics in Columbia University.

In the August number of this Journal for 1901,* I pub-
lished an account of an electrometer intended for measuring
very high electrical resistances, and at the same time described
a set of units of resistance that had been made with its help,
the highest of which reached to 14,000,000 of megohms. Since
then, with a similar set of high resistances, measurements of
the external and internal resistances of certain insulating sub-
stances have been undertaken, which may not be without
interest, in spite of the fact that they are to be regarded as first
approximations. The reason of this limitation is to be found,
not in any especial uncertainty in the resistance of the units
employed, but in another direction, viz: metallic connection
with substance under experimert is necessary, but mere mechan-
ical contact is not sufficient to abolish contact-resistance, which
is apt to be quite large, and various expedients were adopted
to remove this troublesome difficulty. The best of these
appeared to.be the application of an amalgam of tin and mer-
cury, of the consistence of soft butter, to the parts in ques-
tion, but others were also employed; as, for example, in the
case of gutta-percha, it seemed fair to cement the tinfoil to it
with a liquid solution consisting mainly of gutta-percha, while
in the case of ebonite, a metallic powder was attached to its
surface by a very weak solution of shellac, and this again over-
laid by tinfoil. Another source of error lies in the ¢¢me of
exposure to the electrical current, as the resistance of the sub-
stance through which it flows is apt to increase with the time.
The initial resistance, or more properly the average resistance
during the first stroke of the electrometer, was in each case
noted, and an effort was made to adjust matters so that this
should be about one minute, but very often it differed con-
siderably from this amount of time, a drawback which could
only be remedied by a large amount of extra labor and expen-
diture of time. It is owing then to these two factors that the
measurements given below are to be regarded as first approxi-
mations. On the other hand, no difficulty was encountered in
causing the current to pass through glass, quartz, mica, gutta-
percha and ebonite, or to flow over their surfaces.

Mode of expervmenting.—Mr. H. C. Parker determined for
me the resistance of five small units in the ordinary way, and

* This Journal, vol. xii, p. 91.

162 ftood—Electrical Resistance of Glass, Quartz,

these were used to build up a large set of high resistances sim-
ilar to those described in my article already alluded to. All
this was done twice, and during these operations, the resist-
ances of the insulating substances under experiment were
determined, as far as possible, immediately. In the case of
surface conduction, tinfoil was cemented on the ends of the
strip and fringed with a line of metallic powder. For internal
conduction, metallic surfaces were fastened to opposite sides of
the plate, which was provided with a guard-rim of tinfoil,
cemented with a solution of gum arabic and glycerine all
around its edges, and grounded to prevent surface electricity
from reaching its opposite side; a precaution which in the —
case of quartz and ebonite is not necessary, quartz giving the
same result with and without the guard-rim, and the surface
conduction of ebonite in dry weather is again still smaller than
that of quartz.

In order to compare together the results obtained, those for
surface conduction are reduced to the calculated resistance
that would be offered by one surface of one square centimeter
between the terminals. For internal resistance, the calculated
resistance of coatings one square centimeter in area, with a
thickness of one millimeter of substance between them, is
given. A set of one hundred dry cells was employed in each
case ; its electromotive force was 150 volts. The striking dis-
tance of the electrometer was one millimeter, except in the
case of very high resistances, where it was reduced to one-half
or one-quarter of a millimeter.

Internal resistance of quartz.—The thickness of the plate
was 1°71™", the metallic coatings, each, one square centimeter
in area. Calculating the resistance for a thickness of 1™™, I
obtained, on .

Feb. 19th 860,000 Q Hyg. 18°
60 P20 OEE OOO. Fee ee

The coatings were composed of an amalgam of tin and mer-
cury.

Internal resistance of glass.—Very numerous measurements
were made on plates of ordinary glass, and on others colored
by oxide of cobalt, and even one plate of glass 6°2™™ in thick-
ness was employed. The current traversed all of these with a
certain facility, but the resistance of the cement used to attach
the coatings increased from week to week, and consequently it
is not worth while to give any figures at present. The internal
and external resistances of two excellent Leyden jars of recent
German make were also measured; in both cases the external
resistance was very considerably greater than the internal ;
hence, in use, the main leakage would be through the sub-

Mica, Ebonite and Gutta-percha. 163

stance of the glass. It would, therefore, seem undesirable to
obtain capacity by thinness of wall.

Internal resistance of gutta-percha.—Two sheets of gutta-
percha were employed, their thickness being 2°79™". Tinfoil
was cemented on them with a solution of gutta-percha, and
the results were fairly constant as seen below, where the
actual figures obtained are given :

No. 1 6,030,000 Q Feb. 19th Hyg. 18°
ce 2 6,270,000 ce ce 4 ce 6¢
coe DA eS BE isin BFF
6e 2 6,640,000 “ce “c 6 ée ee

6,050,000 “

Calculating this for a surface of one square centimeter and
a thickness of one millimeter, we obtain, 18,500,000 ©.

Internal resistance of ebonite.—Three plates of this material
were used, one having a thickness of 2°29"; the other two
were only ‘5™™ thick. The coatings were made of a metallic
powder backed by tinfoil, all of which probably did not insure
the best possible contact.

Reducing the results to the standard of one square centi-
meter of surface and a thickness of one millimeter, we have

No. 1 76,400,000 Q Hyg. 32°
te a AR aneLOUs) Mein S8 7 |e
ac 12 WA, GOO; 000- Fi. “FS + ¢

55,000,000 ‘*

Internal resistance of mica (muscovite, beautiful large sheets).
—The thickness of No. 1 was -122™"; Nos. 2 and 3 had a
thickness of -°32™™. The coatings were made in the same way
as with ebonite. Reducing to an area of one square centi-
meter and a thickness of 1™™, I obtained for

No. 1 100,000,000 Q yg. 32°
“2? L9O_000,000 © Se s
i 3. 110,000,000... .<68"- &

133,000,000 «

From this it would seem that the internal resistance of mica
is somewhat greater than that of ebonite, though, as will be
shown later on, its surface resistance is much smaller.

Surface resistance of plate glass (window glass, commercial).
—This proved to be unexpectedly regular, in spite of some
variations in the hygrometric condition of the air. Two plates
were employed, each having a length of 23:7™ between the
terminals. ‘The actual figures obtained on three different occa-
sions are as follows:

164 Rood—Electrical Resistance of Glass, Quartz,

No.1 3,400,000 Q Feb. 19th Hyg. 18°
“ 2 3,600,000 “ “ 99d eo
so 8,700,000 3 - March 3d... #9e82e
“ 2 3,600,000 « « ree ©

3,570,000

Calculating this for one surface of one square centimeter
between terminals gives 1,590,000 Q.

Surface resistance of cobalt glass —From previous experi-
ments it was known that the surface resistance of cobalt glass
is considerably greater than that of the ordinary varieties of
window glass, and the results given below confirm this idea.
Four strips were used, with an average length between termi-
nals of 4°. The experiments were all made on the same day,
the hygrometer indicating 34 per cent of moisture. The
actual figures obtained are:

No. 1 24,400,000 Q Hyg. 34°
Si. BO TOOTIO WEE). bees
Heo, SQA ADO 000 FE oll yee
6545” SE N00,00D seo feel

23,000,0V0

Calculating this for one surface of one square centimeter
between terminals gives 22,000,000 0.

Surface resistance of mica.—F our slips of mica were used ;
the distance between terminals was 2°5™; the actual figures
were:

No.1 29,900,000 Q Hyg. 35°
“ 2 95,600,000 “ «© «
A 1. 0 ROOM, ek aren ae
“ 4 31,600,000 “ « «

31,600,000 “

Calculating for one surface of one square centimeter, we
obtain 50,760,000 Q.

Surface resistance of gutta-percha.—Two slips of the mate-
rial were used, the distance between terminals being 1°. The
actual figures were: :

No.1 17,500,000 © Feb. 20th Hyg. Az
¢ ce

ce 9 18,900,000 ce ce ee 4

fhe eo OO OU. : 2 She ~ 35.
(<4 9 7,700,000 ce <4 ee ce ce
1 6,500,000... “March'3d es me
ce 9 19,500,000 «e (73 ce <3 ce

16,000,000 «

Calculating the mean for one surface of one centimeter
between terminals, we have 432,000,000 ©.

i hee oe a
ain ’ ‘

Mica, Hbonite and Gutta-percha. 165

Surface resistance of quartz (crystal). —The distance from
terminal to terminal was 1°5°"; the actual figures were:
76,000,000 Q Feb. 20th Hyg. 17°
75,000,000 <“ Se 27 thi Se SAS
Calculating the mean for one surface of one square centi-
meter, we have 521,000,000 Q.
Surface resistance of ebonite—Two strips were used; the
breadth of gaps was in one case 5°5™”, in the other 6-2™™.
Calculating the resistances actually obtained, for one surface
of one square centimeter we have
No. 1 1,107,000,000 Q Hyg. 34°
“ 2 2 940,000,000 “« “« «

2,000,000,000 “

These results are sufficiently discordant ; those with quartz are
scant, but damp weather prevented additional observations.

In closing this article, I give two determinations of the
resistances employed in the investigation, for the purpose of
showing the degree of accuracy readily obtainable in the
process of building up high units by my method. The two
operations were separated by an interval of about a week ; the
units were one year old.

53,760 53,830 Q
71,860 72,370 “
81,400 80,500 “
244,000 241,000 “
220,000 205,000 “
297,000 292,000 “
389,000 384,000 “
1,161,000 1,163,000 “
585,000 554,000 “
1,565,000 1,504,000 “
1,401,000 1,380,000 “
2,246,000 2,193,000 “
2,891,000 3,140,000 “
2,203,000 2,193,000 “«
3,707,000 3,432,000 *
2,990,000 3,130,000
3,070,000 3,130,000 «
1,958,000 1,803,000 “
7,808,000 8,194,000 “
7,902,000 8,460,000 “
8,720,000 8,540,000 “
5,584,000 5,560,000 “
8,120,000 8,194,000 “
6,780,000 6,906,000 “

New York, June Ist, 1902.

166 Scientific Intelligence.

SCIENTIFIC INTELLIGENCE.

MISCELLANEOUS SCIENTIFIC INTELLIGENCE.

1. American Association.—The fifty-first annual meeting of
the American Association for the Advancement of Science was
held at Pittsburg from June 28 to July 3. The President of
the meeting was Professor Asaph Hall, U.S. N. The attend-
ance was rather large, 431 members registering; the number of
papers read was also very considerable, amounting to some 320,
including the addresses delivered and also papers before the
affiliated societies. The local committee provided some fifty
excursions for the entertainment and instruction of the members
and the many manufacturing establishments of the region offered
a rich field for study and observation. The address by the retir-
ing President, Professor Charles 8. Minot, was upon the subject
‘“The Problem of Consciousness in its Biological Aspects ;” this
is printed in the number of Science for July 4 and the following
issues of the same journal give the other addresses and general
proceedings in great fullness.

The officers elected for the following year are as follows :

President: Ira Remsen, Johns Hopkins University.

Vice Presidents: Section A, G. B. Halsted, Austin, Texas ;
B, E. F. Nichols, Dartmouth College ; C, C. Baskerville, Chapel
Hill, N. C.; D, C. A. Waldo, Purdue University ; E, W. M.
Davis, Harvard University ; F, C. W. Hargitt, Syracuse, N. Y. ;
G, F. V. Colville, Washington; H, G. M. Dorsey, Chicago ;
I, H. T. Newcomb, Philadelphia.

Permanent Secretary, L. O. Howard, Washington ; General
Secretary, H. B. Ward, University of Nebraska; Zreasurer, R.
S. Woodward, New York.

The next meeting of the Association will be held in Washing-
ton, D. C., Dec.-29, 1902 to Jan, 3, 1903, ues the recently
established convocation week.

OBITUARY.

Dr. Carto Riva, Docent in petrography and Assistant in the
mineralogical laborator y of the University of Pavia, was killed
by an avalanche June 3d, during the ascent of the Monte Grigna, —
southeast of Lake Como. Dr. Riva was one of the most promis-
ing of the younger petrographers and mineralogists of Italy and
had already published several papers describing rocks from the
Monte Adamello and other regions in the Alps, from Italy and
Sardinia ; he was also correspondent for the Zeitschrift fir
Krystallographie, and gave promise of a long and productive
career. His American friends, who were familiar with his charm-
ing personality, will learn with deep regret of his untimely end.

JOHN E. WOLFF.

THE

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES.]

Arr. XX.—The Relationships of some American and Old
World Birches; by M. L. Fernatp. (With Plates V-VI.)

[Contributions from the Gray Herbarium of Harvard University, New
Series, No. xxiii. ]

A RECENT attempt satisfactorily to identify some birches
from the alpine summits of New England has made it neces-
sary to study in detail certain species of the Old World. In
this detailed study so many well-known European and Asian
forms have been found identical with trees and shrubs of
America which have ordinarily passed as endemic, that the fol-
lowing notes are offered as a partial solution of the difficulties
which have long surrounded the species of Betula. The
writer has been specially fortunate in having constant access
to a large suite of specimens in the Gray Herbarium which
were examined and labelled by Regel in the preparation of his
monograph on the Betulaceze in DeCandolle’s Prodromus.*
The material in the United States National Herbarium and in
the Herbarium of the Geological Survey Department of Can-
ada has been generously loaned by Messrs. F. V. Coville and
J. M. Macoun, the collections of the Arnold Arboretum have
been placed at his disposal, and valuable specimens and notes
on northwestern forms have been freely furnished by Profes-
sors L. F. Henderson and C. V. Piper. Thus it has been
possible to examine a very complete representation of the

enus.

; The birches are trees of boreal range. Unknown in the
southern hemisphere and the tropics, they abound throughout
the northern and mountainous sections of North America,
Europe and Asia, reaching a more extreme northern range on
both continents than any other trees.t| By Regel and other

* Regel in DC. Prodr., xvi, part 2 (1864), 161-189.
+See F. A. Michaux, Hist. des Arb., ii, 129, and Sylva, ii, 47; Sargent,
Silva, ix, 46; Rehder in Bailey, Cyc. Am. Hort., i, 158.

Am. Jour. Sc1.—FourtH SERIES, Vou. XIV, No. 81.—SEPTEMBER, 1902.
12

168 Fernald—Relationships of some American

early students of the group various European and American
species were considered essentially identical ; but among recent
authors there has been an increasing tendency to regard nearly
all American forms as endemic, and to maintain as organically
distinct trees growing, for instance, on the American and
the Asian sides of Behring Sea.

That there is not only a strong similarity, but a marked pro-
portion of identity, in the herbaceous floras of the cireumboreal
regions has long been recognized; and, although the tendency
to ignore this almost axiomatic truth has been conspicuous in
the work of certain American as well as European botanists, the
increasing evidence of its validity has forced itself upon the
attention of others. This similarity in the boreal floras of the
eastern and western hemispheres was very emphatically pointed
out by Sir Joseph Hooker in his discussion of the Distribution
of Arctic Plants.* Although our knowledge of the distribution
and relationships of plants has very greatly increased since
Hooker’s tabulation was made, his outlines still furnish an
approximately accurate basis for generalizations. If we con-
sider the boreal’ distribution of some well-known northern
genera whose species were then clearly interpreted, we shall
find that of the 61 Arctic species of Carex recognized by
Francis Boott, 58 extend southward into temperate Europe, 51
into temperate America, and 43 into temperate Asia; of the
16 Arctic Junct recognized by Hooker all 16 extend into
temperate Europe, 15 into temperate America, and 14 into
temperate Asia; of Hooker’s 24 Arctic species of Saxifraga,
19 extend into temperate America, 17 into temperate Europe,
and 14 into temperate Asia. This estimate is altered from
that of Hooker only in so far as recent data has been immedi-
ately accessible to the writer; but it does not, of course, take
into account such well-known species as Carex riparia, C.
Pseudo- Cyperus, Juncus tenuis, ete., which occur in temperate
regions of both hemispheres, though not within the Arctic
Circle, but which, if counted, would increase the number of
identical species in the eastern and western hemispheres.

Toward the warmer regions of each continent the number
of identica] plants rapidly diminishes, and the proportion of
endemic species becomes very great. But in view of the
marked similarity of the herbaceous vegetation of northern
Europe, Asia and North America, it seems only logical to ex-
pect a notable proportion of identities between the trees and
shrubs of boreal range, especially when, as in Betwla and Salix,
several species extend nearly or quite to the northern limits of

* J. D. Hooker, Trans. Linn. Soc., xxiii, pt. 2, 251-358; see also Gray,

Mem. Am. Acad., n. s., vi, 377-449, and extract ‘‘Flora of Japan,” in Sci.
Pap. of A. Gray, selected by C. S. Sargent, ii, 124-141.

and Old World Birches. | 169

vegetation and are rapidly spread by means of their thin-
winged or hairy fruit. In fact, this relation of the Alaskan
willows to those of the Old World is brought out by Mr. F. V.
Coville in his scholarly monograph on the Willows of Alaska,*
in which he recognizes 23 species, 2 of them circumpolar in dis-
tribution, and 9 others occurring in Siberia. That is, of the
23 willows recognized by Mr. Coville in Alaska, only 12, or
scarcely more than one-half, are endemic North American
species. A similar detailed study of northeastern Sadices will,
it seems to the writer, likewise show a closer relationship be-
tween certain other American and Old World forms than has
been generally recognized.

The examination of the birches by the writer has led to the
uniting of some well-known American and Old World trees
and shrubs. The conclusions reached in his studies, and cer-
tain details of the studies themselves, may best be presented
by discussing separately the species which have been specially
examined.

S Albee. Trees or shrubs: the wing two or three times as
wide as the achene ; rarely only a little broader.

BEruLA ALBA.

The original Letula alba of Linneeust seems to have em-
braced both the common white birches of northern and central
Europe. In 1788, however, Roth distinguished the two trees:
“ Betula alba. 8B. foliis ovato-acuminatis inciso-serratis scabris,
ramis erectis strictis” ; and “ Betula pendula. B. foliis ovato-
acuminatis inciso-serratis glabris, ramis flaccidis pendulis.” +
This separation of the two trees by Roth has been very gen-
erally ignored by recent European authors, nevertheless, and
the species with pubescent (“ scabrous ”) leaves and upright
branches, the true Getula alba as interpreted by Roth, is uni-
versally known abroad as B. pubescens, Ehrh.,§ while the
smooth-leaved species with pendulous branches, the Betula
pendula, Roth., is ordinarily known by its later name, B. ver-
pucosa, Ehrh. :

By some authors, as Prantl] and Guerke,§ these two white
birches are treated as distinct species, while by others the
entire group is regarded as a polymorphous species with many
subspecies and varieties. Rehder, in Bailey’s Cyclopedia of
American Horticulture, disposes of the forms in this way,

* Proc. Wash. Acad. Sci., iii, 297-362 (1901). + Sp.; ii, 982 (1758).

+t Roth, Fl. Germ., i, 404, 405 (1788). S Beitr., v, 160, vi, 98 (1790-91).

| Prantl in Engler & Prantl, Nat. Pflanzenf., iii, pt. 1, 45.

Si Pi. Ear., ti, 47, 48.

170 fernald—Relationships of some American

recognizing Betula alba, with the two subspecies, B. pendula,
Roth (B. verrucosa, Ehrh.), and B. pubescens, Ehrh., each with
numerous marked variations. Mr. Rehder, however, keeps
separate from the Old World B. alba the American B. papy-
rifera and B. occidentalis. |

The most conservative treatment of the group was, in some
particulars, that of Regel in DeCandolle’s Prodromus, though
many of his varieties have proved of little value. There Regel
recognized two species of white birches (his section Adda), -
Betula alba, L., and B. microphylla, Bunge. B. alba, how-
ever, he divided into nine subspecies :

verrucosa, with numerous varieties, from Europe, Asia
and America ;

populifolia, a well-known strictly American tree ;

mandshurica, a local Asian tree ;

latifolia, with varieties, from central and northern Asia to
Japan and Kamtschatka ;

occidentalis, with varieties, from North America ;

papyrifera, with varieties, from North America and
Siberia ;

pubescens, with varieties, from Europe, Asia and North
America ;

tortuosa, from northern Europe and Siberia ;

excelsa, a doubtful form referred to by earlier authors.

Thus it is clear that Regel, the most devoted student of the
birches, found it impossible to distinguish as clear species the
diverse trees and shrubs which pass as white or canoe birches.

The recently accumulated material shows that Regel’s course
was perhaps the most philosophic of any which has been pro-
posed. Yet the primary characters of the trees, first clearly
distinguished by Roth, are generally so constant that it seems
better to the writer to recognize as species, l.e., as centers of
variation, the two forms designated by him. These were the
true Betula alba, with mostly stiff and ascending branches,
the young branchlets puberulent or pubescent and the ovate
often doubly serrate leaves more or less pubescent beneath, at
least when young; and £B. pendula, with more flexuous
branches, the branchlets glabrous or verrucose with resiniferous
atoms, and the deltoid or rhombic-ovate often simply serrate
leaves glabrous, and more or less glutinous at least when young.
These two trees in their characteristic forms are now very well
understood abroad, though some of their reduced forms present
perplexing problems. ut since the days of Regel the Amer-
ican trees have been very generally regarded as endemic species,
and it is to the discussion of this question that special atten-
tion is here directed.

and Old World Birches. 171

Betula alba (as interpreted by Roth).

As already stated, Betula alba, L. (as interpreted by Roth)
has passed in Europe as B. pubescens, and under this latter
name it is ordinarily distinguished from B. pendula, Roth
(B. verrucosa, Ehrh.). The species is generally recognized as
occurring through northern and central Europe and across
Asia, and by Regel and his followers it was regarded as like-
wise American, though by recent authors it has been excluded
from our flora. The typical form of LB. alba is the large
White Birch of northern Europe, with ascending branchlets,
puberulent or hairy twigs, and ovate or rhombic-ovate leaves
more or less pubescent beneath, especially when young.

The representative of this tree in America is our common
Paper or Canoe Birch, Betula papyrifera, Marsh.* (B. papy-
racea, Ait.+), which was discussed in detail by Michaux in his
Sylva.t There Michaux wrote at such length as “ will not be
deemed superfluous by persons who justly appreciate the im-
portance of precise ideas on subjects like the present,” of the
Canoe Birch (8. papyracea), the European White Birch (B.
alba, including, at least in the illustration, B. pendula, Roth),
and the American White or Old Field Birch (BL. populifolia).

Michaux’s plate of Betula papyracea is of a northern form
very closely approaching var. cordifolia, Regel; and in com-
menting upon this tree he said “the bark is of a brilliant
white, like that of the White Birch of Sweden, and, like that
too, it is almost indestructible... .. . This bark, like that of
the European species, is devoted to many uses,” which are
very fully enumerated. In his discussion of the European
L. alba he said, “ The trunk and limbs of the large trees are
covered with a thick bark, whose epidermis is white and _per-
fectly similar to that of the White Birch [B. populifolia] and
the Canoe Birch [B. papyracea|. The small branches like-
wise resemble those of the species just mentioned [B. papy-
racea|, being slender, flexible, and of a brown color spotted
with white.” He then discussed the uses of the European
White Birch, concluding his remarks with: “Such are the
principal uses of the European Birch, all the valuable proper-
ties of which are completely united in the Canoe Birch of
North America.” In his discussion of B. populifolia, Michaux
further said: ‘“ The trunk of this species is clad in a bark of
as pure white as that of the Canoe Birch and of the European
Birch; but its epidermis, when separated from the cellular
tissue, is incapable of being divided, like that of the two pre-
ceding species, into thin sheets; which constitutes an essential

* Arbust. Am., 19 (1785). + Hort. Kew, iii, 337 (1789).
¢t Michaux, Sylva, ii, 50-57.

172 Fernald— Relationships of some American

difference.” And in his summary of characteristics, he said:
“The White Birch of Europe and the Canoe Birch resemble
each other in their wood, their bark, and their ample propor-
tions, which are perhaps superior in the American species.
They differ in the form of their leaves, and they grow on very
different soils: the Canoe Birch is exclusively attached to rich
lands constantly cool, and capable of yielding an abundant har-
vest of corn or of clover, and it propagates itself naturally only
in that part of North America which corresponds in climate to
the 54th and 55th degrees of latitude in Europe.”

These distinctions pointed out by Michaux are not, however,
of such fundamental importance as to prove the American
Canoe Birch (Betula papyrifera) organically distinct from the
White Birch (B. alba, L., B. pubescens, Ehrh.) of Europe.
Michaux made no distinction in his Sylva between the true
Betula alba and the smooth-leaved and smaller B. pendula,
Roth. His plate was of the latter plant, which has more
deltoid leaves than the true B. papyrifera, but is quite like
another American tree soon to be discussed. The foliage of
the true B. alba is, nevertheless, as will be seen on comparison
of American and European material, so similar as to present no
apparent distinctive feature. The Canoe Birch in its best de-
velopment, which Michaux places at a height of 70 feet, is
found in deep soil on hillsides; but, had that keen observer
been permitted to remain longer in our northern regions and
to have explored the upper slopes of our mountains, he would
have found the same tree there growing from the crevices of
bare ledges—quite as barren a soil as that occupied by the
European tree, which, by his own statement, may be 70 or 80
feet high. The more southern occurrence of the American
tree is at least negative evidence of its identity with the tree
of Scandinavia, northern Germany and Russia, for,as is now
well known, the isotherms which cross New England pass far
to the north in Europe, and the vegetation of northern New
England and adjacent Canada has more than once been com-
pared with that of Scandinavia. In fact, Michaux himself, in
discussing the range of the Canoe Birch, said: ‘ This part of
North America, though situated 10 degrees further south, very
nearly resembles Sweden and the eastern part of Prussia,
not only in the face of the soil, but in the severity of the
climate.” Thus the comparative notes of Michaux leave little
by which to distinguish the American from the European tree ;
and it is further worthy of note that in Boswell Syme’s enu-
meration of the uses of the bark of the European tree he says:
‘In Russia it is applied to the same purposes for which that
of the Canoe Birch is used in North America, boats being
formed of it that are nearly as light and portable as those made

ee) = eon on

ers ee a a a ee Pen

and Old World Birches. 173

by the Red Indians of Canada.”* There is, then, no evidence
from the descriptions and comparisons of Michaux and other
well-informed European authors that the American Betula
papyrifera is separable from the true BL. alba of northern
Europe.

The American tree presents numerous but apparently incon-
stant variations in the size and toothing of its leaves, and the
size of its strobiles and achenes. These variations, however,
are very Closely matched by specimens in American herbaria
of the European Betula alba (B. pubescens, Ehrh.). If, for
example, we compare branches collected by Robert Chalmers
at Campbellton, New Brunswick, in 1887, or Rydberg’s No.
1,005, from the Black Hills cf South Dakota (distributed as
B. occidentalis) with a Bohemian specimen from Tausch and
Russian material from Regel; Macoun’s Cypress Hills (Assin-
iboia) material (No. 5,919) with Christiania (Norway) mate-
rial from Blytt; a sheet in the Gray Herbarium, collected on
Lake Winnipeg by Bourgeau, with Hampe’s No. 1,321 from
the Hartz Mts., in central Germany; or Clement’s No. 2,919
from Nebraska and Medford (Massachusetts) material of Wm.
Boott’s with the 1807 sheet of B. pubescens sent by Regel to
the Gray Herbarium ; we cannot help being impressed by the
identity of the pubescence, leaf-outline, strobiles, ete. Other
comparisons of American and European specimens emphasize
this identity ; so that, in view of these facts and the essential
similarity of bark, branches and stature, as pointed out by
Michaux and others, there seems no question that Betula
papyrifera is the true B. alba of Linneeus.

Betula alba, forma occidentalis.

In the discussion of Betula papyrifera in his Silva, Pro-
fessor Sargent referst to the tree of the northwest coast, a form
which “differs from the eastern [ papyrifera] in its greater
height and rather darker colored bark, in its more pubescent
branchlets, which sometimes do not become glabrous until their
second season, although vigorous shoots of young plants in the
east are often clothed with thick pubescence, and in its rather
larger leaves, which, on the lower surface, are also more pubes-
eent.” Later, however, the same author has identifiedt this
large tree with B. occidentalis, Hooker,§ not the small tree or
half-shrub taken for &. occidentalis by Nuttall and other
American authors, including Sargent, Silva, ix, 65, t. 453;
and he concludes that the bark “is very different from that
of the eastern tree, and it is probably best to consider it a
species.”

* Syme, English Botany, viii, 184. + Sargent, Silva, ix, 59.

¢ Sargent, Bot. Gaz., xxxi, 288 (1901). § FI. Bor.-Am., ii, 155 (1839).

174 Fernald—Relationships of some American

The original material of Betula occidentalis came from the
Strait of Juan de Fuca; and if we examine specimens from
that region we shall find them matching in every character of
pubescence, leaves and strobiles large forms of eastern B. alba
(BL. papyrifera), as already intimated by Professor Sargent.*
Branches collected by Professor L. F. Henderson in 1888 on
the Gulf of Georgia, essentially a continuation of the Strait of
Juan de Fuca, are clearly B. occzdentalis, as shown by a trac-
ing atthe Arnold Arboretum of the original Scouler specimen.
These specimens from the Gulf of Georgia have some of the
branchlets strongly hirsute with sordid hairs mixed with abun-
dant resiniferous atoms, while other (perhaps older) branchlets
appear quite glabrous. In the original description of B. occz-
dentalis, Hooker emphasized the abundance of resin, though
nothing was said of hairs upon the branchlets. This character,
however, is extremely variable, as shown by a large suite of
specimens from Washington and British Columbia; and, ad-
mitting this inconstancy in the amount of resin and pubes-
cence, the Gulf of Georgia branches seem to the writer quite
inseparable from many sheets representing B. alba (B. papyri-
fera) from different sections of America. In such characters
as are shown in herbarium material they are identified with the
following among other specimens: Almota Creek, Whitman
Co., Washington (Piper, No. 3,570), very resiniferous but not
hairy, thus matching the original description; Cascade Mts.
and Sumass Prairie, lat. 49°, British Columbia (Lyall), the
latter the type of B. alba, subsp. occidentalis, B, commutata,
Regel, the branchlets resiniferous but not hairy, thus agreeing
with Hooker’s original description; Pend d’Oreille River,
Washington (Lyall), branchlets quite as hairy as in the Gulf
of Georgia tree; Hill City, South Dakota (V. Bailey, sheet
No. 230,000, U. 8. Nat. Herb.), branchlets hairy; Black Hills,
South Dakota (2?ydberg, No. 1,005), branchlets hairy ; Queens-
town Heights, Ontario (J/acoun, Herb. Geol. Surv. Can., No.
.23,636), branchlets resiniferous, hirsute at tips; Goat Island,
Niagara Falls, New York (Coville, sheet No. 294,829, U. S.
Nat. Herb.), branchlets puberulent ; Bald Eagle Ridge, Center
Co., Pennsylvania (Porter), branchlets as in Gulf of Georgia
specimens; North Conway, New Hampshire (Wm. Soot),
branchlets as in last; Mountain Rock, Ellsworth, Maine (fer-
nald), branchlets resiniferous, slightly hairy; Manuels, New-
foundland (fobinson and Schrenk, No. 139), branchlets as in
last. Other specimens from the coastal region of Washington
and British Columbia are quite like some eastern branches.
Thus Lake and Hull’s plant from Tukanon River, Washington,
and Macoun’s No. 23,639 from Vancouver, are inseparable from

*Silva, |. c.,{Bots Gaz, Ine:

and Old World Birches. Wd

a Niagara specimen collected by Asa Gray; and Piper’s No.

1,128 ‘from Whatcom Co. , Washington, cannot be distinguished
from Lowrie’s material from Bald Eagle Mt., Blair Co., Penn-
sylvania (sheet No. 149,835, U.S. Nat. Herb. ), and the old
specimen of Wm. Oakes’s from - Topsfield, Massachusetts, re-
ferred by Regel to B. occidentalis, var. commutata ; and in
their leaves these trees very closely "approach a Canoe Birch
growing at Oak Island, Revere, Massachusetts, and well-known
to the local botanists from its lustrous brown bark. From
herbarium specimens, then, there is no character by which to
separate the northwestern Betula occidentalis from the eastern
B. alba ( papyrifera).

The important character upon which Professor Sargent lays
stress in his recent note* is the color of the bark, which in the
northwestern tree is usually brown. This character alone seems
hardly sufficient to separate specifically the two trees, espec-
ially in view of such brown-barked trees in the East as that on
Oak Island; a tree growing at the foot of Mountain Rock,
Ellsworth, Maine (represented in the Gray Herbarium), w ith
eray-brown bark strongly tinged with plum-color; and Robin-
son and Schrenk’s No. 139, from Newfoundland, in which the
old bark, though becoming pale, retains much of the brown
which is often seen in young trees. Furthermore, it is worthy
of note that the Pacitic Coast tree is not thor oughly constant
in its bark. The Henderson plant of 1888, from the Gulf of
Georgia (essentially the type locality), was sent to the late
Sereno Watson, who called it B. occidentalis. Subsequently,
however, Professor Henderson, writing under date of Decem-
ber 9, 1897, of the difficulties in studying this group, said:
«Just where B. papyrifera leaves off and B. oceiden talis be-
gins, | am, and always have been, at a loss to say.” He then
comments upon a tree which “is undoubtedly B. papyrifera
in every respect,” adding, “and I have no doubt that my num-
ber 1,712, from the shores of Lummi Island, Gulf of Georgia,
sent Mr. Watson in ’88, and referred by him to B. occidentalis,
is the same thing.” Thus it would ‘seem that Betula alba
(papyrifera) is not constant in its pale bark in the East, and
that L. occidentalis of the Northwest may not always be dis-
tinguished by its dark bark.

The tendency of another dark-barked northwestern tree to
become quite as pale as the eastern Canoe Birch will be noted
in the discussion of a species soon to be considered ; but in this
connection it is worth while to note a tendency to darkening
of color which has been observed in certain other northwestern
plants. It is well known that many species of Carex growing
in shade or in southern areas have pale or hyaline scales, while

hot Gases le ¢,

176 Fernald—Relationships of some American

the same plants in exposed or alpine situations have the scales
brown, chestnut, or even nearly black.* On the Pacifie slope
of North America, especially from the Cascade and Coast
Ranges to the coast of Oregon, Washington, British Columbia
and Alaska, this tendency is likewise very conspicuous. It is
well shown in such plants as Carex praticola, Rydb., in the
Vancouver Island form described as C. pratensis, var. furva,
Bailey; C. pennsylvanica, in the variety vespertina, Bailey;
C. aurea, in the Vancouver plant, and in Suksdorf’s No. 35
from Falcon Valley, and F. Binn’s material from Port Ludlow,

Washington; C. limosa, in the variety stygia, Bailey; C. fila- -

Jormis, var. latifolia, in the Vancouver plant, and Hall’s No.
607 from Oregon, Suksdorf’s No. 51 from Klikitat Co., Wash-
ington, and Geyer’s No. 72 from Oregon ; and C. rostrata, in
the Vancouver plant, and Piper’s No. 994 from Seattle, and
Suksdorf’s Nos. 55 and 56 from the Cascade Mts. of Wash-
ington. The same tendency has likewise been noted in
Fileocharis and Juncus. Without attempting here a discus-
sion of the conditions which tend to produce upon our north-
west coast the darkening of scales or.other chartaceous portions
of plants, it may be suggested that the brown color ordinarily
seen in the Canoe Birch of Vancouver and the coastal region
of Washington and British Columbia is perhaps due to the
same physiological cause.

Emphasis has likewise been laidt upon the great height of
this northwestern tree as compared with the eastern Betula
alba (papyrifera). Yet it is interesting to note that David
Lyall (whose specimens from the Lower Frazer River are a
close match for Hooker’s original description and for the trac-
ing of the Scouler specimen) specially commented upon the
tree as “growing to the height of 60 or 70 feet,’ + while
Piper’s Whitman County trees are much lower. Since the
brown-barked Betula occidentalis differs, then, only in this
somewhat inconstant color of the bark from the ordinarily

pale-barked B. alba, and since it is often no taller than the —

eastern tree, it seems to the writer hardly worthy special
recognition, and that the tree was well treated by Professor
Sargent in the Silva as a local tendency of the white-barked
tree. us

Betula alba, var. glutinosa.

Besides the ordinary Canoe Birch and its brown-barked form
of the Pacific slope, there are many tendencies of Betula alba
in America which deserve special comment. <A single tree

* See, for example, Boott, Ill. ii, 98, ete.; Holm, this Journal, 4th series,

ii, 218; Fernald, Proc. Am. Acad., xxxvii, 500, 504. :
+ Sargent, Silva, 1. c.; Bot. Gaz.,].c. $ Lyall, Jour. Linn. Soc., vii, 134.

se

and Old World Birches. Lil

left standing, after the widespread devastation of forest fire,
by the Wassataquoik River, near Mt. Katahdin, in Maine, is
conspicuous on account of its pendulous “ weeping ” branches,
small leaves and essentially straight peduncles. Though in
habit this tree is strikingly like B. pendula, its leaves place it
nearer £. alba. In fact, the material matches perfectly a
specimen sent to the Gray Herbarium by Regel, from Finland,
of B. alba, var. glutinosa, Trauty.,* the form: ‘pendulat fioured
by Reichenbach (ic. Fl Germ., xii, t. 625) as 3B. pendula,
though not the species of Roth. This Wassataquoik Valley
tree may be known as Betula alba, var. glutinosa, Trautv.
This form (L. glutinosa, Wallr.t) is treated by various Euro-
pean authors as a natural hybrid between B. alba ( pubescens)
and B. pendula,§ but the apparent absence from Maine of
BL. pendula renders this origin of the tree improbable.

Betula alba, var. cordifolia.

Another tree more common in the mountainous portions of
New England than Betula alba, var. glutinosa, differs from
the ordinary Canoe Birch only in its cordate-ovate leaves. In
this character it is very constant, however, and seems to de-
serve the varietal recognition given it by ‘Regel as B. alba,
subsp. papyrifera, B, cordifolia, | (B. cordifolia, Regel, Mon.
28, t. 12, figs. 29-36). The species, B. cordifolia, was based
upon one of de la Pylaie’s specimens from N ewfoundland, and
in his later treatment of the tree Regel cited two sheets in the
Gray Herbarium, one from Mt. Katahdin, Maine, the other
from Lake Superior. The tree is common on the upper
wooded slopes of Katahdin and other mountains of northern
New England, becoming a dwarf shrub at their summits, where
it has passed as LB. papyracea, var. minor, Tuck. Tucker-
man’s original material and description, however, was ot
another form soon to be noted. The tree or shrub of the
mountainous districts of New England, with cordate leaves,
pubescent on the veins beneath, may be known as Betula albie,
var. cordifolia (Regel). Its range is a broad one, from north-
ern Lasrapor (CU. A. Kenaston, sheet No. 25,240, U. S. Nat.
Herb.) and Newrounpianp to New Brunswicx, Marne, the
White Mountains, New Hampsuire, Lake Superior, Jowa
(B. Fink, No. 109), Atperra, Brirish Cotumpta, Ipano, and
WASHINGTON.

* Trauty. ex Regel, Mon. Bet. (1861), 20. + Regel, 1. c., 28.
t Sched. Crit., 497 (1822).

S$ See Koehne, Deutsch Dendr., 109; Guerke, 1. c., 48.

|| Regel in DC., 1. c., 166

178 Fernald—Relationships of some American

Betula alba, var. minor.

The dwarf alpine shrub which was described by Tuckerman
as Betula papyracea, var. minor,* is represented in the Gray
Herbarium by material from Tuckerman himself. It is not,
perhaps, the commonest dwarf Canoe Birch of the White
Mountains, but is a form with the small elliptic- or truncate-
ovate leaves strongly glutinous and quite without pubescence.
From the comparatively narrow wings of the samaras in Tuck-
erman’s specimens, Regel took the plant to be a form of
B. davurica, Pallas, of eastern Asia, and upon the Tuckerman
sheet he based his B. duwvurica, B, americana.+ The shrub
is, however, more properly a phase of B. alba, and it occurs
likewise on the North Summit of Mt. Katahdin, where it
sometimes has the leaves smaller firmer and more glutinous
than on Mt. Washington; in Gaspé and in Labrador; and in
the mountains of Saskatchewan, Assiniboia and Alberta, where
it intergrades with 6. microphylla. Though ordinarily dis-
tinguished from LB. alba, var. cordifolia, by its elliptic- or
truncate-ovate glabrous leaves, smaller erect strebiles, and
smaller samaras, this second dwarf Canoe Birch shows ten-
dencies to integrade with that form. In its most characteristic
development this shrub of the New England mountains
matches B. tortwosa, Ledebour,t described from the Altai Mts.
of central Asia, where it had likewise been mistaken by Lede-
bour for BL. davurica of Pallas. Therange of Betula tortuosa
has since been extended through northern Russia to Silesia,
Finland, Lapland and Greenland, where it is sometimes treated
as aspecies, but by Koehne and by Guerke as a form of JB.
alba (B. pubescens, var. tortuosa (Ledeb.) Koehne).§

If we compare specimens collected by the late Edwin Faxon
at “ Willis’s Seat” on Mt. Washington, New Hampshire, and
Altai material of Betula tortuosa sent to the Gray Herbarium
from the Imperial Herbarium at St. Petersburg, we shall find
the former differing only in its glabrous branchlets.. This
want of pubescence is, however, quite as conspicuous in a
Lapland specimen from Regel of his B. alba, subsp. tortuosa
a genuina. If, again, we compare the original sheet of Tuck-
erman’s BL. papyracea, var. minor from the White Mountains
with the plate of B. tortwosa in Flora Danica (xvii, t. 2918)
we shall find them too close for ready separation. Specimens
from Mt. Katahdin are, furthermore, quite inseparable from Ro-
senvinge’s No. 160 from Greenland, distributed from the Botan-
ical Museum of Copenhagen as B. odorata, var. tortuosa and
represented by sheet No. 2,044 in the Herbarium of the Geo-
logical Survey of Canada; and various Mt. Washington speci-

* Tuck., this Journal, xlv, 31 (1848). + Regel in DC., 1. c., 175 (1864).

t Fl. Ross. iii. 652 (1849). § Deutsche Dendr. 109 (1893).

and Old World Birches. 179

mens are practically identical with Rosenvinge’s Nos. 174 and
210 from Greenland. This shrub or dwarf tree offers consid-
erable variation in the outline and toothing of its leaves, but
no more than does the same form abroad ; and since the young
branchlets and leaves are sometimes pubescent as in B. alba
its treatment as a dwarf variety seems the wiser course, and
the shrub may be known as Betula alba, var. minor (Tucker-
man).
Betula alba, var. carpatica.

A shrub growing on Anticosti Island at the mouth of the
St. Lawrence (J/acoun in Herb. Geol. Surv. Can., No. 25,823)
differs somewhat markedly from other American forms. It is
nearest Betula alba, var. minor, but differs in its broadly
rhombic leaves, mostly cuneate at base and permanently pubes-
cent beneath, especially on the nerves. This shrub is undoubt-
edly the B. borealis, Spach* (BL. pumila, y borealis, Regel,
Mon. Bet. 55, t. 13, figs. 38, 39), described from Newfound-
land, and it is identified with Regel’s description of B. pubes-
cens, € carpaticat (B. carpatica, Waldst. and Kit. in Willd.
Sp. iv, 464). The specimens, except in their more abundant
pubescence (in this matching Regel’s “ folia subtus in venarum
axillis saepe barbata”), are well matched by a Lapland sheet
distributed by Regel as B. pubescens, var. carpatica ; and they
are equally well matched by a sheet distributed by Fries as
B. glutinosa. This plant, B. glutinosa of Fries, not Wallroth,
is treated by Guerke as identical with B. carpatica. The
same form is well represented in Flora Danica (xvi, t. 2851) as
B. odorata, var. rhombifolia, Lange, with B. odorata, vay.
carpatica, Lange (BL. carpatica, Waldst. and Kit.) cited as a
synonym. Guerke, however, does not identify B. rhombifolia,
Tausch, with 3B. carpatica, and considering this doubt it is
least confusing to call the plant Betula alba var. carpatica
(Waldst. and Kit.). The same form occurs in the Herbarium
of the Canadian Geological Survey from Banff, Alberta (San-
son, Nos. 22,298 and 22,444), though in its somewhat resin-
iferous branchlets the first cited specimen tends toward B.
microphylla, Bunge. An old specimen collected by Dr. Rich-
ardson on the Franklin Expedition (sheet No. 23,628, Herb.
Geol. Surv. Can.) is apparently the same.

BETULA PENDULA.
Betula pendula, var. japonica.

Aside from Getula papyrifera and B. occidentalis the only
birch of the section A/bae generally recognized in America
until within three years has been B. populifolia, a purely

* Spach, Ann. Sci. Nat., ser. 2, xv, 196 (1841). + Regel in DC., 1. c. 168.

180 Fernald—Relationships of some American

American species quickly distinguished by its comparatively
low stature and southern range, its chalky white or gray-flecked
close and scarcely defoliating bark, its glossy extremely cau-
date-attenuate deltoid leaves, its compact grayish brown fruit-
ing strobiles averaging 3°" long and 6™™ thick, and the samara
rarely 3°5™™ broad. Very recently, however, three trees have
been described from northwestern America as unique American
species: Betula kenaica, W. H. Evans,* from Alaska; B.
alaskana, Sargent,t from Saskatchewan and Alaska; and B.
fontinalis, Sargent,t the Rocky Mountain half-shrub which has
there passed as B. occidentalis.

The writer has not seen the original material of B. kenaica,
but a large suite of specimens collected by Messrs. Coville and
Kearney in the type region, on the Harriman Alaska Expedi-
tion, has been placed in his hands by Mr. Coville. This series
of 15 sheets in the United States National Herbarium is sup-
plemented by very important notes by Mr. Coville on the color
of the bark, ete.

The first specimens cited by Professor Sargent for his Betula
alaskana are one of Bourgeau’s collected on the Saskatchewan
during the Palliser Expedition, and Macoun’s material from
Prince Albert, Saskatchewan, collected in July, 1896 (not 1876
as originally published). The other material came from Alaska
and the tree is said to be ‘“‘the ‘canoe birch’ of all travellers
in Alaska, and it is the common birch tree of the Yukon
Valley.” The Bourgeau plant, the first specimen cited by
Professor Sargent and consequently (for want of any specially
indicated specimen) to be regarded as the type, is in the Gray
Herbarium. This plant was labelled by NRegel and was
included by him in the Prodromus as B. alba, subsp. verrucosa
6 resiniferda, a form otherwise known to him from Amur, near
Udskoi (latitude 55°) on the Okhotsk Sea, and from the neigh-
boring province, Transbaikalia. BL. verrucosa, var. resinifera
was distinguished by Regel from the typical B. verrucosa by
the densely resiniferous branchlets, a character very conspicu-
ous in the Bourgeau specimen, but less marked in Macoun’s
Saskatchewan material. And although in B. kenaica the
twigs were originally described as “not resin-dotted,” later
collections show the character to be inconstant and that the
twigs are sometimes quite as glandular or resiniferous as in the
Bourgeau specimen of A. alaskana ; and furthermore, BL. alas-
kana has the twigs only “more or less verrucose with conspic-
uous resinous glands,” so that little stress should be laid upon
this character. ;

A comparison of thé original descriptions of Betula kenaica
and B. alaskana shows no point by which they can be sep-

* Bot. Gaz. xxvii, 481 (1899). + Bot. Gaz. xxxi, 236, 239 (1901).

and Old World Birches. 181

arated. Both are trees commonly with red-brown or brown-
gray bark, both have the twigs smooth or resiniferous, the
leaves dark green above, paler beneath and ultimately quite
glabrous. The fruiting catkins are described in practically
the same terms. 6. kenaica is common in the region of
Kenai Peninsula and Kadiac Isiand; while &. alaskana is
characterized as “the ‘canoe birch’ of all travellers in Alaska,”
and the range given includes “the Alaskan coast on the shores
of Lynn Canal... ; and westward.” Since Lynn Canal lies
on the coast in the same latitude as Kenai Peninsula and
scarcely 500 miles to the eastward, the general range covered
by Professor Sargent’s “westward” may well include the
type region of Letula kenazca.

A comparison of specimens of Getula kenaica and B. alas-
kana likewise fails to reveal any distinctions. For instance,
Coville and Kearney’s No. 2,423 of B. kenaica (sheet No.
373,620 in the U. 8S. Nat Herb.) is quite inseparable from
Macoun’s Prince Albert material (No. 12,952a, Herb. Geol.
Surv. Can.), one of the original specimens cited for 3B. alas-
kana; and neither of these specimens can be separated in
habit, twigs, leaves or pubescence from No. 20,327 of the
Canadian Geological Survey Herbarium, material collected by
R. G. McConnell near Dawson City on the Yukon, where B.
alaskana is said to abound. Furthermore, although the Bour-
geau specimen of LB. verrucosa, var. resinifera, the first cited
specimen of B. alaskana, is young, its leaf-outline and very
resiniferous twigs seem quite identical with those of Coville
and Kearney’s No. 2,412 of B. kenaica (sheet No. 373,618,
U.S. Nat. Herb.). There is, then, no question that Betula
alaskana described in 1901 is identical with LB. kenaica, pub-
lished in 1899.

Both Mr. Evans and Professor Sargent had their species
compared at St. Petersburg, before publication, with Siberian
material; and it is certainly very unfortunate that these com-
parisons were not more satisfactorily made, since very many
sheets in the Gray Herbarium, as well as the United States
National Herbarium, show what appear to be identical
branches collected in Asia over a wide range of country,—
from Kamtschatka to northwestern Siberia (Yenisei River),
south to Japan, Mandschuria, Mongolia and Sungaria. This
Asian tree was treated by Regel in the Prodromus as Betula
alba, subsp. mandshurica and latifolia, while the more
strongly resiniferous were included in his subsp. verrucosa, 6
resinifera. The form treated by Regel as subspecies /ati-
folia had, however, been described by Siebold* in 1830 as
Betula japonica.

* Verh. Batav. Gen., xii (1830), 25.

182 Lernald—felationships of some American

In Asia as in northwestern America the abundance of resin
on the branchlets seems to vary without any accompanying
change in leaf-outline or other character perceptible in the
specimens; and by comparison with well authenticated Asian
specimens of Betula verrucosa, Ehrh. (B. pendula, Roth), B.
mandshurica, Regel, and B. latifolia (B. japonica) with its
vars. Tauschi and kamtschatica, essentially all the available
material of the American 6. kenaica (including B. alaskana)
can be exactly matched. Thus such specimens as Coville and
Kearney’s Nos. 1,408, 1,422, and 2,495 of B. kenaica (Nos.
373,608, 373,609, 373,622, U. S. Nat. Herb.) and an old sheet
in the National Herbarium from Fort Simpson (Mackenzie
District) are good matches for Japanese material collected by
Maximowicz and distributed as B. alba, subsp. latifolia u
Tauschii, Regel (represented in the Gray Herb. and by sheet
No. 25,841 in the U.S. Nat. Herb.). These specimens also
match too closely the original description and _ figure of var.
Tauschv, Regel.* Further material of var. Zauschii col-
lected by Maximowicz in Mandschuria and distributed from
the Imperial Herbarium at St. Petersburg (represented in
Gray Herb. and by sheet No. 25,345, U.S. Nat. Herb.) has
the leaves more coarsely and irregularly toothed than the speci-
mens previously cited, but in this it is not distinguishable from
Coville and Kearney’s No. 2,412 of B. kenaica (sheets Nos.
373,618, 373,619, U.S. Nat. Herb.) nor from the Bourgean
specimen, type of B. alaskana; nor does this folage differ
from that of specimens collected by Korjhinsky at Blagovye-
schensk, Amur (represented in Gray Herb. and by sheet No.
273,721 in U. 8. Nat. Herb.). Furthermore, these specimens
cannot be separated by any apparent character from Kamt-
schatkan material sent to the Gray Herbarium by Regel as his
B. latifolia, var. kamtschatica, and they are quite like his
original figure of that variety ;+ while Macoun’s Saskatchewan
specimen (No. 12,952a) and McConnell’s Dawson specimen
(No. 20,327) of B. alaskana, and Coville and Kearney’s Nos.
1,442, 1,444, 1,445, and 1,626 of B. kenaica (sheets Nos.
373,610, 373,611, 373,612, 373,616, U.S. Nat. Herb.) are quite
as inseparable from Regel’s original description and figure of
B. alba, subsp. mandshurica.t It is thus very evident that
Betula kenaica and B. alaskana cannot be treated as local
American trees, but that they are quite identical with the well-
known B. pendula or with the Asian tree which was described
by Siebold in 1830 as B. japonica, and figured in its various
forms by Regel in 1865.

* Bull. Soc. Imp. Nat. Mosc., 1865, 399, t. 7, figs. 11-14 (repr. Bemer-
kungen tiber die Gattungen Betula und Alnus (1866) 12, t. 7, figs. 11-14).

+ Bull. Soc. Imp. Nat. Mosce., 1. ¢c., figs. 16-20.
¢ Bull. Soc. Imp. Nat. Mosc., 1. c., fig. 19.

and Old World Birches. 183

This Asian tree, which presents indiscriminate variation in
the size and toothing of its leaves and the abundance of resin
on its branches, was treated by Miquel* as a variety of the
Eurasian Betula alba ; while by Rehder it is more properly con-
sidered a variety of the subspecies B. pendula (B. verrucosa).
This disposition of the tree seems most advisable, since typical
B. pendula crosses northern Asia and is distinguished from
the trees which Siebold called B. japonica and Regel called

_B. latifolia only by an inconstant tendency of the leaves to
be more cuneate at base. The form of the tree with broad
leaves rounded or truncate at base, common in northern Asia,
and ineluding much of B&B. kenaica and B. alaskana, should be
known, then, as Betula pendula, Roth, var. japonica (Siebold),
Rehder in Bailey Cye. Am. Hort. i, 159.

Reference has been made to the tendency of Letula alba
to develop brown bark when it occurs upon the northwest
coast of America. Similar reddish or brown bark likewise
ordinarily distinguishes the northwestern tree which has been
called B. kenaica (B. alaskana) from the true B. pendula.

But this is not a sufficiently constant character to separate it
specifically from the Asian specimens which so closely match
it in every detail shown in the herbarium. The writer has
been unable to find any clear statement that the bark of B.
pendula, var. gaponica, is either white or brown. Bourgeau,
however, who had previously remarked upon the abundance of
Betula papyrifera on Ile Royale, Lake Superior,+ and who
was consequently familiar with the ordinary white-barked
tree, referred the Saskatchewan tree without question to that
species ;{ and Macoun referred his Prince Albert tree without
comment to L. papyrifera. The fact that these acute collec-
tors of the first mentioned specimens of 4. alaskana saw in
their Saskatchewan trees only the white-barked B. papyrifera
is sufficient evidence that the brown bark of &. alaskana is not
a character to be confidently relied upon. Furthermore, Mr.
Coville writes of B. kenaica under date of March 12, 1901:
“In certain individual trees, however, and perhaps on exposure
to certain climatic conditions the layers of the bark separate
and the bark turns white as in papyrifera. This factor has
made it difficult for the collectors who have observed the tree
to tell whether in upper Cook’s Inlet there is a white-barked
birch distinguishable from kenaica.” Sheets Nos. 373,611,
373,619, and 373,620 in the United States National Herbarium
show strips of lichen-covered bark of B. kenaica no darker

* Miquel, Ann. Bot. Mus. Lugd. Bat. ii, 136.

+ Bourgeau in letter to Sir Wn. Hooker, Palliser, Rep. Brit. N. A. Expl.
Exped. 247.

t Bourgeau, 1. c. 249; see also Hooker, ibid., 260; Gray, ibid., 263; Sulli-
van, ibid., 85, etc.

Am, Jour. Boe OURre SERIES, VoL. XIV, No. 81.—SEPTEMBER, 1902.

184 Fernald—Relationships of some American

than is occasionally seen in B. alba (B. papyrifera) in the
Northeast and already mentioned in the discussion of B. occi-
dentalis. The color of the bark, then, furnishes no satisfac-
tory reason to separate Betula kenaica from the Asian B.
pendula, var. japonica.

Betula pendula (typical).

The broad-leaved var. japonica is not the only form of
Betula pendula which crosses from Europe and Asia into
North America. As already noted, the typical European B.
pendula with leaves cuneate at base extends across northern
Asia. Some of the specimens which have been referred to
B. alaskana are quite inseparable in their leaves, branches,
and strobiles from European specimens; and, furthermore,
this characteristic European form extends southward and east-
ward in America to western Illinois, the Great Lakes, and the
St. Lawrence Valley. Sheet No. 351,017 of the United States
National Herbarium, collected as L. populifolia in woods at
Warren, Illinois, by L. M. Umbach, is not distinguishable from
sheet No. 25,340 from Sungaria in central Asia, nor from
material in the Gray Herbarium collected in Sweden (Slom-
berg, No. 1,691). Sheet No. 261,119 of the U. S. National
Herbarium, collected by Professor James Fowler in Ontario as
B. populifolia, has the strobiles and leaves of B. pendula and
it cannot be separated from material in the Gray Herbarium
from Christiania, Norway (S/yit), and from St. Petersburg,
Russia (2egel). A sheet in the Herbarium of the Geological
Survey of Canada (No. 12,950), collected in the Province of
Quebec as B. populifolia by W. Scott, is quite identical with
sheet No. 149,801 of the U.S. Nat. Herb. from Savoy. Vari-
ous other leaf-specimens from Quebec, Manitoba, and other
regions of temperate North America are probably B. pendula,
but without fruit it is at present unwise so to refer them. The
American specimens cited were gathered as the endemic Betula
populifolia, at least one of them in “woods”; and _ since
undoubted &. pendula is found from the Saskatchewan Plains
northward, there seems little question that the European tree,
crossing Asia, is truly indigenous likewise in the northeastern
sections of America, just as are Coptis trifolia, Drosera
rotundifolia, Viburnum Opulus, Lysiumachia thyrsifiora, and
many other well known species* which occur in northern
Europe, central and northern Asia, Japan, northwestern Amer-
ica, and northeastern America.

* Among them Caltha palustris, Viola Selkirkii, Parnassia palustris,
Potentilla palustris, Circaea alpina, Pyrola minor, Moneses grandiflora,
Menyanthes trifoliata, Rumex persicarioides, Allium Schoenoprasum, Juncus
effusus, Eriophorum gracile, Carex filiformis, Hierochloe borealis, etc., ete.

See Gray, Mem. Am. Acad., n. s., vi, 377-499; and Extract ‘‘ Flora of
Japan” in Sci. Pap. of A. Gray, selected by C. S. Sargent, ii, 124-141.

and Old World Birches. 185

BrETULA MICROPHYLLA.

Professor Sargent has recently referred* to the fact that the
writer identifies the common brown-barked tree or half-shrub
of the Rocky Mountains with Bunge’s Betula microphylla
described from the Altai of central Asia. Professor Sargent,
however, was unwilling to identify the American tree with
Bunge’s species, and he accordingly proposed for it the new
name B. fontinalis. But a further comparison of B. fon-
tinalis and an authentic branch from the Altai of Bunge’s
species sent from the Imperial Herbarium of St. Petersburg
to the Gray Herbarium, some of the original Tyan Shan mate-
rial of Regel’s B. alba, subsp. soongoricat in the Gray Her-
barium and in the United States National Herbarium, the
figures of Regel’s B. fruticosa, var. cuneifolia,t and lastly
the original detailed full-page description by Bunge of his
B. microphylla,§ has more firmly convinced the writer that
all three of these trees described from the Altai and the Tyan
Shan ranges are one species, B. microphylla, Bunge; and that
this species differs from B. rhombifolia, Nutt. (B. occidentalis,
Nutt. and of many other authors, not Hook. B. fontinalis,
Sargent) only in growing among the mountains of centrai
Asia rather than among those of northwestern America.

A comparison of Bunge’s original description of Betula
microphylla with Nuttall’s description of his B. occidentalis, |
the tree subsequently called by Sargent L. fontinalis, shows
no definite character of leaves, strobiles, samaras, nor twigs
upon which the two may be separated; and Bunge’s words,
“ epidermide trunci flavescente neque alba, primo intuito dis-
tinctissima,’”4{ are certainly not inapplicable to the small tree of
northwestern America described in the Botany of California
*‘ with close dark-colored bark (at lenyth light brown).”’** Fur-
thermore, if we compare with the branch of B. microphylla
sent from St. Petersburg to the Gray Herbarium the original
specimen of Nuttall’s 4. rhombifoliat+ (preserved in the Gray
Herbarium and referred by Sargent to B. fontinalis), and
such specimens as Maconn’s from the North Saskatchewan and
Dawson’s from the Columbia River (Herb. Geol. Surv. Can.,
Nos. 23,620, and 23,624) or F. W. Anderson’s material from
McCarthey Mts., Montana (U.S. Nat. Herb. No. 25,261), we
shall be perplexed to make out points of distinction. If we
also compare with some of the original material of Regel’s
B. alba, subsp. soongorica, 8 microphylla from 5,000 ft. in

* Bot. Gaz. xxxi, 239.

+ Bull. Soc. Imp. Nat. Mosc. (1868) vi—repr. Enun. Pl. Semenow. (1869) 99.
¢ Regel, Mon. Bet. (1861) 35, t. 7, figs. 16-23.

S St. Pétersb. Mém. Savans Etrang. ii (1835) 606—reprint, Fl. Alt. Suppl. 84.
| Nutt., Sylva, i, 22, t. 7 (1842). “| Bunge, 1. ¢.

** Watson, Bot. Cal. ii, 79. ++ Nutt., 1. c. 24, t. 8.

186 Fernald— Relationships of some American

the Tyan-Shan Mountains as represented in the Gray Herba-
rium and by sheet No. 25,339 in the U.S. National Herbarium,
such specimens as V. Bailey’s No. 5 from South Dakota,
Coville and Leiberg’s No. 81 from Nevada, and R. 8. William’s
No. 404 from Montana (sheets Nos. 229,983, 275,933, and
290,087, U. S. Nat. Herb.); or Macoun’s Devil’s Lake plant,
Dawson’s South Kootenai Pass Plant, and Macoun’s Crow
Nest Pass Plant (Nos. 2,050, 23,623, and 24,368, Herb. Geol.
Surv. Can.); or Lake and Hull’s No. 790 from Washington,
Rydberg and Bessey’s No. 3,928 from Montana, Rydberg’s No.
1,006 from the Black Hills, and A. Nelson’s No. 1,647 from
Wyoming, we shall be further perplexed in separating the
American species from the Asian. Similar comparisons of
American specimens with the figures of B. fruticosa, var.
cuneifolia (which by Regel was identified with B. microphylla),
lead to the same result. In view of this evidence and the
essentially identical descriptions of Nuttall and Bunge, the
writer is still unable to see in Betula fontinalis, Sargent (B.
occidentalis, Nutt., not Hook. B. rhombifolia, Nutt., not
Tausch) anything but B. microphylla, Bunge, of the moun-
tains of central Asia. |

§ Nanae. Shrubs: wings of the samaras narrower
than, or very rarely as broad as the achenes.

The Dwarf Birches like the Canoe Birches present such
tendencies to intergradation that it is difficult to draw clear
specific lines between them. Yet in America three fairly
marked species or centers of variation can be distinguished.
These are Betula pumila, L., with the young shoots normally
pubescent with long soft hairs, and quite glandless, but in an
extreme form with glands or resiniferous atoms mixed with
the long pubescence; B. glandulosa, Michx., with the young
shoots glandular or resiniferous, at most puberulent with close
short hairs; and B. nana, L., a tiny shrub with the young
shoots puberulent or finely pubescent, but not glandular. In
its typical form confined to arctic and alpine regions of Green-
land, Europe and Asia, 6. nana is represented in America by

Bretunta NANA, var. MicHAvuxit.

A very dwarf birch with cinereous-puberulent glandless
branches and tiny suborbicular or flabelliform leaves has been
collected at various points in Newfoundland, Labrador, and the
Hudson Bay region, and has been referred to Betula nana, L.,
or its var. flabellifolia, Hook. An examination of this material
shows that the strobiles are made up of simple oblong scales,
instead of the deeply three-lobed scales characteristic of B.
nana and most other species of the genus.

and Old World Birches. 187

This dwarf shrub of Newfoundland and Labrador is with-
out question Betula Michaux, Spach,* based upon the B.
nana of Michaux, not L., and made by Spach the type of his
section Apterocaryon. Subsequently the plant was raised to
generic rank by Opiz,t and called Apterocaryon Michauan ;
while, on the other hand, Regel in his first Monograph treated
the plant as a variety of B. nana, though he later recognized§
it as distinct. Habitally the plant 1s quite inseparable from
the European B. nana; and since sheet No. 334,540 of the
United States National Herbarium, from Nugsuak Peninsula
in Greenland, shows strobiles with simple and variously divided
seales, it seems that Regel’s earlier treatment of the plant was
wiser and that the Newfoundland and Labrador representative
of Betula nana is var. Michauxii (Spach) Regel.

BETULA PUMILA.

In its normal form Betula pumila, L.| has the leaves and
branchlets quite glandless and in their youngest stages densely
pubescent with silky hairs. The shrub, however, passes imper-
ceptibly into a state (var. glabrescens, Regel) in which the
branchlets and leaves are quite glabrous. This tendency is
common, but apparently not of such constancy as to merit
special recognition. It is interesting to find, however, that the
specimens which represent this tendency are inseparable from
B. alpestris, Fries,§| of Greenland, Iceland, Scandinavia and
north Germany, an identity which was at least suspected by
Regel.** Such specimens in the Herbarium of the Geological
Survey of Canada as Macoun’s No. 23,840 from Anticosti and
Waghorne’s No. 23,821” from Labrador are quite inseparable
from sheet No. 149,806 in the U.S. Nat. Herb. of B. alpestris
from the Dovre Mts., Norway (AA/lberg) and a sheet in the
Gray Herbarium from Lapland (Andersson), except in the
length of the petiole. This character, however, is very incon-
stant and it seems to the writer an insufficient point on which
to keep apart two plants which are otherwise inseparable.
Nor are the extreme American plants otherwise different from
the plate representing B. fruticosa, var. humilis, Reichenbach
(Ie. Fl. Germ. xii, fig. 1280) and referred by Guerke to B.
alpestris, and the plate of B. alpestris in Flora Danica (Suppl.
t. 87). The name 4. pumila, however, long antedates B.
alpestris, Fries, and it should now be applied to the shrub of
northern Europe as well as America. The true Betula pumila
occurs in swamps from Lasrapor and NewrounDLAND west-

* Ann. Sc. Nat., ser. 2, xv, 195. + Lotos, v (1855), 258.
¢ Regel, Mon. Bet. (1861) 45.
§ Regel in DC. Prodr. xvi, part 2 (1864), 171. || Mant. 124 (1767)..

| Summ. Veg. Scand. i, 212 (1846). ** Regel:in-DC., 1. e. 173.

188 Fernald—Relationships of some American

ward, and locally south to Morris Co., New Jersry, Champaign
Co., Onto, Lake Co., Inprana and McHenry Co., Iturors, in
Greenland, northern Europe and Siberia. The leaves vary
much in outline and in the degree of permanence of the pubes-
cence ; and the samaras in the breadth of the wings and achenes.
Narrowly obovate and orbicular leaves are often found upon
the same shrub, so that species or varieties based upon these
characters (as L. Grayi, Regel, Bull. Soc. Imp. Nat. Mose.
Xxxvill, 406, t. 6, figs. 9-13) have little value.

Betula pumila, var. glandulifera.

In the Great Lake region, however, and from there north-
ward and westward, where Letula pumila meets B. glandu-
losa, it presents a perplexing form. In its long pubescence
(when well developed) the shrub seems to be B. pumala, but
mixed with the pubescence and sometimes upon the leaves are
the characteristic glandular atoms of &. glandulosa. Ordi-
narily, though, the shrub is readily distinguished from the latter
species by the longer pubescence of the young shoots. This
intermediate and transitional form, Betula pumila, var. glan-
dulifera, Regel in DC. 1. c., occurs from western Ontario and
Micuican to Minnesota, SASKATCHEWAN and British Cotum-
BIA, south to Inano and Orxgeon.

BETULA GLANDULOSA.

Betula glandulosa* presents two variations which in their
extreme developments appear very distinct, but which again
so mingle their characters as to be quite inseparable. Typical
B. glandulosa is an upright shrub sometimes several feet high,
but in exposed situations it becomes dwarfed and widespread-
ing. In the typical form of the species the leaves are obovate,
but occasionally on these shrubs orbicular or reniform leaves
are found. This shrub with obovate leaves is most common in the
interior of North America, from the Yuxon and MackEnziz
Rivers to Hupson Srrairs, south in the mountains to northern
CaiForniA, Uran, Cotorapo, and Sourn Dakota; across the
plains to Manrroza and Laxse Superior; and through LasBra-
por and locally to the higher mountains of northern New
Enexianp. It is also in Greentanp, KamTscnatTKa and the
Autat Mountains.

On the eastern mountains (Albert, Katahdin and Washing-
ton), in Labrador, and in the Altai the shrub passes imper-
ceptibly to a form with orbicular or reniform leaves, var.
rotundifolia, Regelt (Betula rotundifolia, Spacht); but on

* Michx., Fl. Bor.-Am. ii, 180 (1803). + Regel in DC., 1. c. 172 (1864).
¢ Spach, 1. c. 194 (1841).

and Old World Birches. 189

the coast of Alaska, the islands of Behring Sea, and adjacent
Kamtschatka this depressed form with small orbicular or reni-
form leaves retains its characteristics in a marked degree, and
were the plant known only from that district it would stand as
an undoubted species.

Mr. Coville has called the attention of the writer to a shrub
four to six feet high which is associated with the depressed
Betula glandulosa, var. rotundifolia, and the Cook’s Inlet
tree which is identified with B. pendula, var. japonica. This
intermediate shrub is well represented in the National Herba-
rium, and it is possible, as Mr. Coville suggests, that it is of
hybrid origin, since similar hybrids of trees and dwarf shrubs
have before been noted.* The material presents a strong super-
ficial resemblance to Greenland specimens of B. alba, var.
minor, although the strongly resiniferous branchlets hardly
place it with that shrub.

In conclusion, it should be emphasized that the specitic lines
in Betula, as in Alnus, Quercus and Salix, are often too vague.
It is quite possible to trace by a series of specimens a direct
connection between the dwarf Betula nana or B. glandulosa
and the tall B. alba. Thus &. nana in its larger development
is separated with difficulty from the Scandinavian B. alpestris.
This shrub, in turn, is quite like glabrate states of the Ameri-
can B. pumila, which, through its var. glandulifera, passes to
B. glandulosa, the larger developments of which pass in the
Cascade Mts. to B. microphylla, and in the Saskatchewan
region to B. alba, var. minor. The latter shrub is often
inseparable on the New England mountains from &. alba, var.
cordifolia, which on the lower slopes becomes a large tree and
passes gradually to the broad-leaved form figured by Michaux
as B. papyracea. <A very similar series is readily made to
include 6. pendula and B. humilis. But since it is obvi-
ously impracticable to regard all these forms as one species, it
seems wiser to recognize the more marked centers of variation
as species which are > admitted to pass by exceptional tendencies
to other forms ordinarily distinguished by marked charac-
teristics.

The American representatives of § Costatue, Betula nigra,
B. lenta, and B. lutea, are represented by related species in
Asia, but none of these trees are of very boreal range, and
they appear well distinguished as endemic species.

* B. pubescens [alba] x humilis, Warnst. Verh. Bot. Ver. Brandenb. xi,
129 (1870).

B. nana x verrucosa [pendula], Sael. Medd. Soc. Faun. et. Fl. Fenn. xiii,
256 (1886).

B. nana x pubescens [alba], Koehne, Deutsche Dendr. 112 (1893).

B. pumila x lenta, Jack, Gard. and For. viii, 243, fig. 36 (1895).

190 Fernald--Lelationships of some American

The American forms of §§ Albae and MNanae now recog-
nized by the writer and discussed in the preceding notes may
be briefly enumerated as follows:

B. arpa, L. Sp. 11, 982 (1753); Roth, Fl. Germ. i, 404. B.
papyrifera, Marshall, Arbust. Am. 19 (1785). B. papyracea,
Ait. Hort. Kew. in, 337 (1789). .B. pubescens, Ebrh. Beitr. v,
160, vi, 98 (1790-91). B. odorata, Bechst. Diana, i, 74 (1797).—
NeEwFounpDianp to ALasKa, south to Pennsytvanta, INDIANA,
Nerpsraska, Wyomine, Ipano and WasuineTon ; passing on
the Pacific coast to the dark-barked forma occidentalis (B.
occidentalis, Hook. Fl. Bor.-Am. ii, 155 (1839)); IceLann,
northern Evropr and Asta, south in the mountains to northern
Spain, Iraty, ete. (Pl. V, figs. 1-6.)

B. ALBA, var. GLUTINOosA, Trautv. ex Regel, Mon. Bet. 20
(1861). B. glutinosa, Wallr. Sched. Crit. 497 (1822). JB.
pendula, Reichenb. Ic. Fl. Germ. xu, t. 625, not Roth.—
Valley of Wassataquoik River, Marinz; Swepen, Finuanp,
GERMANY, SWITZERLAND, AUSTRIA.

B. auBa, var. corpirotia. LB. cordifolia, Regel, Mon. Bet.

28, t. 12, figs. 29-386 (1861). BL. alba, subsp. papyrifera, B,
cordifolia, Regel in DC. Prodr., xvi, pt. i, 166 (1864). B.
pupyriferd, var. minor, Wats. and Coult. in Gray, Man. ed. 6,
472 (1889), in part, not B. papyracea. var. minor, Tuck.—
Lasrapor and Nrwrounpnanp to New Brunswicx, Mains,
New Hampsuire, Lake Superior, lowa, ALBERTA, BrRitisH
CotumetiA, Ipano, and WAsHINGTON.
_ B. avpa, var. minor. B. davurica, Ledeb., Fl. Alt. iv, 245
(1833), not Pallas. LB. papyracea, var. minor, Tuckerman,
this Journal, xlv, 31 (1848). B. tortwosa, Ledeb., Fl. Ross. iii,
652 (1849). B. edorata, var. alpigena, Blytt, Norg. Fl. 402
(1861). &. alba, subsp. tortuosa, Regel, in DCS iter aiae
(1864). B. davurica, B, americana, Regel, |. c. 175 (1864).
B. odorata, var. tortuosa, Lange, Fl. Dan. xvu, 10, t. 2918
(1877). B. papyrifera, var. minor, Wats. and Coult. 1. ¢., in
part. BL. pubescens, var. tortuosa, Koehne, Deutsche Dendr.
109 (1893).—Laxsrapor to the mountains of Mainz, New
HamrsHire and ‘VERMONT; mountains of SASKATCHEWAN,
Asstntpora, and ALBERTA; GREENLAND, IckLAND, LAPLAND,
Finan, northern Germany, Attar Mrs.; dwarf forms from
Auaska resemble this variety, but have the strongly resini-
ferous branchlets of B. glandulosa and B. pendula. (Pl. V,
figs. 7-12.)

Betula ALBA, var. carPATIcA. LB. carpatica, Wald. and
Kit. in Willd. Sp. iv, 464 (1805). B. borealis, Spach, Ann.
Sci. Nat., ser. 2, xv, 196 (1841). B. glutinosa, Fries, Summ.
Veg. Seand., 212 (1846), not Wallr. &B. pumila, y, borealis,
Regel, Mon. Bet. 55, t. 18, figs. 38, 39 (1861). B. alba, subsp.

~~. ps eae

weet. 2) Per

and Old World Birches. 191

pubescens, [ car "‘patica, Regel in DC., 1. c. 168 (1864). B.
odorata, var. carpatica, Lange, Haandb., ed. 3, 708 (1864).
vif odorata, var. rhombifolia, Lange, Fl. Dan., xvi, t. 2851

(187 1).— Anticosti, Quesec; ALBertTa; and northward; Scay-

DINAVIA to Germany, Austria and Russia. (PI. V, figs. 13-14.)

B. penpura, Roth, Fl. Germ. i, 405 (1788). B. verrucosa,
Ebrh., Beitr. vi, 98 (1791). B. alba, var. verrucosa, Wallr.,
Sched. Crit., 495 (1822). B. alba, var. vulgaris, Spach, 1. ¢.
186 (1841). B. odorata, Reichenb., Te. Fl. Germ., xii, fig.
1288 (1850), not Bechst. JB. gunmifera, Bertol., Fl. It., x,
229 (1854). B. alba, subsp. verrucosa, a, vulgaris and 6, resin-
ifera (in part), Regel ; in DC., 1. c. 163, 164 (1864). B. kenaica,
W. H. Evans, Bot. Gaz., xxvii, 481 (1899), in part. B. alas-
kana, Sargent, Bot. Gaz., xxxi, 236 (1901), mostly —QuvEBEc to
Inirots, ALBerTA, Mackenzie and Ataska; Europe and Asia,
widely distributed. (Pl. V, figs. 15-18; Pl. VI, figs. 19-22.)

B. PENDULA, Var. JAPONICA, Rehder in Bailey, Cyc. Aan.
Hort., i, 159 (1900). B. japonica, Siebold, Verh. Batay. Gen.
xl, 25 (1830). B. latifolia, Tausch, FI. Ratisb., 751 (1838).
Bs “alba, subsp. verrucosa, 8 resinifera, Regel in DC., |. ¢. 164,
in part; subsp. mandshurica and latifolia, Regel, 1. ¢. 165
(1864). B. kenaica, W. H. Evans, 1. c. (1899), mostly. B.
alaskana, Sargent, |. c. (1901), in part.—Sunearia to Japan,
north to the Yenisei River, Srperta, and KamrscHatTKa ; east
to Peel River and Fort Simpson, Mackenzie, and along the
ALASKA coast; and apparently on the coast ‘of Washington
Co., Marne (trees sterile). (Pl. VI, figs. 23, 24.)

B. POPULIFOLIA, Marshall, Arbust. Am. ,19 yar: B. alba,
var. populifolia, Spach, Ann. Sci. Nat., ser. 2, xv, 187 (1841).
B. alba, subsp. populifolia, Regel, in DC., 1. ¢. 164 (1864).—
Prince Epwarp Isianp to central Marne, southwestern QUE-
BEC, and western New Yorx, south mostly on the coastal plain
to DELAWARE. :

B. MicropaHyiia, Bunge, St. Pétersb. Mém. Savans Etrang.,
ii (1835), 606-—reprint, Fl. Alt. Suppl. 84. B. occedentalis,
Nutt., Sylva, i, 22, t. 7 (1842), not Hook. B. rhombifolia,
Nutt., 1. ¢. 24, t. 8, not Tausch. B. fruticosa, var. cuneifolia,

. Regel, Mon. Bet. (1861) 35, t. 7, tigs. 16-23. _B. alba. subsp.

soongorica, Regel, Bull. Soe. Imp. Nat. Mose., vi, (1868)—
reprint, Enum. Pl. Semenow (1869) 99. B. fontinalis, Sar-
gent, Bot. Gaz. xxxi, 23 (1901)—-In the mountains from
Aaska to northern CaLIFornta, SASKATCHEWAN, SoutH Da-
Kota, and New Mexico; Atral and Tyan SHAN Mrs., central
Asia. (Pl. VI, figs. 25-32.)

B. nana, L., var. Micwavxm, Regel, Mon. Bet. (1861), 45.
B. Michaucii Spach, ].c.195 (1841). “Apter ocaryon Michauzis,
Opiz. Lotos, v (1855), 258.—Hupson Bay, Laprapor and
Newrounpianp. (PI. VI, figs. 47, 48.)

192 Fernald— Relationships of some American

B. pumiva, L., Mant. 124 (1767). 2B. alpestris, Fries, Summ.
Veg. Scand., i, 212 (1846). SB. fruticosa, var. humilis,
Reichenb. Ie. Fl]. Germ., xii, fig. 1280 (1850). 8B. nana var.
alpestris, Regel, Mon. Bet. 45 (1861). &. Grayz, Regel, Bull.
Soe. Imp. Nat. Mosce., xxxviii, 406, t. 6, figs. 9-13 (1865),—

Lasrapor and.NrwrounpLanp to Ontario, south to Morris

Co., New Jersny, Champaign Co., Outo, Lake Co., Inprana,
and McHenry Co., Intinois; GREENLAND, SCANDINAVIA, north-
ern GERMANY, Fintanp, Lartanp, Russia, Srperta. (Pl. VI,
figs. 33-38.)

B. PUMILA, Var. GLANDULIFERA, Regel in DC., 1. c. 178 (1864).
—Ontario and Micuiean to Minnesota, SaAskKATCHEWAN and
British Cotumpia, south in the mountains to Ipano and
OREGON.

B. etanputosa, Michx., Fl. Bor.-Am., ii, 180 (1808). B.
nana of various American authors, not L. £. Littelliana,
Tuckerman, this Journal, xlv, 30 (1843).—GreEENLAND and
Hupson Srrairs to KamtTscnatKa, and the Altai Mts. of
SIBERIA; south in eastern America to the higher mountains
of Marne and New Hampsuire; in the interior to LaxKsE
Superior and Manrropa; and in the western mountains to
SoutH Dakota, Cotorapo, Utan, and northern CALirornia.
(Pl. VI, figs. 39, 40.)

B. GLANDULOSA, var. ROTUNDIFOLIA, Regel in DC., 1. ¢. 172
(1864). B. rotundifolia, Spach, 1. c. 194 (1841). B. nana, var.
sibirica, Ledeb., Fl. Ross., 11, 654 (1849).—GreEENLAND and
northern Laprapor to KamrscHarKa, and the Altai Mts. of
Stperra, south to the higher mountains of Marne and New
Hampsuire, and along the coast of ALaska. Perhaps hybrid-
izing with B. pendula. (Pl. VI, figs. 41-44.)

tn 6

and Old World Birches. 193

EXPLANATION OF PLATES.*

Puate V, figs. 1-6, BETULA ALBA.

Figure 1.—Leaf of young flowering branch of B. pubescens, Ehrh., from
the Vosges Mts., near Ramberviiles, France (Joad).

Figure 2.—Leaf of fruiting branch of B. papyrifera, Marsh., from Lake
Winnipeg, Manitoba (Bourgeau).

Figure 3.—Leaf of fruiting branch of B. occidentalis, Hook., from Pend
d’Oreille River, British Columbia (Lyatv).

Figure 4.—Leaf of young branch of B. pubescens, Ehrh., from Bohemia
(Tausch).

Figure 5.—Leaf of fruiting branch of B. papyrifera, Marsh., from South-
port, Maine (Fernald).

FicurE 6.—Leaf of fruiting branch from Austria—after Ettingshausen and
Pokorny, Phys. Pl. Aust. iii, t. 201.

Puate V, figs. 7 to 12, B. ALBA, var. MINOR.

Figure 7.—Leaf of fruiting branch of B. tortuosa, Ledeb., from the Altai
Mts., Siberia (ex. herb. St. Petersb.).

Figure 8,—Leaf of fruiting branch of B. odorata, var. tortuosa, Lange,
from the Dovre Mts., Norway—after Lange, Fl. Dan. xvii, t. 2918.

Ficures 9, 10.—Leaves of original fruiting branches of B. papyracea, var.

: minor, Tuck., from Mt. Washington, New Hampshire.

Ficure 11.—Leaf of dwarf form of B. papyracea, var. minor, Tuck., from
Mt. Washington, New Hampshire (Faxon).

Ficure 12.—Leaf of fruiting branch of B. odorata, var. minor, Rosenvinge
in herb., from Greenland (Hartz).

PuaTE V, figs. 13, 14, B. ALBA var. CARPATICA.

Figure 13.—Leaf of fruiting branch of B. borealis, Spach, from Anticosti,
Quebec (Macoun).

Figure 14.—Leaf of B. alba, subsp. pubescens, ¢ carpatica, Regel, from
Lapland (Regel).

PuateE V, figs. 15 to 18, Puate VI, figs. 19 to 22, B. PENDULA.

FicuRes 15, 17.—Leaves of young flowering branch of B. alba, subsp. ver-
rucosa, 0 resinifera, Regel (type of B. alaskana, Sargent) from
Saskatchewan (Bourgeau).

Ficures 16, 18.—Leaves of fruiting branch of B. verrucosa, Ehrh., from
Christiania, Norway (Blytt).

Figure 19.—Leaf of fruiting branch of B. pendula, from Warren, Illinois
(L. M. Umbach, sheet No. 351,017 U. S. Nat. Herb.).

Figure 20.—Leaf of fruiting branch from Gotland, Sweden (Blomberg).

FIGURE 21.—Leaf of B. alba, subsp. mandshurica, Regel, from Mandschuria—
after the original illustration (Regel, Bull. Soc. Imp. Nat. Mosc.,
1865, t. 7, fig. 15).

FIGURE 22.—Leaf of fruiting branch of B. verrucosa, 6 resinifera, Regel,
from Nushagak River, Yukon District, Alaska (McKay).

PuaTE VI, figs. 23, 24, B. PENDULA, var. JAPONICA.

Figure 23.—Leaf of B. alba, subsp. latifolia, a Tauschii, Regel, from Eastern
Asia—after Regel, Bull. Soc. Imp. Nat. Mosc., 1865, t. 7, fig. 11.

Figure 24.—Leaf of B. kenaica, W. H. Evans, from Cook Inlet, Alaska
(Coville and Kearney, No. 2412).

*The plates illustrating this paper have been carefully prepared by Mr.
F. Schuyler Mathews. Unless otherwise stated, the figures are life-sized.

194. . Fernald—American and Old World Birches.

PuaTE VI, figs. 25 to 32, B. MICROPHYLLA.

FIGURE 20.—Leaf of young shoot of B. microphylla from the Altai Mts.,
Siberia (ex. herb. St. Petersberg).

FicurE 26.—Leaf of B. fruticosa, var. cuneifolia, Regel (syn. B. microphylla,
Bunge) from the Altai Mts., Siberia—after Regel, Mon. Bet. (1861)
tii, eS:

FIGURE 27.—Leaf of fruiting branch of B. occidentalis, Nutt. (B. fontinalis,
Sargent) from Idaho Springs, Colorado (Engelmann).

FIGURE 28.—Small leaf from fruiting branch of B. occidentalis, Nutt. (B.
fontinalis, Sargent) from Coulee City, Washington (Lake and
Hull, No. 790).

FIGURE 29.—Leaf of young shoot of B. occidentalis, Nutt. (B. fontinalis,
Sargent) from the Black Hills, South Dakota (Rydberg, No. 1006).

Figure 30.—Leaf of fruiting branch of original material of B. alba, subsp.

soongorica, 8 microphylla, Regel, from the Tyan Shan Mts.,.

Central Asia (Semenow).

Figures 31, 32.—Leaves of fruiting branch of B. occidentalis, Nutt. (B.
fontinalis, Sargent) from Laramie Peak, Wyoming (A. Nelson,
No. 1647).

Puate VI, figs. 33 to 38, B. PuMILA.

Ficure 33.—Leaf of sterile shoot of B. pumila from Anticosti, Quebec

(Macoun).

Figure 34.—Leaf of sterile shoot of B. alpestris, Fries, from Lapland
(Laestadius).

FiguRE 35.—Tip of branch from Bonne Espérance, Quebec (Allen, No. 70).

Figure 36.—Leaves of glabrate form from Bay of Islands, N ewfoundland
(Waghorne).

Figure 37.—Leaves of B. alpestris, Fries, from N BE ee Lange, FI.
Dan., Suppl. t. 37.

Figure 38.—Leaves of young branch of B. alpestris, Fries, from Lapland
(Andersson, No. 186).

PuateE VI, figs. 39, 40, B. GLANDULOSA.

F1GuRE 89.—Leaf of branch from Mt. Washington, New Hampshire (J. A.
Allen).
Figure 40.—Leaves of branch from Hopedale, Labrador (Sornborger, No.

80).

PuatE VI, figs. 41 to 44, B. GLANDULOSA, var. ROTUNDIFOLIA.

Ficure 41.—Leaf from Mt. Washington, New Hampshire.
Ficures 42, 48.—Leaves from Nunivak Island, Behring Sea (J. M. Macown).
Ficure 44.—Leaf from the Altai Mts., Siberia (ex. herb. St. Petersb.).

PuaTE VI, figs. 45, 46, B. Nana.
Figure 45.—Fruiting branch from Grenjadastad, Iceland (Elizabeth Taylor).
Figure 46.—Scale from same, enlarged three times.

PuatE VI, figs. 47, 48, B. nana, var. MICHAUXII.

Ficure 47.—Fruiting branch from Grand Lake, Newfoundland (Waghorne).
Ficure 48.—Scale from same, enlarged three times.

Plate VI.

Am. Jour. Sci., Vol. XIV, 1902.

21

a

19

22

Sellards—Fronds of Crossotheca and Myriotheca. 195

Art. XXII.— On the Fertile Fronds of Crossotheca and Myrio-
theca, and on the Spores of other Carboniferous Ferns,
From Mazon Creek, Illinois; by E. H. SELLARDS. (With
Plate VII.)

INFORMATION in regard to the spore-bearing organs of Car-
boniferous ferns has accumulated slowly and with difficulty.
The parts of the plants are usually disconnected, and more or
less fragmentary. Dimorphic genera are not uncommon, and
specimens connecting the fertile and sterile segments or fronds
are rare. It is usually difficult to correlate genera described
from microscopic structure with others based on plant impres-
sions. And yet, a satisfactory knowledge of a considerable
number of genera and species has resulted from the work of
the various investigators who have taken up this subject since
the time of Brongniart. These investigations indicate that
ferns with the annulus absent or but slightly developed, and
having other Marattiaceous characters, predominated in the
Carboniferous, and included a much greater range of form and
structure than is seen in the living representatives of the
family. Besides the numerous exannulate ferns, others are
known with a well-developed annulus, and are, therefore, pre-
sumably Leptosporangiate. The reference of these fossils to
their respective families of living ferns is attended with more
or less doubt. The Hymenophyllacez, Gleicheniaces, Schize-
acese, and Osmundacez have been recognized with some degree
of certainty.

Prof. Renault, of the Museum of Natural History of Paris,
has recently described Parkeriordea Renault, a genus from the
Coal Measures of Grand’ Croix, near St. Etienne. The char-
acter of the annulus, the form of the spores, and the ornamen-
tation of the exospore have led him to refer this form to the
Parkeriaceze. Certain of the spores show sculpturing, while
others are smooth and have three radiating lines at the apex.
The former are interpreted as microspores, the latter as mega-
spores.”

The same writer had previously detected what he believed
to be indications of heterospory in Pecopteris, one of the
Marattiaceous ferns,t as well as in an extinct family of ferns,
the Botryopterides.{t The evidence of heterospory in the
Botryopterideze has not been fully accepted. The papers on
Pecopteris and Parkerioidea are mentioned more fully later
in the present article, where, in connection with the descrip-

* Comptes Rendus de l’Academie des Sciences, March 10, 1891.

+ Ibid., October 21, 1891. t Bull. Soc. Hist. Nat., Autun. , iv, 1891.

§See Seott, Studies in Fossil Botany, p. 289 ; Zeiller, Eléments de Paléo-
botanique, p. 74.

Am. Jour. Sct.—FourtTH Srrigs, Vou. XIV, No. 81.—SEPTEMBER, 1902.

196 Sellards—Lronds of Crossotheca and Myriotheca.

tion of the spores of some Carboniferous ferns, it is pointed
out that, as far as the radiating lines (“lines of dehiscence ”’)
at the apex are concerned, they cannot be considered character-
istic of megaspores. :

The fructification of several genera of Carboniferous ferns
is well shown in an exceptionally large and complete collec-
tion from Mazon Creek, Illinois, in the Yale Museum. The
sporangia are often preserved, and in many cases the spores
are found in place, and can be removed and studied. In the
present paper the Mazon Creek representatives of two inter-
esting genera, Crossotheca and Myriotheca, will be described,
together with the spores of some other species.

In the earlier American literature many of the fertile ferns
were grouped together without sufficient regard to their true
generic separation. Sorocladus, as established by Lesquereux,*
was so broadly defined as to include more than one natural

genus. WS. stellatus was the first of the five species described .

under the genus. Four of the species referred in the “ Coal
Flora” to Sorocladus had been previously placed by Les-
quereux in the Tertiary Staphylopteris Presl. It was probably
Schimper’s objection to considering these forms under Presl’s
genus that led Lesquereux to create aseparate genus for them.t

One of the five original species of Sorocladus, S. ophioglos-
soides, has been referred by David White to Crossotheca, and
S. sagittatus was recognized as falling naturally within the same
genus.

Crossotheca.

Zeiller, Ann. Sci. Nat. (Bot.), ser. 6, vol. xvi, 1888.

Crossotheca is a genus of more than ordinary interest, because
of its dimorphic fronds, its large marginal sporangia, and large

spores. Several species are known in Europe, all of Coal —

Measure and Permian age. Besides the two species just men-
tioned from this country, a third is added in the present paper.

Crossotheca sagittata.
Plate VII, figures 1-3c, 8.

Staphylopteris sagittatus Lesq., Geol. Surv. Il., vol. iv, p. 407, pl. xiv,
figs. 4-6, 1870.

Pecopteris abbreviata Brongn. ? Lesq., Geol. Surv. II1., vol. iv, p. 403, 1870 ;
Second Geol. Surv. Penn., Description of the Coal Flora, vol. i, p. 248,
pl. xlvi, figs. 4-6a, 1880. |

Sorocladus sagittatus Lesq., Coal Flora, vol.i, p. 329, 1880; Atlas, pl.
xlviii, figs. 10-100, 1879 ; vol. iii, p. 761, pl. C, figs. 4-5, 1884.

Pecopteris Fontainei, Lesley’s Dict. of the Fossils of Penn., p. 606, 1889
text figure. See, also, Lesquereux, unpublished manuscript.

* Second Geol. Surv. Penn., Description of the Coal Flora, vol. i, p. 327,
1880.

+ Paléont. végét., vol. iii, p. 512.

tMon. U. S. Geol. Surv., No. 37, Flora of the Lower Coal Measures of
Missouri, pp. 60-64, 1899.

or"

Sellards—Fronds of Crossotheca and Myriotheca. 197

Crossotheca sagittata preserves the details of fructification
much better than the other American species, and illustrates
well the characters of the genus. The large fertile pinnules
are expanded at the base in the form of an arrow, thus allow-
ing greater area for the attachment of the sporangia. The
small pinnules are slightly or not at all enlarged at the base.
The upper surface of the pinnule is flat with a distinct median
line, and with the lateral veins obscured. The sporangia are
unusually large, measuring 23 to 47™ in length and $ to ?™
wide. They are placed as seen in figures 1, 2, and 3, in a single
row around the entire border of the pinnule, free nearly or
quite to the base, and are often seen filled with spores.

Figure 2 gives a side view of the pinnule as partly freed
from the matrix, and showing the full length of the sporangia.

From a study of the type of the genus, Prof. Zeiller thought
it probable that the sporangia were united in little clusters at
the ends of the nerves. The specimens figured and others in
the Yale collection indicate that in the case of C. sagittata, at
least, the sporangia. are attached side by side in a single row,
without any tendency toward grouping. The same specimens
confirm the statements of Zeiller that these are individual spo-
rangia, since, in the specimens at hand, they are often filled
with spores, in contradistinction to Stur’s interpretation of the
fringed pinnules as dehisced sporocarps.*

Some of the best preserved sporangia show a slit on the
outer side, as seen in figure 3a, which probably indicates the
place of dehiscence.

The sterile part of the frond is very different from oe fer-
tile, so much so that if not found in direct connection their
relation would hardly be suspected. The pinnules are small,
rounded, close, oblique, connate, and decurrent at the base, the
smaller entire, the larger becoming lobate. The ultimate
pinnee are broadly linear-lanceolate, alternate, oblique, and
close, often touching. The rachis is large and round. The
midrib of the pinnule-is broad, shallow, and decurrent. The
lateral veins curve regularly to the border, and fork once,
twice, or three times, according to the size of the pinnule.
The surface i is rough, appearing “minutely scaly. Some of the
veins are heavier than others, giving the venation an irregular
appearance.

The extreme apex of the frond is often sterile, the fertile
pinnules and pinne appearing at some distance below. This
is not always the case, however, since in the frond figured. by
Lesquereux (“ Coal Flora, ” volume i iil, pl. C, figure 4) the entire
apical part is fertile. It is not possible to state, on any evi-

* Abhandl. d. k. k. Geol. Reichsanstalt, Wien, 1885-87, Flora der Schatz-
laren Schichten, Part I, pp. 273-275. aa fee

198 Sellards—Fronds of Crossotheca and Myriotheca.

dence at hand, whether the entire frond below the apex was
fertile, or only a few of its segments. There is no indication
of sterile segments below the fertile ones, although some of
the incomplete fronds reach a length of 15 or 16. It is prob-
able that some of the fronds were entirely sterile, and that
others were mostly sporangia-bearing, the apical part only
being, in most cases, sterile.

The spores of this species are large, from *056 to :060™™,
round, and marked at the apex by three distinct radiating
lines. The exospore is thick, resistant, brownish, and marked
by minute warty thickenings.

The sterile fronds were at first doubtfully referred by Les-
quereux in the “ Coal Flora” to Pecopteris abbreviata Brongn.*
In volume ili of the same work, however, a small part of the
sterile apex is figured in connection with the fertile frond, and

David White states that Lesquereux’s unpublished manuscript ©

contains descriptions and figures of the two parts in connec-
tion.t In the Yale collection the fertile and sterile parts are
shown in direct connection in no less than nine instances.
(See figure 8.)

Crossotheca trisecta sp. 0.

PuatE VII, figures 44c, 9.

A second and apparently new species is present in the mate-
rial from Mazon Creek. The sterile part of the frond is much
like that of C. sagittata, but the fertile pinnules are entirely
different. The latter are usually trisectate. The central lobe
is elongate-ovate, or nearly round, and borne at the end of a
slender stalk. The lateral lobes are smaller, round, and borne
on short lateral stalks. A second pair is sometimes borne by
the larger pinnules. Lateral lobes may be lacking in one or
two pinnules near the apex of the pinna. The sporangia are
probably smaller than those of C. sagittata, and are not dis-
tinctly preserved on either of the two fertile fronds in the
present collection. The sporangia-bearing lobes have a form
much like that of the type of the genus C. Crépinz, but the
type species lacks the trilobate appearance of the pinnules,
and has more finely divided sterile fronds having a different
type of venation.t C. ophioglossoides from Clinton, Missouri,
has narrower and longer fertile pinnules.

The lines on the upper surface of the pinnule, present on
other species of the genus, are much more distinct than on C.

* Second Geol. Sury. of Penn., Description of the Coal Flora, vol. i, p.
248; Atlas, pl. 46, figs. 4-6, 1880.

+ Mr. White informs the writer that the name Pecopteris Fontainet, sp.
novy., is given to the sterile fronds of this species in Lesquereux’s manuscript.

t See the figures of C. Crépini given by Zeiller, Ann. Sci. Nat. (Bot.), ser. 6,
vol. xvi, 1888, and by Stur under the name of Sorotheca Crépini, Flora der
Schatzlaren Schichten, pl. xxxv, figs. 3, 4.

Sellards—Lronds of Crossotheca and Myriotheca. 199

sagittata, and sometimes branch. The figured specimen is
15™ long; the first 5 or 6™ are sterile, the remaining pinne
being partly or entirely fertile. When detached and in fragments
the sterile part of the frond is distinguished with difficulty
from that of C. sagittata. The pinnules are perhaps more
finely lobate.

The spores are smaller than those of C. sagitéata, measuring
from -030 to :036™". They are somewhat triangular, with a
smooth, thin exospore.

The name Crossotheca trisecta is suggested for the species.

| Myriotheca.

Zeiller, Ann. Sci. Nat. (Bot.), ser. 6, vol. xvi, 1883.

Myriotheca has numerous independent, sessile, round or
ege-shaped sporangia, covering the entire lower surface of the
pinnule. The genus is represented at Mazon Creek by a single
species, which apparently is the fern described by Lesquereux
from Morris, Ill., as Sphenopteris scaberrima,* although the
rachis is smooth or striate, not punctate as given for that
species. The round sporangia are very numerous, close, or
almost contiguous, half immersed in the leaf substance, and
cover the entire lower surface without any kind of regularity
of arrangement or grouping. The spores are of medium size,
measuring from °036 to -040™", triangular, with the sides some-
times slightly concave. The genus is a rare one both in
Europe and in America. No other species has been reported
from this country. The genus was founded by Zeiiler on a
single fragment from the. Coal Measures of France. The
European specimen representing the type species, JZ. Desazllyi,
has smaller pinnules with a tendency to become lobate. The
sporangia of the American species are nearly round, and larger
than those of the European species, measuring ‘40 to *50™™.

Because of the absence of any indication of an annulus, Prof.
Zeiller included both Crossotheca and Myriotheca with the
Marattiaceze. The large size of the spores and comparatively
small output to the sporangium are, however, characters not
met with in the living representatives of that group. The
position of the sporangia, marginal in Crossotheca, and cover-
ing the whole lower surface of Myriotheca, is unusual for
Marattiaceous ferns.

Spores of other Ferns from the same Locality.

At least four other species of ferns in the Yale collection
have the spores preserved. All retain their natural brown
color, and something of their food contents, and, as far as
appearances are concerned, might be spores from living plants.

* Geol. Surv. of Illinois, vol. iv, p. 408, pl. xv, figs. 1 and 2, 1870.

200 Sellards—Lronds of Crossotheca and. Myriotheca.

Some of the larger spores have granular bodies within or cling-
ing to them, which show a dark spot at the center and some-
thing of the structure of concentric starch grains.

The spores from a large number of fronds of two species
have been examined in order to find whether or not there were
indications of more than one kind of spores. The species
studied were Pecopteris (Ptychocarpus) unita Brongn., and
the form referred doubtfully by Lesquereux to P. villosa
Brongn.,* which probably belongs to the Asterotheca division
of Pecopteris. Both species are extremely abundant at Mazon
Creek, and the sporangia-bearing fronds numerous. The spores
of P. villosa are small, measuring only -013 to -016"", smooth
and spherical. The exospore is very thin. The spores of the
European examples of P. unita have already been made
known by Renault.t The spores of the specimens at hand are
‘016 to -018™" long, and -010 to ‘011 wide, being elongate or
bean-shaped when seen from the side. The exospore is thick
and smooth. The three radiating lines at the apex can be seen
on some spores. As in other Marattiaceous ferns, the spore
output to the sporangium in both species was evidently very
great. The spores from many specimens of both species and
from various parts of the same specimen present no differences
in structure, size, or sculpturing, that could be interpreted as
indicating two kinds of spores. It is, therefore, practically
certain that both were homosporous, and this is exactly what
might be expected in typically Marattiaceous ferns. The fact
is of interest in connection with Renault’s paper referred to
above, in which unusual conditions ‘are observed in one of the
European species of Pecopteris. The fern described by
Renault, which is of the Asterotheca division or sub-genus of
Pecopteris, has one set of spores which are smooth and marked
at the apex by three radiating lines. These are considered
megaspores. In the same sorus, and possibly in the same spo-
rangium, are other spores of about the same size, thought to be
microspores, which lack the lines at the apex and show struc-
tures interpreted as the mother cells of antherozoids.

The second paper by the same author describes somewhat
similar appearances in Parkerioidea. The megaspores are
smooth and show the triradiate lines at the apex. The micro-
spores, found in the same sporangium, lack the lines, and are
sculptured with a polygonal network.

The three radiating lines are seen on the spores of all the
ferns examined by the writer, when viewed from the apex, and

* According to Mr. Robert Kidston, Fossil Flora of the Radstock Series,

Trans. Roy. Soe. Edinburgh, vol. XXxili, Part II, 1886-7, p. 37, P. villosa has a
doubtful existence, having "been established in all probability on a villous
specimen of P. oreopteridia or a Closely related species.

+ Bassin Houiller et Permien d’Autun et d’Epinoe, Pt. II, p. 10.

Sellards—Fronds of Orossotheca and Myriotheca. 201

on both micro- and megaspores of Selaginella, and the Car-
boniferous Lycopods, as well as on such heterosporous living
-ferns as Marsilea. The tetrahedral division of the spore mother
cell, of which the three radiating lines are indicative, is well
_known to be extremely constant, not only for the Pterido-
phytes, but for all those plants commonly grouped under the
Archegoniatez, and for the microspores of most of the flower-
ing plants.*

The lines, therefore, cannot be considered characteristic of,
or in any way distinguishing megaspores. Their absence in
some cases may be due to imperfect preservation, or they may
be obscured by the view of the spore presented. In the case
of Parkerioidea, there is some doubt as to whether it is not
possible that the spores when seen from the apex present the
Jines and a smooth face, and when seen from the base are sculp-
tured, the lines being obscured by the thickness of the spore.

The recent studies of Prof. Bowert have directed attention
to the importance of the size of the spores and the number to
the sporangium. Bower’s investigations show that among liv-
ing ferns an increase in the size of the spores, correlated with
a decrease in the output to the sporangium, accompanies, in a
general way, the advance in development and specialization
from the Marattiacesze through the various families of the
Leptosporangiate ferns. The little that is definitely known of
the spores of fossil ferns supports Bower’s conclusions. It isa
question, however, how far the size and number of the spores
may be relied upon to separate Marattiaceous from non-Marat-
tiaceous ferns.

The spores of Pecopteris villosa are smaller than those of
such living Marattiacezee as Angiopteris evecta, Kaulfussia,
Marattia Douglassii, or Danea moritziana. Crossotheca and
Myriotheca have much larger spores, comparable in size to
many of the Leptosporangiate ferns. It is hardly possible,
with fossil ferns, to count the number of spores to the spo-
rangium, but it is evident that in the case of Pecopteris unita
and P. villosa, the output to the sporangium was very great,
yee that of Crossotheca and Myriotheca was comparatively
small.

Geological Department, Yale University Museum.

* See Campbell, University Text-book of Botany, 1902, pp. 199 and 323.

+ Studies in the Morphology of Spore-Producing Members, Parts III and
IV, Phil. Trans. Roy. Soc., vol. clxxxix, 1897, pp. 35-81 ; vol. excii, 1899,
pp. 29-138.

202 Sellards—Fronds of Crossotheca and Myriotheca.

EXPLANATION OF PLATE VII.

Figure 1.—Crossotheca sagittata ; part of a fertile frond. The lower and
the upper pinnules on the right hand side show the expanded
base. The second pinnule from the base on the same side gives
the full length of the sporangia. x11.

FIGURE 2.—Same species ; a pinnule with the matrix removed to show the
full length of the sporangia. x 2.

FIGURE 3.—Same species; a medium sized pinnule seen from the base and
side. x2,

FIGURE 38a.—A single sporangium. x4.

FIGURE 3b.—Spores of the same species. x 240.

FIGURE 3c.—Cross section of a fertile pinnule with sporangia. x 2.

FIGURE 4.—Crossotheca trisecta sp. n.; fertile and sterile parts of frond. x 44.

Ficuress 4a, 4b.—Fertile pinnules of C. trisecta.

FicureE 4c.—Spore of C. trisecta. x 240.

FIGURE 9.—Myriotheca scaberrima ; fertile pinnules. x 2.

FIGURES 5a—5c.—Spores of the same species. x 240.

Ficures 5d, 5e.—Starch grains from the spores of the same species. x 240.

FIGURES 6a-6d.—Spores of Pecopteris villosa. x 240.

FIGURE 7.—Pecopteris unita ; fertile pinna. x 2.

FIGuREs 7a—7d.—Spores of the same species. x 240.

FIGURE 8.—Crossotheca sagittata; fertile and sterile parts in connection.

FIGURE 9.—Crossotheca trisecta; sterile pinnule. x 2. [ x14.

CARBONIFEROUS FERNS.

Sellards— Validity of Idiophyllum rotundifolium. 208

Art. XXILI.—On the Validity of Idiophyllum rotundifolium
Lesquereux, a Fossil Plant from the Coal Measures at
Mazon Creek, Illinois ; by E. H. SELLARDS.

THE genus /diophyllum was proposed by Lesquereux for a
species of Carboniferous plants represented by a single speci-
men from Mazon Creek, Illinois.*

The species /. rotundifoliwm appeared to present characters
so peculiar that the genus was placed by Lesquereux in his
classification with the “‘ Ferns of Uncertain Relation,” and was
compared with Dictyophyllum, and with dicotyledonous leaves.

Figure 1.—Obverse of the type of Idiophyllum rotundifolium Lesqx.

In working over a large collection of plants from the same
locality, in the Yale Museum, the writer has found the obverse
side of Lesquereux’s type. The obverse (figure 1) is well pre-
served, and completely demonstrates the true nature of the
plant. The fossil represents the distorted apical part of a young
and not fully expanded fern frond. The tip is pushed to one side,
-erowding the lateral pinnz and partly obscuring the outline of
the pinnules. The pinnules, however, are much more distinct
than is represented in Lesquereux’s figure of the counterpart.
They are 6 to 10™™ in length, 3 to 34™" in width, alternate,
obtuse, and somewhat cordate at the basal attachment. The
rachis is longitudinally striate; the lateral pinnz are close,
oblique, and alternate. In all these characters /diophyllum
rotundifolium agrees with Neuropteris rarinervis Bunb., a

* Second Geol. Surv. Penn., Description of the Coal Flora, vol. i, p. 159 ;
Atlas, pl. xxiii, fig. 11, 1880.

204 Sellards— Validity of Idiophyllim rotundifolium.

species common at the locality. The outline and cordate base of
the pinnules, and especially the faint indication of lobation of
the upper border, just above the base (figure 2), are characters
of LV. rarinervis. Although the lateral nervation is mostly
obscured, there can be little doubt regarding the specific iden-
tity. The lateral venation shown in figure 2 is taken from
several pinnules. The lateral pinnze are somewhat more
crowded than is usual for JV. rarinervis, a condition explained
by the immaturity of the frond. The plant is a difficult object:
to photograph, but the figure here given will perhaps help to
clear up some of the peculiarities attributed to the fossil. The
appearance of nerves “sometimes crossing each other in con-
trary directions and forming by intervention regular quadrate
or rhomboidal meshes ” is caused by the crowding and overlap-
ping of the pinnules. :

FIGURE 2.—A single pinnule ; showing the characters of Neuropteris rari-
nervis Bunb. ; taken from the right side of the specimen. x2.

Altogether, it seems evident that the species Jdiophyllum
rotundifolium is a synonym of WVeuropteris rarmervis, and
that the genus Jdiophyllum has no status in systematic fossil
botany.

Phidach the courtesy of Mr. David White, the writer has
recently had an opportunity of examining the type of Jdzo-
phyllum rotundifolium, now in the Lacoe Collection of the
United States National Museum. The plant is not entirely
freed from the matrix, the tips of the lateral pinnez on the
right side and at the top being still partly covered. Mr.
White had also recognized that the fossil represents an imma-
ture frond, and agrees with the writer that it is Wewroptercs
ravrinervis. |

Yale University Museum, Geological Department.

Gooch and Gilbert—Ammonium Vanadate. 205

Art. XXIV.—The Precipitation of Ammonium Vanadate
by Ammonium Chloride; by F. A. GoocH and R: D.
GILBERT.

{Contributions from the Kent Ghemical Laboratory of Yale University—CX.]

BERZELIUS was the first to point out and utilize in analysis
the fact, that when a vanadate in concentrated solution is treated
with a saturated solution of ammonium chloride, white insolu-
ble ammonium metavanadate is deposited.* The method of
treatment was modified by v. Hauert in that solid ammonium
chloride was added to the solution of the vanadate, as concen-
trated as possible, until it failed to dissolve, the mixture allowed
to stand and then treated with a large amount of strong alco-
hol, the precipitate filtered off, washed with alcohol, dried,
ignited in a covered platinum crucible until all ammonium
chloride was volatilized, and the residue ignited carefully with
ammonium nitrate. Roscoet was unable to obtain exact qnan-
titative results by this method on account of the solubility of
the ammonium metavanadate in the alcoholic mixture and the
danger of mechanical loss during the ignition. Holverscheit
also, in testing v. Hauer’s method,§ noted loss of vanadium, due
to solubility of the ammonium vanadate in the alcoholic liquid,
finding in every case vanadium in the filtrate by means of
ammonium sulphide or by hydrogen dioxide. In the average
-of six determinations Holverscheit found a loss of 0°0015 grm.
ealeulated as V,O,. Another modification of the method was
proposed by Ditte,| who, finding that precipitation by an excess
of ammonium chloride was complete in a solution made color-
less by boiling with free ammonia, attributed the deficiency in
the amount of V,O, found after precipitation under the pre-
scribed conditions, to volatilization of the vanadium during igni-
tion under the influence of ammonium chloride, and endeavored
to avoid liability to such error by making it certain that no
ammonium chloride should remain with the metavanadate at
the time of ignition. To this end, the solution made colorless
by boiling with ammonia was cooled to 30°-40°, nearly satu-
rated with ammonium chloride and finally treated with four
or five times its volume of alcohol; the precipitate thus thrown
down was filtered off, washed with absolute alcohol until free
from ammonium chloride, dried and ignited, the carbonized
residue having been washed with nitric acid before the final
ignition. Holverscheit’s criticism of Ditte’s process consisted
in showing that losses which may occur in the ignition, and

* Ann. Phys. xeviii, 54, 1831. + Jour. Prakt. Chem., lxix, 388.

¢ Ann. Chem. Suppl., Vili, 101, $ Inaug. Diss., Berlin, 1890, p. 11, et seq.
| Ee pe- Rend., civ, 982. .

206 Gooch and Gilbert—Ammonium Vanadate.

which were found to be trifling when care was used, are
mechanical and not due to volatilization ; and, secondly, that in
his experience the average error of seven trials resulted in a
loss of 0:0029 grm. whether upon 0:1398 grm. or 0:2796 grm.
of V,O,. Holverscheit rejects both the method of v. Hauer
and that of Ditte as inexact on account of the solubility of the
ammonium metavanadate.

In 1883,* previous to the proposal of Ditte, Wolcott Gibbs
applied in the determination of vanadic acid in the vanadio-
molybdates (and sometimes in vanadio tungstates), another and
much simpler modification of the method of Berzelius. Gibbs’

method consists in boiling the double vanadate with ammonia _
(to convert the complex salt into a mixture of vanadate and —

molybdate), adding a saturated solution of ammonium chloride
in large excess, concentrating the liquid kept alkaline with
ammonia to a small volume, standing twenty-four hours, col-
lecting the precipitated ammonium vanadate upon an asbestos
filter in a perforated crucible, washing with a cold saturated
solution of ammonium chloride, and, either igniting the pre-
cipitate and weighing the residue of V,O, upon the asbestos,
or dissolving the precipitate with boiling water, reducing to
the condition of V,O, and titrating by permanganate. This
method is called in question by Rosenheim,t who gives analyt-
ical results of certain experiments which, according to Rosen-
heim’s description, are not at all upon the lines laid down by
Gibbs. In the first place nothing is said by Rosenheim as to
concentrating the mixture toa relatively small volume after
adding a saturated solution of ammonium chloride in very large
excess; and in the second place Gibbs did not make the final
washing with dilute alcohol, as Rosenheim says he did, but
finished the washing with a cold saturated solution of pure
ammonium chloride. Rosenheim’s variation of the experi-
mental procedure in these important particulars, vitiates his
conclusions with regard to the precipitation of vanadie acid
from solution by ammonium chloride, that “small amounts
nevertheless remain, as Roscoe rightly affirms, in solution.”
Rosenheim’s opinions of the method are echoed by Milch,t
Liebert§ and Euler.|

We have thought it desirable, therefore, to investigate anew
the question as to whether the precipitation of ammonium
metavanadate by ammonium chloride is complete enough to
place that mode of separating vanadic acid from solutions of
its salts within the list of good analytical methods.

The ammonium vanadate used in our experiments was ana-

* Proc. Am. Acad., x, 242, 249; Am. Chem. Jour., v, 371, 378.

+ Inaug. Diss., Berlin, 1888. t Inaug. Diss., Berlin, 1887.
§ Inaug. Diss., Halle, 1891. || Inaug. Diss., Berlin, 1895.

Coos ——

Gooch and Gilbert—Ammonium Vanadate. 207

lyzed by the iodometric method of Holverscheit,* and found to
contain 76°14 per cent of V,O,.

The ammonium chloride employed was shown to be pure
and free from iron by heating a solution of it to boiling, adding
bromine water and a slight excess of ammonia, and filtering.
No residue whatever remained upon the filter.

In each experiment a weighed portion of ammonium vana-
date was put in a small beaker and heated with about 25° of
water and a few drops of ammonia upon the steam bath until
solution was complete. To the solution were added 25° of a
cold saturated solution of anmonium chloride and a few drops
of ammonia. The mixture remained upon the steam bath,
with addition from time to time of a little ammonia to keep
the metavanadate of normal composition and colorless, until
the volume had been reduced to about 25°, and was then cooled.
On cooling, a small amount of ammonium chloride crystallized
out, but only a little if the proportion had been properly
adjusted. If too large an amount of ammonium chloride
erystallized out, it was nearly redissolved by the cautious addi-
tion of ammonium hydroxide. The mixture was allowed to
stand twenty-four hours to insure complete crystallization of
the ammonium metavanadate, and was then filtered on a weighed
asbestos filter and perforated crucible, the precipitate being
transferred and washed with a cold saturated solution of
ammonium chloride. Crucible and precipitate were heated, at
first very gently to drive off the ammonium chloride without
occasioning mechanical loss of the vanadium, and finally to
redness and fusion of the pentoxide remaining. At the outset
some difficulty was occasionally found in removing from the
walls of the beaker the adherent crystals of ammonium vana-
date, but this difficulty was overcome, in the experiments
recorded, by the device of forming upon the walls of the
beaker before using it a film of paraffin of extreme thinness by
rinsing the beaker with a dilute solution of paraffin in naphtha
(0°5 gram. of paraffin in 300° of naphtha) and allowing the
naphtha to evaporate. Crystals of the vanadate adhering to
the walls of the beaker thus previously prepared, are easily
removed by means of the ordinary rubber or “ policeman.”
Table I contains the record of six consecutive experiments
made after some preliminary study of the method. The
washings and filtrate were in several instances acidified with
hydrochloric acid and tested with hydrogen dioxide without
giving indication of the presence of vanadium.

These results are sufficient to show that the method of Gibbs
is capable of yielding an analytical separation of value, but as
Gibbs pointed out it is ordinarily preferable to estimate the

* Inaug. Diss., Berlin, 1890, p. 49.

208 Gooch and Gilbert—Ammonium Vanadate.

renee h f “TaBLe I. eta

NH; VO; taken. V.O; present. V.O; found. Error.

grams. grams. grams. grams.

0°5 0°3807 0°3814 0:0007 +
0°5 if 0°3818 — 0°0011+
0°5 ip C°3813 0°0006 +
Oro. es 0°3808 0°0001 +
0°5 os 0°3808 0°0001 +
0°5 = 0°3799 ~ 0°0008— |

vanadium by volumetric means rather than to go through the
tedious and exacting process of ignition to recover the vana-
dium pentoxide. Gibbs used the method of reduction by
hydrogen sulphide, and titration of the tetroxide by perman-
ganate, but for the purpose of testing the method we have
thought it best to use the same method for determining the
metavanadate separated which was used to determine the com-
position of the vanadate taken, viz., Holverscheit’s iodometric
method. In the beginning some trouble was experienced in
the proper handling of the separated vanadate for this purpose.
To dissolve the precipitate in hot water introduced too much
hot water into the distilling flask; filtering on paper and then
putting paper and vanadate into the flask together introduced
an error, due probably to the action of the freed bromine upon
the paper. The difficulty was tinally overcome satisfactorily
by collecting and washing the separated ammonium metavana-
date upon asbestos felt deposited upon a perforated platinum
cone of considerable size,*.rolling up the felt enclosing the
crystals of vanadate, then putting asbestos and vanadium into
the distillation flask, without addition of any water, ready for
the addition of potassium bromide and hydrochloric acid
according to Holverscheit. The use of so much asbestos made
it difficult, however, to boil the mixture of acid, bromide, vana-
dium and asbestos, so recourse was taken to heating this mix-
ture in the hot air chamber of a high temperature parafiin bath.
The mass of asbestos made it necessary to use more hydro-
chloric acid to overcome the viscosity of the mixture, and the
use of so much acid introduced the element of danger that the
volatilization of the acid to the receiver containing potassium
iodide might, in presence of air, set free iodine outside the
reaction of the process; so the flask was connected with a car-
bon dioxide generator, and the operation was carried on in an
atmosphere of carbon dioxide. With the exception of the
paraffin bath substituted for the burner as a source of heat, the
arrangement of the apparatus is shown in the accompanying
figure. Following with 1:5 grams of potassium bromide, the
introduction of asbestos and vanadate, the distillation flask B

* Am. Chem. Jour., vol. i, p. 320.

_ Gooch and Gilbert—Ammonium Vanadate. 209

was connected, as shown in the figure, with the receiver charged
with a solution of 2°5 grms. of potassium iodide in boiled water,
and the apparatus was filled with carbon dioxide. The stop-
cock of A was closed, the bulb of A was charged with 50™ of
strong hydrochloric acid, the connection with the carbon diox-
ide generator was again made, with care to displace all residual
air from the bulb. The acid was introduced into B by opening
the stopcock and, in a gentle current of carbon dioxide, the
heating was continued an hour after the development of color
in the receiver showed that bromine was coming over from the

flask. The iodine set free in the receiver was titrated with
sodium thiosulphate standardized against iodine of value known

by comparison with a arsenious acid. ‘The standard of the

ammonium vanadate taken, as determined by the Holverscheit
method in its simple form, was found to be unchanged when
portions of the salt were treated according to the modification
necessitated by the presence of so much asbestos in the trial
tests, viz., In presence of more acid and in an atmosphere of
carbon dioxide.

In this manner the determinations of the following table
were made:

TaBLeE II.
Standard

NH,VOs; taken. Na.S.0; used. V.O; found. Error.
grams. em? grams. grams.
0°1000 8°35 0°0760 0:0001—
0°1000 8°39 0:0764 0°0003 +
0°1000 8°38 0:0763 0°0002 +
0°1000 8°37 0°'0962 0:0001 +
0°1000 8°32 0°0757 0°0004—
0°1000 8°25 0:0751 0:0010—
0°1000 8°36 00761 0°0000

‘0°1000 8°33 0°0758 0°0003—

210 ~ Gooch and Gilbert—Ammonium Vanadate. -

These results, representing determinations of the vanadium
pentoxide in the ammonium metavanadate after separation by
the same method used in determining the amount of pentoxide
in the vanadate taken, abundantly confirm the conclusions
reached by a consideration of the results of Table I.

So it appears that the criticism of Rosenheim, obviously
made undera misapprehension of the details of the Gibbs
method, to which Milch, Liebert and Euler give their acqui-
escence, is unfounded. The process of Gibbs gives a practically
complete precipitation of ammonium vanadate, when to the
solution of the soluble vanadate such an excess of ammonium
chloride, with a little ammonia, is added, that the solution, after
concentration and cooling, deposits ammonium chloride, and
the mixture is allowed to stand twenty-four hours. Should
too much ammonium chloride for convenient handling erystal-
lize out on cooling, this is to be redissolved by the careful
addition of ammonia; but care should be taken that after
standing, a little solid ammonium chloride and free ammonia

should still remain in the mixture. The precipitated ammo- -

nium vanadate is to be washed with a cold saturated solution
of pure ammonium chloride, and the vanadium in the vanadate
determined by any appropriate means. Volumetric processes
are to be preferred to the slow ignition. We do not recom-
mend as a suitable procedure for ordinary use the complicated
modification of the Holverscheit method, which was employed
in this investigation in order that the same method of determi-
nation of vanadium might be used before and after the separa-
tion process. Reduction as suggested by Gibbs, and titration
by permanganate or the iodimetric estimation of Browning,*
should be of service in ordinary cases.

* Zeitschr. Anorg. Chem., vii, 158.

4
“
“
9

Hillebrand and Penjfield—TLhe Alunite-Jarosite Group. 211

Art. XX V.—Some Additions to the Alunite-Jarosite Group
of Minerals ; by W. F. HILLEBRAND and 8. L. PENFIELD.

CONCERNING two new varieties of jarosite which will be
described in the present communication, one is from Nevada,
and was collected by Mr. H. W. Turner of the United States
Geological Survey and sent to the survey laboratory at Wash-
ington for identification; the other is from New Mexico, and
was sent by Mr. J. H. Porter of Denver, Colorado, to the
Mineralogical Laboratory of the Sheffield Scientific School.
Except for slight differences in color the two minerals look
exactly alike, each consisting of minute, isolated, tabular crys-
tals, which, as may be seen with the microscope, consist of com-
binations of a rhombohedron with largely developed basal
planes. By chance it happened that the present writers dis-
covered that they were both engaged in the investigation of
compounds belonging evidently to the same group, and it was
_ decided to bring the results together into one paper.

Natrojarosite.

The material collected by Mr. Turner was obtained on the
east side of Soda Springs Valley, Nevada, on the road from
Sodaville to the Vulcan Copper i
Mine. It consists of a glistening
powder, made up of perfect crystals
having the habit shown in figure 1,
although generally only one rhom- .
bohedron, 7, is present instead of
two, as shown in the figure. The
largest crystals observed were 0°15™™
wide and 0:025™™ thick, and. the
general average would not be over
half that size. In spite of being so
minute, however, it was possible to
measure the angles of the crystals
with the reflection goniometer, the
chief difficulty arising not so much
from their small size as from the
vicinal character of the basal planes.
After repeated trials a crystal was
- found having a fairly good basal
plane, and from this crystal the following angles were obtained :

Caleu-

Measured. Measured. lated.

Car, 0001, 1011=51° 53 cAs, 0001 ,0221=68° 42’ 68° 35’
CAT , 0001 A. 1101=51- 53 CNS GCOOKA 202 = 68) 4852 - Sees

eanr, OOOLA O111=52 26 fete wot A VIO S86 55.8554
Am, Jour. Sc1.—FourtTH SERIES, VoL. XIV, No. 81.—SEPTEMBER, 1902.
15

212 Hillebrand and Penfield—Additions to the

The crystals belong to the rhombohedral division of the
hexagonal system, and the angle cA7, 51° 53’, which is prob-
ably very nearly correct, has been assumed as fundamental,
and from it the following axial ratio has been calculated :

C= 1°104.

That the axial ratio as given is very near the truth is shown
by the fact that the measurements of cas and rar do not
vary many minutes from the calculated values; while on a
number of other crystals, measurement of the angle cay,
though varying considerably, was found to be not far from 52°.
The angles of car and 7A?’ of the ordinary potassium jarosite
are 55° 16’ and 90° 45’, respectively.

Under the microscope the crystals exhibit normal optical
properties. Using a high power lens and convergent light, the
thicker crystals show the dark cross and the beginnings of the
first ring of the interference figure. The birefringence is
negative. ‘The color of single crystals, when seen under the
microscope in transmitted light, is golden-yellow. Many of
the crystals show numerous brown inelusions. The color
shown by a mass of the erystals is yellowish-brown, and the
material glistens, owing to reflections from the basal planes of
the minute crystals.

The material used for the chemical analysis was the purest
that could be obtained, although crystals containing the brown-
ish inclusions just aS, Peould not be avoided, and there
were occasional brown ferruginous particles mixed with the
crystals. The specific gravity of the material was found to be
3°18 at 380°5° C. The results of the analysis by Hillebrand are
as follows:

Ratio.

Ke.0.. 5. 2a are 0-319 3-29
Na. Of 7. poet eee ae a
BOS ea 0°35 0-004
SO, Oe ae a eee 0°387 4°00
ER O below W0S os 43e"2 Ore
HH. O above 108° 2: : 11°03 0°613 6°33
igs eee: pools see
BRO OAs ee 0°23
Oe rr re ye 0:04

99°94

The ratio of Fe,O,: Na,O:SO,:H,0O is evidently 3:1:4:6
as in ordinary jarosite, where the alkali is potash instead of

*Of the soda °22 per cent is not extracted by hot water after full ignition
of the mineral and hence may belong to a feldspar or some other foreign
mineral. Only 5°81 per cent is assumed to belong to the jarosite and used
in deriving the molecular value.

apts

Liha. até, ote ee

f
LL

ee __ * «i

“.

a
4
4
5
r

Alunite-Jarosite Group of Minerals. 218

soda. The slight excess of Fe,O, and H,O, as indicated by the
ratio, is evidently due to some ferric hydroxide; probably the
dark ferruginous impurities seen under the microscope are in
part responsible for this, and there are also traces of some
arsenate and silicate present. Regarding the excess of Fe,O,
and H,O as due to impurities, it is found that 94 per cent of
the material analyzed may be regarded as pure natrojarosite,
as indicated below :

After deducting Theory for
impurities. NaeFe.[/OH]i2[SOz]..
eh ) eee 46°43 or 49°39 49°49
AO er. Sah PO ss BOGE S 6°39
Te Ore ite od OS ae ot OSG eee
2 a 30°96 “ 32°94 32°99
Oe NOra br Sk bE TD 11°13
94°00 100°00 100'00

That six per cent of impurities should be present in a crys-
talline powder such as analyzed is not surprising, when it is
taken into consideration that it would require something like
2,500,000 erystals to make one gram of material, the estima-
tion being based on the assumption that the crystals are 0°-10™™
in axial diameter and 0:02" thick, which is certainly above
their average size.

Among the specimens from Cook’s Peak, New Mexico, sent
to the Sheffield Laboratory by Mr. Porter, were sorne masses
of a rather firmly cemented aggregate of minute crystals of a
mineral of the jarosite group. The specimens are of a brown-
ish-yellow color, and have in places the glistening appearance
of a mica-schist. They also look as though they had been sub-
jected to pressure and had been somewhat sheared. The
material is rather easily crushed, and the powder when exam-
ined with the microscope exhibits the properties of the natro-
jarosite just described. The crystals are associated with a little
limonite and quartz, and pure material for analysis could not
be obtained. Only a partial analysis, therefore, was under-
taken with the following results :

Tick Os Soe ce 55°60
INGOM Se rea LS 4°49
109 (se a SR ee 0°77
AIO) ps SL 8 0°96

SO, and H,O were present but not determined. The results
are sufficient to indicate that the material is essentially natro-
jarosite.

Plumbojarosite.

_ This material is from Cook’s Peak; New Mexico. It occurs
as a glistening, crystalline powder and as loosely cohering

214 Hillebrand and Penjfield—Additions to the

masses which may easily be crushed by pressure between the

fingers. The crystals are very symmetrical, and are exactly
like those of natrojarosite, figure 1, although generally only
one rhombohedron » is present. On the average the crystals
are a trifle smaller and noticeably thinner than those of natro-
jarosite. A number of crystals were measured on the reflect-
Ing goniometer, the chief difficulty arising rather from the
vicinal character of the faces than from their small size. One
unusually large crystal, 0°28" broad and 0:015™™ thick, was
finally found, having the development shown in figure 2, which
is unusual, for generally 7 (1011) and
not s (0221) is the prevailing rhom-

bohedron. Fortunately the crystal
koe > was so taken up on a minute point
Qs os of wax that the measurement of

sas in three rhombohedral zones
was possible. The results of five measurements of sas over
the upper and lower pole edges varied between 109° 5’ and
109° 30’, the average being 109° 16’; while six measurements
over the middle edges varied between 70° 10’ and 71° 00’, the
average being 70° 36’. The average of the two supplemen-
tary values gives sas, 22010221 = 109° 20’, which has been
assumed as fundamental, and from it the following axial ratio
has been calculated :

9

~

C= W216:

On the crystal from which the foregoing measurements were
obtained the basal plane was vicinal and hence no reliable
measurements of cas could be had from it. On a number of
other crystals, however, the angle of cx~7 was measured with
varying results, the variation resulting from the uncertainty of
the reflections from the basal planes. Four measurements of
car, which were recorded in the note-book as derived from
the best reflections, varied between 54° 15’ and 54° 44’, the
average being 54° 30’, while cn7, 0001, 1011, by calculation
from the fundamental measurement, is 54° 82’. Hence it may
be assumed that the axial ratio as established is reasonably
exact. The calculated value of rA7, 1011, 1101, is 89° 42".

In polarized light the crystals exhibit normal optical prop-
erties and negative birefringence. Being on the average thin-
ner than crystals of natrojarosite, it is seldom that, with the
highest powers and convergent light, even the beginning of
the first ring of the uniaxial interference figure is visible.
Individual crystals show under the microscope in transmitted
light a golden-yellow color. A mass of crystals has the appear-
ance of a glistening dark-brown powder, the color being
decidedly darker than that of natrojarosite.

Alunite-Jarosite Group of Minerals. 215

The analysis of the mineral was made on the very best
material, having a specific gravity of 3°665 at 30°C. The
results are surprising, and were wholly unlooked for, since it is
found that this jarosite contains lead in the place of alkalies.
The results by Hillebrand are as follows:

15> He Wi. IV. = Mean. _—_—Ratio.

Mewes |. 49°36 42°38 42°37 0°265 ) ...
Paes (fo) soeOeeey 2b 10 0:001 :
a 19°69 19:99 19°89 19°79 19°84 0°089 1-05
a ory 17
139 “21+ =r
_ =a 27-65 2e-07 27°06 0°338 4-00
H,O below 105° = 02 : 02
H,O above 105° 9°59 — 9:49 9°54 0530 6:27
2 a 56 51 “47 51
CL eee "27 “27 “21
2 "05 05
UL ‘Ol ‘O1

100715

The ratio of Fe,O,: PbO: SO,: HO is very close to 3:1:4:6,
indicating that the mineral is a variety of jarosite, and the slight
excess of Fe,O,, H,O, and PbO+ alkalies, may be accounted
for by assuming that slight impurities are present, partly
ferric hydroxide, in part some lead salt, and perhaps a soluble
silicate, as shown by the complete solubility of the silica in

_acids. Assuming that the ratio is exactly 3:1:4:6, it is found

that 4°36 per cent of impurities are present, and the remaining
95°64 per cent may then be regarded as plumbojarosite, as
follows :

Theory for PbFe.[OH]:2[SO.]..

ety 83 420°s9 © or) 42°44
12) 3 cae aie 18°86 = 19°72
Se et 2706 = = 28-29
ERO. gpg - |) 9°55

95°64 100°00

Since it took probably 2,500,000 crystals of natrojarosite to
make one gram of material, it certainly must have taken fully
4,000,000 to make a gram of plumbojarosite, for the crystals of
the latter mineral, though somewhat heavier, are decidedly
thinner than those of the former; hence the presence of 4°5

*The presence of alumina was not definitely proved. The figures here
given are the differences between the several weights of the ammonia pre-
cipitates and those of the ferric iron in them, as determined by permanganate
after reduction by hydrogen sulphide.

+ Probably somewhat high.

216 Hillebrand and Penfield—Additions to the.

per cent of impurities in such a crystalline product is not to be
wondered at.
Jarosite and Alunite.

As seen from the analyses of these minerals which have been
published, the alkali metal they contain is almost always potas-
slum, though sodium is at times present. -The formulas assigned
to the two minerals are therefore K ,0 +3Fe,0,+480, +6H, O
and K,0+3AI],O, + 48O, + 6H,O, ‘which may be variously
expressed, as will be indicated later.

A mineral corresponding to natrojarosite of this article,
though containing a little potash, has been described by W. P.
Headden* from the Buxton mine, Lawrence Co., 8. D. The
crystals are described as scales, consisting of a combination of
base and rhombohedron. The material analyzed was evidently
somewhat impure, as quartz and some As,O, are reported. As
the As,O, evidently does not belong to jarosite, the assump-
tion may be made that some scorodite, FeAsO,-2H,0, is present,
and the results of Headden’s analysis: may then be interpreted

as follows:
Original Scorodite Natro-

analysis. and quartz. jarosite. Ratio.
| Wel @ PRMGpd Saree sag 46°27 1°60 44:67 or 50°10 3°13
bcs 6 Slane y 4°35 4°35. 35 4°86
LGA epee) 1°47 1°47 ed 1°65 1:04
EE sa Sha gee RA 0°39 USo ae 0-44
Se gt ENN Ce ai 28°46 28°46 .: 31°98 4°00
Eira! teas! 10°55 0°72 9°83 * 11:02 6°13
62 4358 AM SRD, 2°36 2°36 bee ae
Guariaties hss th 6°10 6°10 feat eee

99°95 10°78 89°17 100°00

Thus, assuming the presence of 4°68 per cent of scorodite and
6°10 of quartz, and deducting them, the remainder agrees very
closely with natrojarosite, giving a good ratio, very near
pede ass.

Alunite ‘containing considerable soda has been described by
W. Crosst from Rosita Hills, Colorado, and by E. B. Hurl-
burtt from Red Mountain, Colorado, and analyses of both
minerals show about equal percentages of K,O and Na,QO, or a
molecular ratio of K,0:Na,O = 4:7. The occurrence, there-
fore, of sodium in the jarosite- -alunite group is in accordance
with previous observations, but the case is quite different with
lead. As far as the present writers are aware, this is the first
instance on record where lead has been observed isomorphous
with the alkali metals. It is interesting to note that the alunite
from Red Mountain, Colorado, occurs as a crystalline powder,

* This Journal (8), a p. 24, 1893. + Ibid. (8), xli, p. 472, 1891.
t Ibid. (8), xlviii, p. 180, 1894.

—— ee

Se

Alunite-Jarosite Group of Minerals. 217

the crystals being exactly like those of natrojarosite and plum-
bojarosite, except that they are «a trifle smaller and white, or
colorless when seen under the microscope.

From a chemical standpoint the most interesting feature
of the new minerals is the light they throw upon the iso-
morphism of potassium, sodium and lead. Ordinarily, even
potassium and sodium are not isomorphous, as shown by the
fact that their simple salts seldom crystallize in the same form.
Although KCl and NaCl both erystallize in cubes, it is not
certain that both salts belong to the same group of the iso-
metric system. It has been shown, for example, by etching,
that KCl crystallizes like NH,Cl in the plagihedral group of
the isometric system, while the etchings produced on halite |
seem to indicate that it crystallizes in the normal group.
Again at Stassfurt, Germany, sylvite and halite both occur
erystallized side by side upon the same hand specimen, instead
of mixing as isomorphous molecules. Even in such complex
molecular compounds as the feldspars, the potassium and
sodium salts crystallize as orthoclase and albite, rather than as -
isomorphous mixtures. Lastly potassium has a strong tend-
ency to form alums which is not shared by sodium. In con-
trast to these differences in chemical nature, we have in the
jarosite-alunite group of minerals not only the alkali-metals,
potassium and sodium, but, what seems still more remarkable,
lead, playing the same role in the compounds, and yielding
erystals which are surprisingly alike in all their physical prop-
erties. The writers can at present offer no other reason for
the isomorphism in the group of minerals under consideration
than that the alkalies and lead play so small a role, and the
remaining constituents so prominent a part in the complex
chemical molecules, that the latter control or dominate the
erystallization by virtue of what may be called their mass effect.

The alunite from Red Mountain, described by Hurlburt,
was analyzed in the Sheffield Mineralogical Laboratory under
the direction of one of the present writers, and it was found
that water was first expelled from the compound at a rather
high temperature, thus indicating that the mineral contains
hydroxyl and no water of crystallization: accordingly it was
shown that the seemingly complex formula of the mineral,
expressed by the ratio Al,O,: K,0:SO,: H,O = 8:1:4: 6, may
be much simplified to K[ Ai(OH),],[SO,],.. In the light of the
present investigation it now seems best to abandon the above
simple formula and adopt one containing double the number
of atoms, in order to make clear the isomorphism between K,,
Na, and Pb. The formulas of the minerals of the group
would then be expressed as follows :

218 LMillebrand and Penfield—Additions to the

Alunite, K, |; Al(OH);|,[80,],° or “K,AlfOH] son:
Natroalunite, Na,[Al(OH),|,[SO,], or Na,Al JOH] [SO,
J arosite, K,| Fe(OH).),(S0,], or KFe lO sem)
Natrojarosite, . Na,[Fe(OH))|[S0,],: or Nae) OR eisams
Plumbojarosite, Pbj|Fe(OH),|,,[SO,|, or PbFe,OH],,[/S50

In the case of the lead compound one atom of lead, and in
the others two atoms of either potassium or sodium, are com-
bined in complex molecules containing fifty other atoms;
hence that the complex of fifty atoms, to the right of the K,,
Na, and Pb in the foregoing formulas, should control or domi-
nate crystallization by virtue of mass effect, and condition an
isomorphism between such unlike elements as sodium, potas-
sium and lead, is not so surprising as would at first appear.

Having adopted the double formulas, as given above, there
are numerous ways of writing developed formulas, of which
the following are perhaps the simplest and most satisfactory :

4

)
—

414

OS =O
HO WAN OR
Ho>Fe—O O ae Oar
HO a 3 _OH
Ho>Fe—9 So. b= Oe Ss — O—Fe< og
HO OH
Ho>Fe—O . y 0 Oi <oH
O—S =O
E> FeO 8-0 er,
eG
HO Bt ( 6)! 2
Ho>Fe—-ON At \ 4ZO0O—-K eon

a Ou IOs
LOS Sf aa
a es aR EWE 0

_-OH
OH

It is interesting to note that although K,, Na, and Pb play so
small a rdle in the alunite-jarosite molecules, the substitution
of Na, for K, is attended by quite a marked variation in the.
angles of the crystals, greater in fact than is generally observed
in isomorphous replacements. That alunite and jarosite con-
taining potash would be nearly alike in their angles is expected,
since crystals of corundum and hematite are surprisingly alike
as shown by the following comparison :

NEY
Foe Oe

Axial length. PAT CAT
CorundumyAlO. oc.” 13630 93° 56’ 57° 34!
Hematite, Fe,O,-.--- 1°3656 94 00 57 37

Ad
£

Alunite-Jarosite Group of Minerals. 219

The relations of the minerals of the alunite-jarosite group
are as follows:

Axial length. ae Gar Birefringence.
Ai 1°252 90° 50’ . 55° 192’ positive
Jarosite _.------ 1°245 90 45 55. 16 negative
Natrojarosite _.. 1°104 85 54 51 58 negative
Plumbojarosite_- 1°216 89 42 54 32 negative

From the foregoing table it is seen that the substitution of
sodium for potassium in jarosite has brought about greater
variation in the angles of the crystals than the substitution of
the bivalent metal lead for potassium.

The three minerals, natrojarosite, plumbojarosite and the
Na-K-alunite from Red Mountain, are very interesting when
studied together as microscopic mounts, the crystals being
practically alike in size and development, aud illustrating very
beautifully on the one hand the isomorphism of aluminum
and iron, on the other the isomorphism of potassium, sodium
and lead. The three substances must have formed under like
conditions, and it is believed that they are solfataric products,
formed under the combined action of heat and pressure.
Being difficultly soluble, they have formed, like many precip-
itates, as fine erystalline powders.

The three products just mentioned, when heated in closed
tubes behave alike; they suffer no change on gentle heating,
but when the temperature is sufficiently high to decompose
the chemical molecules, the crystals break up into fine powder
or dust, which is carried along by the escaping vapors and
deposited for a considerable distance along the sides of the
tubes. In addition to water, SO, and SO, are copiously given
off during decomposition. In the case of natrojarosite, and
the same would doubtless hold true for the Na-K-alunite, it is
found that after ignition, one-fourth of the sulphate radical has

been retained by the alkali metal, and may be extracted by

water. In the case of plumbojarosite, however, all of the sul-
phate radical is expelled by ignition, doubtless because the
ferric-oxide present serves to decompose any lead sulphate
which might have a tendency to form. Anglesite, PbSO,,
when heated alone in a closed tube suffers no decomposition,
but when finely triturated with limonite and beated, acid water
is given off. Finely powdered natrojarosite and plumbojaro-
site are slowly but completely soluble in boiling hydrochloric
acid. Plumbojarosite when fused with sodium carbonate on
charcoal yields globules of lead and a coating of lead oxide.

It has seemed to the writers best to designate the new com-
pounds described in this article as natrojarosite and plumbo-
jJarosite, the names signifying their relation to a well known

220 Hillebrand and Penfield—The Alunite-Jarosite Group.

species. Other members of this group will doubtless be
found, and the name natroalunite might be employed to
designate the two varieties of alunite from Colorado men-
tioned on page 216, where the proportion of the soda to the
potash molecule is 7:4. It is highly probable that a series of
alunite-jarosite compounds could be made artificially.

It is with pleasure that the writers acknowledge their indebt-
edness to Messrs. Turner and Porter for calling attention to
the interesting compounds described in this article.

Laboratories of the U. S. Geological Survey, Washington, and of the

Sheffield Scientific School, Yale University, New Haven,
; Feb., 1902.

Kindle— Niagara Limestones of Hamilton County, Ind. 221

Art. XX VI.—The Niagara Limestones of Hamilton County,
Indiana; by EDWaRD M. KINDLE.

HAMILTON County is located slightly north of the geographi-
cal center of Indiana. The drift in this part of the State is so
deep that the Paleozoic rocks are rarely exposed at the surface.
Only a few outcrops occur in the county. The most extensive
exposure is the one at Connor’s mill on White river, five miles
above Noblesville. Two small quarries southwest of Fishers-
ville afford an equally good opportunity to study the Paleozoic
rocks of the county. Several days were spent by the writer in
collecting from the beds exposed at these two localities.

Connor’s mill.—The outcrop at Connor’s mill consists of a
hard, light buff dolomite, which is exposed for two or three
hundred yards below the dam. The beds show a dip of 20° to
40° to the southwest. Above the dam one-third of a mile the
limestone outcrops again, dipping 30° to the northwest.

The See! is a list of fossils collected from beds exposed
below the dam :*

MERC RIRIISHS 22 ee ead Oe.
PeeIsinGh CF, COMCUMM ..-..-----------4--- Ts
EMMMIRM OS SMAIRETOE certo ui el se ee MS > 8
MEMUPLOCTINUS Cf. CTASSUS __ =. ..-- -.-- --2-2- Z
Somehsocunm multicostatum_.....-- ..-=-~=122 a.
Conchidium sp. -------- ean. tS ot ee eee t
UNMEELACIREES | ee. et Ie
paeeuMcodonta projunda@ .-.- 2... --.---.-.-- e.
mepemcne TROMMOIdalIs._:..2--.-.----- 2-42 e
EES SCE EES ae

PMEREGEIS 862 oe fer ees ee eke
MP OITO CUINUTG 22 ee
BERS MIGUUTENSIS 22-0" aL |
DRE MCCEI SHG! 8 GOO ee r
Mmnmnclan Clogs 6 2. ek gt.
amen prites Cy. sericea... 2. 22-2 2-2 T..
EET SSS 7B ed a ee ot
Platyostoma cf. niagarense .....-.---...---- C.
Spherexochus CIN SS Se a a:
EennmCnE NIT GOKEHSIS (222 is lk Ll C.
EME eer ee le.
MOIS Ch IPIGUIOMOUS PS Cc.
ETE ET Se ey ee rr

Fishersburg quarries.—Southwest of Fishersburg one-half
mile a very pure white sandstone has been quarried for glass

* The species listed will be described and figured in a future paper.

222 Kindle—Niagara Limestones of Hamilton County, Ind.

Figure 2.—Fishersburg quarry, showing unconformity between Niagara
limestone and underlying sandstone.

Kindle—Niagara Limestones of Hamilton County, Ind. 223

making. It isa fine-grained, massive, loosely-cemented rock,
erumbling easily. A buff dolomite, having the same lithologi-
cal characteristics as the outcrops at Connor’s mill, rests upon
the sandstone.. The line of contact between the two forma-
tions is clearly shown in the quarry and is seen to be a very
irregular one, resembling unconformity. The sandstone is
believed however to be a local lense. Such lenses are known
at other localities in the State where both the upper and lower
contact with the Niagara limestone is clear. The limestone
on either side of the projecting mass of sandstone extends
below the surface of the pool which fills the quarry. The
limestone beds show a dip of about 35° to the north.

A careful search failed to discover any fossils in the sandstone.

The fauna of the limestone as well as its physical characters
indicate that it belongs to the same formation as the beds at
Connor’s mill.

A comparatively short time was spent in collecting from the
Fishersburg quarries, and for this reascn the following list of
fossils from that locality includes fewer species than the pre-
ceding list:

masemclasme Gf. calicula 222... .222-------- r.
EMIS TIES HUGO OTENStS = 225 ssl le e
Pep TAO TObUS 8 . olke ns ie
Conckwavum multicostatum ....---.-----<---~- a.
mnewcemarid Gf. OICOSIGLG ... 2 222 22 -- 2222 225-- Ki
Prevamicies sulplanus. ..- 2.13.22. 2.22 2222. - r.
EME NOONGO SP... 325 2 es ea tu. C.
Men RG ae Pr ue See ae, Gi
Mince gerd pisi forms... 2 ee 2-2 ee - rs
PL piOena TROMOOTAGIS .......)2---.--. 224. - ¥
Me mmanslis 2) oe A
WOH ee IG es el F.
MEPeRCMOCNUS TOMIMGEN. ..-. 2-25 .--. 422 5-4 4 F.
Beemmeapaie). brisuleadtus 2... 22-2. .2- 122-22) - Fe:
MRE Soe Oe RE oe OE Lk YT,

Correlation.—Richard Owen described the outcrop at Con-
nor’s mill in his report published in 1863,* but offered no
opinion as to the age of the beds. The earliest reference to
the age of these beds occurs in a report on the geology of
Hamilton countyt by Dr. R. T. Brown, who considered them
to be of Devonian age. No paleontological evidence was
offered in support of this opinion, the author of the report
stating that “the outcrops of rock in Hamilton county are

* Rep, Ind. Geol. Surv. for 1859-62, p. 102.
+ 14th Ann. Rep. Ind. Geol. Surv., 1884, p. 27.

224 Kindle—Niagara Limestones of Hamilton County, Ind.

quite barren of fossils.’ In 1901* the limestones at Connor’s
mill and near Fishersburg were referred by the writer to the
Niagara, but the paleontological evidence for this determina-
tion was not given. The faunal lists here given clearly show
the Niagara age of these beds. The Lockport (Niagara) lime-
stone is their probable equivalent in the Niagara group.

The Hamilton county outcrops are the most southern expos-
ures in the State which show highly tilted Niagara strata.
The orogenic disturbances, which caused a general tilting of
the Niagara rocks in northern Indiana previous to the begin-
ning of Devonian sedimentation, did not affect the southern
portion of the State, where they lie nearly horizontal, and are
conformable with the Devonian rocks. ‘North of the Ohio
river eighty miles the Niagara rocks are slightly unconforma-
ble with the Devonian,t+ but nearly horizontal. The Devonian
rocks have not been observed in contact with the Niagara in
Hamilton county, but it is very probable that they are uncon-
formable as they have been shown to be further north in the
Wabash valley.t

U.S. Geol. Survey,
New Haven, Conn., June, 1902.

25th Ann. Rep. Dept. Geol. and Nat. Res., Ind., p. 559.
+ Kindle 25th Ann. Rept. Dept. Geol. Nat. Hist. enh , plate 16.
¢ Ibid., p. 562.

C. Barus— Velocity and Structure of the Nucleus. 225

Art. XXVII.— On the Velocity and the Structure of the
Nucleus ; by C. Barus.

1. Nucleation produced by shaking and tts velocity.—Table
I contains the results for the absorption velocity, #, of nuclei
shaken out of dilute solutions. The data were obtained by
measuring the condensational coronas in a spherical vessel
(diameter, 2/2 = 30™) in the lapse of time. If the loss of
nuclei be regarded as taking place at the walls of the vessel
and to constitute a drain on the whole nucleation (7 particles
per cub. em.) which moves radially outward at the rate #, then

the equation* n= n,e—*"/F may be assumed. The time
elapsed, ¢, is usually given in minutes. To deduce from £& the
velocity of the nucleus, «, I have for the present regarded it
as sufficient to multiply by 6 or preferably by 15/7, so that
«=51k. The concentration, c, is given in per cents or
grams of dry salt in 100% ™ of solution. It is essential that
a definite bulk (usually 500% °™) be used throughout, if the
data are to be at once comparable.

TaBLeE I.—Nuclei shaken out of solutions by 10 jerks. Bulk of solution

usually 500°™?, Nucleation, n = no eg < R= 15",
Concen- Nuclea- Absorption Nuclear
Solvent. Solute. tration, c. tion,n. velocity, k. velocity, x.
% em/min. em/min.
Water Water 0° 32 1°5 7°6
85 6°7 34°
26 9°4 48°
Water HCl 9) 160 "06 30
"005 140 "02 "10
"00005 80 2°3 E2°
Water. H,SO, 6 215 ‘03 15
"003 130 ‘O7 36
Water NaCl. 2 175 ‘01 05
Water CaCl, 2 240 04 20
1 240- a it
(double bulk) 1 460 "044 20
Water FeCl, 2 230 us -
has, 180 ee ad
Water Fe3NO, 12 175 03 "15
"12 160 03 15
Water Al3NO, E2 200 07 "36
"12 80 03 "15
Water Ca2NO, 1-4 260 ‘04 "20
‘O14 90 ‘02 10
‘0014 60 "10 “51

* This Journal (4), xiii, pp. 91, 94, 1902.
+ Note the doubled n and single k.

C. Barus— Velocity and Structure of the Nucleus.

226
TABLE I (continued).
Concen- Nuclea- Absorption Nuclear
Solvent. Solute. tration, ¢. tion,n. velocity,k. velocity, x,
Lb cem/min. em/min.
Water Ca2NO, 00014 60 1:4 71
Water Alum 5 200 05 “25
(dry) O15 90 i 56
Water Na,SO, 9 450 ‘07 "36
"009 110 “le “61
‘00009 50 Fad | ne
Water H,N NO, 2 190 "03 "15
‘02 129 “11 56
Water K,SO, 2 280 06 31
"02 -140 07 °36
Water Na,PO, 9 195 06 31
Water Sucrose 2 160 “ts 66
"02 51 19 97
Water Glucose 1°6 150 4 aah
‘016 50 1:0 onl
Water Glycerin 26 95 ia 5°6
"026 50 2°3 12°
Water Urea 2 110 19 97
Water Alcohol 0 55 4°8 24°
1°6 43 1°8 9°
Water Tartaric 2 130 02 10
acid 02 100 02 10
‘0002 95 “20 bga
Benzol Naphtha- 2 90* "04 “20
lene 02 90 ‘03 “5
none 80 "02 "10
Benzol _Paraffine if 130 02 10
O 80 "04 20
_ Benzine none at 2 16 "82

2. Leemarks on the data.—The main deductions from these
tables have been briefly given elsewheret and need not there-
fore be detailed here. Great caution is necessary because of
the inevitable irregularities of the results. The dependence of
the nucleation, n, on the concentration ¢, agrees fairly well
with the equation, n=n,+A/(log(B/c)), where A and B are
constants. Hence in making comparisons, it will be con-
venient to refer ail data to the logarithmic concentration, log
c, from the peculiar character of the results. In a general
way the number of nuclei evolved, caet. par., depends on the
mass of solute dissolved per cub. em., and for 1 per cent solu-
tions is within rather narrow limits independent of the saline

* These data are computed as if the solvent were water. They are thus

partly relative. The data needed to find n for benzol are not consistently

forthcoming.
+ Science, xv, pp. 912-914, 1902.

C. Barus— Velocity and Structure of the Nucleus. 227

or acid solute taken. For neutral organic solutes, the number,
mn, is little more than half as large. This seems then to be in
the nature of a class distinction. Pressure decrements of less
than 2™ suffice to precipitate the nuclei; but for the small
supersaturations, the greater number re-evaporate to the nuclear
stage and many exhaustions are needed to throw them out.
For the high degrees of supersaturation all nuclei are precipi-
tated at once. The number of nuclei generated depends on
the bulk of solution shaken and on the intensity of the agita-
tion. Hence they seem to arise in the solution itself and not
by friction with the walls of the vessel. The attempt to find
a limiting number has not yet been successful. Incidental
conditions which have not all been traced to their source seem
to have considerable influence on n.

3. Persistence of nuclei.—A comparison of the data for n
and those of & for the same concentration shows that for the
same solution m decreases with log ¢ while # increases with log
c, the arena of greatest variation being the very dilute solu-
tions which merge into water. Usually the growth of & con-
tinues even after the decrease of m has appreciably subsided.
Corresponding to the greater m which characterizes the saline
solutes as compared with the neutral organic solutes, the veloci-
ties of the nuclei of the latter are greater than those of the
former, under otherwise like conditions. Nuclei with neutral
organic solutes in water are less persistent. Extreme per-
sistence may be obtained with the volatile hydrocarbon sol-
vents with an appropriate solute, like benzol-paraffine, and
often the solute seems to be spontaneously nuclei-producing, ©
like benzol-naphthalene. In general persistence for a given
nucleus is a question of the mass dissolved per cub. em. of the
solution ; but cases like water-tartaric acid occur in which a
low order of nucleation, m, is associated with exceptional per-
sistence, or low values of &.

4. Structure of the nucleus.—From experiments like the
present I recently came to the conclusion that all condensation
nuclei are concentrated solutions; that the increment of vapor
pressure due to increasing convexity is eventually, but slightly
before molecular diameters are reached, compensated by a
decrement of vapor pressure due either to concentration* or
to electric charge.t The latter case is interesting inasmuch as
the nuclei are just about large enough that a single electron
spread out over the surface would suffice to equilibratet the
increment of vapor pressure due to surface tension. The con-
centration hypothesis is more straightforward and requires

* Science, xv, pp. 912-914, 1902.
+ This Journai (4), xiii, pp. 400-402, 1902 ; ibid., p. 478.
¢ This Journal (4), xiii, p. 473, 1902.

Am. Jour. Scl.—FourtTH Series, Vou. XIV, No. 81.—SEPTEMBER, 1902.
16

228 0. Barus— Velocity and Structure of the Nucleus.

less imagination, and it will therefore be chiefly referred to in
the following discussion. The nucleus is to be regarded small
enough that it is not symmetrically bombarded by the mole-
cules of vapor. Hence its velocity is conceived to be inereas-
ingly larger as it is smaller, i. e., as the conditions favorable to
non-symmetrical bombardment increase. If too large it will
be stationary, apart from gravity; if small enough, it must
eventually acquire the molecular velocities themselves.

5. The relations of equilibrium here involved are peculiar
and need a more detailed elucidation. If vapor pressure
increases with increasing convexity, for capillary reasons, but
eventually decreases again as a result of the concentration

; i

the normal vapor pressure at a flat surface, it follows that as
the size of the drop continually decreases the vapor pressure
at its surface must pass through a maximum. The accompany-
ing diagram” is an attempt to represent the case graphically for
two given solutions, by making the vapor pressures the ordi-
nates, and the radii of the given droplet of solution the
abscissas. The line ad indicates the normal vapor pressures.
All particles whose sizes taken from the curve 6m’s’ correspond
to the abscissas between s’ and 6 therefore evaporate in the
lapse of time, those lying near the maximum, m’, fastest, those
lying near s’ or 6 with proportionate slowness, while the latter
are also lost by subsidence and may be dismissed from con-
sideration. On the other hand, a particle whose radius is.
smaller than the abscissa corresponding to s’ will grow so that
as’ is the stable radius of the nucleus obtained by shaking the
given dilute solution.

If the solution is weaker, the droplet shaken out of it will
have to evaporate further to reach the critical density of the
stable nuclear state, and the increment of vapor pressure due
to surface tension will also be larger or the maximum, m, will
be higher. The curve bms now represents the conditions and
is to be similarly interpreted.

* The two curves should have been drawn preferably without intersecting.

C. Barus— Velocity and Structure of the Nucleus. 229

If the solvent is pure or contains only a trace of solute, the
nucleus will vanish completely, or else the particle left may be
too small to serve as a condensation nucleus for a given pres-
sure decrement of exhaustion. The size after the lapse of
time depends on the fixed quantity of salt originally entrapped.

If the vapor is not quite saturated, the chances for evapora-
tion will be enhanced. ‘The line a will be correspondingly
lowered, but equilibrium may result for a smaller size of
nucleus, until eventually the solid saline residue of the nucleus
alone is left. In so far as these concentration nuclei occur in
the atmosphere, one is justified in concluding that their size
(apart from the effects of temperature and barometric pressure
on surface tension and vapor density*) will increase under
mean atmospheric conditions as they are suspended at higher
distances above the earth’s surface, until the levels of perpetual
saturation are invaded.

6. There is one outstanding question relating to the time
losses which must now be considered. These coefficients,
dn/dt, are much smaller for concentrated than for weak solu-
tions. This observation was referred to the diffusion of the
nucleus and its absorption at the walls of the vessel with dif-
ferent velocities, #. The diagram shows, however, that near
the points s there must be retarded evaporation for all parti-
cles, because of the small differences of vapor pressure remain-
ing. Hence the persistence of nuclei shaken out of solutions
might be ascribed to this effect. True, no reason is evident why
stronger solutions should differ from weak solutions in their rela-
tive time losses, d log n/dt. Referring to Table 1 again, solu-
tions of CaCl,, H,SO,, which can not dry in a saturated atmos-
phere, are seen to show nothing exceptional in their behavior.
It may be computed that the stable nucleus is, in this case, not
even very concentrated. Special experiments are nevertheless
needed to clear up the matter, and they must be so devised as
to give direct evidence of the occurrence of diffusion or motion
of nuclei, and the value of its amount. If this is large enough
to be compatible with the data for & in this chapter, then the
hypothesis of retarded evaporation may be dismissed.

It is with this end in view that the experiments of the next
table are contrived and the results show that the motion of the
phosphorus nucleus, as actually observed, is considerably faster
than the average case computed for the nuclei in the present
chapter, and consequently the interpretation here accepted is
corroborated.

7. Motion of nucler directly observed.—Table II shows the
velocity of the nuclei as found directly in a tower-like receiver
1 meter high, into the bottom of which the nuclei have been
introduced. The height of the fog bank, «’, seen on exhaustion

* Note that temperature and elevation produce opposed effects.

230 @. Barus— Velocity and Structure of the Nucleus.

after the lapse of a definite number of minutes, give the
velocity sought at once, after correcting for the instantaneous
elevation of the plane of demarcation due to the withdrawal of
non-nucleated air from the top of the apparatus. The table
gives the ratios p’/p of pressures after and before exhaustion,
respectively. The experiments are very trying, chiefly because
of the difficulty of finding the initial position (time ¢ = 0) of
the plane of demarcation between clean and nucleated air. For
this reason I will also give the height of the fog bank seen on
exhaustion after 50 minutes. The table shows that incidental
conditions (whether the inflowing filtered air be quite dry or
not, ete.), often very subtle, largely influence the results, but a
detailed explanation, for which there is no room here, must be
given elsewhere.

TaBLE II.—Apparent and corrected rates of motion (x! and x) of phosphorus
nuclei in different saturated vapors.

Height of

Pressure- fog bank

ratio. Vapor. Apparatus, etc. mx 10°. «x10?. ° after 50™,
: em/min. em/min. cm.
"83 Benzol Tower 75 61 AQ
"78 Toluol Globe 80 62 40
90 ee Tower - 43 39 ei
83 Acetone Tower 58 48 32
4 s Initial rate 75 §2 38
<< j{ Amyl Tower, initial ; 100 83 51
e | Alcohol apparent 75 62 46
“ Ethyl Tower 106 88 62
RF Alcohol Dried air 72 60 ee |
. Methyl Tower 24 20 29
- Alcohol Dried air 64 53 42
r¢ Water . Estimated* 10,000 8300 —

8. General comparison of nuclear velocities.—It will now
be opportune to make a comparison of all the nuclear veloci-
ties made in widely different experiments throughout my
work. I will begin with the results of my first memoirf on
the subject, in which the velocities of phosphorus nuclei in
ordinary atmospheric air were studied both by mechanical
methods (steam jet and absorption tubes), and by the electrical
condenser methods. The mean value of & = 18 em./min. may
be taken. Hence in air, « = 90 cm./min., nearly. The num-
ber of particles was of the order of x = 10* per cub. em. of air.

In more recent experiments and by a methodt of comparing

* Rise or fall of the strands or fog filaments.
+Experiments with Ionized Air; Smithsonian Contributions, pp. 1-93,
01

t This Journal (4), xiii, p. 92, 1902.

C. Barus— Velocity and Structure of the Nucleus. 231

coronas of different orders, for phosphorus and other nuclei,
present in saturated water vapor to the average extent of 10°
per cub. em., the value &= °1 cm./min. was ascertained. This
is equivalent to «=°5 cm./min., in air saturated with water
vapor.

This datum may now be compared with the diffusion veloci-
ties of Table II of this chapter, in which fresh phosphorus
and other nuclei were used, densely distributed and tested in a
variety of vapors, alcoholic, hydrocarbon, ete. The usual
values of diffusion velocity lie between « = ‘5 ecm./min. and
‘9 cm./min., being thus of the order of the preceding case.
Water vapor itself did not admit of measurement. The value
estimated from the filamentary advance immediately after the
nuclei enter, 80 cm./min., agrees more closely with the case for
atmospheric air at the beginning of the paragraph. -As stated
elsewhere, there is much in the behavior of water which is
left unexplained. When the nuclei are first introduced into
the mixture of air and saturated water vapor, the air effect
does not seem to be negligible.

Finally the experiments on the evanescence of the nuclei
produced by shaking solutions lead to a series of values of &
as follows. A few hundred nuclei per cub. cm. were usually
present after shaking the solutions, and less than 50 (usually)
after shaking pure water.

For the saline solutions of 1-3 per cent, of -01 per cent and
of -0001 per cent, respectively, « = °25, ‘40, and 10. For pure
- water « = 25 or even 50.

For aqueous solutions of solid neutral organic solutes of 1-3
per cent, and ‘01 per cent, « =°8, and 3, respectively. The
acid organic solutes like tartaric acid seem to behave quite
differently. The solutions of this body of 2 per cent, -02 per
cent and -0002 per cent, showed diffusion velocities of « = ‘1,
‘1, and 1-0, respectively, thus evidencing uniformly greater
persistence of nucleation than even the saline bodies.

For neutral liquid organic solutes in aqueous solution of 1-3
per cent and -Ol per cent e=6and 12. Nuclei from these
bodies are thus very fleeting.

Finally for hydrocarbon solution of solid hydrocarbons of
1-3 per cent, « = ‘10 or ‘20, not differing much in persistence,
etc., from the salt solutions.

The total range found for the diffusion velocity of nuclei
produced by shaking solutions of less than 3 per cent is thus
from « = ‘1 to 50 cm./min., increasing with the degree of
dilution-of the solution, the largest value cited (pure water)
closely approaching the datum for phosphorus nuclei in atmos-
pheric air, « = 90 cm./min.

These variations of the velocity of the nuclei produced by

232 =. Barus— Velocity and Structure of the Nucleus.

shaking and its relation to the order of values found in the
direct experiments with phosphorus nuclei given in Table II,
is accounted for by the thermodynamic hypothesis for their
occurrence, which makes them particles of concentrated solu-
tion. The size of the nucleus for a given solvent depends
essentially on the extremely small mass of solute which it hap-
pens to contain. It is larger when shaken from the more
concentrated dilute solutions than from the weaker solutions,
because the critical density is reached in the former case with
less evaporation and the capillary increment of vapor pressure
to be compensated is at the same time smaller.

Thus it is quite reasonable to suppose that the nucleus
obtained from phosphorus or other emanations will be smaller
and therefore more mobile than the nucleus shaken out of the
more concentrated dilute solutions, but not so small in general
as the nuclei shaken out of pure water or any other pure sol-
vent, in which the amount of solute is an actually vanishing
quantity. To carry out this comparison one should eliminate
the peculiar features of water and use the same neutral solvent
in both cases: but if the above results for benzol in Tables I
and II are brought together, the same inferences follow.

9. Conclusion.— Among the tenable hypotheses, each of
which has some peculiarity in its favor, the above paragraphs
are believed to sustain the inference that condensation nuclei
in a nearly saturated medium are concentrated solutions. In
proportion as the medium is less saturated, they may pass into
the dry solute if such a one is present. The conditions of
equilibrium which are supposed to intervene are given in § 5.
Evidences in favor of this point of view are varied. In the
first place the velocity and therefore the size of the nucleus, no
matter of what origin, varies with the medium in which it is
suspended or is generated. Two mechanical suggestions may
be offered in explanation: either the nucleus condenses more
or less vapor spontaneously as assumed above, or the nuclei
themselves cohere into greater or smaller clusters, depending
on the medium. I pointed out that the latter case calls to
mind the suspension of clays, ete., in water or other liquids.
The endeavor to refer the differences to specific inductive
capacity or to the ionizing properties of the liquids breaks.
down with such solvents as acetone.

Of all the nuclei tested, those produced by shaking seemed
to be simplest as to their origin and therefore best adapted to
throw light on the subject. At first sight such a nucleus
should be the dry residue left after evaporation ; but the effi-
ciency of gaseous solutes like HCl, etc., shows that a solid
solute is not necessary. In view of the astonishingly small
quantity of solute which suftices to produce persistence, how-

C. Barus— Velocity and Structure of the Nucleus. 238

ever, it is a question whether the total absence of solid solute
in experiments with acid solutions can be guaranteed. The
next evidence is perhaps better; if it were a mere question of
fixed residue, then the neutral organic solutes like the sugars,
urea, glycerine, etc., should be quite as effective in producing
persistent nuclei as the saline solutes. Table I shows that this
is not the case. Finally since the pressure decrement of much
less than 2° of mercury prodnees precipitation, the vapor pres-
sure excess at the surface of the nucleus is small, and quite
within the limits of reduction produced by solution. The
evidence from hygroscopic bodies in which dry residues are
out of the question, has already been given. § 6.

Turning now to the electrical point of view, one notices at
the otitset a marked similarity in the trend of the above results
to the researches of Lenard* on the electricity produced by
waterfalls and jets. He showed that pure water when prop-
erly comminuted by spraying was made electrically positive
while the surrounding air became electrically negative, that
these charges originated in the liquid in contact with air and
could be indefinitely increased with the intensity with which
the spray was projected against a solid obstacle. This is in
harmony with the conditions for producing the above nuclei.
Again, the electric manifestations stated for pure water were
totally changed in character by the addition of mere traces of
solute to the water. Thus in case of dilute salt solutions the
liquid on comminution became negatively charged, and the air_
positively charged. As little as 005 per cent of salt was
sufficient to nearly wipe out the water effect. This then is
quite similar to the conditions of persistence of nuclei instanced
in the above tables. |

Lenard’s explanation is in terms of the ‘ Doppelschicht,”
which for pure water is conceived to be negative outward, and
for aqueous solutions and other bodies usually positive out-
ward, and due purely to voltaic contact. If this view is cor-
rect, i. e., if air in contact with pure water is invariably
negative by contact action no matter whether the surface is
plane or markedly convex, and if air in contact with salt solu-
tions is invariably positive, the solution being negative, then it
is the negatively charged nucleus (solution) which persists and
the positively charged nucleus (pure water) which is fleeting
or virtually non-existent. It follows, therefore, that condensa-
tion with subsidence is equivalent to a removal of negative
electricity provided the air charge is not simultaneously
removed. Of. § 5.

Brown University, Providence, R. I.

* Lenard, Wied. Ann. , Xlvi, pp. 584-636, 1892.

234 Hmerson—Corundum and a Graphitic Essonite.

Art. XX VIII.—Wote on Corundum and a Graphitic Essonite
Jrom Barkhamsted, Connecticut ; by B. K. Emerson.

SoME years ago I received from Mr. W. E. Manchester of
Pleasant Valley in Barkhamsted who is well acquainted with
the mineralogy of the region, a large block of a heavy bluish
black rock which proves to be corundum, and several speci-
mens of a very peculiar large garnet which are penetrated by
quite thick sheets of graphite.

The accompanying rock, which seems to be the country rock
and which occupies a large area, is a coarse mica schist, which
in a northwest direction for two miles abounds in smaller
garnets 2-5™™ across, of red color, accompanied by dark red-
dish brown staurolite crystals an inch across, and also cyanite in
blackish crisscrossed blades an inch wide and two inches long.
Near where the cyanite occurs there comes in above it a fibro-
litic gneiss which I have, farther north, associated with the
Algonkian. This latter rock contains layers, an inch thick ; of a
compact or finely fibrous fibrolite (faserkiesel) containing many
large grains of magnetite. The mineral proves under the
microscope to be made up of almost pure matted fibrolite
needles.

Garnet.—The remarkable garnets occur in well-formed dode-
cahedrons above two inches in diameter and extend over a broad
area forming a very coarse continuous drusy surface, some erys-
tals rising so as to show almost all their faces, others fused into
large groups which rise several inches above the surface of attach-
ment. They do not, however, present the aspect of crystals which
have grown freely into a cavity, since the faces are not pol-
ished and continuous but rather dull and often deeply exea-
vated with deep and irregular cavities from which some mineral
has been removed by solution. In the least weathered crystals
there are still large grains of included calcite visible, and we
may assume that the mineral which has been removed from
the depressions was calcite, that the whole broad surface of
great crystals has grown from a base of mica schist up into a
bed of crystalline limestone, which has been removed by solu-
tion from the surface of the’ crystals.

The character of the crystals agrees with this derivation.
They are pale honey-yellow to almost colorless, agreeing in
tint and general appearance closely with the crystals from
Gatineau, Canada, except that they are simple dodecahedrons,
are much larger and lack the polished faces, though this may be
due to a subsequent weathering. They are doubtless a quite

Emerson—Corundum and a Graphitic Essonite. 235

pure essonite. Because of this weathering they have now a
dull gray and unattractive appearance, and this is increased by
the fact that about half the surface of each crystal face is occu-
pied bya dull black minerai in irregular blotches, arranged in
a somewhat pegmatitic way. This sometimes increases in
amount so as to occupy almost the whole surface of each
face of the large crystals, and these faces are then irregular,
and in the great mass of aggregated crystals where this is car-
ried to the extreme it is continued up to a rather sharp plane
beyond which the crystals are quite free from the black min-
eral, and this plane passes through the middle of the largest
crystals.

In a section cut through a crystal where the surface was
almost all black, the dark color disappeared rather suddenly
about half an inch from the surface, but the dark mineral
penetrated in narrow wedge-like plates far into the compara-
tively pure garnet.

The mineral is a pure soft graphite, especially in the wedge-
like projections last mentioned, or a very fine granular mix-
ture of graphite and garnet over the gray areas.

When a fractured surface on a slide is examined with a lens
it has nowhere a homogeneous or crystalline appearance, but
looks more like a fine-grained white aplite, and this character
is maintained under the microscope which shows a brightly
polarizing mosaic in which one can, with difficulty, detect here
and there a nonpolarizing grain of garnet.

Large grains or phenocrysts of twinned calcite are very
abundant, and the rest of the field is mostly made up of large
feathery aggregates of coarse columnar wollastonite, full of
minute blebs of quartz and quite large anhedra of diopside.

It is remarkable that such large well-formed crystals have
so small a fraction of garnet in their composition. It is only
at the surface that a thin Jayer of nearly pure garnet can be
found.

Corundum.—The corundum block is from a very pure bed
about two and a half inches thick, with the glistening luster of
the Ceylon massive emery. It is a dark blue over most of its
surface with small irregular patches of pistachio-green.

The specific gravity of the rock is 3°64. It scratches topaz.

Under the microscope the rock is made up of stout colorless
grains two to four times as long as wide, with straight, dis-
tinct, almost cubical fracture lines.

These grains have a very low polarizing color, gray to white
of the first order. They have a much higher index of refrac-
tion than the cyanite which is scattered in them, in porphyritic
plates.

236 Emerson—Corundum and a Graphitic Essonite.

The cyanite is in rather square blades with good cleavage
and basal parting and around its borders has broken up into an
alteration product.

The whole field is filled with coaly matter in trains, aggre-
gates and groups of round balls which often occupy the center
of the corundum grains. This carbonaceous matter has plainly
been introduced in an oily or tarry condition and has been
inspissated in place, and the abundant graphitic matter in the
garnet gives indication of the same origin.

Amherst College, Amherst, Mass.

THE

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES. |]

Arr. XXIX.—An Experimental Investigation into the Exist-
ence of Free Ions in Aqueous Solutions of Electrolytes ;
by Juzius Oxsen.*

Van’t Horr in 1887+ showed that the gas laws of Boyle,
Gay-Lussac, and Avogadro, hold also for dilute. solutions.
From considerations of osmotic pressure, theoretically and
experimentally, he was able to apply Avogadro’s law to solu-
tions, and gave it as follows:

“The pressure which a gas exerts at a given temperature, if
a definite number of molecules is contained in a definite vol-
ume, is equal to the osmotic pressure produced by most sub-
stances under the same conditions, if they are dissolved in any
given liquid.” ;

These laws may be combined in the well known expression

PN =P
which holds for most substances, where P is the osmotic pres-
sure and V, T, and R, as in the gas law, are volume, absolute
temperature, and = a for a perfect gas, respectively, and in-
cludes Avogadro’s law if we consider kilogram-molecules of
the substances, as Horstmann has shown.

Van’t Hoff found that the above expression held for most
substances, but there were many important exceptions. These
exceptions were the acids, bases, and salts; in these the osmotic
pressures were greater than the law required. In other words,
he found that the solutions with abnormal osmotic pressures
were the electrolytes. He accordingly introduced into the
expression the coeflicient ¢ and wrote it

Ba Vee oe ava
* An abstract of a thesis presented to the Graduate Faculty of Yale Uni-

versity for the degree of Doctor of Philosophy.
+ Zeitschr. Phys. Chem., i, 4815 1887.

Am. Jour. Sc1.—FourtuH Series, Vout. XIV, No. 82.—OctToBErR, 1902.
17

238 J. Olsen—Investigation into the Existence of

This coefficient would be one for “ideal” solutions and greater
than one for the acids, bases and salts.

Arrhenius* showed, also in 1887, how these exceptions
might be explained; and thus the law hold for a// substances.
The osmotic pressure, other things being equal, depends upon
the number of particles present in a given volume of the solu-
tion, in the same way as the pressure of a gas depends upon
the number of its particles. If then the osmotie pressure is
too great, there must be more particles present than were put
in. These substances must therefore be dissociated, i. e., the
molecules split up into two or more parts. Arrhenius rea-
soned in this manner.

This same assumption had been made use of to explain the
abnormal pressures of some gases; but on account of the
chemical objections had not before been applied to solutions.
Arrhenius accordingly brought forth again the electrolytic
theory of Clausius, and established it, seemingly, by showing
how to caleulate the amount of dissociation.

Clausius had believed that “some of the molecules of an
electrolyte are dissociated into their ions, which move about
independent of each other.” During electrolysis these mo-
mentarily free ions would be guided to their respective poles.
According to this theory there is no direct decomposition of
the electrolyte by the current, as was held by Grotthiiss in his
theory. It has been shown since the time of Grotthtss that
electricity moves with the same ease in electrolytes as in
metals, so there is no place for such work as decomposition.

From the activity coefficient (a), which he considers the ratio
of the actual molecular conductivity to the maximum con-
ductivity, i. e., the conductivity at infinite dilution, Arrhenius
proceeds to show how to determine the coefficient 2 of the
formula, and thus test the theory. By a comparison of the
lowering of the freezing point of a liter of water, in which a
gram molecule of the substance is dissolved, and 7 as caleu-
lated from the conductivity method, he found the agreement
to within tHe limits of experimental error.

The theory has since been tested by a great many men, by
the method of the lowering of the freezing point, the rise of the
boiling point, osmotic pressure, conductivity, etc. There are
seemingly some exceptions as shown by Kahlenbergt and
others, but the preponderance of evidence is certainly in its
favor so far. Some of these exceptions may be due to the
somewhat limited knowledge we have of the true composition
of some substances. And the law has its limitations, in the
same way as the gas laws, and holds only for dilute solutions.
The theory, at least, explains a great many phenomena that
cannot be explained, at present, by-any other means.

* Tbid., i, 631, 1887. + Am. Jour. Phys. Chem., June, 1901.

i a

Free Ions in Aqueous Solutions of Electrolytes. 289

No direct experimental proof was attempted until Ostwald
and Nernst,* m 1889, performed the following experiment.
It will be given substantially in their own language.

A glass tube about 30 or 40™ in length, provided with a
stopcock, was drawn out to a fine capillary at one end. The
tube was then partly filled with mercury and hung upright,
with the capillary in a solution of dilute sulphuric acid. By
suction at the upper end the mercury was drawn up and the
sulphuric acid after it. When at a convenient height, about
the middle of the capillary, the stopcock was closed and the
liquid thus held in place. A platinum wire fused into the
tube connected with the mercury.

A large glass flask was now filled with dilute sulphuric acid
and insulated by placing it on a disk of hard rubber. The
outer surface of the flask was coated with tinfoil and its neck
varnished with shellac. The contents of the flask were con-
nected with the sulphuric acid into which the capillary tube
dipped, by means of a wet string. The positive pole of a
small electric machine was brought in contact with the tin-
foil, and the mercury of the capillary electrometer connected
with the earth.

As soon as the electric machine was set in motion the mer-
eury of the electrometer rushed up and at the same time bub-
bles of gas separated, which broke the thread of mercury in a
number of places.

This is the explanation given.—By charging the outer cov-
ering of the flask with positive electricity, the negative elec-
tricity on the inner side would be attracted and held, while
the positive would be repelled. The latter would go by the
wet string to the capillary electrometer and then by the plati-
num wire tothe earth. There is no closed current; the entire
movement of electricity is produced by induction. This proves,
they say, that free ions were present and that they moved.

They performed several other experiments, but all are on the
same principle, and the same conclusions are drawn.

There seems to be good reason for the belief that the con-
clusion arrived at does not necessarily follow, but that we need
further proof.

The potential of an electric machine is very high and the

flask they used in connection with the capillary electrometer

is really a condenser. As no substance insulates absolutely
there must be in such an arrangement some current, and in
such a restricted path there is a possibility of a considerable
difference of potential. They do not say that the potential
difference, between the two ends of the capillary electrometer,
was measured. If this exceeded about 1°22 volts there was
electrolysis and consequently a formation of gas bubbles.

* Zeitschr. Phys. Chem., iii, 271, 1889.

240 J. Olsen—Lnvestigation into the Kxistence of

J. H. Pratt* shows that no current passes through a capil-
lary electrometer below 1:22 volts. If no current can pass
through a Lippmann electrometer before about the electrolytic
limit of decomposition is reached, then this instrument is not a
good one for this experiment.

In order further to test the capillary electrometer when
connected for some time to voltages below 1:22, 1 made an
instrument as follows:

A glass tube about 15°" in length was drawn to a capillary
at about 5™ from one end. Into this shorter portion a platinum
wire was fused. The tube was then filled with mereury and
dilute sulphuric acid and inverted into a beaker containing
dilute sulphuric acid and mereury. A cork was put on the
upper end, and by releasing this a little the boundary between
the mereury and sulphuric acid could be brought into the
capillary. A space was left above the mercury so that it could
recede when connected up to a Daniell. <A platinum wire
connected with the mercury in the beaker.

In this form the instrument, so far as its effective action is
concerned, is exactly the same as that used by Ostwald and
Nernst.

It was connected up to a Daniell of about 1:05 volts for
hydrogen polarization and watched every few minutes for five
hours, but no bubbles of gas were observed. It was afterwards
connected up at the same voltage for eight hours with the
same result.

Incidentally it may be stated that it was found that, if the
tube was sealed at the upper end and entirely filled with mer-
cury so there was no chance for it to recede, bubbles were
formed in a few minutes, first at 1°05 and later with a new
tube and new mercury at about °88 volts.

In 1888 Ostwaldt+ cites the following as a proof of the
existence of free ions in solutions, but does not perform the
experiment.

Let two vessels, A and B, be filled, for example, with a solu-
tion of potassium chloride, insulated and connected by a siphon
filled with the same liquid. Now let a negatively charged
body, K, be brought near to A; it will act inductively upon the
system. A will become positive and B negative. If now the
siphon be removed and then the body K, A will remain posi-
tive and B negative. To discharge A we can connect it with
the earth, and the potassium ions, which are positive, will give
up their charge and become atoms. These would act upon the
water, forming potassium hydroxide, and hydrogen, which
would escape from the solution as a gas.

* This Journal, 1888 + Zeitschr. Phys. Chem., ii, 1888.

0 ee ———

~ Free [ons in Aqueous Solutions of Hlectrolytes. 241

Ostwald says that the amount of gas liberated would be too
small to be seen, and so, as before stated, he does not attempt
the experiment, but a year later, together with Nernst, gives
the experiment previously mentioned.

It was believed that an electrostatic charge would attract
and repel the unlike and like charged ions respectively. And
as it is the ions which in giving up their charge produce a
current, then if these ions could be guided by an electrostatic
charge we should get a current with electrodes indifferent to
the solution, and be able to detect it with a sufficiently delicate
galvanometer.

These were the beliefs upon which the following investiga-
tion was founded.

The experiments were performed on a brick pier capped
with a granite slab and entirely independent of any vibration
of the building. A delicate Rowland galvanometer together
with a telescope and scale was used. The scale was about
156:0™ from the mirror of the galvanometer. The deflections
were read to 0-1™™.

A U-tube was first used, into which were placed platinum
wire electrodes. The wires were wound into a spiral in order
to increase the surface exposed. Before putting them in, the
electrodes were heated to a white heat in a blowpipe flame.
Distilled water and a trace of sulphuric acid was used. It was
found that the cell always gave a deflection; however, it was
intended to take the deflection first, before an electrostatic
charge had acted upon the solution, and then taking out the
electrodes, let a charge from an electrophorus act upon the solu-
tion for some time and then put in the electrodes and take
the deflections again.

It was found that an electrode taken out of dilute sulphuric
acid and exposed to the air while the other remained in the
solution was, when replaced, strongly negative. As this effect
was so decided, and the exposure of the two electrodes could
not be made of exactly the same duration, nothing could be
gained from this method.

A cell was now made from a low round beaker about 53-0
in diameter. Platinum foils 22°5°" were welded to plati-
num wires which were run up through corks in a piece of
wood that served as a cover for the cell. The foils were bent
so as to fit around the curvature of the beaker and be about
2°0"™ distant from its edge. The cover of the cell was held in
place by small nails driven into it from the under side, and it
could be turned around so as to bring the electrodes into any
desired position around the edge of the cell.

The cell was placed on a wooden stand in such a way that the

242 J, Olsen—Investigation into the Existence of

disk of an electrophorus could be brought as close as I pleased
to either electrode. 10,5, 1 and 1 per cent solutions of sul-
phurie acid were tested in this cell. The deflections just after
the electrodes were put in the solution were very large and the
cell could not be worked to any advantage until after a couple
of days. The deflections with the more : dilute solutions were
not so large, and thus it did not take so long for them to get
into working condition.

The electrophorus disk did not hold the charge over 8 or
10 minutes and had to be recharged. Every time a charged
body was brought near to the cell, although the galva-
nometer circuit was open there was a considerable deflection
due to ordinary electrostatic induction, and this would show
itself also on closing the circuit to take a deflection. It was
necessary then to avoid this as much as possible, and this
could be done if the charge could be retained for a con-
siderable time. The knobs of two Leyden jars were then
placed close to the opposite electrodes and charged, the one
- positive and the other negative, by a friction machine. The
jars of course hold the charge much better, but the effect
did not seem to be very decided, which is perhaps due to the
small surface exposed by the knobs.

In order to have a charge that would waste but slowly and
at the same time have a considerable surface exposed, I con-
nected the knobs of two Leyden jars to the disk of an electro-
phorus. The disk was about 21:0™ in diameter and was gener-
ally brought to within about 2-0 or 3-0°™ of the cell and the knobs
of the jars charged with the friction machine. The intensity
of the electrostatic charge was judged, partly, by the induction
produced while charging, and as this arrangement gave by far
the greatest induction, it was believed to have the greatest
effect on directing the ions, as was found later to be true.
The friction machine gave a spark of from ? to 1 inch in
length. The induction while charging is considerable, and
although it seems to be about the same in whatever position
the electrodes or disk are placed, so long as the charge is of
the same sign, and although the recharging, with the jars,
need be done only every 20 or 30 minutes, yet it was thought
best to charge the jars only when the electrodes and jars are
always in the same relative position.

The following positions were used and will be designated as
follows :—Positions (1), (2), (8) and (4), figure 1.

The T-shaped figure represents the disk,which can be charged
by the jars and the electric machine. The cell is represented
as seen from above. The ares of circles represent the plati-
num electrodes.

°° ewe eee

a ae

Free Ions in Aqueous Solutions of Electrolytes. 248

The jars in connection with the disk were charged only
when in positions (1) or (4), and discharged before the elec-
trodes were turned into positions (2) or (8).

The galvanometer circuit is kept open except at the moment
of taking a deflection. The direction of the galvanometer
needle was determined beforehand.

The time between each deflection in the different series was
generally 2 or 5 minutes. I will describe a 20-minute series,
deflections being taken every 2 minutes, starting from posi-
tion (1).

Before the jars were charged, the deflections were taken
in position (1) every 2 minutes for 20 or 30 minutes or until
the deflections became about uniform; then the cover of cell
together with the electrodes were turned through 90° into

eee

Position 1. Position 2.
| Sa
Position 3. Position 4.

position (2), say. The turning was done very slowly and care-
fully so as not to disturb the liquid more than necessary, and
care was also taken to avoid shaking or turning the cell. The
deflections were then taken in position (2) every 2 minutes for
20 minutes, and then the electrodes were turned 180° through
the first position into position (3), and the deflections taken in
that position every 2 minutes for 20 minues. These readings
would indicate the condition of the liquid at positions (2) and
(3) before being acted upon by an electrostatic charge, and
comparing these readings with the readings after the solutions
has been acted upon by a charge would show the effect of the
charge.

After the condition of the lianid was determined as above
stated, the electrodes were turned into position (1) and the jars
in connection with the disk charged positive or negative by
the friction machine. The charge was allowed to act for the
same interval of time as was used before, namely, 20 minutes,
and the deflections taken in this position also, in order to keep
the cell always in the same condition, that is, deflections were
taken every two minutes throughout the entire experiment. If
the cell is allowed to rest even a few seconds over the regular
interval, a change in the amount of the deflection is noticed.
The cell runs down quickly and recovers quickly.

244 J. Olsen—ILnvestigation into the Existence of

If the deflections in any direction are large at the begin-
ning of the observations, the direction camot be reversed by
this charge, and this method and the effect is only a differential
one, but if the cell is in a more stable condition the directions
can be reversed, and this was done scores of times.

The induction produced when the jars are discharged | is of
course in the opposite direction from what it is when charg-
ing, but it is not large and does not appear in the third reading
of 2-minute intervals. Diffusion acts against the heaping up
of positively or negatively charged ions in any part of the
liquid, and this often appears in the last observations of a 20-
minute series. ;

A solution of potassium hydroxide and a solution of potas-
sium sulphate were also tested by this method, using platinum
electrodes as before, with the same results as with dilute sul-

phuric acid.
ne / Jef vuLw

Incidentally it may be mentioned that the induction pro-
duced, in the galvanometer circuit, by bringing up a charged
body to the cell was of about the same order of magnitude
whether there was any liquid in the cell or not.

One acid, one salt, and one base were used in these experi-
ments. The only reason for choosing these particular ones
was, that it was necessary to employ some substance which
on being decomposed and set free would not affect the pla-
tinum electrodes, and it was desirable also to have substances
whose ions were known to have a considerable velocity.

Distilled water was always used.

It was now thought that it would be interesting to study the
effect of an electromotive force, less than the decomposition
value, upon an electrolyte. Accordingly the following experi-
ments were performed :

No electrostatic charge was here employed.

A solution of different percentages of sulphuric acid was
put into a cell having four platinum electrodes. The cover of
the cell was so made that the inner or middle electrodes could
be rotated independently of the outer.

The two inner electrodes @ and a were connected to the
galvanometer circuit, and the two outer electrodes A and A’
were to be connected to a Daniell. qaand A were about 8:0 ™™
apart, and @ and @’ about 50:0™™ apart. The circuit to the

Free Ions in Aqueous Solutions of Electrolytes. 245

galvanometer was closed only while taking a deflection, and
this was never done while the circuit of the Daniell was closed.

The connections were generally made as follows (fig. 2) :

If a of the galvanometer circuit was found to be negative
and @’ positive when beginning the observations, then A was
connected to the positive pole and A’ to the negative pole of
a Daniell. After the Daniell had been connected thus for a
few minutes, it was disconnected and the sign of @ and a’
tested by closing the galvanometer circuit. @ might still be
negative and a’ positive, but by connecting up the Daniell the
same way and repeating this, @ could always be made of the
same sign as A. Then keeping the circuit of the Daniell
open and leaving everything else as before except turning the
electrodes a and a’ through an are of 180°, so that they ex-
changed places, and after a few minutes closing the galva-
nometer circuit, it was found that the direction of the deflection
was reversed. Turning back to the first position would reverse
it again. This could be repeated several times without connect-
ing up the Daniell again. |

This would show that the liquid in the neighborhood of a
positive pole connected to a Daniell is charged with positive
ions, or around the negative pole is charged with negative
ions, or both.

The electrodes connected to the Daniell became polarized
and remained so for some time. The same results were
obtained, however, if after disconnecting the Daniell, the
electrodes were discharged by short circuiting them. In this
case the Daniell had to be connected for a longer period to
obtain the same result, but this was all.

By connecting the Daniell to the znmner electrodes and the
galvanometer to the owter electrodes, the results were the same
but it took a longer time.

This would show that the ions diffuse in all directions, but
that they are more numerous between the electrodes.

Richarz and Lonnis* have shown, by means of the reaction
with titanic acid, the presence of hydrogen peroxide at the
cathode in dilute sulphuric acid at electromotive forces too
small to produce an evolution of gas. But no hydrogen peroxide
was discovered at voltages below one Daniell.

The same results as those above mentioned with one Daniell
were obtained in my experiments by using 4, 4 or ? the voltage
of a Daniell, but it took a correspondingly longer time. When
the electrolytic limit of decomposition was reached, the deflec-
tions became very large and the solution responded more
quickly to the influence of an electrostatic charge, but the

* Zeitschr. Phys. Chem., xx, 1896.

246 J. Olsen—Investigation into the Existence of

direction of the deflections were the same as
3 before. Inthe experiments of the effect of an
electrostatic charge on a solution which has
been subjected, first, to an electromotive force
less than the electrolytic limit of decomposi-
tion, and then on the solution after the decom-
position limit has been exceeded, a cell of the
form shown in fig. 3 was used.

a and @’ are the electrodes connected to the galvanometer
circuit ; A and A’ are the electrodes which may be connected
to the Daniell, and T is the disk which may be charged elec-
trostatically.

As has been stated, a larger and more prompt effect was
noticed when an electromotive force, although below the
decomposition value, was connected up to a cell and then an
electrostatic charge ‘brought up than when no current had
previously been allowed to act upon it. Also that a pair of
electrodes became of the same sign as the nearest electrode
connected to a Daniell, this also although below the decompo-
sition value of the solution.

This may be explained as follows :

Either the ions already free are heaped up around the elec-
trodes as they become charged by the current, or the current,
although below the decomposition value for the evolution of
gases, yet decomposes the electrolyte. Accordingly the fol-
lowing experiment was performed.

According to H. C. Jones,* strong acids and bases are com-
pletely dissociated at dilutions of about one one-thousandth nor-
mal solutions. If there is complete dissociation, then there
are of course no more molecules to be decomposed by the
current unless the water decomposes. I took, in order to be
sure of complete dissociation, a one-fourteen-hundredth nor-
mal solution of sulphuric acid, and using the cell shown in
figure 2, subjected the outer electrodes A and A’ to-a potential
difference of about 0°8 volt.

After taking several deflections by closing the aa’ or galva-
nometer circuit, in order to get the cell into a somewhat steady
condition, and noting the direction of the deflections, the AA’
circuit was connected up to a potential difference of "about 0:8
volt from a Daniell, in such a direction as to increase the
deflections in the galvanometer circuit. It did increase the
deflections and on reversing the direction of the current from
the Daniell, the deflections in the galvanometer circuit
decreased and finally reversed.

Here no decomposition of the sulphuric acid could take
place as it was completely dissociated before, and the voltage

* Theory of Electrolytic Dissociation, p. 190.

Free Ions in Aqueous Solutions of Electrolytes. 247

was too low to decompose the water; hence it would seem
that there are free ions present and that they are heaped up
around the electrodes when they are charged by a Daniell.

Some time after the preceding experiments the following
were performed.

It is known that when two metal electrodes, as for example
copper, are placed in a solution of their salt and the circuit
closed, that the smallest potential difference between them will
give rise to a current and the metal will be dissolved from the
anode and deposited on the cathode.

If then we take two copper electrodes, weigh them carefully
and they on being placed in a solution of copper sulphate do
not give too decided a deflection, we may be able to change the
direction of the deflection by an electrostatic charge and then
on short circuiting them and keeping up the charge we should
get a diminution in weight of the anode and a gain in weight
of the cathode. |

Copper wires were bent into a flat zigzag form, in order to
give a larger surface exposed and at the same time so that
they might be placed near to the edge of the cell and conse-
quently as near as possible to the electrostatic charge.

The electrodes were marked, cleaned in nitric acid, washed,
dried and carefully weighed. The electrodes on being placed
in the solution gave a decided deflection, but I was able to
reverse this direction with two ordinary sized Leyden Jars
connected to a disk of an electrophorus and charged by a fric-
tion machine. The circuit was then closed through the galva-
nometer, and the electrostatic charge renewed occasionally by
giving the electric machine a few turns. If the circuit was
closed without any external resistance as by a short thick wire,
the cell ran down quickly and sometimes gave a deflection in
_ the opposite direction, so the circuit was made through the
galvanometer. The circuit was kept closed for about three-
quarters of an hour, breaking it only while recharging jars.
The electrodes were then taken out, washed with distilled
water, dried, and again weighed. It was found that the
cathode had gained about 1:2™" and the anode lost about 0°47".

It had been found throughout these experiments that the
negative charge was not as effective as the positive, which is
perhaps due principally to its wasting faster than the positive.
However, having four Leyden jars connected to the disk so as
to hold the charge longer, I found the negative charge the
more effective, and the direction of the deflection would often
be reversed at the first observation taken after char ging. The
first observation was not generally taken until about 10 min-
utes after charging so as to allow the induction to disappear.
Having reversed the direction with a negative charge, the jars
were discharged. Now leaving everything as before but

248 J. Olsen—FLiree Ions in Aqueous Solutions.

charging the jars positive, the direction of the deflection was
again reversed. |

The experiments were repeated with pure zinc amalgamated
electrodes in zine sulphate with the same results. The zine
cell did not, generally, give as iarge a deflection as the copper
and so could more easily be reversed.

The greater effectiveness of the negative charge when it is
sufficiently large may be due to the fact that the positive ions
move faster than the negative, and the attracted ions would
be guided more effectively than the repelled.

Toward the close of this work the humid weather interfered
with the experiments somewhat, but results were obtained
which indicated that it would be possible to entirely waste
away the anode, by letting the electrostatic charge act on the
closed circuit for a sufficiently long time.

It was intended also to test a non-electrolyte by this method,
using a sugar solution, but the distilled water I was able to
obtain always gave a deflection in the galvanometer, showing
that it contained some dissociated electrolyte, and therefore
this test could not be made.

These experiments have shown :—

1. That, when a cell containing an electrolyte is connected
_ to a Daniell, although the electromotive force is below the
decomposition value of the electrolyte, the solution in the
neighborhood of the anode becomes positive, or in the neigh-
borhood of the cathode becomes negative, or both.

2. That, after an electrolyte has been electrolyzed by a cur-
rent, it is affected by an electrostatic charge in such a way that
the liquid nearest the charge becomes of the opposite sign, or
the liquid farthest from the charge becomes of the same sign
as the charge, or both.

That is, an electrolyte after having been electrolyzed acts as
if it contained particles charged with electricity which are free
to move and may be directed by an electrostatic charge.

3. That an electrolyte subjected to an electromotive force
less than the decomposition value of the solution behaves in a
similar manner.

4, Finally, that an electrolyte which has never been acted
upon by a current also behaves in a similar manner when
acted upon by an electrostatic charge.

That is, it behaves as if it contained particles charged with
electricity, which are free to move, and these particles have
not been produced by a current.

This corresponds to the definition of free ions.

Tn conclusion I wish to thank Professor Wright, who sug-
gested the subject and a large number of the experiments, and
whose advice and assistance has always been cheerfully given.

Sloane Physical Laboratory, New Haven, June, 1102.

|

Penfield—Solution of Problems in Crystallography. 249

Art. XX X.—On the Solution of Problems in Crystallography
by Means of Graphical Methods, based upon Spherical and
Plane Trigonometry; by 8. L. PenFrexp.

Iyrropuction.—It is believed that all who have occasion
to teach mathematical crystallography must have had experi-
ences similar to those of the present writer. There are occa-
sional students who find the subject very easy. These are
naturally persons who are proficient in mathematics, and have
good imaginative powers anda quick eye for grasping geo-
metric forms. A little guidance from the instructor and some
experience in handling crystals and apparatus are all that such
persons require in order to obtain a very satisfactory grasp of
the subject. The majority of students, however, find the
subject somewhat difficult, one of the chief reasons for this
being that they are troubled by the mathematics, although
the latter embraces, for the most part, only simple principles
of geometry and spherical and plain trigonometry. Prob-
ably the majority of those who undertake the study of erys-
tallography are especially interested in the subject because of
its practical bearings on chemistry, mineralogy and_petro-
graphy. They are not especially mathematical in their tastes,
and the calculations needed for the solution of the numerous
problems which arise seem difficult and distasteful, not so
much because of want of suitable training, but because, in all
probability, mathematical principles have not been used very
much in their other studies, perhaps, for a number of years ;
hence it is thatinnumerable dithculties and discouraging mistakes
are often encountered when numerical calculations are made by
means of formulas and logarithms. Probably all crystallog-
raphers have found numerical calculations burdensome, and
the only real way to absolutely avoid mistakes is either to do
all work in duplicate, or to apply some other method of
checking results. The main object of the present communi-
cation is to set forth and explain certain graphical methods,
based chiefly upon the stereographic projection, which have
been found very useful in the writer’s laboratory, and it is be-
lieved that they are especially applicable not only for the
presentation of crystallography to beginners, but also for the
solution of many complicated problems which may arise.

All who have worked much at crystallography have doubt-
less made use of graphical methods to a greater or less ex-
tent, and many descriptions of such methods may be found,
but for the purposes of the present communication it does not
seem necessary to review the work already done by others, as

250 Penfield—Solution of Problems in Crystallography

the methods to be presented, though based on simple prin-
ciples, are worked out, for the most part, along lines differing
from those generally recommended. The principles involved
in solving problems in spherical trigonometry by graphical
_-methods have been already fully explained by the present
writer in a communication entitled, ‘‘ Zhe Stereographie Pro-

jection and its Possibilities from a Graphical Standpoint,” * °

to which frequent reference will be made. The advantages of
the methods there described are that the mathematical prin-
ciples involved may be comprehended easily, and only a simple
and inexpensive outfit is needed for carrying on the work. With
a little study, the methods may be mastered by any une, whether
well equipped in mathematics or not, and for the solution of
most problems the results are sufficiently exact. Whenever
numerical] calculations must be made, the processes of solving
the problems graphically furnish the key to the methods of
making needed caleulations.

From an educational standpoint it is always a gain to do
things understandingly, rather than mechanically, and the
advantages of the graphical methods about to be described are
that approximate solutions of all kinds of problems in erystal-
lograpby may be made quickly without the use of tables or
formulas of any kind, while at the same time an accurate and
useful projection, or map, has been constructed.

In the actual study of crystals, after having made all neces-
sary measurements and determined the system of crystalliza-
tion, three processes must generally be followed: (1) The
determination of the symbols of the several forms observed.
(2) The determination of the axial ratio from certain se-
lected fundamental measurements. (8) The calculations of
a number of interfacial angles from the fundamental meas-
urements. For the first of these processes graphical methods
are quite sufficient, especially when it is borne in mind that,
owing to the numerous imperfections of crystals, the best
measurements which may be had are at times only approxi-
mately near the truth. The two remaining processes must be
carried out by means of numerical calculations, unless the
methods suggested by Fedorowt are employed, but the

_ graphical methods, if likewise followed out, will furnish a—

ready means of checking numerical work. It may also be
pointed out that as examples multiply of minerals and other
chemical compounds whose crystallographic properties have
been carefully studied, it becomes less often necessary to make
calculations; for the work on any compound, when once well

* This Journal (4) xi, pp. 1-24 and 115-144, 1901.
+ Zeitschr. fiir Kryst., xxxii, 464, 1890.

’

‘be ep ee eS

— Se ees he, lL eee ace aes

by Means of Graphical Methods. 251

done and recorded in the literature, need not be repeated.
Hence a student of crystallography, working as would natu-
rally be the case, on well-known compounds, may be able to
solve the majority of the problems presented for study by
means of graphical methods, and thus be spared much of the
time-consuming labor of making numerical calculations.
Outfit.—For executing the graphical methods described in
this paper the following articles are needed: (1) Paper upon
which a graduated circle and certain stereographic scales
have been printed. (2) Stereographic protractors for plotting
and measuring. (3) Beam-compass. (4) Curved ruler. (5)

Two triangles and a set of drawing instruments, including a
needle-point and hard pencils.

The uses of most of the above-mentioned appliances have
already been described, and only a few suggestions concerning
them are now needed. The paper mentioned has a circle of 14™
diameter, graduated to degrees, printed on it, the circle shown
in figure 1 being a reduced copy. Four scales, of which fre-
quent use will be made, are also printed on the paper. No.
1 gives the lengths of radii for describing arcs of great
circles; for example, if it is known that the point J), figure
1, is 32° from P, the radius 7 for describing the great circle
WV DS may be taken directly from the scale. Scale No. 2 gives
the lengths of radii for describing ares of small circles about
points which are located on the divided circle; for example,

it being known that the point A, figure 1, is 25° from S, and

252 Penjield—Solution of Problems in Crystallography

through & a small circle 4 BA’ is to be drawn, every point of
whichis 25° from S, the radius 7’ is given by the scale. Ona .
sphere, the center for describing such a circle would be at S;
in the stereographic projection, however, the center c’ lies be-
yond S, on the diameter VS. Scale No. 3 is a stereographic
scale, giving the distance in degrees from the center C to any
desired point; for example, in figure 1, to the point x, which
is 51° from C. About # a small circle has been drawn, every
point of which is 46° from z.. By means of scale No. 3 the
needed points were located; a, 51°; p, 5°=51°—46°; and p’,

97°=51°-+46°, all measured from (’; and the radius 7” for
describing the circle is half the distance from p to p’, the cen-
ter c’’ falling somewhat beyond z Scale No. 4 gives decimal
parts of the radius of the divided circle, and is used for meas-
uring the lengths of axes, assuming the radius of the circle as
unity.

The stereographic protractors needed are three in number.
One of these (referred to as protractor No. I and shown in
figure 4 of the former article) is printed on cardboard, and
has along its base line a graduation which will be referred to
in this paper as the stereographic scale. The scale is the same
as a portion of that designated as No. 3 on the drawing paper;
hence this protractor is not essential, but its convenience will

ee se eS OC

by Means of Graphical Methods. 253

recommend it. The second protractor, figure 2, is a measuring
scale, consisting of a series of stereographically projected
small circles drawn about the 0° and 180° points. It was
described in the former paper as protractor No. I, having a
circular form, but one-half of the circle, which is stgloeie
ous, has since been cut away from the original plate. It will
be referred to in this paper as the small circle protractor.
The third protractor, figure 3, consists of a series of stereo-
graphically projected great circles, and will be referred to as
the great circle protractor. Purposely no numbers are printed
on this protractor, and those using it will do well to lay it
printed side down and number the graduation with India ink,
as shown in figure 3. If a number then happens to interfere
with the uses of the protractor, it may be easily washed away.
The uses of the protractors may be explained by reference to
figeure 1. If y and 7’ are two points on a stereographic pro-

~

jection, the great circle passing through them may be found
by centering the great circle protractor, figure 3, over the pro-
jection and turning it until an are is found which passes
through, or proportionately near, y and vy’. The position of
the protractor then determines the points a and / on the per-
iphery, and its graduation gives the angle at a, 33° 20’, or the
distance z to 2’. Knowing the distance z to 2’, the radius for
constructing the great circle may be taken directly from scale
No. 1 of the drawing paper. By now applying the small circle
protractor, figure 2, with its 0° and 180° points at @ and 4, the
distances @ to y and @ to y’ may be determined from the grad-
uation, and the difference gives the distance y to y’. Thus a
great variety of measurements may be made with facility and
rather astonishing accuracy. 7

A beam-compass is an essential for accurate work, as ordi-
nary dividers are not stiff enough to be relied upon for draw-
ing circles having long radii. A compass constructed on the
principles shown in figure 4 may be recommended, the idea
having been suggested by Professor Charles B. Richards of the

Am. Jour. Sor.—Fourta SERIES, VoL. XIV, No. 82.—OoTOBER, 1902.
18

254 Penfield—Solution of Problems in Crystallography

Engineering Department of the Sheffield Scientific School.
The beam is constructed from a round stick of hard wood,
# inch diameter, sawed lengthwise. The needle-point p is pre-
pared by gluing the surfaces of two pieces of hard wood,
placing a needle between them and pressing ina vise. After
hardening of the glue, the piece is trimmed down to con-
venient size, and is held firmly in place between the half-
rounds by means of two screws. Two rings 7 and 7 (obtained
from a harness shop) serve to grip the pencil firmly between
the two sticks. The block A is not essential, serving simply
to hold the pencil (fastened by an elastic at ¢) when its posi-
tion is shifted. A fine adjustment is had by moving the upper
end of the pencil, which gives a much slower motion to the
point. In order to adjust the pencil quickly with reference to
scales 1 and 2 of the drawing paper, it is convenient to have
these scales pasted on the beam. The writer has found com-
passes of this description most convenient, and rather prefers
them to the beam-compasses listed by dealers in drawing in-
struments at a rather high price.

A curved ruler is indispensable for drawing some very flat
circular ares. The form first suggested by Wulff* and modi-
fied. by the present writerf is recommended. The ordinary
drawing instruments need no especial comment. Two tri-
angles of transparent celluloid (45° and 60°), measuring
about 12 inches along the hypothenuse, are recommended.
Their edges serve as rulers, and when held one against the
other, the edge of the farther triangle being set at any desired
angle, the direction may be transferred to any point by sliding
the farther triangle along the edge of the near one, held firmly
against the drawing paper.

As the scale upon which the plotting is done is rather small,
a lens is recommended, and it is especially convenient to have
one mounted ona stand. The kind suggested by the writer,
and shown in figure 33 of his earlier paper, though expensive,
has proved most useful.

Having described the necessary outfit, attention may next
be directed to consideration of the principles upon which the
methods of plotting are based. Re:

The Spherical Projection.—The stereographic is often called
the spherical projection, but it is believed that it would be best
to restrict the use of the latter term, in so far as it concerns |
erystallography, to the projection of the faces of a crystal upon
asphere. This latter conception is a very simple one: A
crystal is supposed to be located at the center of a sphere,

* Zeitschr. ftir Kryst., xxi, p. 253, 1892.
+ The Stereographic Projection and its Possibilities, loc. cit., p. 188.

by Means of Graphical Methods. 255

figure 5, its vertical axis orientated so that it corresponds with
the axis joining the north and south poles; then from the cen-
ter of the sphere lines, normals are supposed to extend out,
at right angles to the several crystal faces, until they touch the
surface of the sphere, where they locate points, known as the
poles of the faces. Figure 5 was plotted with much care in
clinographic projection, and was made partly with the idea of
putting on record a good figure illustrating the simple prin-
ciples of the spherical projection. A combination of cube a,
octahedron 0, and dodecahedron d is shown, with normals ex-
tending to the surface of an imaginary sphere. The projected

great circles, uniting the poles of the faces, are shown in the
figure as ellipses. The figure well illustrates the important
principle, that the normals to a series of faces which are in a
zone all lie in one plane, and touch the sphere on an are of a
great circle. The length of are between the poles of any two
faces, as measured on a great circle, is the same as the angle
between the normals, hence the same as between the faces as
measured with the reflection goniometer. Wire models con-
structed on the principles shown im figure 5 are‘very helpful
to beginners.

For studying crystals, it is convenient to make use of the
stereographic projection. Generally the poles of one hemi-
sphere only (the upper) are represented, the projection being
made on the horizontal plane, or plane of the equator. The

256 Penfield—Solution of Problems in Crystallography

point of vision is then at the south pole, 001, of figure 5. A
model, similar to one made by the writer and illustrated in an
article on Map Projection,* is a help in conveying to begin-
ners correct ideas of the principles upon which the stereo-
graphic projection is based.

The fundamental proposition of mathematical crystallog-
raphy.—That property which must first be ascertained for
every crystallized substance is its system of crystallization,
and then follow the determinations of the inclinations and
lengths of the axes. Figure 6 represents a general proposition
(the axes shown are those of rhodonite, triclinice) in which the
quantities sought are the axial inclinations a, 8 and 7, and the
lengths of the axes a and ¢, it being assumed that 6 is equal to
unity. The angles a, z and oe of figure 6 are those of a plane
triangle with the axes 4 and ¢ for its sides, and it is evident
that after having determined @ (the 6 axis being unity) it is

only necessary to find either z or pe in order to solve the tri-
angle and obtain the length of the ¢ axis. Likewise, 7 having
been determined, it 1s only necessary to find either ct or o in
order to obtain the length of the @ axis. With @ and the
length of one axis, either @ or c, determined, the length of the
other axis may be found if either » or yis known. For the
complete solution of a problem in the triclinic system, there-
fore, the axial inclinations a, § and 7 must be determined, and
in addition only two other angles are needed (7 and z, for ex-
ample) in order to find the lengths of aandc. As will be
shown, all of the desired parts a, z and p; 7, o and 7; and
8, wand » are to be found on astereographic projection, where
angles are preserved so that they may be measured.

Figure 7 represents a hypothetical case, a combination of
the three pinacoids a, 100; 6, 010; and ec, 001 of rhodonite,
in combination with the pyramid p, 111. Five measurements
made on such a crystal suffice for the solution of the problem,
and for rhodonite they may be as follows:

* This Journal, xiii, p. 247, 1902.

—— a
.

wa ae ee er eee ts

a a ee

by Means of Graphical Methods. oe

ab, 100010 = 94° 26’ aAp, 100A 111 = 56° 19’
ae, 100A 001 = 72 364 cAp, OO1A111 = 29 49
bAc, OLOA001 = 78 424

‘The lengths and inclinations of the axes derived from the

foregoing measurements are as follows:
mbes G= BWE285.5-1 9-06 212 Gs-.:
a=103° 18’; B= 108° 44’; 7 = 81° 39’

From the measurements given above the poles a, 6, ¢ and p
of rhodonite have been accurately located in the stereographic
projection, figure 8, and through the poles the principal zones,
corresponding to those of figure 5, have been drawn. The

17100

relation between figures 6 and 8 is a most important one.
The angles a, z and p, made by the meeting of the great circles
at the pole a, figure 8, are identical with a, z and p of figure 6,
and a like correspondence holds true at the poles 6 and ¢ of
figure 8, as indicated by similarity in the lettering. In order
to make clear that the angles a, z and p of the stereographic
projection, figure 8, are identical with a, z and p of the rhodo-
nite axes, which, it must be remembered, are shown in clino-
graphic projection in figure 6 and therefore somewhat dis-
torted, let it be assumed that a model corresponding to figure 5,
but constructed so as to represent the triclinic system, is at
hand. Since the front pinacoid @ of figure 7 is parallel to the
band ¢ axes, its normal is at right angles to the plane of these
axes ; hence if we imagine ourselves as looking at the 6 and ¢
axes of a triclinic crystal in the direction of the normal to the

258 Penfield—Solution of Problems in Crystallography

face a, 100 (the axes being properly orientated within a sphere,
and the latter standing sufficiently far away to give the effect
of orthographic projection), the relations shown in figure 9
would be seen. The outer circle represents the bounding sur-
face of the imaginary sphere. The } and ¢ axes are seen
within the sphere with their true inclination and lengths;

= 108° 18’, and 6: ¢ = 1.00: 0°621. Normals to all erystal —

faces parallel to the ¢ axis (compare figure 5) are located in a
plane at right angles to the ¢ axis, which plane would be fore-
shortened in orthographic projection to the line CC’. Like-
wise normals to all faces parallel to the } axis lie in the plane
represented by the line BA’. Since the planes represented in
figure 9 by CC’ and BB’ are at right angles, respectively, to

fe
the c and } axes, they must make with one another an angle
equal to that of the axes, that is a. All forms intersecting the
6 and c axes at their unit lengths, therefore, having the-direc-
tion of their intersections represented by the line de, figure 9,
would have their normals in a plane at right angles to the
direction bc. . This plane seen in orthographic projection is
represented by the line PP’, and, therefore, owing to the con-
struction, the planes BB’ and PP’ make with one another an
angle equal toz. In like manner the planes CC’ and PP’
make with one another an angle equal to p.. Hence it is that
the angles made by the planes in which the normals are located,
represented by CC’, BB’ and PP’ in figure 9, are to be found
in the stereographic projection, figure 8, at ¢, a line from a
to the center of the sphere being the common intersection edge
of the three planes. In like manner it may be demonstrated

eal ee -_

Ay eae Se

by Means of Graphical Methods. 259

that at the poles } and «¢, figure 8, where normals to the pina-
coid 010 and the base 001 meet the sphere, the plane angles
8, wand » and y, o and cz of figure 6 are to be found. How
these facts are to be made use of, will subsequently be shown
by examples.

Another principle which may be made use of for the solu-
tion of many problems is that based upon the conditions estab-
lished by three faces in a zone, known by some as the cotan-
gent and tangent relation.* For the purpose of illustrating
this principle, reference is made to figure 10, which represents
a regular arrangement of dots conditioning angles like those

10

found in the prismatic zone of crystals of rhodonite. The
dots may be regarded as representing crystal particles or mole-
cules, and the forms shown by the heavy outline of the figure
are the planes @ (100), 6 (010), m (110), J/ (110) and f (130),
as seen in orthographic projection. It may be assumed that
the crystal is made up of layer upon layer of particles like the
one shown, figure 10 being an orthographic projection of such
a system of particles, as seen in the direction of the vertical
axis of the crystal. Since rhodonite is triclinic, a single layer
of particles, represented by the dots, does not le in a plane
parallel to the paper, but in a plane tilled down to the front -
and right by an amount indicated by the axial inclinations

*Dana’s Text Book of Mineralogy, edition of 1898, p. 31.

260 Penfield—Solution of Problems in Crystallography

8 = 108° 44’, and a=1038° 18’. The lines w, —w# and y, —y,
therefore, are not the true crystallographic @ and 6 axes, but
the somewhat foreshortened orthographic projections of the
axes. Notwithstanding foreshortening, however, for all pur-
poses of calculation in the zone under consideration, the lines
x,—x and y,—y are lines of reference, regarding which the law
of definite mathematical ratio of crystal faces holds true, the
same as for real crystallographic axes. The angle between
x,—x and y,—y is determined by the measurement of 6 (010)
on @ (100), and the angles of a third form, m (110), establishes
the relative lengths of the lines of reference Ox and Oy. Any
fourth face in the zone, therefore, when referred to Ox and Oy
must have an inclination in conformity with the fundamental

law of definite mathematical ratio. For example, in figure 10
the face f is parallel to the line x z which intersects the line of
reference Oy at one-third its length; the symbol of fis there-
fore 180, and the angle 5A f is wholly dependent upon the
two angles bA.a@ and bAm. Hence it is, that knowing the
symbols and angles of three faces in a zone, the angle of a
fourth face may be found if its symbol is given, or its symbol
may be determined if its angle is given.

Figure 10 furnishes a key to the solution of problems like
the one just given. Let it be assumed that on a rhodonite
crystal two angles have been measured, as follows: DA@=
94° 26’ and bAm = 45° 53’, and that the poles 010, 100 and
110 are plotted on a divided circle, figure 11. Lines of ref-
erence 7,—az and y,—y are then drawn through the center,
parallel, respectively, to the 6 and @ faces, therefore intersect-

by Means of Graphical Methods. 261

ing the divided circle at 90° from the poles 010 and 100. A
line parallel to the face m (at right angles to the normal to 77)
is next drawn. To do this easily use may be made of the
principle of geometry that the angle between two lines meet-
ing at the circumference of a circle is measured by half the
are intercepted on the periphery; hence the line may be drawn
from x to 91° 46’ (twice bAm, 45° 53’) on the graduation of
the divided circle, measured from —a. This line determines a
point y on the second line of reference, and establishes the
ratio of Ox to Oy, the numerical value of which, however,
need not be determined. Designating the distances Ox and Oy
as w and y, respectively, the law of definite mathematical ratio
demands that all faces in the zone under consideration must
intercept the two lines of reference at definite multiples or
fractions of e and y. In figure 11 the angle of DA fis 18° 4’,
and by drawing a line from «to 36° 8’ (double the angle) on
the periphery it is found that it intersects at $y; the symbol
_ of the face 1 is therefore 130. Likewise by drawing lines from
ato —ty, —y,and —3 y, the angles of the forms 130, 110
and 310 (measured from 010) are found, the values being most
easily obtained by halving the ares intercepted on the per-
iphery, as indicated in the figure. The case just cited,
although a simple one, is general, for it is wholly a matter of
choice in the orientation of a rhodonite crystal that the forms
designated as 6, 7, m, a, ete. are taken as parallel to the vertical
axis.

Cases may arise where reliable measurements may be had
from three known forms, for example, 0, f and m, figure 11,
and it is desired to find the angle of a fourth form a, 100,
which may either be absent, or be of such poor quality as to
yield no satisfactory teens bat To solve such a problem
construct the line 7,—# parallel to 6, and from a draw lines
parallel to fand m. The direction of @ is then found by
constructing a line through the center which will be intersected
by the lines previously drawn in the ratio $:1. With the aid
of a straight edge and a pair of dividers such a line is easily
found.

Again, it will often happen that the zone under considera-
tion is of such a nature that the ratio of the intercepts on the
line of reference Oy can not be told by simple inspection, in
which case the following procedure may be resorted to. Any
four faces taken in succession may be designated as A, B, CO
and J), and their indices, respectively, as Akl; mno; pqr,
and wyz. Taking the first line of reference (corresponding
mw, — oe, eure 11) as parallel to the face A, the ratio of the
intercepts on a line parallel to the fourth face 'D may be found
from a system of left to right, right to left, cross-multiplica-

262 Penfield—Solution of Problems in Crystallography

tion and subtraction, carried out according to the following
scheme :*

A, hkl 12 1—3 2—3

B,mno _hn-—km , ho-—lm , ko-ln

B,mno my—nx’ mz—-ox’ nz—o0y

D, xyz

D, aye

C, par _ &9-yp . er—ep, yr—2q
CO, pqr pk—qh’ pl—rh’ qi-rk
A, Akl 7

The result of multiplying one of the upper expressions by
that directly beneath gives the desired intercept. The result

in some cases is indeterminate, as, for example, schemes 1—3
and 2—3 applied to prismatie zones, where the third index of
the forms is zero, or scheme 1—2 applied to a zone between
some prism and base, where the first and second indices of the
forms maintain the same ratio. When a numerical solution is
desired the following equation is generally made use of:

Cot AC — cot AD = rational intercept
Cot AB — cot AD .

Provided AD equals 90°, its cotangent equals zero, and the
problem is then much simplified. It may then be easier to

* Nore.—The scheme of cross-multiplication and subtraction here sug-
gested is essentially like that given in Dana’s Text Book of Mineralogy, page
31, Groth’s Physikalische Krystallographie, page 584, and other works, but
by adopting, as has been done, a somewhat more symmetrical arrangement
and procedure, the operation is more easily remembered and carried out.

by Means of Graphical Methods. 263

consider the relation in terms of tangents rather than cotan-
gents. The tangent principle, though simple, is so exceed-
ingly important that it is thought best to introduce an
example indicating a graphical method of solution. The illus-
tration, figure 12, is chosen from barite, the forms shown being
those of the macrodome zone, with c=001, and z=100. It is
assumed that the angles have been measured as follows:
eye —2' bbs 3 cA d=38" 514’, and cA u=58° 102’.. The
poles are easily located on the graduated circle, and the slopes
of the several faces plotted by drawing lines from « to points
on the periphery equal to twice the value of the angles, meas-
ured from —#. It is then evident that if w is taken as the
unit dome 101, the intercepts of d and / on the vertical axis
are respectively 4 and 4, and their symbols therefore 102
and 104. If the length of the @ axis of barite were
known (a@=0°815), the slopes of uw, d and 2 would naturally
be plotted from the unit length of a, when the slope of w

would determine the length of the vertical axis (e=1°'314).
For plotting and measuring the lengths of axes, the radius of
the divided circle may be taken as wnzty, and measurements of
axial lengths made by means of scale No. 4 of the sheets on
which the divided circles are printed.

Another procedure applicable to all cases where the axial
inclinations are 90° is illustrated by figures 18, 14 and 15. In
figure 13 three axes meeting at 90° at O, and their intersec-
tions with a pyramid abc, are shown in clinographic projection.
The line Oz, at right angles both to the vertical axis and the
edge wb, is an important factor to consider, and may be desig- ©
nated as the base-line of the triangle cOx. Another important
factor to be considered is the slope of the pyramid, or its in-
clination, along the line ac, with the base-line Ov. From
figure 14, which represents a section along the plane ¢Oz of
figure 13, it may be seen that the slope of a pyramid, or the
angle made by the line cw with the base line Ox, is the same as
the interfacial angle of base on pyramid, 46° 6’ in the figure.
Knowing the base-line Ox and the slope of any pyramid, there-

264 Penfield—Solution of Problems in Crystallography

fore, the problem may be plotted to scale and the length of
the vertical axis Oc determined. The foregoing simple trigo-
nometric relations may now be applied to figure 15, where the
divided circle is employed for the double purpose of a stereo-
graphic projection and for solving problems in plane trigo-
nometry. Let it be supposed that the poles of the three pina-
coids 100, 010 and 001, and of a pyramid p have been accu-
rately located in stereographic projection. Through 001 and p
a line is drawn which is the projection of a great circle (a meri-
dian) determining the point m on the divided circle. Since
the axial inclinations are 90°, the value of ct (compare figures
6 and 8) is measured by the are 100/Am, which in figure 15 is
39° 11’. Considering the radius of the divided circle as wnzty

15

78°227"|

00

and as equal to the 0 axis, the tangent of z (100A 2) gives the

intercept on the @ axis; accordingly this may be found by
drawing a chord from 010 to a point on the divided circle
equal to twice 100 Am, measured from 010. Applying scale
No. 4 of the printed sheets along the direction of the @ axis,
the length of the axis is found to be 0°815. The chord da,
figure 15, determines also the point # on the radius passing
through p, and the distance from x to the center is the base-
line corresponding to Ox of figures 138 and 14. The numeri-
cal value of the length of the base-line, which is equal to the
sine of z, need not be determined, but the distance may be
transposed to #’ on the horizontal line by means of dividers.
Applying the stereographic scale to the radius passing through
p, the angle of 001A p (the slope of the pyramid) may be
measured, equal to 46° 6’ in figure 15, and a line drawn at this

a. ay

ae tee

OP

ae ae

ia ee

by Means of Graphical Methods. 265

angle from «' intersects the vertical line at a point ¢, which,
when measured with scale No. 4 of the printed sheets, gives
the length of the vertical axis, c=0°66. The whole procedure
of determining the’ lengths of the @ and the ¢ axes 1s exceed-
ingly simple both in its principles and in its execution, and
with slight modifications may be applied even to the mono-
clinic system where the # axis is at right angles to the @ and ¢
aXes.

Instead of laying off the base-line Ox’ to the right of the
center by means of dividers, as shown in figure 15, it may be
plotted to the left, by making use of the principle that Ow
equals the sine of tr. Accordingly locate m to the left on the
divided circle, and draw a vertical chord ; then plot the slope
of p from the center, making use of the graduation of the
divided cirele, and measure the length of the vertical axis on
the chord a’c. On one divided circle, therefore, a stereo-
graphic projection may be made in three quadrants, reserving
the upper ‘left hand quadrant for plotting the length of the
vertical axis. It may be noted that almost at a glance at such
a chart the two important angles of the two-circle goniometer,

and p, may be told from the graduation; for example, the
radius through p, figure 15, determines m, and, therefore, the
angle g, measured from 010, which is nearly 51° in figure 15 ;
and the slope of p, plotted from the center and intersecting
the vertical chord at 0°66, meets the divided circle at a little
over 46°, which is the angle o, measured from 010. For
plotting measurements made with the two-circle goniometer
and interpreting the results, the method is most admirably
adapted.

Having explained at considerable length the principles in-
volved in solving problems graphically, some examples of the
application of the methods to specitic problems in the several
erystal systems will next be given. The examples cited are
problems which have been encountered in the writer’s labor-
atory, and in all cases it will be assumed that the fundamental
measurements were made to the nearest minute, hence the
results of plotting will indicate the accuracy of the graphical
methods. Generally, the scheme indicated in figure 15 will be
followed, that is, the stereographic projection will be made on
the lower half of the divided circles, reserving the upper half,
or the upper left hand quadrant, for plotting the intercepts on
the vertical axis. The horizontal radius, from the center to
the right, will indicate the 4 axis and the lower radius the @
axis. Lase-lines corresponding to Ox of figures 13, 14 and 15
will be transposed to the horizontal diameter, as in figure 15,

‘and indicated by correspondence in lettering. The graphical.

solutions were all executed carefully with fine pencil lines and

266 Penjfield—Solution of Problems in Crystallography

subsequently traced with India-ink in order to adapt the work
to photo-engraving. The illustrations, therefore, are all from
the original sheets on which the plotting was done.

Isometric System.

In this system the axes are of equal lengths, and it is neces-
sary to determine from measured angles the symbols of the
forms, or, having the symbols given, to find the interfacial
angles. The first problem that will be presented is one which
might, for example, be given to a beginner for the purpose of
introducing the subject of mathematical crystallography and
showing the relations between crystal forms and their angles.
It may be supposed that the student is supplied either with
models or crystals from which measurements are derived.

Starting with a simple combination, the poles of the cube
are first located at 100, 010 and 001, figure 16, and those of
the dodecahedron at distances of 45°, midway between them,
the positions of 101 and 011 being determined by the stereo-
graphic scale. By means of the small circle protractor it may
now be found that the distances between the poles of the
dodecahedron are 60°. The octahedron truncates the solid
angles of a cube, and is in the zones between cube and dodeca-
hedron (compare figure 5); its poles may therefore be located
by drawing the necessary great circles. The pole 111 being
thus located, it may be found by the stereographic scale that
the distance from 001 to 111 is 542°, theory 54° 44’, and from
111 to 110 is 354°, while like results will be obtained by meas-
uring from 100 or 010 with the small circle protractor. Thus
a conception is quickly gained, not only of the relations of the
important forms, cube, octahedron and dodecahedron, but also
of the values of their interfacial angles.

For simplicity of plotting, the tetrahexahedron is the next
best form to consider, and a fluorite crystal with bevelled edges
may be chosen as an example for study. The angle of one of
the bevelling face to cube is 18° 26’, and the poles appear on
the stereographic projection in the zones between cube and
dodecahedron. ‘To determine the symbol of one of the poles,
for example, 310, figure 16, it is only necessary to note the
distance from 100 (18° 26’) and draw a line from 010 making
the same angle with the diameter, which line intersects the
front axis (radius) at $; the ratio on the first and second axes
is therefore 4 @: @, and the indices of the pole are 310. The
interfacial angle of the tetrahexahedron, the distance from 310
to 301, measured with the small circle protractor, was found
to be 25° 50’, theory 25° 50’. |

4
nt
i

by Means of Graphical Methods. 267

The trisoctahedron bevels the edges of an octahedron; its
poles, therefore, are always to be found on the zones between
octahedron and dodecahedron. Taking a crystal of galena, an
octahedron with bevelled edges, as an example, the angle of
the trisoctahedron, measured over an octahedral edge, is 38° 56’.
One-half of the foregoing angle, 19° 28’, is therefore that of
trisoctahedron on dodecahedron, and the pole between 110 and
111 is easily located by means of the stereographic scale. To
determine the symbol of the form, which, being in the zone
between 110 and 111, must have a 1:1 relation on the first

16

To 2a \ Lo2G, 20. 3a:

and second axes, draw the line from 010 to 100 and thus de-
termine the base-line Oy. The slope of the face, its angle on
001, is 70° 32’ (complement of 19° 28’), hence from 7’ (the
base-line transposed to the horizontal diameter) construct a line
at 70° 32', which, if continued, will meet the vertical axis
(radius) at 2a. The form intersects the axes, therefore, at
a@:a: 2a, the indices being 221. To fix the pole 212 between
111 and 101, place the small circle protractor over the projec-
tion, its zero at 010, and space off the distance 19° 28’ from
101 with dividers ; then transpose to the paper. Another and
better way is to lay off the distance 19° 28’ from 001 on the
horizontal diameter, and construct a small circle with radius

268 Penfield—Solution of Problems in Crystallography

70° 32’, taken from scale No. 2 of the printed sheets. To find
the interfacial angle from 221: 212, make use of the great
circle protractor for finding the great circle between the poles,
construct the great circle and measure with the small cirele
protractor as indicated on page 253. From 221 to 212 was
found to be 27° 35’, and from 221 to 122, 27° 15’, the ealeu-
lated distance being 27° 16’.

The poles of the trapezohedron are found on the great circles
between octahedron and cube. With garnet as an example,
measurement may be made over an axis, and the angle is 70°
32’, one-half of which is the angle of cube on trapezohedron.
The pole between 001 and 111 may then be located by means
of the stereographie scale, at 35° 16’ from 001, and the slope
plotted from the base-line Oy’ goes to 4 on the vertical axis ;
_ the indices are therefore 112. By applying the great cirele
protractor it is found that 112 is in the zone of the dodeca-
hedrons 101 and 011, and this relation may be made use of for
locating other poles of the trapezohedron, 211 and 121. Other
interfacial angles may then be measured; for example, 112 A112
and 211 A211 were found to be, respectively, 48° 30’ and 48°
10’, theory 48° 11’, and 211 A121 and 121A112, 33° 45’ and
338° 15’, theory 33° 34’.

The hexoctahedron most frequently met with on garnet
bevels the edges of the dodecahedron, the angle between the
bevelling faces being 21° 47’, or 10° 54’ from the trapezo-
hedrons. To locate the pole between 101 and 112, place the
small circle protractor over the projection with its zero at 110,
then with dividers space off 10° 54’ from 112 and: transfer to
the paper. A radius drawn through the point thus found in-
tersects the divided circle at a little over 264° from 100. A
line from 010, drawn at right angles to this radius, intersects
the first axis (radius) at 3, and establishes the relation ga: a
on the first and second axes and also the baseline Ox. The
distance from the hexoctahedron to 001, measured with the
stereographic scale, was found to be just short of 37°, and this
slope plotted from «@’ intersects the vertical radius at 3. The
form is therefore 4a : a : 4c, the indices being 213. The poles
of the hexoctahedron under consideration are especially easy
to locate by means of zones, as indicated in figure 16, and to
show the accuracy with which the plotting was done some dis-
tances measured on the projection are given, as follows:
111 A231=22° 5’, theory 22° 12’; 010 A\231=36° 457, theory
36° 42’; 312 \312=31° 10’, theory 31° 00’; and 001A 231=
74° 25’, theory 74° 30’. The method of finding the symbol
of the pole 231 (a: a: 38a) by means of the base-line Oz
and slope 74° 25’ is indicated in figure 16, but needs no dis-
cussion. Attention is called to the six poles in the zone

SO 26 6! op Oe ra

Bit =

ee pl i rn. a) 5 ee el

by Means of Graphical Methods. 269

between 001 and 210, determined by the intersections of zones.
Their distances from 001 being measured with the stereo-
graphic scale, the slopes may be plotted either from z’ to the
right or, better, from the center, and their intersections with
the vertical radius, or the perpendicular erected at «’ to the
left, then determined. The distances from 001 to the several
poles, measured with the stereographic scale, were found to be
as follows:

Form 215 214 213 212 211 42]
(eased. 924°°5’ 29° 5’ .36°.55! 48° 10’. 66° 00’ 77°. 15’

Calenlated, 24 5 2912 36 42 48 11 65 54 7 238

Two methods for determining the symbol of a hexocta-
hedron, observed in combination with the cube on crystals of

17

fluorite from Weardale, Durham, England, are indicated by
figure 17. No zones were encountered in measuring the crys-
tal, hence two measurements were necessary for determining
the form, those made being between two hexoctahedron faces
adjoining a cube, 21° 13’, and from cube to hexoctahedron,
24° 19’. Starting from the center of a-stereographic projec-
tion, the poles of eight faces of the hexoctahedron must lie on
a small circle drawn about 001 with a radius of 24° 19’, which
may be taken from the stereographie scale. Moreover, the
poles of the hexoctahedron must lie symmetrically on either
side of a radius passing through 111, distant 10° 36’ (half of
21° 13’) from the pole of a possible truncating face, or 79° 24’
(the complement of 10° 36’) from 110 and 110; accordingly
the point p was located at 10° 36’ from the center on the radius
to 110, and, taking the radius 79° 24’ from scale No. 2 of the

Am. Jour. So1.—Fourts Series, Vout. XIV, No. 82 —Octoper, 1902.
19

270 =Penfield—Solution of Problems in Crystallography

sheets, a small circle through.» was constructed which inter-
sected the small circle previously drawn at the point marked
317. Through the point thus found a radius was drawn, which
intersected the divided circle at 18° 25’, and from 010 a line
at right angles to the radius was drawn, thus determining the
intercept % on the first axis, as also the base-line Oy. The
angle of enbe on hexoctahedron is 24° 19’, hence a line drawn
from the center at this angle was found to meet the perpendic-
ular from y’ at 0-141, determined by scale No. 4 of the printed
sheets, or practically ith of the radius, the true value being
0-143 ; hence the axial relation of the face under consideration
is $a: a: 4a, the indices being 317.

It is always an advantage to plot near the periphery of a
stereographic projection, where the spaces between degrees
are larger than near the center; therefore a solution, similar to
the one just given, but made about 100, is also shown in figure
17. The small circle with radius 24° 19’, taken from seale
No. 2, was first drawn about 100. All points on the zone 100
to 011 are 90° from O11, hence the desired pole may be located
by means of a small circle constructed about 011 at a distance
of 79° 24’. To do this, two points were located on the hori-
zontal diameter at 79° 24’ from 011, to the right and left,
making use of scale No. 3 of the printed sheets. One of these
points is at s, figure 17, the other too far to the left to be
shown, and through them the small circle s¢ was con-
structed, which intersects the small circle previously drawn at
the desired point, marked 713. To find the symbol, the radius
through the pole 713, and the chord from 010 at right angles
to it, were drawn; thus the intercept on the first axis was found
to be 0:144 or 1. By means of the stereographie scale the dis-
tance from 001 to the pole 713 was found to be 67°, and this
angle plotted from the center intersects the perpendicular from ~
x’ at 0°333 or 4. The axial relation is therefore ta: a: ga,
the indices being 713. The distance from 713 to 317, meas-
ured with the small circle protractor, was found to be 43° 15’,
theory 43° 13’.

TETRAGONAL SYSTEM.

In this system the problem presents itself of finding from
one fundamental measurement the length of the vertical axis,
and from interfacial angles the symbols of occurring forms.
The example chosen for illustration is a crystal of octahedrite
from Brazil, given to the writer by Professor Groth of Munich.
The crystal, shown in figure 18 in orthographic* and clino-

* The orthographic projection is represented as revolved 18° 26’ about the

vertical axis, the reason for which will be explained in an article now in
preparation.

by Means of Graphical Methods. 271

graphic emeenian; was studied and figured by Mr. P. B.
Condit.

The priionehion is a simple one, base ¢(001L), pyramid p (111)
and ditetragonal pyramid s. As fundamental measurement
the angle of p/A p over the base, 136° 36’, was chosen, from
which ¢A p=68° 18’ is obtained. The poles of the pyramid p
are easily located, figure 19 » by means of the stereographic scale.
Erecting a perpendicular at x” to represent the vertical axis,
and plotting the angle 68° 18’ from the center, the length of
the vertical axis was found by means of scale No. 4 to be

19
C-L22F 0

BO.

1-774, the calculated length bemmg 1°777. By means of the
small circle protractor the distances between the poles of the
pyramid, 111 A111 and 111 A111 were found to be 82° 20’ and
82° 15’, the calculated value being 82° 9’.

For the determination of the symbol of the ditetragonal
pyramid, the angles between two adjacent s faces and from ¢
to s were measured, as follows: sAs=9° 40’, cAs=25° 30’.
About the center a small circle with radius 25° 30’, taken from
the stereographic scale, was first described, upon which the
poles of the pyramid must be located. Since ss is 9° 40’,
four of the poles (those lettered sin figure 19) must be sym-
metrically located at 4° 50’ on either side of the vertical diam-

272 Penfield—Solution of Problems in Crystallography

eter: accordingly p’ was located at 4° 50’ from 001 on the
horizontal diameter, and by means of the curved ruler a small
circle was drawn through p’ and points on the divided circle
4°50’ from 100and 100. The intersections of the small circles
shown in figure 19 locate the poles of four of the s faces.
Through the poles thus found diameters were drawn, which
were found to intersect the divided circle at about 11° 10’ from
100, and a chord from 010 at right angles to one of these
diameters intersects the front axis (radius) at just a trifle short
of 0°2, or 1. The distance 100 to 510 by calculation is 11° 18”.
Erecting a perpendicular at y’ and plotting the angle ¢As,
25° 30’, from the center, the intersection on the perpendicular,
measured with scale No. 4, was found to be 0:093 or ;5th of
the vertical axis found, 1°774+19=:0933. The relation of
the s faces on the axes is therefore 1@: @: ;'5¢, giving 5°1:19
as the indices.

HEXAGONAL SYSTEM.

The example chosen for illustrating this system is a crystal
of calcite from Egremont, England, shown in figure 20. The
angle 7/7, 74° 55’, was selected as the fundamental measure-
ment, and the poles of 7 were located as follows: Since 7A7
=74° 55’, a7, figure 21, must equal the complement of one-
half 7A 7, or 52° 324’; accordingly, on the radii Oa, and O—a,
the points s were located at 52° 32’, respectively, from a, and -
—a,, and, using scale No. 2 of the printed sheets, small circles
se were drawn through the points s, which intersect at 7. The
distance Or, measured with the stereographic scale, was found
to be 44° 40’, caleulated 44° 37’. The inclination thus found,
plotted from O, meets the perpendicular from @ at 0°854, cal-
culated 0°854. Having located the poles 7, the great circles
uniting them were constructed. The two scalenohedrons v
and ¢ are in the principal zones of the rhombohedron, between
7 and a, and were located on the stereographic projection by
the measurements @/\v=238° 31’ and aAc¢=10° 34’, the small

-eireles c's’ and c’’ s’’, respectively, 23° 31’ and 10° 34’ from
—a,, figure 21, serving the purpose. The radius through w
and the chord from a, at right angles to it establish the rela-
tion on the horizontal axes, $a : @ —4a, and also the length of
the base-line Oy. The slope of the face, the distance Ov,
measured with the stereographic protractor, was found to be
69° 00’, calculated 69° 2’, and this angle plotted from the cen-
ter intersects the perpendicular from v’ at just a trifle short of
0-854, or unity, the discrepancy being due to slight errors im
plotting ; for example, the radius through v should have inter-
sected the divided circle at 19° 6’ from m, whereas, as plotted,
it fell just short of 19°, an error perhaps of 8’. The sym-

by Means of Graphical Methods. 273

bol of the scalenohedron, as indicated by the method, is
4a: a@:—4a:¢,or 2131. The radius through ¢, and the chord
at right angles to it from @,, establish the relation on the hori-
zontal axes, #a:a@ —2a, and also the length of the base-line Oz.
The slope of the face, measured with the stereographic scale,
was 80° 35’, calculated 80° 32’, and this angle plotted from
the center intersected the vertical form 2’ at a distance 2°993ce,
practically 3c. The symbol of the scalenohedron is therefore
$a:a:—3a: 3c, or 4371. The rhombohedron M/ was deter-
mined by the measurement 7A M=31° 10’, giving 0001 A W/=
75° 47’. This angle, plotted from the center, would intersect
at 4c, measured from ~’, or, instead of extending the drawing
to such an extent, the slope may be plotted from «¢, as indicated
in figure 21, and its intersection with the base-line, equal 4<’0,
determined. The symbol of J/ is therefore 4041.

21
To 3c,from z?

It seems worth while to call attention to a procedure em-
ployed in solving a complex problem encountered in the study
of calcite crystals from Union Springs, N. Y.* The form to
_ be considered is that of a scalenohedron v,, in the same zone
as v and ¢, figure 21, but just a trifle steeper thanv. The
angle » on v, measured 3° 55’; hence since @\v=23° 31’,
-&@Av,=19° 36’. Employing the same method of plotting as
indicated in figure 21, the radius through the pole of v, was
found to intersect the divided circle at 21° from m (theory
21° 3’), and a chord drawn from a,, at right angles to it, gave

* This Journal (4) x, p. 237, 1900.

274 Penfield—Solution of Problems in Crystallography

the following intersections with the horizontal axes, as deter-
mined by scale No. 4 of the printed sheets, 0°573a : a: —0°365a.
The foregoing figures, which indicate a parameter relation, are
not easily interpreted, but their reciprocal values (1°746:1:
—2°740), which are equivalent to the indices, make it clear
what the relation is, namely, 12:1: —22. The distance 0001
to v,, measured with the stereographic scale, was found to be
72° 35’ (theory 72° 30’), and the intercept on the vertical axis,
plotted in the same manner as, for example, v in figure 21,
was found to be almost exactly 4c, the reciprocal (index)
being #. The indices expressed as fractions are therefore
12:1-— 22-3, or, cleared of fractions, 7:-4°11:°3.

ORTHORHOMBIC SYSTEM.

The example chosen for illustrating this system is a crystal
of anglesite from Eureka, Utah, selected from a suite of beauti-
ful specimens sent to the present writer by Mr. Maynard Bixby
of Salt Lake City. In figure 22 the crystal is shown both in
orthographic and clinographic projection, the figure having been
drawn by Mr. W. E. Ford of the Sheffield Laboratory. The
method employed in plotting is indicated by figure 23. The
positions of the poles of the prism m, 110, and brachydome oa,
011, were first located from the fundamental measurements,
mm, 110A110 = 76° 16’ and oAo, O11 A011 = 104° 24".
These measurements are especially favorable for determining
the lengths of the @ and c¢ axes, which are the tangents, re-
spectively, of z and z, figures 6 and 8, or the tangents of one-
half the given angles; hence it is only necessary to draw two
chords from 6, intersecting the graduated circle at 76° 16’ and
104° 24’, measured from 010, figure 23. Thus the lengths of
a and c, measured on the respective radii by scale No. 4, were
found to be 0°786 and 1:289, calculated 0°786 and 1°289. The
slope of the brachydome, cA o=52° 12’ (4 of o Ao), may also
be plotted in the upper left-hand quadrant, and the length of
the ¢ axis measured on a perpendicular erected at 010, as indi-
cated. The macrodomes d and 7 will be considered next.
These were located on the vertical diameter by means of the
angles c\d=389° 23’ and cAl=22° 19’, and lines plotted at
these angles in the upper left-hand quadrant intersect the per-
pendicular from @’ (¢ a’ equals the length of the @ axis) at dis-
tances equal respectively to $ and + the length of the c axis,
thus determining the symbols d=102 and 7=104. The posi-
tion of the pyramid p is determined by the intersections of the
zones @:o0andc¢:m. Thus plotted, the slope of py, measured
with the steresgraphic scale, was found to be 64° 30’, calculated
64° 24’, and a line drawn at this angle from the center meets

by Means of Graphical Methods. 275

the perpendicular from x’ at a distance equal to the length of
the c axis; hence the symbol of pis 111. The pyramid y, in
the zone 0: p, was determined by the angle oA y=26° 42’,
from which aA y=63° 18’; hence a small circle sc, drawn at
63° 18’ from a, figure 23, fixes the position of y. A radius
through y was found to intersect the divided circle at 57° 30’
from @, 100, which corresponds with the pole of the prism 1,
and a chord from 4, at right angles to this line, intersects the
front radius at a distance equal to 2a. The intercepts on the
first and second axes of n, y, and a third face in the zone, yp,
are therefore 2a : 6, the indices of n being 120. The slope

of y, measured with the stereographie scale, was found to be
56° 45’, calculated 56° 48’, and a line at this angle intersects
the perpendicular from y’ at a distance equal to the length of
the ¢ axis; hence the parameter relation of y is 2a: 6: ¢, the
indices being 122. The pyramid # was located by the meas-
urement ¢\4“=387° 23’, and a line at this angle meets the
perpendicular from y’ at a distance equal to 4c; hence the
parameter relation of w is 24:6: 4¢, the indices being 124.
The pyramid 2, in the zone d : p, was located by the measure-
ment dA z=24° 49’. This was done by first determining the
zone n, p, d, 4 and o by means of the great circle protractor,

276 Penfield—Solution of Problems in Crystallography

and then constructing a small circle s’ c' about d at a distance
of 24° 49’. A chord from 6 at right angles to the radius
through z was found to intersect the front radius at 2a.
The slope of z, measured with the stereographic scale, was
54° 15’, calculated 54° 16’, and a line at this angle intersects
the perpendicular from z’ at a distance equal to 4¢. The
parameter relation of 2 is therefore 3a: 6: 4c, giving as the
indices 324. In addition to the lines needed for the graphical
solution of the foregoing problem, numerous zones are shown
in figure 23 which were easily determined by means of the
great circle protractor. The figure indicates that, starting
with the poles of m (110) and o (011), it would have been a
simple matter to have determined the location of the poles and
the symbols of all the forms by no other means than zonal
relations. |

Monocuinic System.

The graphical methods employed in this system may be
illustrated by pyroxene, a crystal of which is represented in
orthographic and clinographic projection in figure 24. The
figure, which represents a common habit of pyroxene from
northern New York and Canada, was drawn’ by Mr. H. H.
Robinson. As fundamental measurements the following were
chosen: mA m, 110A 110=92° 50’; a\c=100 A001=T74° 15’;
and d@As, 101 A111=29° 35’. In stereographic projection,
figure 25, the location of the poles of the prism m and of the
base c, the latter by means of the stereographie scale, need no
special comment. The construction of the great circles through
‘m and ¢ is then an easy matter, and the angles which they
make at ¢ with the vertical diameter are equivalent to ct of
figures 6 and 8. The measurement of c, 47° 30’, was made on
the great circle GC at 90° from ¢.* Since 7, figure 6, is 90°
in the monoclinic system, the tangent of 7 must equal the
length of the @ axis; hence c was laid off from 6, and a deter-
mined as 1:093, calculated 1:092. Since dAs=29° 35’, BAS
must be the complementary angle, 60° 25’; hence the pole of
s was located by the intersection of a small circle (60° 25’
from 6) with the great circle through ¢ and m. Great circles
through @ and s and 6 and s were then drawn, thus determin-
ing the poles of p andd. Theanglez, compare figures 6 and 8,
was measured on the great circle (diameter) at 90° from a, and
found to be 30° 85’; hence, referring to figure 6, since @ is
90°, the tangent of z must equal the length of the ¢ axis,
which, when plotted from 6, was found to be 0°592, calculated
0589. The ares mp and mc, measured with the small circle

* This Journal (4), xi, p. 18, 1901.

by Means of Graphical Methods. 2UE

protractor, were found to be 45° 10’ and 79° 20’, calculated
45° 20’ and 79°10’. By means of the stereographie scale cA d
was determined as 31° 10’, calculated 31° 20’. The poles of
the pyramid o were located by means of the measurement
m \.0=35° 30’, a small circle from m, 110, being used. The
ereat circle go made an angle, z’, of 49° 45’ at a, which, when
plotted from 6, was found to intersect the vertical radius at:

1180, practically 2c, the calculated value being 1178; hence
the axial relation of 0 is —a:6: 2c, the indices being 221.
The poles of 0 could also have been located by constructing a
great circle through m and s, 110 and 111, thus determining
201, and then a second great circle through } and 201.

For the purpose of showing the adaptability of the stereo- .
graphic projection to measurements made with the two-circle
goniometer, and also for illustrating another method of treat-
ing the monoclinic system, a problem encountered in the study
of azurite crystals from Broken Hills, New South Wales, will
be presented. The crystals, which are exceptionally beautiful,
were measured and figured by Mr. Ralph G. VanName.
They are lengthened in the direction of the ortho axis, like the
ordinary development of epidote, and the habit is well illus-
trated by figure 26, drawn in clinographic projection with the
} axis to the front. The crystal was orientated on the goni-
ometer with the orthodome zone horizontal and the pinacoid a,
100, vertical, thus making an imaginary clino-pinacoid the

polar face, although the form was not present on the crystal.

278 Penfield—Solution of Problems in Crystallography

In the stereographic projection, figure 27, the angles g, deter-
mined by the readings of the vertical circle, were laid off on
the divided circle, starting from a, 100, and the angles p, given
by the readings of the horizontal circle, were plotted on their
respective diameters by means of the stereographic scale. The
angles for making the projection shown in figure 27 are as
follows:

26

? P “

8,21 00, ¢00° 007% - 90R00' D,. 104, 102°, 154 See
m, 110, sé 49 39 0, 134, Ke bin an
w, 120, oh 30 29 Ay S188, 1220 oe 12 43
DeQoi 25) O18 53 02 6, P01 Asa ae 90 00
a del TAD! + S51) 59 06 4, 302, S146 32 <
6,001, 8736 90 00 ©, 2015 Tos" Pag e

/, 1028, sk 501° 39 Fe D1 ie 52:08
7 LOL, at 48 40 0, 241, ey 32 44
seo... FS 29 36

Having located the poles of the several faces, the zones
shown in figure 27 were easily found by means of the great
circle protractor. Starting, then, with the assumption that the
symbols of two poles are known, for example, m, 110, and p,
111, it is possible to determine the symbols of the remaining
forms by no other means than that afforded by zonal relations.
As it is the object of this paper, however, to indicate the
methods of graphical solution, the problem may be dealt with
in a manner analogous to that employed in the orthorhombic
system, since the 6 axis is at right angles to the plane of the a

by Means of Graphical Methods. 279

and c axes. In order not to complicate figure 27 to too great
an extent, the solutions depending upon plane trigonometry
were worked out on a second divided circle, shown in figure 28.
Having located the poles a, 100, and c, 001, the vertical and
clino axes are represented by diameters at 90° from them, as
indicated by broken-dashed lines; the lower half of the verti-
eal diameter, however, may be employed for plotting the rela-
tions of the 6 axis, the length of which is unity. Referring to
figure 27, the angles z and t (compare figures 6 and 8) were
measured, respectively, by the stereographic scale as 41° 20’
and 40° 20’, and these inclinations plotted from @ and «¢, figure
28, determine the lengths of the ¢ and @ axes, which were

|
|
|
¢ 0.587- $C

) ae za
otek [12° #3”
° ,
sie 48 I ov
aU .

found to be as follows: c=0°880 and a=0°850, calculated,
e=0°880 and a=0°850. In order to determine the relations of
the prisms m and w, through unity on the @ axis, a line was
drawn at right angles to the radius through m and w, thus fix-
ing a base-line Oz; then from @ the inclinations of m and w
(the p angles), 49° 39’ and 30° 29’, are plotted, and, as shown
in figure 28, they intercept the lower radius at unity and one-
half, respectively. Since the lower radius is taken to represent
the 6 axis, the axial relation of m is a: 6: mc, and of w, a:
4b : we, the indices being 110 and 120. Naturally the reverse
of this method may be employed to advantage for determin-
ing the length of the @ axis in this system, when the inclina-
tion # and the prismatic angle are given. For determining
the symbol of a clinodome, for example s; through unity on
the ¢ axis draw a line at right angles to the radius through s,

280 Penfield—Solution of Problems in Crystallography

thus determining the base-line Oy, and from y’ (the distance
Oy transferred to the horizontal diameter) lay off the o angle
of s, 29° 36’, which intercepts the lower radius (6 axis) at one-
half ; hence the axial ratio of s is wa: 40: ¢, the indices being
021. For a pyramidal. face, for example A, through unity on
the @ axis, a line at right angles to the radius through 4 was
drawn, which in figure 28 was found to intercept the vertical
axis at 0°587, or exactly 3c. Thus the intercept on the vertical
axis and the base-line Oz are determined. From 2’, on the
horizontal diameter, a second line was drawn at an inclination
of 12° 43’, the o value of A, which intersects the lower radius
at 0°111, or 1b. The axial ratio of A is therefore —a@: 1b: Ze,
the indices being 2°18°3. Thus it makes no difference how
complex the symbol may be, it is only necessary to keep
in mind a few simple principles and the result is quickly
obtained.
Finally, having made a projection such as shown in figure

27, the small circle protractor may be applied to it and the
distances between any desired poles determined. Some results
actually obtained in this way are as follows:

Meas. Cal. Meas. Cal. Meas. Cal.

CA, 88° 10' 88710’ enk, 71°25’ 71° 25’ GAp, a0 oO sai
CAP, 52 30 52 28 cAo,77 24 17 25 anh, 40 0a meeeea
cah, 68 10 68 12 map,35 45 35 42 a’Ao, 61 05 60 59

Trictinio System.

For illustrating the application of graphical methods to this
system a somewhat complex crystal of anorthite has been
chosen, shown in orthographic and clinographic projection im
figure 29. The crystal was studied and figured by Mr. J. C.
Blake. The following were selected as fundamental measure-
ments for making the stereographic projection, figure 30, and
solving the problem :

bAm, 010, 110=58° 04’ MAC, 110,001=65° 53’
DAC, 010A 001=85 50

Having located 6, m and M on the divided circle, the posi-
tion of 100 was first determined by the relation of four faces
in a zone, explained on page 260, 6, m and J being, respect-
ively, the first, second and fourth faces, and 100 the unknown
third face. This problem, as solved by itself on a separate
sheet, is shown in figure 31. The lines of reference Ov and
Oy are at right angles respectively to 0b’ and MJ’. A line
drawn from 2, at right angles to the radius through m, inter-
sects Oy at a point marked 3. By solving the equation given
on page 262 the ratio of the cotangents is found to be 3: 1,
hence a line from ~z to the point 1 on the line Oy determines
the direction of the pinacoid 100. A line from the center at

by Means of Graphical Methods. 281

right angles to xy fixes the pole of 100, which was found to be
87° 5’ from 4, caleulated 87° 6’. Having located the pole 100,
the direction Oy’, parallel to zy, may now be taken as the sec-
ond line of reference. Lines from « to $Oy’, both to the right
and left of the center, intercept the divided circle at 58° 50’
and 62° from —z, and angles equal to one-half of these values

29 30

—¥6

he Sone Sl ty’ j

(page 260) give the inclinations 6A fand b’ Az, namely, 29° 25’
and 31°, caleulated 29° 29’ and 30° 58’. The indices of ,f and
z are therefore 130 and 130. The pole e¢, figure 30, was next
located from the measurements 6A¢ and m/c, small circles
about the poles ) and m serving for this purpose. The small
circles are not shown in the figure, but that about 4, having a
long radius, was constructed by means of the curved ruler,*
_ that about m by means of a radius taken from scale No. 2 of
the sheets. Having located ¢, the great circles b,¢,b' ; m,e,m’;
and 100,¢, 100 were drawn, thus determining the angles a, §

>
* The Stereographic Projection and its Possibilities, loc. cit., p. 12.

282 Penfield—Solution of Problems in Crystallography

y and 7; compare figures 6 and 8. The angle a was measured
by means of the stereographic scale on the diameter 1 to 2,
figure 30, and found to be 98° 25’, calculated 93° 14’. Sim-
ilarly 8 was measured on the diameter 4 to 5, and found to be
116°, caleulated 115° 56". To measure 7 and t, a great circle
at 90° from ¢ was first constructed.* The supplement of 7,
from 6 to 7, was then determined as 88° 55’ by means of the

small circle protractor, giving as the value of 7, 91° 5’, caleu-

lated 91° 12’.. In like manner the are from 7 to 8 was meas-
ured, giving 32° 5’ as the value of r. On still a third sheet,
shown in figure 32, the angles a, 8 and 7 were plotted, and
also t, the latter determining the length of the brachy axis,
a=0°636, calculated 0°635. It is still necessary to make use of
the fifth fundamental measurement, from 6 to e, in order to
determine the length of the vertical axis. By means of a small
circle about 6, figure 30, the position of e, on the great circle
b,c, was located, and a great circle through 100, e and 100 de-
termines z. The value of z, measured on the diameter from
2 to 8, was found to be 45° 45’, and when plotted as shown in
figure 32, the length of 2c was found to be 1-095; hence
c=0°5475, calculated 0°550. The value obtained directly from
z in this case is 2c, because the symbol of e is 021.

Having located, as in figure 30, the poles 6, m, 100, WZ, ¢—

and e, and also. fand z from figure 31, the remaining forms of
the crystal were easily identified by means of zonal relations
and the measurement of a few angles. The construction of
the zones shown in figure 30 would have proved a laborious
task had it not been for the great circle protractor. This was
used not only for determining the zones, but, far more impor-
tant, its graduation furnished the means of getting the radii of
the circular ares from scale No. 1 of the sheets.

Meas. Cal. Error
Cal, 001A 201, 41°° 30’ 41° 28’ + 2
Cag, . O01LA 84 45 34 46 — 1
CA. | Ot 01: 51 40 bil 26 +14
cay, 001A 201, 81 30 Si +16
CAD, (Ol yn1 14, 33 pao 33 17 — 7
eR a. 40Ol ALT, 34 tae 34 10 +25 -
Cnn AOU AL, 5d CahO 5A 17 — 7
Co, (00k F1b, 58 00 512 + 8
cau, O001A221, 85 00 84 50 +10
enka UOL,A0b1, T 20 514 MO +10
can, 001,021, 47 00 46 46 +14
Cade. JMG 18 . 55 18 38 +17
O 207, 7) ONO 52020 ek Wl + 9
b nn, OWA 34 10 34°20 10
On”G,- O10> 88 05 38>) .1G —11
Aw, O10A aishe, De 38 42 +13

* The Stereographic Projection and its Possibilities, loc. cit., p. 18.

ee ae SS ae =e"!

by Means of Graphical Methods. 283 -

A list of measurements made with the small circle protractor
is given in the table on the foregoing page. The average of
the errors is 11’, with errors of more than a quarter of a degree
in only three instances. In figure 30, the angle between the
great circles b,c and 6,p determines » (compare figures 6 and
8), which was found to measure 29° 25’. The angle »v, plotted
from unity on the @ axis, as shown in figure 32, gives another
method for determining the length of the ¢ axis; in this case
¢@ was measured as 0°545, caleulated 0°550.

CoNCLUSION.

In the foregoing pages the attempt has been made to indi-
cate not only the methods of plotting problems in the several
systems, but, also, to give an idea of the accuracy of the results
thus obtained. In judging the accuracy of the results it must
be kept in mind that the scale employed was quite small, the
engraved circle being only 14°™ (54 in.) diameter. Each prob-
lem presented was worked out from fundamental measure-
ments, as would have been the case if done by numerical cal-
culation, and the results indicate that for all practical pur-
poses graphical methods are in every way satisfactory. The
lengths of axes have frequently been obtained correctly to the
third place of decimals, and never varied more than one in the
second place, and measurements of arcs and angles by the
stereographic protractors have been reasonably close to the
calculated values in all cases. After becoming familiar with the
principles involved in dealing with the projection—and they
are not difficult to understand—any problem may be worked
out without the use of. prescribed formulas, or tables of any
kind, and this the writer believes is one of the great advan-
tages of the method. Then, too, it is a very simple step to
proceed from a graphical to a numerical solution, for the prin-
ciples involved are identical, and it is only necessary to apply
the formulas of spherical and plane trigonometry in order to
obtain the desired results by calculation. The writer has never
made use of formulas for solving even the most complex prob-
lems of crystallography, other than those needed for the gen-
eral cases which arise in dealing with spherical and plane
triangles, and it is his belief that the too general use of pre-
seribed formulas is, if anything, a hindrance to true progress
in crystallography. Graphical methods will also be found to
have very decided value as giving a ready means of checking
the results of numerical calculations. During the past two
years, since they have been in use, numerous cases have arisen
where mistakes in calculation have been made, at times
amounting to less than a degree, but they have been almost

284 Penfield—Solution of Problems in Crystallography.

instantly detected by applying the protractors to the stereo-
graphic projections. Generally it has been found that mis-
takes were due to errors (carelessness, perhaps, is a better term)
in such simple processes as addition and subtraction.

In an earlier communication* a description was given of
blackboard appliances which have proved most useful for dem-
onstrations of the stereographic projection on a rather large
scale before a class. Thus, in the discussion of any problem
which may arise, a projection may be made in a few minutes,
from which the zonal relations, symbols and angles of the
forms may be determined. As a demonstration of the prac-
ticability of the method, a projection of pyroxene, like figure
25, page 277, was constructed from fundamental measurements,
and the following determinations made, the time consumed
being nineteen minutes:

Given, mA mm’ =92° 50’; cA p=88° 49’ and aAc=TA4® 10’.

Axes:determined; a :6::.¢ =:1-ils: 1:2:0-585: ;

Axes calculated; a: 0::'c = 1-092 21 08a

' Angles. Meas. Cal. Angles. Meas. Cal.
i Cy VO aA OOT=79": 79°, 9! 0A 0, 221 a 29184) 4a
crd, 001A101=31 31 20 CA0, OO1LA321=65 65 21
PAD, VILAI1II=48} 48 29 app, 100, 111=5ae Sas
Sas. 111 Adi 595 59041 GAS, 100A 111=763 ieee
Cas, 001A TII=494 42-2 a2A0, 100, 221=625. soles

It should be stated that the foregoing problem was exe-
cuted with blackboard crayon, and with beam-compass and
scales graduated to every fifth degree only.

Finally the writer would state that in his own laboratory the
graphical methods have proved most serviceable, and it is be-
lieved that by means of them a far better msight into mathe-
matical crystallography and the meaning of zonal relations
and angles has been gained by the students, than when other
methods were employed. There is no wish to discourage
numerical calculations; such must be made, and facility in
making them must be acquired, but the graphical methods, if
properly interpreted, are, if anything, a decided help, for they
give meaning to the calculations, which, far too often, are
made by prescribed formulas, wholly meaningless to the
average student.

* The Stereographic Projection and its Possibilities, loc. cit., p. 140.

Sheffield Laboratory of Mineralogy and Petrography,
Yale University, New Haven, May 1902.

a

ewes! eer oe ©

Gooch and Blake— Estimation of Bromic Acid. 285°

Arr. XX XI.—The Estimation of Bromic Acid by the Direct
Action of Arsenious Acid; by F. A. GoocH and J. C.
BLAKE. 3

[Contributions from the Kent Chemical Laboratory of Yale University—CXI. ]

In a former paper from this laboratory* it has been shown
that iodic acid may be reduced quantitatively by arsenious
acid. In the work of which the present paper is an account
the attempt has been made to apply arsenious acid similarly to
the quantitive estimation of bromic acid, according to the
equation

3H,AsO,+HBrO, = 3H,AsO,+ HBr.

In the following series of experiments, made to discover the
limits within which regularity of action might be expected,
definite amounts of arsenious oxide dissolved in acid potassium
carbonate were mixed with measured portions of a solution of
potassium bromate, sulphuric acid was introduced in the
amounts indicated, and, after standing either at the ordinary
temperature, on the steam-bath or at the boiling temperature,
potassium acid carbonate was added and the arseniate remain-
ing was titrated with iodine to color, the indication being
confirmed by addition of starch. The conditions of acidity,
the excess of arsenious oxide, the dilution, the time of action,
and the temperature were varied within wide limits. The
potassium bromate for this particular series of experiments
was thrice recrystallized from water, dried at 110°, and made
up in solution containing 2°8 grms. to the liter. Of this solution
25 or 50° were measured out for each experiment.

TABLE I,

Volume not exceeding 200°™%,

H.SO, As2O3 Error in
KBrO; As2,0O; (£1): ~.'Fime of As2Os oxi- terms of
taken. taken. taken. diges- unchanged. dized. KBrOs.
grm. erm, em*. tion. erm. grm. erm.

Heated at the boiling temperature.
0°1400 0°4950 10 30 min. 0°2476 0°2474 0°0008—
071400 0°4950 10 30° “ 0°9480 0°2470 0°0010—
0°1400 0°6188 10 20 SG LOS 0°2480 0°:0006—
0'1400 0°7425 8 25 1/59 OA GAT 0°2484 0°0003—
0°1400 0°7425 8 ZK Vek Se: 04943 0°2482 0°0004—
0°1400 0°6188 7 20 ei O3704 0°2484 0°0003— -

(

0°1400 0°6208 20 “ 0°3720 0°2483 0°:0004 —
* Gooch and Pulman, this Journal, xii, p. 450 (1901).

Am. Jour. Sci1.—FourtH SERIES, Vou. XIV, No. 82.—OcrTossr, 1902.
20 .

‘286 Gooch and Blake—Estimation of Bromic Acid

KBrO;
taken.
grm.

-0°1400
071400
0°1400

[ 0-1400
0°0700
0°0700
0°0700
0-0700
0:0700
0:0700
0°0700
0°0700
0°0700
0:0700
0-0700
0-0700
0:0700
0°0700
0°0700
0-0700
0:0700
0-0700
0-0700

[ 0-0700

071400
0°1400
0°1400
0°1400
071400
071400
071400
071400
[ 0-0700

0°1400
0°1400
0°1400
0°1400
01400

As.Os
taken.
erm.

0°6188
0°4950
0-4950
0°4950
0°2475
0°2475
0°2475
0°2475
0°2475
0°2475
0°1584
0°1584
0°1584
0°1584
071881
0°1881
0°1881
0°1881
0°1881
0°2475
0°2475
0°2475
0°2475
0°2475

0°4950
0°4950
0°4950
0°4950
0°4950
0°4950
0°4950
0°4950
0°2475

0°4950
0°4950
0°4950
0°4950

0°4950

TABLE I—COontinued.

Volume not exceeding 200°™°.

H.SO.4 As2O3
(1 >i)" Aimeot As2O3 oxi-
taken diges- unchanged. dized.
cms. tion. grm. grm.

Heated at the boiling temperature.

a) 15 min. 0°3712 0:2476
Bb Lb = *'O-2880 M024 7m
3°5 15 “ -0°2482 0°2468
2°5 LT t OFDEQO 0°2321
i) Peto ess OueAO 0°1235
) Lo O39 0°1236
9) nS Se ORL Qos ODRaE7.
5 15 .“- .0°1244 Onan
5 LOT ro dee 0°1236
5 DO eh itastanl 0 8 be 21S 97 0°1238
5 LQ ee O:0nee 0°1239
5 TOO OOOaa 071227
O 10. 500356 0°1228
5 10°. *2) 2O°0S5s 0°1231
8 10 “ ~0:0649 0°1232
5 10 ‘* 0°0640 0°1241
5 10 ‘* 0:0649 0°1232
5 5 O06S2 0°1229
5 5 “ °0:0652 0:1219
2°5 10,2, ¢8> “O:238 0°1237 |
2 LOc. MS S3Ob285 0°1240
TLS 70 3 Sede 4D 0°1233 |
1515) dO 55%, HOS Ag 0°1235
11D Da WOO 0°1175

Heated on the steam-bath —80°.

i0 150 min. 0°2476 0°2474

10 150 2 FJ 0e24g6 0°2474
+ 240 “ 0:2478 0°2477
= 240 “ 0°2475 0°2475
4 180 “ 0°2472 0:2478
a 180 ‘ 0:2470 0°2480
3°O 90"! nO 2475 0:2475
3°5 9020S “RO QAT6 0°2474
1715." BO SCO 1622 0°0853

Digested at the ordinary temperature.

4: 46 hrs. 0°2480 0:2470
3°5 AY “o> 0°24 75 0°2475
3°9 41° * 02470 0°2480
3°5 17 ~ "5, 0:2490 0°2460
3°5 17 072492

0°2458

Error in

terms of

KBrO; =
erm.

0:0008 —
0-0010—.
00011 —
0:0097— |]
0:0005 —
00004 —
0°0015 —
0:0007—
0:0004—.
0-0003 —
0:0003 —
00009 —
0:0009—
0:0007—
0‘0007~——
0°0002 —
0:0007—
0:0008—
0:0008—
0:0004 —
00002 —
0°0006 —
0:0005 —
0:0038— |

00008 —
00008 —
0:0007—
00008 —
0°0006—
0°0005 —
60008 —
0:0008—
0°:0217— |

0°0005—
0°0008 —
0°0005—
0:0016—

0:0017—

a

by the Direct Action of Arsenious Acid. 287

An examination of these results discloses the fact that for
conditions varying within a rather wide range the oxidation of
the arsenious oxide reaches a fairly definite limit. The
bromate effects practically the same proportionate oxidation of
arsenious oxide in a volume of 200° or less, whether the sul-
phuriec acid of half-strength present amounts to 3°5°™ or 10™,
and whether the time of digestion is 15 minutes or 30 minutes
at the boiling temperature, one and one-half or four hours on
the steam-bath, or two days at the ordinary temperature.
Setting aside those experiments in which the addition of acid
did not exceed the equivalent of alkali carbonate present by
more than 1°’, the average absolute variation from theory of
the entire list of forty-two experiments amounts to 0:0007—
grm. in terms of bromate, individual variations departing from
the average by about the same figure. The obvious meaning
of this fact seems to be that the slight error is due to impurity
in the potassium bromate employed and not to incomplete
oxidizing action on the part of the bromate.

In the following series of experiments another preparation
of bromate was employed, and its value was fixed by reduction
with acidified potassium iodide and titration of the iodine set
free according to the method formerly proposed by Kratchmer.*
The rate at which the action proceeds, according to the
equation

6HI+HBrO, = HBr+3H,0+1,,

has been investigated by Ostwald,t by Noyes,t and by Judson
and Walker.§ Time of action, proportion of iodide to bromate,
excess of acid, and dilution are all, within limits, determining
factors in the reaction; but in the analytical process it is
usually assumed that the reaction goes soon to completion if
free acid and a moderate excess of potassium iodide are
present. In the following table are recorded the results of
experiments in which measured amounts of standard solutions
of potassium bromate (approximately 2°8 grms. to the liter)
were treated with potassium iodide and hydrochloric or
sulphuric acid for definite times in glass-stoppered bottles, the
iodine liberated being determined by titration with sodium
thiosulphate standardized against nearly decinormal iodine of
value fixed by comparison with decinormal arsenious oxide
dissolved in acid potassium carbonate.

* Zeit. Anal. Chem., xxiv, 546 (1885).

+ Zeit. Phys. Chem., ii, 127 (1888).

t Zeit. Phys. Chem., xix, 599 (1896).

$ Jour. Chem. Soc., Ixxiii, 410 (1898).

288

@

KBrO; = Iodine

taken.
erm.

) 071410
) 071410
8) 0°1400
) 0°1400
) 0°1400
) 071408
) 01408
) 0°1408
) 0°1400
)

2) 0:1400
3) 01400

(14) 0°1400
(15) 071400
(16) 0:1400
(17) 0°1400
(1
(19) 0:1400

(20) 01400

8) 0:1400

(21) 0:1400

(22) 01400

considerable.
reaction is very incomplete.

taken. taken.

erm.
0°6423
0°6423
0°6378
0°6378
06378
0°6416
0°6416
0°6416
0°6378
0°6378
0°6378
0°6378
0°6378

0°6378
0°6378
0°6378
0°6378
0°6378
0°6378
0°6378
0°6378
0°6378

Pasun 1.
A
Time
of
KI H.SO, stand-
(i), ine,
orm, pce. Thrs:
uy 5 none
ik 5 none
3 5 4
3 5) oo
3 5 4
3 5 22
3 5 29
3 5) 22
3 hues and
3 YA >
3 2°) 4
3 ee >
3 0:5 4
B
Het
Sp. cris
cem?,
3 8 4
Bie 548 4
Bn ate 2
3 4 4
a 4 4
3 4 4
3 4 if
3 4 li
3 4 l

Approxi-
mate

volume.

cm®,

150
200
100
100
100
100
100
100
100
100
,100
100
100

100
100
100
100
100
100
100
100
100

Todine
found.

grm.
0°6223
0°5929
0°6343
0°6329
0 6329
0°6396
0°6364
0°6381
0°6340
0°6336
0°6336
0°6343
0°6331

0°6336
0°6329
0°6336
0°6336
0°63338
0°6340
0°6336
0°6333

‘0°6336

Gooch and Blake—Estimation of Bromic Acid

Error in
terms of
KBrO;
grm.
0°0044 —
0°0108—
0°:0008 —
0:0011 —
0:0011—
0°0004 —
0:0011—
0:'0008 —
0:0008—
0°0009—
0:0009—
0°0008—
0°0010—

0:0009—
0.0011—
0°0009—
0.0009—
0°0010—
0°0008—
0:0009—
0°0010—
0°0009—

In experiments (1) and (2) the excess of potassium iodide
was only about 20 per cent over the amount demanded by
theory, the time of standing was only the few minutes required
for the manipulation of ‘the process, and the dilution was
It is obvious that under these conditions the

On the other hand, a glance at

the table makes it evident that the oxidation in all other
experiments, whether in A or B, proceeds to practically the
So it appears that the
variation in the amount of acid above the minimum, the kind
of acid (whether sulphuric or hydrochloric), and the time
above the minimum half-hour, within the limits defined, are

same point,

showing similar errors.

without apparent effect upon the reaction in the presence of
the amount of potassium iodide used, approximately four times

by the Direct Action 6f Arsenious Acid. 289

the amount required by theory. The balancing error due to
the evolution of iodine by the action of atmospheric oxygen
upon the acidified solution of the iodide was found by experi-
ment to vary with the strength of the acid and the time of
exposure, from 0°0001 grm. to 00003 grm. expressed in terms
of the bromate, and these values are not greater than the
differences observed between parallel determinations of the
same sort. Probably, therefore, all the errors as shown in the
table should really be increased a trifle to approximate the
truth, notably those of the experiments allowed to stand the
longest period, twenty-two hours.

The average apparent error of the process as applied to this
particular sample of bromate is 0°0009- grm.; and 2°5™* of
sulphuric acid of half-strength (1:1) or the equivalent amount of
hydrochloric acid, 4™ of the acid of sp. gr. 1:18, in the
presence of about 3 grms. of potassium iodide, complete the
action within a half hour, at a dilution of 100, as far as it
will go under any of the conditions tried. The phenomenon
noted by Ostwald,* that small amounts of hydrochloric acid
tend to force the reaction more rapidly than equivalent
amounts of sulphuric acid, does not appear in these experi-
ments, no doubt because the action was pushed to the limit by
the smallest amount of the weakest acid employed.

The error of deficiency shown again appears to be due to
impurity in the sample of bromate rather than to incomplet-
ness of the reaction. If the reaction were incomplete it would
be natural to look for the cause in the possibility of the
inhibiting influence of the iodine set free,t+ but it was
found in three parallel experiments that the introduction
of ‘5 grm. of free iodine dissolved in potassium iodide
failed to influence the error appreciably. For the present pur-
pose, however, the absolute purity of the bromate is not a
matter of moment, if, as seems to be the case, Kratchmer’s
reaction indicates its value with accuracy, as the basis of experi-
mentation. It seems safe to assume, then, that the average
result of experiments (8) to (22) will give a very fair valve for
the sample of bromate investigated. That is to say, the
experiments show an average deficiency in bromate amounting

to 0°0009 grm., or to 0-64 per cent.

Portions of this bromate were taken for reduction by
arsenious acid. Solutions not exceeding 200° in volume,
containing the bromate and a considerable excess of arsenious
oxide acidified with sulphuric acid of half-strength, were boiled
for periods varying from ten to forty-five minutes, neutralized
with potassum acid carbonate, and titrated with iodine. The
resnits are shown in Table III.

= koe: cit., p. 151. + Judson and Walker, loc. cit., p. 411.

290 Gooch and Blake—Estimation of Bromie Acid

Tas eE III.

Error in
KBrO; As,O; H.SO, Time As.O3 As,O3 terms of
taken. taken. (£21): in unchanged. oxidized. KBrOs.
erm. erm. cm*®. minutes. grm. erm. grm.
( L)yO°eTOk: 01881 10 0°0661 0°1220 0°0014—
( 2) 0:0701 01881 10 0°0650 0°1231 0°0009—
( 3) 0°0701 0°2475 10 0°1232 0°1248 0°0002—
( 4) 0.0701 0°2475 10 0°1236 0°1239 0°:0004—
( 5) 0°0701 0°2475 25 0°1234 0°1241 0°0003—
( 6) 0°0701 0°2475 0°1234 0°1241 0.0003—
( 7) 071402 0°4950 15 02479 02471 0°0012—
( 8) 071402 0:4950 15 0°2476 0°2474 0°00 LO—
( 9) 071400 06188 20 0°3708 0°2480 0°0004—
( 20 0°3710 0°2478 0°0005—
( 20 0°3706 0°2482 0°0003—
( 30 0°3708 0°2480 0:0004—
( 45 0°3711 0°2477 0°0006—

10) 071400 0°6188
11) 0:1400 0-6188
12) 0°1400 0°6188
13) 071400 06188

7 oT wT ST Tr Oo W& OFT OF St ST OT or
bo
Or

Here again, as in the experiments of Table I, the indications
of the process of reduction of the bromate by arsenious acid,
point to a slight deficiency in the oxidizing power of the
bromate. The mean deficiency, 0°0006 grm., differs from the
indications of the process of reduction of the same sample by
the potassinm-iodide method by about 0:0003+ grm.

The question now arises as to what is the impurity in the
bromate. In a product recrystallized several times, and in one
in which no chloride can be detected, as was the case with
the bromate of these experiments, the impurity most natural
to look for is potassium chlorate, which, might resist removal
in the process of purification by recrystallization. The pre-
paration of bromate upon which these last experiments were
made was therefore tested by igniting it and treating the resi-
due with potassium bichromate and sulphuric acid, volatiliz-
ing and collecting any chloro-chromic anhydride thus formed,
and converting the last into lead chromate.* Traces of
chlorine were thus found, which, not appearing in a similar
test upon the unignited bromate, must have had their origin in
chlorate intercrystallized with the bromate. Ina former paper
from this laboratory+ it has been shown that a chlorate
may be determined by adding to it in solution potas-
sium iodide in known amount and an excess of an arse-
niate and sulphuric acid, boiling the mixture between definite
limits of concentration, determining by titration with iodine
the amount of arsenious oxide produced, and calculating
the amount of chlorate present, from the difference between
the amount of arsenious oxide thus produced and that which

* Gooch and Brooks, this Journal, xl, p. 287 (1890).
+Gooch and Smith, this Journal, xlii, p. 220 (1891).

by the Direct Action of Arsenious Acid. 291

should be produced if the whole amount of iodide added were
allowed to act upon the arseniate alone. There appears to be
no reason why’a bromate treated by this process should not
leave a similar record of its oxidizing power. A mixture of
bromate and chlorate should, therefore, in this process, leave a
record of the full amount of oxidation of which both together
are capable. The following table contains the account of
experiments made in this manner upon the sample of bromate
the action of which in the iodide method and in the arsenious
acid method is recorded in Tables II and ILI.

Papi 1V-

(H2SO.) (1:1) 20°™3; initial volume 105 to 170 ; final volume 35°™°,

Iodine corre-

Ivalue sponding to Jodine corre- Error in
KBrO; H.KAsO; of KI As203 sponding to terms of
taken. taken. taken. produced. KBrOs. KBrOs.
erm. erm. erm. erm. erm. erm.
(1) 0:0700 2 0°4146 0°0948 0°3198 0:0002+
(2) 0:0700 2 0°4146 0°0954 0°3192 0°0000
(3) 0°0700 2 0°4146 0°0969 0°3177 0°0003—
(4) 0:0700 2 0°4146 0°0975 0°3171 0°0004—
(5) 0°1400 2 0°7832 0°1458 0°6374 0°0001—
6) 0°1400 2 0°7832 0°1465 0°6569 0°0002—
+ 0°1400 2 0°7832 0°1462 0°6370 0°0002—

The mean error of these determinations, in which: all
oxidizing material is calculated as bromate, is not far from
0-0001— grm.

In the case of still another preparation of potassium bromate,
made by acting with commercial bromine on potassium
hydroxide, crystallizing and _ recrystallizing several times,
determinations by means of arsenious acid and’ by the arseniate-
iodide process resulted as shown in the following statement :

TABLE V.

B
Arsenious Acid Method.
(Volume not exceeding 100°™?.)

AsoO3 As2O3 Error in
KBrO; As.O3; H.~SO, Time un- Oxi- terms of
taken. taken. a in changed. dized. KBrO;
erm. erm. erm. minutes. erm. erm. erm.
At the boiling temperature.
mea,O4 § O°2475 5 10 O° E241 0°1234 0°0010—
m0704 0°24'75 5 15 0°1239 0°1236 0:0009 —

0°0704 0°2475 oe 15 — 0°1256 0°1239 0°0007—

292 Gooch and Blake—Estimation of Bromic Acid.

TABLE V—Continued.
1
Arsenious Acid Method.
(Volume not exceeding 100°™3,)

As2O3 As2Os Error in
KBrO; As2,O; H2SO, Time un- oxi- terms of
taken. taken. Cs) in changed. dized. KBrO;
erm, erm, erm. minutes. erm. erm. erm.

On the steam bath,—80°.
0°0704 0°2475 5 30 0°1241 0°1229 0°0015—
O°0704 0°2475 5) 30 0°1246 0°1234 0°0010—
At atmospheric temperature.
0:0704 0°2475 9) 10 -0°2066 0°0409 0°0474—
0°0704 0°2475 5 10 0°1991 0°0488 0°0426 —
O°0704 0°2475 5) 30 0°1533 0°0922 0°0185—
00704 0°2475 5) 60 0°1231 0°1244 0°0004 —
00704 0°2475 i) 120 0°1242 0°1233 0°0011—
0°0704 0°2475 5 120 0°1238 0°1237 0°0009 —
B

Arseniate-Iodide Method.
(H2SO,) (1:1) 20°™* ; initial volume 110°™?; final volume 35°™3,
Iodine corre-
sponding to Jodine corre- Error in
KBrO; H2KAsO, I value of As2O3 pro- sponding to terms of

taken. taken. KI taken. duced.. KBrOs. KBrO3.
grm. erm. erm. erm. erm. erm.
0°0704 2 0°3812 0°0607 0°3205 0°0001 —
0°'0704 pi 0°3812 0°0588 0°3224 0'0003+
0°0704 2 0°3812 0°0595 Ors2ay 0°0001+-
0°0704 2 0°3842 0:0590 0°3222 0°0002+-
0'0704 2 0°3812 0°0615 0°3197 0:0003 —
0°0647 2 0°3812 0°0839 0°2973 0:0004+

The mean error of the arseniate-iodide process applied to
this particular sample of bromate is 0°0001+ grm. in terms of
bromate ; the arsenious acid method shows a mean deficiency
of about 0:0009 grm. for the smaller amount of bromate
employed, if, as is obviously reasonable, those results are
omitted from the averages which were obtained by standing
at the ordinary temperature for periods less than one hour. -

It appears, therefore, that the deficiency in the indications
of the iodide process and the arsenious acid process are
satisfactorily accounted for by the presence in the bromate of
traces of chlorate; and that the oxidizing power of a bromate
may be determined by boiling it in solution with a known
excess of arsenious oxide and an excess of sulphuric acid,
and determining the amount of arsenious oxide remaining
unchanged. A chlorate, as we have found by direct experi-
ment, is scarcely affected by this treatment.

os

Speyers—Solubilities of Some Carbon Compounds, ete. 298

Art. XXXIL.—Solubilities of Some Carbon Compounds and
Densities of their Solutions ; by CLARENCE L. SPEYERS.

At the present time we are more or less inclined to find in
solution an action corresponding to vaporization, and as vapori-
zation under ideal conditions is generally considered indepen-
dent of the contents of the space into which the substance
volatilizes, chemical action of course being excluded, so in
forming dilute solutions we generally expect ‘the act of solution
to be independent of the medium into which the solute passes,
that is, independent of the solvent. In some respects our
expectations are justified and find their expressions in the van’t
Hoff laws; but on the other hand, we cannot fail to see that the
solvent plays a very important part, for many substances refuse
altogether to dissolve in certain solvents, and consequently the
simplicity of the gaseous condition cannot be carried over
directly to solutes without running the risk of 1
straying away from the path leading to a satis-
factory explanation of solution.

It was in the hope of finding that the different
molecular aggregations of the solutes would
explain their differences in solubility that the
following experiments were made. It was in
the hope of finding that in a saturated solution
the osmotic pressure of the solute in the form
of simple molecules would be the same in all
solvents because the liquid phases are all bal-
anced by the same solid phase, that is by the
same undissolved solute. The large number of
measurements of the solubilities of metallic
salts were not available on account of ionic
decomposition.

The same solutes and solvents were chosen
which were used in determining heats of solu-
tion* and in determining molecular weights in
concentrated solutions.t

The apparatus shown in figure 1 was used; the tube A was
about 17° long and 25™ wide; B was about 12° long and
about 2°0™ wide; © was about of the same size as B. Solu-
tion, excess of solute and thermometer were in A. At the
bottom of B was a plug of filter paper, ); otherwise B was
empty, and C wasempty. After the tubes were char ged, they
were sunk in a water bath nearly to the tops of B and C. The
solution and solute were kept stirred by air blown through O,

the air escaping around the loosely fitting stopper of A. When

* Journ. Am. Chem. Soc., xviii, 146, 1896. + This Journal, xiii, 213, 1902.

294 C. L. Speyers—Solubilities of Some Carbon

the solvent was other than water, the air was dried by calcium ~
cehlorid.

About half an hour was needed to get the temperature of the
water bath constant to 0°1° for ten minutes. When that was
done, at the end of the ten minutes, the solution with undis-
solved solute was drawn over into B, and through the filter
plug into ©. When the solute was a nitrogenous body, the
solution was analyzed by the Kjehldahl method; by specific
gravity, when the solute was non-nitrogeneous. "Evaporation
to dryness and weighing the residue was altogether unsuitable,
too much decomposition.

The solutions were believed to be in just the right condition,
since the large mass of crystals present when the solution was
drawn over must have prevented appreciable supersaturation
on the one hand, and on the other, as the temperature was
always falling at this time, though only very slightly, the solu-
tion must have been completely “saturated,

The thermometer was carefully calibrated. It was divided
into tenths and the temperature is probably correct to this
quantity, but hardly correct to a smaller quantity on account
of the uncertainty of the temperature of the exposed thread.

The examination and purification of solvents and solutes
have already been described.*

The following tables give the results caleulated to per cent
gram-molecules, that is, to the number of gram-molecules of
solute in 100 gram-molecules of solution at the accompanying
temperature. The usual chemical formule were used in the
calculations.

WATER.
Chloral
Urea. Urethane. hydrate. Succinimid. Acetamid. Resorcinol.:
0-07: 16°77 O:0 Bb 0°0° 20°66 0°0° 1°58 0°0° 29°64 3500 eee
11°0° 20°82 10°3°. 6°09 11°3° 30°23 11°3° 2°74 10°6°° 34°35) TO aie
19°8° 22°69 “11°1° 6°62 23°7° 45°86 20°7°° 4°23 19°9° 40°72. oe
31°7° 28°24 23°5° 43°70 38°1° 59'41 °33°3° 9:91 39°9°° 40:62) anes eee

51-40 S667 el4o soe el 69°3° 27:14 45°5° 60°14 VWoer som
69°5° 4915) 1870-075 °58 63°0°.7 740
MeETHYL ALCOHOL.
Urea. Urethane. Acetanilid. Naphthalene. Acenaphthene.
O°0° (eas OO". S1-18 F077 Se eacss 00" =) Ored 00% §«=0°3 9%

10°S° “Heaval 10°6° 41°70 115? E709 146°" 4-68 12-4". 0°38
21°7° 10°81 99°5° 58:58 99°38? Te 1e 31°89 210-57 30°7°: 0°Tam
404° 15°96 40°9° 90:00 33°6° 1396 AS-Oon se sae A6°0°. —-:1°5S

612°. ar ge 402 Ya 17-05 59°9° (12°34 62°3° 29-46%
ATA? > “DB-79
60°9° 35:24
63°35 .83°38

* This Journal, 1. ¢.

Compounds and Densities of their Solutions. 295

ETHYL ALCOHOL.

Urea. Urethane. Acetanilid. Naphthalens. Resorcinol. Acenaphthene.
freon 23°91 00°) 501 . 0-0". 1°80 0°0° 34:37 0°0° 0°57
3°04 10°5° 36°86 10°8° 6°84 86. Stee o-2>. 37-69 “Oe 0-84
meeeestigeas ol) 49°57 17-93 318° 489 31°8° 41°84 30°3°. 1°70
BGbs0-9- 72°35 43°5° 15°40 46°9° 9-70 50°6° 47°07 49°8° 3°86
ies S032 61°6° 31°36 69°8° 64°92 73°1° 58°03. 71°6° 12°94

_ Phenan- Chloral |

_ threne. hydrate. Succininid. Acetamid. Benzamid. p-Toluidin.
Os2— 0-0° 34°35 00°. Ges 900° 185305200" S06 0°0° 20°72
eee 41-08 11-1°.. 1-36 -18°6° 32°87 10°4° 4:25 11°7° 33°88
meee S316. 24°2° 2°44 49°5° 56°06 32°6° 8°72 . 22°1° 50°90
BG ASA? 92°32 43°7° 5°63 62°0° 78°92. 50°4° 14°44

ae bea 58°6° 11°49 722°. 20°86

PrRopyL ALCOHOL.

Urethane. Naphthalene. Acenaphthene.
00° 19°48 0:0° 2°09 0:0° 0°88
mee. 59°97 10°4° 2°70 10°5° 0°97
91 53°31 30°3° 5°34 Sib 1°88
30°4° 68°75 50°3° 15°34 50°3° 4°37
40°7° 85°74 68°5° 62°9 Pa 4 9-8
CHLOROFORM.
Chloral
Urethane. Acetanilid. Naphthalene. Acenapthene. hydrate.

Pee see 0) 63-24 «070° 19°58 0-0° 12°72 0:°0° 2°67
weet oe 106 7-32. 10°6°°23-714. 11°2° 16°54 12°5° 4-293
Reeeteroa abs 15°64 30°3° 38°53 29°8° 54-62? 27-7° 31-93
311° 69°92 45°9° 24:07 52°5° 57°40 52°7° 42°19 44-0° 100°0

epee test 61°4° 37°55 44°4° 100°0
46°3° 100°0
TOLUENE.
Phenan- Chloral
Urethane. Naphthalene. Acenaphthene. threne. hydrate.

ore O 12°82 0:0°' 788 ~0-0° 11-88 0:0° 1°78
eared 25:2" 23-96 10°3° 10°76 13°9° 17°40 10°0° 4:24
S39 Ga60 46°3° 47°37 24°1° 16°53 30°8° 26°90 20°7° 11°42
4A0°7° 87°86 69°5° 84°43 41°6° 29°29 54°9° 53°25 29°6° 31°18

61°5° 45°08 78°3° 81°98 42°5° 89°86

It is unnecessary to plot these data, for with the exceptions
of urethane in water and chloral hydrate in chloroform, the
curves are quite regular. In the cases of urethane and chloral
hydrate, the curves change direction very suddenly at about
12°, showing a very sudden decrease in solubility as the tem-
perature decreases from this point.

296 C. L. Speyers—Solubilities of Some Carbon

Plots showing the variation in solubility as we pass from
one solvent to another are perhaps of some interest. They are

GAA

genie

CHL,
) /0 Zo 30
Sutcinimid

ee

NeANE

—!

WEN

°
S
S
w
8

N
NN
La
a
iA
eat
ea

os: | Lith slene :
= Hey Rie hy B 5 a
pe be Ho B22
=" 10 20 30 0 50 60 70 o /0 ZO JO 7) 50 se
Aulamnia

Resoreinst
given in figures 2 and 3. The solubilities are expressed in per-
centage gram-molecules* along the horizontal line in that sol-

* According to the ordinary chemical formule.

Compounds and Densities of their Solutions. 297

vent whose formula is.given at the commencement of the line.
We notice that the relative solubility of a solute in several
solvents can change very much with the temperature. For
instance, acetamid is less soluble in alcohol than in water
between 0° and about 50°, but above 50° it is more solable in
aleohol than in water. This is analogous to the relative
changes in vapor pressures of some liquids. We notice the
small increase in solubility of urethane in water as we pass
from 0° to 10° and the rapid increase as we pass beyond that
temperature. There seems to be some slight error in the
solubility of the aqueous solution at 30°, but there is none at
10°, for the solubility at that temperature was determined
twice. We notice also that there is no regularity in change of
solubility as the molecular weights of the alcoholic solvents
increase, whether we consider the molecular aggregations of
the solutest or not.
M. Schroedert has deduced the equation

n
ae

= A(t,-t)

in which 7 is the number of gram-molecules of solvent in a
saturated solution; A is a constant, the same for each solute
whatever the solvent may be. 7, is the fusion temperature of
the solute and z is the temperature of saturation. How this
equation is obtained need not be considered. According to it,
the curves of solubility of any solute in all solvents should
coincide. Schroeder tested the equation with p-dibrombenzene
in ethyl alcohol, propyl alcohol, isobutyl alcohol, ethyl ether,
earbon disulphid, benzene, and brombenzene; with naphthalene
in benzene. chlorbenzene, and carbon-tetrachloride;. with
m-dinitrobenzene in benzene, brombenzene, and chloroform.
His results justify it, but mine do not; only for solutions of
acenaphthene, naphthalene, and phenanthrene are the coinci-
dences at all good.

The specific volumes of the solutions must be known before
we can proceed to determine the osmotic pressures of the
solutes, and therefore we now pass on to the densities.

These were easily obtained by sinking a pycnometer with a
very small mouth into the saturated solution and drawing this
in, the very small mouth keeping out particles of solute.
Since the solution was kept stirred by air and contained an
abundant excess of solute and since the temperature was kept
constant for some ten minutes before drawing into the pyc-
nometer, the solutions are to be considered in just the right

+ This Journal, 1. c.
¢ Zeitschr. Phys. Chemie, xi, 449, 1893.

298 C. L. Speyers—Solubilities of Some Carbon

condition, neither supersaturated nor undersaturated. The
thermometer was the one used for the solubilities and the tem-
perature is certain to 0°1°.

The uncertainty in the density is about 3 units in the third
piace in extreme cases; in general it is 2 or 3 units in the
fourth place, the pycnometer holding between 8°5 and 9:8"
and the weighings being made to milligrams.

The following tables give the temperatures and correspond-
ing densities; namely, weights in grams reduced to vacuum of
one cubic centimeter of solution.

WATER.
Chloral

Urea. Urethane. hydrate. Succinimid. ©
BO. a~d9T 0:07 1/1028 0°02 T=438 0°0° °1°025
19-0" nei Ay 110° 12025 15:3". Tage 15°9°" "104s
39°2° 1:7158 04-0 "07s 31-0" 1578 36°6° 3 rad
463 2 ae 38°8° 1:°064 , 466° 1623 50°3° > 1s140
6475" .4°181 39-7". 2065 65°0°.) 15481
84-91%) 047196 84°4° 1°208

Acetamid. . Resorcinol. Mannitol.

0:0° 1°055 00° 108 00° ~ 1°044

15°6° 1:046 14-4? 1125 15°2°° 1050

B16) 1087 B1-1° > 449 ST? ee

49°" “1 O22? AS O° FGI 47°7° 1:099

68°4° 1°011 Go'47 ea ye 68°0° 1°148

85°4° 1°179 85°9° -1:2907

METHYL ALCOHOL.

Urea. Urethane. Acetanilid.

070° 0°8612 0°0°. 0°9565 00° 0°8602
17°4° 0°8674 15°5° 0°9902 16°7° 0°8698
29°7° 0°8764 28°2° 1 oza 29°9° 0:8924
50°5° =0°9086 39°5° 1:044 43°9° 0°9206
66°8° 0°9534 61°7° 0°9596
Methyl Alcohol

Naphthalene. Acenaphthene. (pure).
00° 0°8194 00° 0°8133 0°0° 08110
16°6° 0°8088 12°5> -O0B030 18°3° 0°7938
29°0° 0°8048 31:6, 0°7900 29°4° 0°7834
47°0° 0°8086 45°6° 0°7823 46°2° 0°7675
61°7° 0°8436 64°1° 0°78438 59°8° 0°7540

68°0" ,.0°9022

Compounds and Densities of their Solutions. 299
EtHyL ALCOHOL.
: Chloral
- Urea. Urethane. hydrate. Succinimid. Acetamid.
7 (08915 0:0° 0°8914 0°0° -£-110 0:0° 0°8150 0:0° 0°8562
Pee obs) b4°1° 0°9443. 16-7° 1319 17°5° 0°8063 17°8° 0°8696
31°6° 0°8060 30°9° 1:004 33°5° 1:°560 33°2° 0°8052. 35:0° 0°8974
aie 2 O-S0at .43°7°° 1°044 40°8° 1°589 57°5° 0°8292 54°4° 0°9416
71°5° 0°8124 79°64. 0°9552> 70:3° 0°9815
80°6° 0:9563
Resorcinol. Benzamid. Acetanilid. p-Toluidin. Naphthalene.
07 1°033 0:0° 0°83381 0-0° 0°8420 0°0° 0°8884 0°0° 0°8175
wine? O96 141° O'8328 17-2° 0°8472 15°6° 0-9168 —17:0° 0°8104
SG°h 1-054 §36-2° 0°8434 39°0° 0°8721 28°4°.0°9458 31°2° 0°8084
G29 077) 572° 078754 58°1° 0°9156 40°6° 0:9636 51°0° 0°8230
ie tr 107, 72°8° 0°9226 76°7° 0°9596 72°4° 0:9563
23 Sy sein it E
Ethyl alcohol

Phenanthrene. Acenaphthene. (pure).

0:0° 0°8141 0:0° 0:8108 0:0° 0°8074

15'6° 0°8035 15°0° 0°8013 - 19°1° 0°7921

35°3° 0°7960 36°3° 0°7910 35°3° 0°7780

52°2° 0°7941 53°4° 0°7890 --52:3° 0°7633

76°4° 0°8654 73°0° 0°8186 - 72°8° 0°7448

PRoPpYL ALCOHOL.
Propyl alcohol

Urethane. Naphthalene. Acenaphthene. (pure).

“QV? ° “—H)2 . 5 “Hj? , te) , :
ics nase: isp 0.8111 .« 18 0° 0.8007
29°1° 09804 30°7° 0°8206 26°6° 0°8110 28°0° 0°7991:-
fet 0353 41°8° 0°8247 47°4° 0°8063 44°4° 0-7854

59°7° 0°8634 64*7— -0°8217 65°1° 0°7678
FAS 0-958 -< >8as3 - Or071sGy .S0-G° .0-7538
CHLOROFORM. |
Chloral

Urethane. hydrate. Acetanilid. p-Toluidin.

“Me 7 Yate 1°52 “He “no Or
a es 2s care 00 es
aE2° £IES1 34°4° 1°565 SAD? L381 31°4° 1:066
43°4° 1°097 - 44°6° 1°615 AT:3° 1-326 40°8° 0°9906

62°47 1-261
69:9° 1:227
Chloroform
Naphthalene. Acenaphthene. (pure).
0°0° 1:393 0°0° 1°438 00° 1°526
90°25 "E304 16°5° 1°378 18°2° 1°492
S52 e250 29°1° 1:°328 33°4° 1°464
53°2°) 1°E36 40°3° 1°28) 44°6° 4448
64°5° 1°072 55°8° 1:209 58-60. bab 7

300 C. L. Speyers—Solubilities of Some Carbon

TOLUENE.
Chloral

Urethane. hydrate. p-Toluidin. Naphthalene.
0°0° 0°8872 0:0° 0:8978 0°0° 0:9110 0:0° 0°9124
18°4° 0°8792 18°5° 0°9328 16°6°, 0°9228 | 173° aSise
34°0° 0°9570 28°4° 1:069 29°4° 0°9430 36°2° 0°9224
45°2° 1:042 40°8° 1°448 40°7° 0°9616 51%") 0:93 74
42°0° 1°445 73°9° 0°9692

Phenanthrene. Acenaphthene. Toluene (pure).

0:0° 0°9254 0:0° 0:9066 0°0° 0°8840

18°3° 0°9337 13°3° 0:9076 17°6° 0°8678

37°9° 0°9515 32°3° 0°9196 32°1° 0°8544

60°1° 0°9890 57°2° 0°9502 49°8° 0°8381

86°3° 1:038 83°5° 1:006 65°9° 0°8232

85°0° 0°8048

Of these solutions, so far as they could be examined, only
the following contain solutes with normal molecular weights.*

Urea in water at 32° to 41°
a “methyl aleohol “ 28°
““c cc ethyl “ “c a7S
Acetanilid + ‘methyl > * © 31°(?)
= «ethyl Fé “21 toes
Naphthalene ee att 3 6 98° 6
# “«" toluene “.* ‘99°. “Aa
Acenaphthene “ propylalcohol “ 45°(?)
4 *« “ toluene © 91" 10a

Selecting 27° as the temperature and taking solubilities at
that temperature from solubility curves and densities from
density curves, we get the following table: |

Osmotic

Solution. Solubility. Density. Pressure.

Urea in water "222s. 22222 26 grmmols 1155 10:4 atmos.
“ _-gmethyl alcohol 22 - 12 i 0°873 3 Oe
« ethyl eaeerare 5 ae 0805 «og
Acetanilid-in methyl alcohol 12° “ 0°891 Dae ee
«“ ethyl ec 14400 0°858) > a-qe eee
Naphthalene inethyl alcohol 4 “ 0°810 Ota
° “ toldaene -_-- - 25 “4 0-918 a
Acenaphthenein propylalcoh’] 1 ‘ 0'810 Che
sd tolavene —_.. = 18 oF 0-918 Le

The osmotic pressure is not at all constant for each solute,
and we see that the undissolved solute does not balance the
dissolved solute to an equal extent in each solvent though the
molecular weights of the solutes are simple. So the analogy
of solution to vaporization is seen to be far from complete.

* This Journal, 1. c.

Compounds and Densities of their Solutions. 301

In conclusion a few words ahout the densities and molecular
volumes. There is no need of putting the data given above
into plots. There is no peculiarity in any of them, not even
in the aqueous solution of urethane nor in the chloroform
solution of chloral hydrate, but the plets showing variation in
molecular volume as a solute passes from one solvent into
another are instructive. They are given in figure 4.

AoW

IN| CHOH Rik

The formula at the beginning of the horizontal line gives
the solvent, and the numbers along the horizontal give the
relative volumes of one gram-molecule of solute at the temper-
ature stated on the lines.

For example, the relative volume of one gram-molecule of
acenaphthene in chloroform at 10° has been obtained in this
way :—

Solubility of acenaphthene from solubility curve=16°0% gram-
mols. .
Proportion of chloroform=84°0*119°3/ (84°0° 1193 + 16°0° 154).
Volume of one gram chloroform at 10° from density curve=
1/1°51ce.
«© chloroform in one gram of solution is [84°0*119:3/
(84:0 *119°3 +16°0* 154)] 1/1°51cec=0°5327ce.
« one gram of solution at 10°=1/1°40ce.
« © acenaphthene in one gram of solution=1/1°40—0°5327
=0°1811ce.
«< _ corresponding to one molecule of acenaphthene=0°1811/16
=0°'0113.

Am. Jour. Sc1.—FourtH SERIES, Vout. XIV, No. 82.—OcToBsr, 1902.
21

302 Speyers—Solubilities of Some Carbon Compounds, ete.

For convenience, the values are all multiplied by 1000 before
plotting.

We notice the following :—

The molecular volumes of the solutes are larger in water
than in the other solvents and decrease in the order of methyl
alcohol, ethyl alcohol, propyl! alcohol, toluene and chloroform,
the only exception being perhaps choral hydrate in ethyl
alcohol compared to its solution in chloroform.

The molecular volumes in chloroform show decided constancy
for every temperature and concentration observed, and to a
lesser extent in toluene. :

In general, the molecular volumes of the solutes decrease as
the temperature rises and the concentration increases. The
exceptions are urethane and chloral hydrate in chloroform.

Rutgers College, April, 1897-June, 1902,

Chemistry and Physics. 303

SGLtENTIFIC INTELEIGENCE.

I. CHEMISTRY AND PHysIcs.

1. Radio-active Bismuth.—Polonium, the radio-active sub-
stance occurring with bismath in pitchblende, was the first of
these substances discovered (by Mr. and Mrs. Curie). Subse-
quent investigations by Giesel led the latter to believe that
polonium was nothing more than bismuth made active by induc-
tion. Marck wa p, however, has recently obtained results which
indicate that polonium is a distinct element. From a by-product
obtained from pitchblende, by the usual methods, bismuth oxy-
chloride was prepared which was strongly radio-active, and
which showed no decrease in activity after several months.
Upon subjecting a solution of this substance to electrolysis, it
was found that the metal which was deposited at first displayed
much greater activity than the final product. This result led to
the attempt to precipitate polonium, from hydrochloric acid solu-
tions of the oxychloride, by means of a polished stick of metallic
bismuth. As a result, the metal became coated immediately with
a fine black deposit, which gradually increased, and when it had
been removed from the solution and washed, it showed a surpris-
ingly strong effect on the electroscope. At the distance of a
decimeter the leaves of the charged electroscope collapsed in a
moment, and even a gutta-percha rod which had been well rubbed
with fur was immediately discharged upon the approach of the
bismuth stick. The important fact was noticed that metallic
bismuth is thus able to remove practically all of the active
material from a solution. The powder could be scraped from
the bismuth rod, and the amount thus obtained from 8§ of bis-
muth oxychloride was about 5™%. From this result it was calcu-
lated that a ton of pitchblende would contain not more than one
gram of the substance. The powder was found not to be a pure
metal, but to contain some chlorine. Upon heating, a small por-
tion volatilized, probably as chloride, and the residue fused to a
white, exceedingly brittle, metallic bead. This was soluble in
nitric acid, and the solution, as far as it could be tested, gave
reactions for bismuth. The salts are as strongly active as the
metal, and the rays are characteristically different from those of
radium in being unable to penetrate any intervening substance.
Even a piece of filter-paper wrapped around a stick of bismuth
coated with the active metal causes it to lose all its effect. The
experiments are to be continued with larger quantities of material,
and it is hoped that an atomic weight determination may be
made.— Berichte, xxxv, 365. . H. L. W.

2. A Thermochemical Constant.—F. W. CLARKE, in a pre-
liminary paper, has made a generalization from a consideration
of Thomsen’s work on the heats of combustion of certain organic
compounds. In the first place, he has re-caleulated Thomsen’s

304 Scientific Intelligence.

results so that they represent the formation of gaseous instead of
liquid water. He then finds for the aliphatic hydrocarbons and

their non-oxygenated derivatives that ___ +4 ee
12a+6b—c—8n

stant, where A represents the heat of combustion of the com-
pound, @ the number of molecules of CO, produced, 6 the num-
ber of molecules of H,O, ¢ the number of oxygen molecules
involved, and 2 the number of atomic unions or linkings in the
compound burned. In fixing the value of n, double and triple
linkings are not distinguished from simple ones, so that for all
chain molecules m is equivalent to the number of atoms minus
one. The coefficients 4, 12, 6, | and 8 appear to have been
chosen arbitrarily, but the results obtained with them in the
cases of fourteen hydrocarbons are remarkably uniform, and the
author evidently considers it important that the constant, which
averages 13,878 in this case, is very near the neutralization con-
stant of acids and bases, which is 13,700 calories. The same
uniform constant was obtained with many other organic com-
pounds by modifying the formula in accordance with the presence
of several other elements : thus for halogen compounds were used

4K 4k

and
124+6b+h—c—8n 12a4-664+2h,—ce—8n

the formulas

41
12a+6b+4h,—ce—8n’
of chlorine, bromine and iodine molecules, supposed by the author
to be formed by the complete combustion of the halides. For
nitrogen compounds two formulas were used ; for cyanides and
nitrites, where m represents the number of nitrogen molecules
set free, the divisor is 124+6)6+3m—c—8m, while for amines 3m
is replaced by 9m. For sulphur compounds the term 9s is intro-
duced, while c,, the number of oxygen molecules present in the
compound, is subtracted when the latter element is present. In
this way conformable results were obtained with nineteen halogen
compounds, four cyanides, ten amines, eight sulphur compounds,
and eleven alcohols. The ring compounds have not yet been
worked out in this way, and there are a few substances in most
of the groups which fail to give the usual result. A considera-
tion of the exceptions is deferred until the complete memoir is
published. The conclusion is reached that, in any class of com-
pounds, the heat of formation is proportional to the number of
atomic linkings within the molecule, and seems to bear no rela-
tion to the masses of the atoms which are combined.

It is to be hoped that Professor Clarke has discovered a new
and important thermochemical law, but from the arbitrary nature
of the formulas used the suspicion arises that the law may be a
very obscure one.—Jour. Amer. Chem. Soc., xxiv, 882. H.L. W.

3. The Heatless Condition of Matter.—In the June number of
this Journal mention was made of Brinkworts and MarrTin’s
theory that pressure may be able to prevent moleeular vibration

where h, h, and h, represent the numbers

.

Chemistry and Physics. 305

and thus cause matter to assume a condition which is devoid of
heat. In regard to that notice the editor has received from one
of the authors a letter which is too long to be inserted, but which
should be acknowledged here. It does not appear that objection
is made to our presentation of the main points of the theory, but
in relation to our opening remark, “ Brinkworth and Martin,
with apparent seriousness, have made a curious extension of the
kinetic theory,” the letter says, “ Your abstractor appeared to
doubt the seriousness of the paper. I write at once to say that
in this matter we are in deadly earnest.” The fact is, we
intended to convey the idea that the seriousness of the article
impressed us. Objection is made to our opinion that “Their
deductions are based upon an entirely unproven assumption,”
and, in this connection, the letter refers to some points which are
to be brought out in a future article. It seems to us that no
injustice has been done to the authors as far as their preliminary
publication is concerned, and that further discussion of their
theory should be postponed until their more elaborate paper has
appeared. H. Ti. W.

4. Elementary Chemistry, by F. W. Criarke and L. M._—
Dennis. 12mo. Pp. 340. New York, 190% (American Book
Company).—In preparing this little text-book the authors have
aimed especially to make it a means of training in the interpreta-
tion of evidence. Each generalization is made to follow the
evidence upon which it rests. The work is accompanied by a
laboratory manual, but a reasonable number of experiments,
nearly all of which are of the simplest character, are described
throughout the text. The book appears less dry and more inter-
esting than most of the works on the subject on account of the
constant attention which is paid to the application of chemistry
to human affairs and its utility in modern life. The authors have
done no more with the theory of ionization than to give it a very
brief mention. Although they consider this theory unsuitable
for discussion in a work of this kind, it seems probable that its
use In a simple manner would present no more difficulty to the
beginner than some of the theories which are introduced. It is
satisfactory to notice that the book contains a short treatise
(about 60 pages) on organic chemistry, a subject which is too
often omitted in books on elementary chemistry. Very few things
have been noticed in the work which seem unsatisfactory, but it
appears that the somewhat imaginary structural formula for the
double sulphate of magnesium and potassium (p. 202) is hardly
appropriate in such a book, and objection may be made to the
equation KCIO,=KC1+0,, because O, is ozone. A mistake is
made in characterizing lithium as the lightest solid known, for
solid hydrogen is mentioned in the book. H..L.. We

5. Liquid Hydrogen and Helium; from the inaugural address
delivered at Belfast by Prof. James Dewar, President of the Brit-
ish Association.— * * * From this speculative divergency it is clear
no definite conclusion could be reached regarding the physical prop-

306 Scientific Intelligence.

erties of liquid or solid hydrogen, and the only way to arrive at
the truth was to prosecute low-temperature research until success
attended the efforts to produce its liquefaction. This result I
definitely obtained in 1898. The case of liquid hydrogen is, in
fact, an excellent illustration of the truth already referred to, that
no theoretical forecast, however apparently justified by analogy,
can be finally accepted as true until confirmed by actual experiment.
Liquid hydrogen is a colorless, transparent body of extraordinary
intrinsic interest. It has a clearly defined surface, is easily
seen, drops well, in spite of the fact that its surface tension is
only the thirty-fifth part of that of water or about one-fifth that
of liquid air, and can be poured easily from vessel to vessel.
The liquid does not conduct electricity, and, if anything, is
slightly diamagnetic. Compared with an equal volume of liquid
air, it requires only one-fifth the quantity of heat for vaporization;
on the other hand, its specific heat is ten times that of liquid air
or five times that of water. The coefficient of expansion of the
fluid is remarkable, being about ten times that of gas; it is by
far the hghtest liquid known to exist, its density being only one-
fourteenth that of water; the lightest liquid previously known
was liguid marsh gas, which is six times heavier. The only solid
which has so small density as to float upon its surface is a piece
of pith wood. Itis by far the coldest liquid known. At ordi-
nary atmospheric pressure it boils at minus 252°5 degrees or 20°5
degrees absolute. The critical point of the liquid is about 29
degrees absolute and the critical pressure not more than fifteen
atmospheres. ‘The vapor of the hydrogen arising from the liquid
has nearly the density of air—that is, it is fourteen times that of
the gas at the ordinary temperature. Reduction of the pressure
by an air-pump brings down the temperature to minus 258
degrees, when the liquid becomes a solid resembling frozen foam,
and this by further exhaustion is cooled to minus 260 degrees, or
13 degrees absolute, which is the lowest steady temperature that
has been reached. The solid may also be got in the form of a
clear, transparent ice, melting at about 15 degrees absolute,
under a pressure of 55™™, possessing the unique density of one-
eleventh that of water. Such cold involves the solidification of
every gaseous substance but one that is at present definitely
known to the chemist, and so liquid hydrogen introduces the
investigator to a world of solid bodies. The contrast between
this refrigerating substance and liquid air is most remarkable,
On the removal of the loose plug of cotton-wool used to cover the
mouth of the vacuum vessel in which it is stored, the action is
followed by a miniature snow-storm of solid air, formed by the
freezing of the atmosphere at the point where it comes into con-
tact with the cold vapor rising from the liquid. This solid air
falls into the vessel and accumulates as a white snow at the bot-
tom of the liquid hydrogen. When the outside of an ordinary
test-tube is cooled by immersion in the liquid, it is soon observed
to fill up with solid air, and if the tube be now lifted out a

Chemistry and Physics. 307

double effect is visible, for liquid air is produced both in the
inside and on the outside of the tube—in the one case by the
melting of the solid, and in the other by condensation from the
atmosphere. A tuft of cotton-wool soaked in the liquid and then
held near the pole of a strong magnet is attracted, and it might
be inferred therefrom that liquid hydrogen is a magnetic body.
This, however, is not the case: the attraction is due neither to
the cotton-wool nor to the hydrogen—which indeed evaporates
almost as soon as the tuft is taken out of the liquid—but to the
oxygen of the air, which is well known to be a magnetic body,
frozen in the wool by the extreme cold.

The strong condensing powers of liquid hydrogen afford a
simple means of producing vacua of very high tenuity. When
one end of a sealed tube containing ordinary air is placed for a
short time in the liquid, the contained air accumulates as a solid
at the bottom, while the higher part is almost entirely deprived
of particles of gas. So perfect is the vacuum thus formed, that
the electric discharge can be made to pass only with the greatest
difficulty. Another important application of liquid air, liquid
hydrogen, etc., is as analytic agents. Thus, if a gaseous mixture
_ be cooled by means of liquid oxygen, only those constituents
will be left in the gaseous state which are less condensable than
oxygen. Similarly, if this gaseous residue be in its turn cooled
in liquid hydrogen, a still further separation will be effected,
everything that is less volatile than hydrogen being condensed to
a liquid or solid. By proceeding in this fashion it has been
found possible to isolate helium from a mixture in which it is
present to the extent of only one part in one thousand. By the
evaporation of solid hydrogen under the air-pump we can reach
within 13 or 14 degrees of the zero, but there or thereabouts our
progress is barred. This gap of 13 degrees might seem at first
insignificant in comparison with the hundreds that have already
been conquered. But to win one degree low down the scale is
quite a different matter from doing so at higher temperatures;
in fact, to annihilate these few remaining degrees would be a far
greater achievement than any so far. accomplished in low-
temperature research. For the difficulty is twofold, having to do
partly with process and partly with material. The application
of the methods used in the liquefaction of gases becomes con-
tinually harder and more troublesome as the working tempera-
ture is reduced ; thus, to pass from liquid air to liquid hydrogen
—a difference of 60 degrees—is, from a thermodynamic point of
view, as difficult as to bridge the gap of 150 degrees that sepa-
rates liquid chlorine and liquid air. By the use of a new
liquid gas exceeding hydrogen in volatility to the same extent
as hydrogen does nitrogen, the investigator might get to within
five degrees of the zero; but even a second hypothetical sub-
Stance, again exceeding the first one in volatility to an equal
extent, would not suffice to bring him quite to the point of
his ambition. That the zero will ever be reached by man is

308 Scientific Intelligence.

extremely improbable. A thermometer introduced into regions
outside the uttermost confines of the earth’s atmosphere might
approach the absolute zero, provided that its parts were highly
transparent to all kinds of radiation, otherwise it would be
affected by the radiation of the sun, and would therefore
become heated. But supposing all difficulties to be overcome,
and the experimenter to be able to reach within a few degrees of
the zero, it is by no means certain that he would find the near
approach of the death of matter sometimes pictured. Any fore-
cast of the phenomena that would be seen must be based on the
assumption that there is continuity between the processes studied
at attainable temperatures and those which take place at still
lower ones. Is such an assumption justified? It is true that
many changes in the properties of substances have been found to
vary steadily with the degree of cold to which they are exposed.
But it would be rash to take for granted that the changes which
have been traced in explored regions continue to the same extent
and in the same direction in those which are as yet unexplored.
Of such a breakdown low-temperature research has already yielded
a direct proof at least in one case. A series of experiments with
pure metals showed that their electrical resistance gradually
decreases as they are cooled to lower and lower temperatures, in
such ratio that it appeared probable that at the zero of absolute
temperature they would have no resistance at all and would
become perfect conductors of electricity. This was the inference
that seemed justifiable by observations taken at depths of cold
which can be obtained by means of liquid air and less powerful
refrigerants. But with the advent of the more powerful refrig-
erant liquid hydrogen it became necessary to revise that conclu-
sion. A discrepancy was first observed when a platinum resistance
thermometer was used to ascertain the temperature of that liquid
boiling under atmospheric and reduced pressure. All known
liquids, when forced to evaporate quickly by being placed in the
exhausted receiver of an air-pump, undergo a reduction in tem-
perature, but when hydrogen was treated in this way it appeared
to be an exception. The resistance thermometer showed no reduc-
tion as was expected, and it became a question whether it was

the hydrogen or the thermometer that was behaving abnormally. .

Ultimately, by the adoption of other thermometrical appli-
ances, the temperature of the hydrogen was proved to be lowered
by exhaustion as theory indicated. Hence it was the platinum
thermometer which had broken down ; in other words, the elec-
trical resistance of the metal employed in its construction was
not, at temperatures about minus 250° C., decreased by cold in
the same proportion as at temperatures about minus 200°. This
being the case, there is no longer any reason to suppose that at
the absolute zero platinum would become a perfect conductor of
electricity ; and in view of the similarity between the behavior
of platinum and that of other pure metals in respect of tempera-
ture and conductivity, the presumption is that the same is true of

Chemistry and Physics. 309

them also. At any rate, the knowledge that in the case of at least
one property of matter we have succeeded in attaining a depth of
cold sufficient to bring about unexpected change in the law express-
ing the variation of that property with temperature, is sufficient
to show the necessity for extreme caution in extending our infer-
ences regarding the properties of matter near the zero of tempera-
ture. Lord Kelvin evidently anticipates the possibility of more
remarkable electrical properties being met with in the metals near
the zero. A theoretical investigation on the relation of “elec-
trions” and atoms has led him to suggest a hypothetical metal
having the following remarkable properties: below 1 degree
absolute it is a perfect insulator of electricity, at 2 degrees it
shows noticeable conductivity, and at 6 degrees it possesses high
conductivity. It may safely be predicted that liquid hydrogen
will be the means by which many obscure problems of physics
and chemistry will ultimately be solved, so that the liquefaction
of the last of the old permanent gases is as pregnant now with
future consequences of great scientific moment as was the lique-
faction of chlorine in the early years of the last century.

The next step towards the absolute zero is to find another gas
more volatile than hydrogen, and that we possess in the gas
occurring in cleveite, identitied by Ramsay as helium, a gas which
is widely distributed, like hydrogen, in the sun, stars and nebule.
A specimen of this gas was subjected by Olszewski to liquid air
temperatures, combined with compression and subsequent expan-
sion, following the Cailletet method, and resulted in his being
unable to discover any appearance of liquefaction, even in the
form of mist. His experiments led him to infer that the boiling-
point of the substance is probably below 9 degrees absolute.
After Lord Rayleigh had found a new source of helium in the
gases which are derived from the Bath springs, and liquid hydro-
gen became available as a cooling agent, a specimen of helium
cooled in liquid hydrogen showed the formation of fluid, but this
turned out to be owing to the presence of an unknown admixture
of other gases. Asa matter of fact, a year before the date ot
this experiment I had recorded indications of the presence of
unknown gases in the spectrum of helium derived from this
source. When subsequently such condensable constituents were
removed, the purified helium showed no signs of liquefaction,
even when compressed to 80 atmospheres, while the tube contain-
ing it was surrounded with solid hydrogen. Further, on suddenly
expanding, no instantaneous mist appeared. Thus helium was
definitely proved to be a much more volatile substance than
hydrogen in either the liquid or solid condition. The inference
to be drawn from the adiabatic expansion effected under the
circumstances is that helium must have touched a temperature of
from 9 to 10 degrees for a short time without showing any signs
of liquefaction, and consequently that the critical point must be
still lower. This would force us to anticipate that the boiling-
point of the liquid will be about 5 degrees absolute, or liquid

310 | Scientific Intelligence.

helium will be four times more volatile than liquid hydrogen,
just as liquid hydrogen is four times more volatile than liquid
air. Although the liquefaction of the gas is a problem for the
future, this does not prevent us from safely anticipating some of
‘the properties of the fluid body. It would be twice as dense as
liquid hydrogen, with a critical pressure of only 4 or 5 atmos-
pheres. The liquid would possess a very feeble surface-tension,
and its compressibility and expansibility would be about four
times that of liquid hydrogen, while the heat required to vaporize
the molecule would be about one-fourth that of liquid hydrogen.
Heating the liquid 1 degree above its boiling-point would raise
the pressure 1# atmospheres, which is more than four times the
increment for liquid hydrogen. The liquid would be only seven-
teen times denser than its vapor, whereas liquid hydrogen is sixty-
five times denser than the gas it gives off. Only some 3 or 4
degrees would separate the critical temperature from the boiling-
point and the melting-point, whereas in liquid hydrogen the sepa-
ration is respectively 10 and 15 degrees. As the liquid refractivi-
ties for oxygen, nitrogen and hydrogen are closely proportional
to the gaseous values, and as Lord Rayleigh has shown that
helium has only one-fourth the refractivity of hydrogen, although
it is twice as dense, we must infer that the refractivity of liquid
helium would also be about one-fourth that of liquid hydrogen.
Now hydrogen has the smallest refractivity of any known liquid,
and yet liquid helium will have only about one-fourth of this
value—comparabie, in fact, with liquid hydrogen just below its
critical point. This means that the liquid will be quite excep-
tional in its optical properties, and very difficult to see. This
may be the explanation of why no mist has been seen on its adia-
batic expansion from the lowest temperatures. Taking all these
remarkable properties of the liquid into consideration, one is
afraid to predict that we are at present able to cope with the
difficulties involved in its production and collection. Provided
the critical point is, however, not below 8 degrees absolute, then
from the knowledge of the conditions that are successful in pro-
ducing a change of state in hydrogen through the use of liquid
air, we may safely predict that helium can be liquefied by follow-
ing similar methods. If, however, the critical point is as low as
6 degrees absolute, then it would be almost hopeless to anticipate
success by adopting the process that works so well with hydrogen.
The present anticipation is that the gas will succumb after being
subjected to this process, only, instead of liquid air under exhaus-
tion being used as the primary cooling agent, liquid hydrogen
evaporating under similar circumstances must be employed. In
this case, the resulting liquid would require to be collected in a
vacuum vessel the outer walls of which are immersed in liquid
hydrogen. The practical difficulties and the cost of the operation
will be very great; but, on the other hand, the descent to a
temperature within 5 degrees of the zero would open out new
vistas of scientific inquiry, which would add immensely to our

Chemistry and Physics. 311

knowledge of the properties of matter. To command in our
laboratories a temperature which would be equivalent to that
which a comet might reach at an infinite distance from the sun
would indeed be a great triumph for science. If the present
Royal Institution attack on helium should fail. then we must
ultimately succeed by adopting a process based on the mechanical
production of cold through the performance of external work.
When a turbine can be worked by compressed helium, the whole
of the mechanism and circuits being kept surrounded with liquid
hydrogen, then we need hardly doubt that the liquefaction will
be effected. In all probability gases other than helium will be
discovered of greater volatility than hydrogen. It was at the
British Association Meeting in 1896 that I made the first sug-
gestion of the probable existence of an unknown element which
would be found to fill up the gap between argon and helium, and
this anticipation was soon taken up by others and ultimately con-
firmed. Later, in the Bakerian Lecture for 1901, I was led to
infer that another member of the helium group might exist having
the atomic weight about 2, and this would give us a gas still more
volatile, with which the absolute zero might be still more nearly
approached. It is to be hoped that some such element or elements
may yet be isolated and identified as coronium or nebulium. If
amongst the unknown gases possessing a very low critical point
some have a high critical pressure instead of a low one, which
ordinary experience would lead us to anticipate, then such diffi-
cultly liquefiable gases would produce fluids having different phys-
ical properties from any of those with which we are acquainted.
Again, gases may exist having smaller atomic weights and densi-
ties than hydrogen, yet all such gases must, according to our
present views of the gaseous state, be capable of liquefaction
before the zero of temperature is reached. ‘The chemists of the
future will find ample scope for investigation within the appar-
ently limited range of temperature which separates solid hydrogen
from the zero. Indeed, great as is the sentimental interest
attached to the liquefaction of these refractory gases, the import-
ance of the achievement lies rather in the fact that it opens out
new fields of research and enormously widens the horizon of
physical science, enabling the natural philosopher to study the
properties and behavior of matter under entirely novel conditions.
This department of inquiry is as yet only in its infancy, but
speedy and extensive developments may be looked for, since
within recent years several special cryogenic laborator! ies have
been established for the prosecution of such researches, and a
liquid-air plant is becoming a common adjunct to the equipment
of the ordinary laboratory.— Nature, Sept. 11, 1902.

6. Vapor-pressures of Liquid Oxygen and of Liguid Hydro-
gen.—In an abstract of a paper by M. W. Travers, G. SENTER
and A. JAQuEROD read before the Royal Society on June 19
(Nature, vol. lxvi, p. 382), the results contained in the following
table are given :

312 Scientific Intelligence.

Vapor Pressures.

LIQUID OXYGEN. Liquip HyDROGEN.
Pressure Temp. on Temp. on Temp. on Temp. on
inmm. hydrogen scale. helium scale. hydrogen scale. helium scale.
fe) fo) le) (e}
800 90°60 90°70 20°41 20°60
760 90°10 90°20 20°22 20°41
700 89°33 89°43 19°93 20°12
600 87°91 88°01 19°41 O56:
500 86°29 86°39 18°82 19°03
400 84°39 84°49 LS ks 18°35
300 82°09 BZ k8 17°36 Lad
200 ROZO7 G17 16°37 16°57
100 fs Bele 14°93 15°13
50 ee OF oes She 14°11

The authors add, in conclusion :

“Though the pressure coefficients of hydrogen and helium
between 0° and 100° C. show no appreciable difference, measure-
ments of low temperatures on the scales of the two thermometers
are not identical. It is probable that at the normal temperature
both gases may be considered as so nearly perfect that the differ-
ence between the gas scale and the absolute scale is insignificant.
As the critical point of helium les much lower than that of
hydrogen, measurements of low temperatures on the helium scale
should approach more closely to absolute temperatures than mea-
surements on the hydrogen scale. It is pointed out that helium
should replace hydrogen as the normal thermometric substance.

The melting point of hydrogen was found to be 14°'10 on the
helium scale.

The pure helium used in the thermometric measurements was
obtained by passing purified cleveite gas through a coil cooled
to 15° in liquid hydrogen boiling iz vacuo. An unsuccessful
attempt was made to liquefy this gas, which could not be con-
densed at 13° under a pressure of 60 atmospheres.

The vapor pressures of solid neon were measured at tempera-
tures corresponding to 20°:4 (12°8™™) and 15°°65 (2°4™™). It was
shown that the vapor pressure did not change as the solid evapo-
rated, proving that neon is a homogeneous substance.”

Il. GEoLogcgy AND MINERALOGY.

1. Martinique-—The July issue of the National Geographic
Magazine is called the ‘‘ Martinique number,” and is devoted to
descriptions and discussions of the volcanic eruption of Mt. Pelée
which began in May last, together with some references to the
contemporaneous eruption of La Soufriére on the island of St.
Vincent. The number is of double the usual size and is a valua-
ble contribution to the study of the recent eruptions in the
Lesser Antilles, containing as it does the reports of Robert T.
Hill, of the United States Geological Survey, and Israel C.
Russell, of the University of Michigan, two of the representa-

Geology and Mineralogy. 3813

tives commissioned by the National Geographic Society to inves-
tigate the phenomena attending the cataclysms. The number is
enriched by the notes by J. S. Diller on the volcanic rocks collected
by Hill and Russell; by a chemical discussion by W. F. Hille-
brand of analyses of ejecta from the two islands, and by a com-
pilation by James Page of the reports of vessels as to the area over
which the dust from the eruptions was distributed. The number is
well illustrated with four maps and nineteen photographs. Four-
teen of the latter are from excellent negatives taken by Russell.

Hill’s report opens with a brief general statement of the geog-
raphy and geology of the whole chain of the Windward Islands
or Lesser Antilles, which are almost entirely volcanic in charac-
ter, with the exception of Barbados. He holds that the volcanoes
of these islands date back to Eocene time, at least; that the
greater masses of the present volcanic heights were piled up
before the Pliocene, and that “the present craters are merely
secondary and expiring phenomena.” In discussing the present
eruption, after relating the premonitory events of the preceding
fortnight, Hill gives in detail the history of the great outburst of
May 8 as derived from the accounts of eyewitnesses, and then
elaborates his personal observations, which were made between
May 21 and 30. He holds that the eruption which destroyed St.
Pierre came from a crater low down on the southwest slope of
Mt. Pelée, two miles from the northern limit of the city and a
mile and a half from the sea, which he calls the “Soufriére” or
“ tang Sec” crater. He considers the mud-flows to be primary
phenomena of the eruption and to have come from this crater
and from several “ mud-craters” on the slopes of the mountain,
the sites of some of which had long been known as thermal
springs.

Hill calculates that only 12°5 square miles, or one-twentieth of
the total area of Martinique, have been seriously affected by the
eruption. No great geological changes have been brought about,
but the configuration of the summit and the outline of the sea-
coast have been changed somewhat. He says, ‘‘ There have been
no lava flows whatsoever... . No true bombs have been ejected
or molten rock in any form.” His conclusion is that conflagra-
tion and death in St. Pierre may ultimately be explained by either
of two theories :

(1.) The heat-blast theory. This hypothesis assumes that the
lapilli, gases, and steam of the ejected cloud were sufficiently hot
to have inflamed the city and destroyed the people by singeing,
suffocation, and asphyxiation. It does not account for the forces
exerted radially and horizontally, nor for the flame.

(2.) The aérial-explosion theory. The explosion of gases within
the erupted cloud after their projection into the air would account
for all the phenomena observed.

Russell’s report takes the form of a letter to the Society. He
does not subscribe to the opinion that the inhabitants of St. Pierre
were asphyxiated by noxious gases or killed by a gas explosion.

314 Scientific Intelligence.

He holds that the general cause of death and destruction was a
blast of steam charged with hot dust, which passed through the
city with hurricane force, and that gases, probably in part inflam-
mable, were present, but played only a secondary part in the dis-
aster.

Most of Russell’s report is devoted to St. Vincent. He holds
that the destruction on this island was due to dust, lapilli and
stones which fell on the land while yet hot ; but that a hurricane
blast of steam charged with burning dust did not sweep down
from La Soufriére as it did from Mt. Pelée. Thearea of devasta-
tion on St. Vincent was much greater than on Martinique. Atten-
tion is called to the violent secondary or superficial eruptions due
to rain or river water coming into contact with the still heated
interior of the great deposits of recent ash in the gorges of the
Wallibou and Dry Rivers; to the pulsating flow of the Wallibou
river due to overloading by voleanic sand; to the canyon-like
dendritic drainage forms already produced in the coating of fresh
dust and lapilli by the rains ; and to the fact that houses standing
on the windward (east) slope of the volcano had suffered most
on the sides farthest from the crater. :

Five miles from the crater Russell found the level fields coated
with a layer of new volcanic débris about two feet thick. This
would be a minimum measure and the average thickness of the
deposit would be several times as great. The greater part of the
débris consists of gray scoriaceous andesite and came from the
columns of fresh lava that rose in the conduit of La Soufriére,
This material was sufficiently cooled to become solid before it was
blown into the air, and to a great extent was reduced to dust by
the sudden expansion of the steam it contained. In addition to
the fragments of fresh lava the fields were strewn with angular
masses of older and much more compact lavas torn from the
throat of the volcano.

Diller describes the older lavas of Martinique, collected by
Hill, as being hypersthene andesite, hornblende-hypersthene ande-
site, hornblende andesite and dacite (quartz andesite), and the
products of the May eruption of Mt. Pelée as belonging to the
hypersthene-andesite class. He says that the specimens from St.
Vincent are of hypersthene andesites, remarkable for their con-
tent of olivine.

Hillebrand concludes that, while the ejecta from the two vol-
canoes are of the same general type, and while the material from
the same vent may vary in composition within limits, according
as it is collected near to or far from the vent, and in coherent or
finely divided form, yet there are characteristic differences by
which it appears easy to distinguish the product of one volcano
from that of the other. Diller’s report is accompanied by seven
analyses of pumice, sand and dust from these eruptions and
Hillebrand’s discussion by three such analyses. E. O. H.

2. The Report of the Geological Survey of Louisiana: G. D.
Harris, A. C. Veatcu, J. Pacneco.—The papers contained in

Geology and Mineralogy. 315

this report are of especial economic value to this and neighbor-
ing States, including as they do reports on the “Salines of North
Louisiana,” A. C. Veatch, ‘Subterranean Waters of Louisiana ”
and ‘Oil in Louisiana,” by G. D. Harris.

The map of the Mississippi Embayment shows the Jackson
Stage for the first time with its most probable northern distribu-
tion. The statement with regard to the close of the Cretaceous
is of interest at this time because of its importance to the oil
industry: “In Louisiana we have reason to believe that the rais-
ing and depression of the Cretaceous beds was of a much more
violent nature [than farther north], that folds and faults were
numerous and on a large scale and a great irregularity of surface
feature characterized the newly formed rocks.” Professor Harris
differs from R. T. Hill and others who believe that there are no
“structural complications” in this formation.

Doubt is thrown on the usually accepted theory concerning
the origin of the “Mud Lumps” of the Mississippi. No new
theory is advanced, but the statement is made that, “that they
rise in domes or anticlines and preserve their regular bedding is
proven by their structure.” “So far as we observed, none were
formed as volcano-like cones.”

The report on the Salines of North Louisiana, A. C. Veatch,
contains:a history of the early operations of the various salt wells
and licks together with well sections and a discussion of the
geology. It is shown that “the principal brine springs are to
be regarded as Cretaceous outcrops.” ‘The dome structure of
these and other north Louisiana Cretaceous outcrops is accepted.
The maps of this report and of Reports III and IV are excellent.

The Reports on the Geography and Geology of the Sabine
and Ouachita Rivers contain discussions and explanations of the
‘“‘Jandslip islands” of the Ouachita River and the shoals of the
Sabine River which are of especial interest to physiographers.

One of the most valuable reports, economically, is that on the
Subterranean Waters of Louisiana.

It is to be regretted that a volume containing so much valuable
matter and so well illustrated by maps and half-tones is not
printed on better paper and bound more substantially.

3. Die Alpen im Hiszeitalter ; von ALBrecut PrENcK und
Epuarp Brickner. Parts 1 and 2. 224 pp., with many charts
and figures. Leipzig. (Tauchnitz.)\—After fourteen years of
investigation Professors Penck and Briickner have begun the
publication of their views regarding glaciation in the Alps. The
completed work will consist of about six parts, dealing with the
entire Alpine system throughout the Pleistocene with especial
reference to the topography resulting from ice action. In the
two parts already issued Professor Penck discusses the general
character of glaciers and of glacial deposits, and explains with
great detail the glacial phenomena of the northern and eastern
Alpine border region, particularly the area between the Iller and
the Lech.

316 Scientific Intelligence.

4. Bacubirito, the Great Meteorite of Sinaloa, Mexico.—An
interesting account is given in vol. iv (pp. 67-74) of the Proceed-
ings of the Rochester Academy of Science, by Henry A. Warp,
of a visit to the gigantic mass of meteoric iron discovered in
1876 in the State of Sinaloa, Mexico. This is perhaps the largest
of all known meteorites ; its weight is roughly estimated at 50
tons, and its only rival is the meteorite of Anighito, Greenland,
to which a recent estimate has assigned a weight of 464 tons.
The three meteorites most nearly approaching these are those of
Bemdego, Brazil, 53 tons; of San Gregorio, Mexico, 114 tons,
and Chupaderos, Mexico, 152 tons. The dimensions of the Bacu-
birito meteorite were measured, after extensive excavations of
the soil in which it was nearly imbedded, as follows: length, 13
feet 1 inch; width 6 feet 2 inches; thickness, 5 feet 4 inches.
The shape is extremely irregular, being compared to that of a
ham; the cubic contents consequently could be only approxi-
mately estimated. In composition, an analysis by Whitfield has
shown it to contain about 7 per cent nickel, and its structure is
eminently octahedral. The exterior shows little oxidation and
the pittings are clean and sharp in outline. It is to be hoped
that the Mexican Government, which with admirable liberality
has already transferred to the School of Mines, Mexico, five of
her largest meteorites, may also undertake the similar protec-
tion of this unique mass.

5. The Origin of Eskers; by W. O. Crossy. Baston Soc.
Nat. Hist., vol. 30, No. 3, pp. 375-411.—It is an accepted
truth among glacialists that eskers are formed either by super-
glacial or subglacial streams, and the large majority of geologists
believe they are of subglacial origin. Professor Crosby has
reéxamined the whole subject and concludes that streams flowing
upon the ice sheet rather than those flowing under it are respon-
sible for most of the esker ridges.

6. Ueber Hussakite (Xenotim) und einige andere seltene
gesteinbildende Mineralien ; by H. Roster. Zeitschr. f. Kryst.,
XXXvl, pp. 258-267.—Prismatic xenotime (hussakite) has been
identified in the heavy residues of a large number (52) of
European rocks and kaolins amongst the crystals usually referred
to zircon, from which it was distinguished in part by the applica-
tion of the magnesium test for phosphoric acid ; in part by the
hepar reaction for sulphuric acid with soda on charcoal, but for
the most part by the difference in the strength of the double
refraction in microscopic preparations. Owing to the elusive
nature of all these tests when applied to microscopic crystals in
microscopic quantities, it seems desirable to carefully verify these
identifications before conclusions of such far-reaching conse-
quences be definitely accepted. |

Contrary to the observations of Derby, who only identified
prismatic xenotime in a single case in scores of residues examined
and who found the octahedral forms almost invariably accom-
panied by monazite, the author finds, in the rocks examined, the

Miscellaneous Intelligence. ; 317

prismatic forms of xenotime more abundant and widespread than
the octahedral one and than monazite as well.

In addition to zircon, xenotime and monazite the author found
anatase, regarded as secondary, in about half of the residues
examined ; dumortierite in two kaolins from 2-mica granite ;
chrysoberyl in seven kaolins (6 2-mica granites and 1 syenite)
and one fresh 2-mica granite; staurolite in two kaolins from
aplite, and andalusite in five kaolins from 2-mica granites.

III. MisceLLANEOUS SCIENTIFIC INTELLIGENCE.

1. International Catalogue of Scientific Literature. First
Annual Issue. M. Borany, Part I, May, 1902. Pp. xiv, 378.
D. Cuemistry, Part I, June, 1902. Pp. xiv, 768.—These two
volumes of the first annual issue of the International Catalogue
of Scientific Literature, commencing with the year 1901, have
recently appeared. This Catalogue is published for the Inter-
national Council by the Royal Society of London, being an out-
growth of the well-known Catalogue of Scientific Papers relating
to the scientific literature of the Nineteenth Century, published
by the same Society. In the form which the plan has finally
taken, each complete annual issue of the Catalogue will consist of
seventeen volumes, one for each of the sections of science sepa-
rately recognized, The set will be sold to the public for £18,
individual volumes costing, according to size, from ten to thirty-
five shillings. The director is Dr. H. Forster Morley, whose
address is at the Central Bureau, 34 and 35 Southampton st.,
Strand, London, W.C. ‘Twenty-nine Regional Bureaus have
been arranged for, which are to furnish the material for the Cata-
logue ; for the United States, communications are to be sent to
Prof. 8. P. Langley, Smithsonian Institution, Washington.

The method of classification and indeed all the details of the
entire plan have been very carefully worked out, but only the
outline can be given here. Each of the volumes issued consists
of three parts: (a) Schedules and indexes in four languages,
English, French, German and Italian; (0) an author’s catalogue ;
(c) a subject catalogue. The subject catalogue is divided into
sections, each of which is denoted by a four-figure number between
0000 and 9999. ‘These numbers follow one another in order but
all are not used, space being left for such additions to the system
of classification as may be found necessary in future years. In
the case of the two volumes now published, the material for 1901
being yet incomplete, the first part will be followed by a second
part in a few months; in future, however, when the organization
is complete, it 1s planned to issue a single annual volume only
for each subject.

The breadth and completeness of this great scheme for putting
within the reach of every worker in science a catalogue of all
original contributions on the subject in which he is interested, in
whatever form or place published, is worthy of the century with

Am. Journ. Sct.—Fourta Series, Vou. XIV, No. 82.—Octoprr, 1902.
22

318 Scientific Intelligence.

which it commences and the Society under whose auspices it is
published. It is to be hoped that the undertaking may not only
receive all needed financial support, but also that editors and
authors may inform themselves minutely as to the details of the
plan, so that they may prepare subject-indexes for all papers at
the time of their publication, thus diminishing very largely the

subsequent labors of the collaborators. ,

2. British Association.—The annual meeting of the British Asso-
ciation for the Advancement of Science was held at Belfast dur-
ing the week beginning September 10th. The meeting is de-
scribed as having been very successful, about equal in numbers
to that held in the same place in 1874, when the attendance was
1,951. The address of the President, Professor James Dewar,
was of the highest interest (see Nature of Sept. 11), dealing, in
addition to some general topics, with the subject of low tempera-
ture and the liquefaction of gases, to which he has personally
contributed so much. We quote at length from the address in
another place.

3. Haperiments in Aerodynamics ; by 8. P. Lanatry. Second
edition, pp. 115 with ix plates. Washington, 1902. (Smith-
sonian Contributions to Knowledge, No. 801.)—This volume is
essentially a reproduction of the first edition of this interesting
memoir, which was issued in 1901 and noticed in this Journal in
November of that year (vol. xlii, p. 427). Since its first publi-
cation, machines made on the principles here described have
actually flown, which fact gives a practical interest to the sub-
ject; a description of the heavy steel flying machines is promised
for some time in the future.

4. Elementary Physical Geography ; by Witt1am Morris
Davis. Pp. x, 401, with numerous maps and illustrations (Ginn
and Company.)—It is a pleasure to note the publicat.on of a
text-book which presents a subject in an interesting elementary
form without the loss of scientific accuracy and a rational mode
of treatment. Professor Davis’s new text reduced from his
«Physical Geography ” is just such a book.

5. The International Quarterly.—The International Monthly,
which since January, 1900 has had a highly successful career
under the editorship of Mr. Frederick A. Richardson (Burling-
ton, Vt.), will hereafter be continued as a quarterly with the
same editorial direction. Each number will be more than double
the size of the monthly, and will present articles of live interest
by able writers at home and abroad. ‘The September issue (pp.
1-214) of The International Quarterly begins vol. vi, of the
series.

OBITUARY.

Proressor RupotF Vircuow, the illustrious German anato-
mist, physiologist and anthropologist, the founder of cellular
pathology, died at Berlin on September 5, in his eighty-first year.

Plate VIII.

XIV, 1902.

iz Vol;

Se

Am. Jour.

NN

‘48800 SUOT[V SepI[spuv] [BOOT
‘ALVIS GNOWHORT WOWd ‘LNHONTA “LG ‘aUgIuaAOg wT *1O[ABT, “WT ‘O AQ ydvisojoug

Pe Ei

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES.]

Art. XX XIII.— Observations on the Evuptions of 1902 of La
Soufriere, St. Vincent, and Mt. Pelee, Martinique ;* by
-Epmunp Oris Hovey.

THE chain of islands bounding the Caribbean Sea on the
east, and known as the Leeward and Windward Islands, the
Lesser Antilles or the Caribbean Islands, are almost wholly of
voleanic origin, the most important exception to this rule
being Barbados. From Grenada northward the chain of vol-
canoes extends in a grand curve for about five hundred miles,
with its convex side toward the east, indicating a line of weak-
ness in the earth’s crust comparable with those which are out-
lined by the festoons of volcanoes along the northern and
western coasts of the Pacific Ocean. The volcanic nature of
these islands has been known ever since they first were explored,
but only a few eruptions have been recorded during the past
four centuries—the most important being those of La Soufriere
on St. Vincent in 1718 and 1812, with Mt. Pelée on Martinique
in slight eruption in 1851. Continuous solfataric action, which
sometimes has been quite violent, is known in the crater of Mt.
Misery on St. Kitts, the “ Soufrieres” of Guadeloupe and St.
Lucia and the Boiling Lake of Dominica. Hot springs have
been a feature of several localities on Martinique, St. Vincent
and other of the islands. :

La SouFRIERE, St. VINCENT.

At least as early as April, 1901, earthquakes became more
frequent and noticeable than usual in St. Vincent, but, with

* The author was sent to the islands of Martinique and St. Vincent as the
representative of the American Museum of Natural History, New York, to
study the phenomena connected with the recent eruptions. The article here
published is a condensation of the ‘‘ Preliminary Report” prepared for the
Museum authorities and published in the Bulletin of the Museum, vol. xvi,
pp. 3383-372, pls. xxxili-li. The author’s field work on the islands covered
the period from May 21 to July 6, inclusive, and his Report and this article
pertain almost solely to the personal observations made during that time.

Am. Jour. Sc1.—FourtH Series, Vou. XIV, No. 83.—NovemBeEr, 1902.
23

320 FE. O. Hovey—Eruptions of 1902 of La Soufriere,

oe

Petit Balejnéi4
DuVal lgs Pt( OX
7 BS

9 1 Rabaka
Sepa eal) Gy R j ey SOK, 4 Dry River.
| Le

Chateaubelait\

SS) Georgetown.
Y/

f May 27
Zax June 3m 10

5 English Miles.
Fic. 1. Map or tHe Isuanp or Sr. Vincent, B. W. I.

The cross-lined area shows approximately the portion of the island which
was devastated by the May, 1902, eruptions of La Soufriére. The violent
outburst of September 3-4 deposited much additional material on the lee-
ward (west) side, and extended the zone of devastation about four miles
south of the boundary here indicated for that side of the island. The dot-
and-dash lines show the principal routes traversed by the author while on
and near the voleano.

St. Vincent, and Mt. Pelée, Martinique. 321

the exception of Mr. F. W. Griffith, of Kingstown, no one
seems to have connected them with an impending eruption of
the Sonfriére. In December of that year, however, the people
living on the western slopes of the voleano began to feel
anxious on account of the subterranean noises heard occa-
sionally. By February, 1902, the rumblings had become so
frequent that the inhabitants were very uneasy and began to
leave the district, so that but one person was left within the
fated area when the great eruption of May 7 occurred. The
rumblings were less distinctly heard on the eastern or wind-
ward side of the island, and the warnings were not heeded so
generally, because it was: supposed that the prevailing winds
would carry to westward any ejecta from the crater, in case of
an eruption. The loss of life was confined to the windward
side of St. Vincent. ;
The first ascent of the Soufriére, since the eruption of May
7, 1902, was made on Saturday, May 31, by Dr. T. A. Jaggar,
Jr., George Carroll Curtis, T. MacGregor MacDonald* and
myself with six porters. We went up from the site of Walli-
bon village, on the leeward (west) side, following the remains
of the old trail to the rim of the crater at 2790 feet above the
sea, an elevation obtained by taking the mean of the readings
of three aneroid barometers.t We found the crater unchanged
in diameter, as nearly as Mr. MacDonald could tell, and there-
fore to be about nine-tenths of a mile in diameter from east to
west and eight-tenths of a mile from north to south, judging
from measurements made on the map. The beautiful crater
lake, for which the Soufriére was famous before the eruption,
had disappeared of course, but there was a small Jake of boil-
ing water in the bottom of the pit, from the southeastern
quarter of which steam was ascending in a strong column.
See figs. 4 and 5, p. 351. This column at intervals was carry-
ing up quantities of black sand with it to moderate heights
above the bottom of the crater. We estimated the surface of
the boiling lake to be about 1600 feet below the point on

* Mr. MacDonald is the owner, with his brother, of several estates on the
leeward side of St. Vincent. One of these was destroyed in the May erup-
tions, and three others suffered the loss of their growing crops through the
outburst of September 3. Mr. MacDonald had the presence of mind to
remain in one of his honses, the Richmond Vale estate, from which there
was an uninterrupted view of the npper portion of La Soufriére, and to take
notes in detail of what happened on May 7 up to the time of the great erup-
tion which took place at 2 Pp. M., when he fled for his life. His notes have
been published in full in the Kingstown Sentry of May 16, 1902, and in the

- Century Magazine, vol. lxiv, pp. 638-642, August, 1902.

+ All the altitudes recorded in this article were obtained by means of
aneroid barometers, except as otherwise stated in the text.

{ For convenience of printing, the views iliustrating this article have been
pJaced together on pp. 351 to 358.

322 E. O. Hovey—Eruptions of 1902 of La Soufriére,

which we were standing, and 2400 feet below the highest point
of the rim. The lake seemed to be shallow, judging from
some nearly flat ground in the bottom of the crater northeast
_of the water. The surface of the old crater lake was 1930
feet (chart) above tide. Its depth in the center was 874
fathoms, according to the statements of P. F. Huggins, engi-
neer, of Kingstown, St. Vincent, who told me that he sounded
it in 1896. His line was too short to reach bottom in the
northwestern part of the lake.

Almost directly opposite the point where we first reached
the rim was the saddle between the “Old” crater and the
crater of 1812, apparently unbroken by the eruption. From
the lower third of this nearly vertical rock-face there issued a
strong stream of water which cascaded down the precipices
and flowed across a rather narrow strip of nearly level ground
in the bottom of the crater and emptied into the boiling lake.
It seemed as if this stream must be the discharge of the waters
now collecting in the crater of 1812, itself the possessor of a
little lake before the eruption of the present year. The
western side of the crater rim showed a gash on its western
side, leading into the Larakai Valley, but the bottom of the
gash was more than a thousand feet above the bottom of the
crater. Mr. MacDonald said that the gash was there before
the eruption took place, but that it seemed to him to have
increased in size since the outbursts began. The gash is very
much smaller than that in the southwest side of Mt. Pelée, and
it does not seem to have had any appreciable, or, better, any
determinable, effect in concentrating the force of Soufriére’s
voleanic hurricanes. Tremendous avalanches of rocks and
earth descended the inner precipitous slopes of the crater at
intervals during our stay on the rim. They made a great deal
of noise, and probably occasioned some of the “groaning” of
the volcano reported by the islanders.

On June 4 Jaggar, Curtis and I made an attempt at the
ascent from the windward side. We reached the altitude of
3200 feet, but turned back without getting to the crater itself,
on account of dense trade-wind clouds. On June 9 Curtis and
I made our third ascent, alone, except for one guide, and
reached the rim of the crater on the southeastern side two or
three hundred yards beyond the spot at which we had turned
back on the preceding occasion. For fifteen or twenty yards
back from the edge of the rim in the ground there were crevices
many yards long and up to three inches wide, which formed
lenses with the edge itself and indicated the imminence of
landslides into the crater. We pushed along the rim north-
ward, until, at an altitude of 3550 feet above the sea, we stood -
between the large crater and the crater of 1812. The summit

St. Vincent, and Mt. Pelée, Martinique. 328

of the Soufriére east of the large crater and south of the small
one is formed bya rather small plateau which slopes gently
toward the southeast, closely analogous in position to the small
high plateau on Mt. Pelée. This platean was covered with a
bed of dust, Japilli and bowlders which was ten and fifteen
feet thick in places, and the trenches cut by recent rains made
traveling very laborious, except near the edge of the crater.

In spite of clouds and rain, this visit, through occasional
glimpses of the interior, enabled me to determine that the
crater of 1812, which for nearly a century has gone by the
name of the “ New” crater, took no active part in the erup-
tions of May of the present year, a conclusion based on the
following considerations: the saddle between the two craters
appeared to be intact, confirming the observations made from
the other side of the large crater; a knife-edge ridge which
ran at a steep incline from the saddle to the bottom of the
small crater and formed the pathway for descent into it before
the eruption, was still there, and had on its slopes bare trunks
of trees standing; in the bottom of the crater along the base
of this ridge one could see talus slopes of dry (?) dust and
Japilli which had slid and rolled down its sides; although the
roaring of the steam and boiling water nearly half a mile
below us in the large crater was obtrusively discernible, no
sound whatever came from within the crater of 1812; the rim
of the small crater showed less and less dust as one receded
from the edge of the great crater. Samuel Brown, a ranger,
or caretaker, on the Lot 14 estate on the southeast slopes of the
Soufriére, who was our guide when we reached the small
crater,, told us that he watched the eruption of May 7 until
the oreat outburst at two o’clock and that no cloud of steam
or “smoke” rose from the small crater. Furthermore, at the
time of my leaving the island, June 10, no column of steam had
risen above that crater since May 7. Brown was at the sugar
factory of the estate, three and one-half miles in a straight line
east-southeast from the crater, a most favorable spot from
which to observe what was going on at the summit of the
mountain, and he saved his life by running into the rum cellar
of the factory and closing the door and ‘the window shutters
just before the volcanic blast swept over the building. Inquiry
in Georgetown found persons who had watched the eruption
from the town and had noted the fact that no column of steam
rose from the small crater.

The Soufriére, and, in fact, the whole of the island of St.
Vincent, is made up of ancient lava flows alternating with vol-
canic fragmental deposits or tuff agglomerates.* These ancient

* The alternation of lava beds and tuffs is well illustrated in fig. 5, p. 351.

324 EL. O. Hovey—Eruptions of 1902 of La Soufrisre,

agelomerates show that there have been many eruptions of the
volcanoes of St. Vincent of the same character as that of 1902.
They contain bombs as well as blocks. The beds of solid rock
on the island show that many of the ancient eruptions were
accompanied by extensive flows of molten lava. The porous
agolomerates have suffered much from the decomposing action
of percolating waters, and the lava beds show extensive altera-
tion due to the same agency. Beautiful spheroidal weathering
is common in the basalts of the southeastern part of the island
and in the elevated beach conglomerates of the windward coast.

Although there are many ancient lava beds in the composi-
tion of the mountain, no stveam of melted lava has issued from
the Soufriére during the present eruption. The “ bread-crust ”
bombs, however, which occur plentifully on the mountain sides,
especially on the windward slopes, show that during the present
eruption molten lava has been present in the throat of the vol-
cano, and that many lumps of half-melted rock were thrown
into the air. Besides the bombs, the volcano ejected blocks of
ancient andesitic lava of several kinds and of varying degrees
of coarseness of grain, and of all sizes up to masses six or eight
feet across, and vast quantities of coarse and fine lapilli and
dust. Most, if not all, of the blocks were thrown out at very
high temperatures, as is shown by their cracked condition,
though they were not actually fused. Although a few bombs,
some of which were twelve to fifteen inches across, were found
on the leeward side as far away from the crater as the site of
Richmond village, three and one-half miles distant, by far the
largest number of both bombs and blocks, as well as the largest
specimens, were found on the windward side, bombs fifteen to
eighteen inches in diameter being common in the bed of the
Rabaka Dry River.

The area of devastation on St. Vincent is very large in pro-
portion to the total area of the island. After plotting it out
carefully on the British Admiralty chart and measuring the
area with a planimeter, I find it to be forty-six square miles,
practically one-third the entire area of the island. From
much of this devastated area, however, the ashes are being
washed off so rapidly by the rain that vegetation is already
asserting itself, and within another year crops will be growing
there again.*

Extensive landslides have taken place on the western side
(see Plate VIII), removing a strip of coast, in places one

* Newspaper reports and private advices from St. Vincent show that the
area of devastation has been extended on the leeward side of the island by
the tremendous eruption of September 3-4 about four miles south of the
boundary indicated on the map herewith presented (fig. 1), while the whole
western portion of the devastated area got a heavy additional coat of lapilli.
The windward side did not suffer materially from this eruption.

St. Vincent, and Mt. Pelée, Martinique. 325

hundred yards wide, continuously from the mouth of the
Wallibon River to Morne Ronde village, a mile and a half to
the north, and at intervals for two miles farther north. These
landslides have left precipitous walls along the shore-line, and
deep water is found where villages stood and prosperous plan-
tations existed before the eruption. We had no sounding line,
but our boatmen could not touch bottom with a twelve-foot oar
three feet from shore on the site of Morne Ronde village. The
sections left by the slides show that the land which has disap-
peared consisted of delta and coast-plain deposits, material
which would easily be shaken from the more substantial lava
flows and agglomerate beds by the vibrations due to the erup-
tions. The eastern, or windward, side of the island is not
nearly as steep as the leeward, and landslides have not occurred
there as features of this eruption. On the contrary, the wind-
ward shore-line from Black Point, a mile south of Georgetown,
northward almost to Chibarabou Point, more than six miles
distant, has been pushed out by the vast quantities of fresh
Japilli which have been brought down from the slopes of the
volcano by the rivers and the heavy rains, during and since the
eruptions, and distributed by the ocean currents.

A large amount of material, too, was brought down by the
Rabaka Dry River an hour in advance of the great outburst of
May 7, which seems to have been due to the bodily discharge
of a portion, at least, of the old crater lake into the headwaters
of that stream. Survivors who attempted to cross the Rabaka
Dry River toward noon of that day report that they were
prevented by a torrent of “boiling hot” water and mud rush-
ing down the valley, and that.a wall of water and mud fifty
or more feet high (they compared it with the height of a fac-
tory chimney) came out of the upper reaches of the river and
swept out to sea. There was no heavy rain that day before
the eruption took place, but the lake still was in the crater
early in the day, according to the tale of a fish-woman who
had ascended the mountain from Gecrgetown that morning
on her way home to Chateaubelair. The trail led along the
rim of the crater for half a mile. The woman reached the rim
at nine o’clock and found that fissures had appeared in the
ground and that the lake was at a higher level than usual and
boiling. She rushed back to Georgetown to warn the people,
but her tale was discredited. Mr. MacDonald’s notes contain
the entries: “12.55 p.m. Enormous discharge to windward
side, color darker. 1 P. M. Tremendous roaring, stones thrown
out to windward thousands of feet.”* While this does not
prove the bodily outthrow of the lake, it shows that there was

* Century Magazine, vol. lxiv, p. 630, August, 1902.

326 EL. O. Hovey—Eruptions of 1902 of La Soufriére,

a great outburst from the crater just in advance of the flood in
the Dry River Valley.

It is evident that there was a blast or a series of blasts of
hurricane violence from the crater of the Soufriere as well as
from that of Mt. Pelée, as a feature of the eruptions of 1902.
The effects were not so appalling, however, on St. Vincent as
on Martinique, because no large city was destroyed there.
The overturned trees constitute the principal evidence on the
island of St. Vincent. They all point away from the crater,
except for slight modifications due to local topography (see
fig. 10). The blast extended radially in all directions from
the crater, suggesting the explanation that some great vol-
ume of steam, rising from the throat of the volcano, could
not find room for expansion upward, on account of the column
of steam and ashes which had preceded it, and the ashes fall-
ing therefrom, and that it expanded with explosive violence
horizontally and downward, following the configuration of the
mountain. This accords with the testimony of Mr. MacDonald
and other eye-witnesses of the eruptions, who say that they
saw the clouds of “smoke” (dnst-laden steam) rushing down
the sides of the mountain with terrific speed. This dust-laden
steam was able to do much work of erosion, as is shown by
the horizontally scoured sides of some of the exposed cliffs and
by the trunks and roots of trees. The roots particularly have
been charred by the heat and have been carved into fantastic,
pointed shapes, as if they had been subjected to the action of a
powerful sand-blast. Erosion has not materially affected the
original surface of the ground as yet, because almost every-
where one can find the living roots and charred blades of
grass and other vegetation beneath the covering of dust and
lapilli, the first of which acted as a protection against the heat
of the rest. Now, however, the heavy rains “take up vast
quantities of the loose lapilli for use as a powerful scouring
agent in attacking the denuded hillsides, and thus old valleys
are being deepened and widened.

The particular feature of the May eruptions of the Soufriere
was the enormous amount of dust* which was thrown into the
air and distributed over a vast circle or ellipse the area of
which cannot yet be calculated for lack of data. The British
steamship Coya had an eighth of an inch of volcanic dust

* The following chemical analysis is of dust from the May eruptions
which I collected May 27 in a room in the Langley Park estate house, ahout
one mile north of Georgetown, St. Vincent, in which twenty-one dead bodies
were found after the eruption of May 7. The analysis was made by Dr. W.
F. Hillebrand of the United States Geological Survey, to whom my acknowl-
edgments are due, and is the unpublished analysis referred to in his article
in the National Geographic Magazine for July (vol, xiii, p. 297) as empha-
sizing the greater amount of sulphur present in the ejecta of La Soufriére

St. Vincent, and Mt. Pelée, Martinique. = — 327

from this voleano fall on her deck when she was two hundred and
seventy-five miles east-southeast of St. Vincent. The steamer
encountered the dust at 10.30 Pp. m., May 7, eight and one-half
hours after the eruption of the Sonfriere began, indicating trans-
port against the prevailing surface wind at more than thirty-two
knots per hour. Reports of vessels from the west (leeward)
of the island are curiously lacking. The dust was spread like
a gray mantle over the island, generally diminishing in thick-
ness from the crater outwards, but collected in vast deposits in
certain valleys on the sides of the mountain, where the condi-
tions seem to have been particularly favorable. The chief of
these beds were formed in the Wallibou, Trespé and Rozeau
valleys on the leeward side, and in the valleys of the Rabaka
Dry River and its tributaries on the windward slope, with by
far the greatest thickness along the Wallibou and Rabaka Dry
Rivers. In the valley of the Wallibou the deposits were not
less than sixty feet deep in places, while in the Rabaka Dry
River the fresh material filled a gorge which is said to have
been two hundred feet deep before the eruptions began (see
fig. 9). From a distance this deposit looks as if it were a
glacier coming out of the mountains.

Such great aceumulations of hot Japilli and dust retain pee
heat for a long time and they have given rise to secondary, or
superficial, eruption phenomena of striking character and con-
siderable interest. The river water and the water from the

than in those of Mt. Pelée. The absence of chlorine is interesting as indi-
cating fresh waters as the source of the steam of the eruptions.

<i: eg ee 55°08

Al.O; ee a es Sa a 18-00

i. oe 2°46 ) (Only approximate, because of effect
Jf een 4-575 of pyrrhotite, 0°91 ¢—see below.)
Li. i. 2 eee 3°34

Li oe ee 7°74

Na,.O ee eS ey eee eS 3°45

ww ES ee 0°65

eth Hn ee 0°66

HO above 100° °C. _____- 1°39

TiO. pepcnoee sg: Sexes. ($0

ZrO2 re ay ee en ee” Us a 4

2S eee none

PO: Lit a 2 (oe ep, a ee 0°17

SO; ee ee es ee Se 0:24

in oe eee eee none or faint trace

5 J. eae (0°36) included in pyrrhotite below
Lo ee none

wih EG asi area eee ree 0-21

22 eR eck ss een trace
SE ee eS ees none

“SS eee aa OS pegs faint trace

Fe-Ss Oye. ee gga hs & 0°91

99°67

328 L. O. Hovey—Eruptions of 1902 of La Soufriére,

tropical showers percolating through the beds have come into
contact with the still highly heated interior, causing violent
outbursts of dust-laden steam. We saw one of these outbursts
from the Wallibou Valley send up a column of such vapor fully
a mile in height. The action lasted for nearly an hour. The
secondary eruptions illustrated by figures 7 and 8 took place on
a clear, dry morning and must have been caused by the perco-
lating river waters. On May 30 we witnessed the throwing of
a dam across the stream and the formation of a temporary lake
by a heavy secondary outburst of dust-laden steam from the
lapilli-bed in the Wallibou Valley. This eruption is illustrated

in fig. 7. After the eruption ceased the little lake soon rose:

to the top of the dam and quickly cut its way down to the
old level, sending a “mud-flow” down the gorge to the sea.
Such a lake in the valley of the Rabaka Dry River cut its new
outlet through a narrow ridge of the old agglomerate constitut-
ing the wall of the canyon, forming as it did so a beautiful series
of channel-bowls, pot-holes and scratched corkscrew channels.

When we first reached St. Vincent, the dust, especially that
covering the Richmond estate, showed in marked manner the
wind-drift surface so familiar in the case of freshly-fallen snow,
and in many places these drifts were from three to four feet
deep (see fig. 6). There were several heavy rains between
May 24 and 29, so that the appearance of the surface was very
different on May 30 from what it was when I first saw it. Its
drifted character was not nearly as evident, and the beautiful
dendritic drainage, which was already in evidence on May 24,
had been greatly extended and intensified. Geological opera-

tions, which under ordinary conditions are performed so slowly —

as to be imperceptible, were being carried forward rapidly
under our very eyes. One item of interest was the action of
the Wallibou River itself under the influence of the loose dust
and lapilli along its banks.* Its waters became so overloaded

with sediment that they could only flow in pulsations, showing »

that intervals of time were needed by the stream to gather
strength to force its way along with its load. On May 24
these waves or pulsations were from fifteen to forty seconds
apart. Such mud streams carried large bowlders down the
river bed to the sea.
When the great cloud of ejecta rose from the Soufriére at
2 p.M., May 7, the portion which was traveling eastward
seemed suddenly to split, according to the accounts of eye-
- witnesses, when it was some distance beyond the island, and
to send a part back to the land. This accounts for the fact
that unprotected windows in the eastern side of houses in the

* First described by the author in the New York Times, June 29, 1902.

—— ss tS

St. Vincent, and Mt, Pelée, Martinique. 329 -

devastated district along the windward coast were all stripped
of their glass.

An official’s estimate of the loss of life on St. Vincent by
the eruption places the number of killed at 1350. The actual
number of bodies buried was 1298, including those of the
wounded who died in the hospitals. Almost all of the people
who passed through the fury of the eruption and escaped
uninjured had taken refuge in cellars the only openings into
which were on the side farthest from the crater and were,
moreover, tightly closed with wooden doors or shutters. The
most striking example of such protection was at Orange Hill,
on the windward coast, two and one-half miles north of George-
town, where one hundred and thirty-two persons were saved
unharmed in an empty rum cellar. This cellar, which is only
partly underground, is part of a sugar factory situated on a
rather flat divide between two ravines which may have tended
to separate the volcanic storm somewhat, though the roof of
the building over the cellar was demolished by the ejecta. The
only openings into the cellar were a door and two windows on
the side opposite the crater, and these were provided with
heavy wooden shutters which were kept closed during the fury
of the eruption. The experiences of the people in these cel-
lars suggest the great desirability of constructing similar places
of refuge for use in time of hurricane as well as of volcanic
eruption.

The deaths on St. Vincent seem to have been due, princi-
pally, to the following causes: (1) asphyxiation by hot, dust-
laden steam and air, (2) burns due to hot stones, Japilli and
dust, (8) blows by falling stones, (4) nervous shock, (5) burn-
ing by steam alone, and (6) strokes of lightning. The last
mentioned cause is perhaps somewhat doubtful, for though it
is very generally named by the survivors, there has been no
substantiation mentioned beyond the fact that there was a
great deal of extremely vivid lightning during the eruption.
The action of steam would account for burns received under-
neath the clothing where the clothing was not even charred.
Sulphur dioxide, SO,, and hydrogen sulphide, H,S, were
observed in troublesome quantities in the steam coming from
the crater, and it is more than probable that these gases, espe-
cially the former, added very materially to the deadly char-
acter of the dust-laden steam. Not an autopsy was made on
any of the hundreds of victims of the catastrophe, so that it
never can be known definitely what part was played by these
or other poisonous gases in the destruction of human life.

Under date of September 5, Wm. J. Durrant, druggist, of
Kingstown, St. Vincent, writes me that great volumes of

330 £. O, Hovey—Eruptions of 1902 of La Soufriere,

“smoke” and steam began rising from La Soufriére at 1 P. M.
of September 8, but that the violent outburst did not begin
until 9.30 that night. Three hours later the eruption was at
its height and the last explosion occurred at 5.40 A.M. The
roaring of the volcano from midnight onward was continuous
and was terrifying even at Kingstown, while the electric dis-
play about the great column of dust-laden steam surpassed
those of May 7 and 18. The matter ejected by this last erup-
tion is described by Mr. Durrant as being “a heavy, black
sand* of the coarseness of blasting powder, with pleaty of
pumice, but very few stones.” Very little light-gray ash like
that of the May eruptions fell this time. Richmond Vale
estate received about eight inches of ash, Chateaubelair about
six inches, Petit Bordel about four inches. Southward the
coat of ash diminished to Peter’s Hope, an estate on the west
coast about ten miles southwest of the.crater, where it ceased
to be of importance. The beginning of this eruption was a
mud-flow toward the site of Morne Ronde village.

Under date of September 26 Mr. T. MacGregor MacDonald
writes me that three men, two brothers and a cousin by the
name of Richards, visited the summit of La Soufriere August
19. From their accounts it was evident that relations within
the crater had not changed materially from what we found on
May 31. On September 17 Messrs. J. Adams and W. Cum-
mings made the ascent to see what changes had been wrought
by the tremendous eruption of September 3-4. They reported
that the crater was filled up to about the level of the surface

* A sample of the material thrown out by the eruption of the Soufriére
September 3 was received from Mr. Durrant September 25 and has been
examined under a hand-lens. It consists of fine and coarse volcanic sand
- and gravel, apparently for the most part comminuted ancient lavas of the
volcano. The fragments from 3 to 15 millimeters across show the coarsely
erystalline structure of the old lavas and many of them show that they are
parts of weathered masses. Olivine, pyrite (pyrrhotite ?) and porphyritic crys-
tals of feldspar, hypersthene and hornblende are abundant in these frag-
ments and the separated minerals make up a large proportion of the particles
about 2 millimeters in diameter. A comparatively large fragment (20™™ in
diameter) shows phenocrysts of feldspar imbedded in dark brown and light
brown scoriaceous glass which is apparently fresh. All the fragments and
the particles of sand are coated with dust which seems to be as fine as-any
that fell during the May eruptions, so that the explanation of Mr. Durrant’s
statement regarding the relative absence of fine dust from the ejecta of Sep-
tember 3-4 may be that the wind carried most of such material northward
and westward away from Kingstown, his point of observation. The cloud
from this eruption of La Soufriére is reported to have produced darkness for
about six hours on September 4 in Fort de France, Martinique.

The dust-coated sand is dark gray when dry, but is almost black when
wet, justifying the description quoted from Mr. Durrant’s letter. Compari-
son of this new material, however, with that collected by myself, May 23-
June 10, indicates that there is no essential difference between the ejecta of
the earlier and the later eruptions.

St. Vincent, and Mt. Pelée, Martinique. 331

of the old lake (the one existing prior to May 7) with ash, and
that there was a sloping surface from all sides down towards a
depression in the middle. The slope was such that from
almost any point. they could have descended to the bottom.
Steam was rising from several points, the most vigorous being
from the spot most active on May 31. Around the central
opening there was mud, and from the opening “fire” was
being thrown up every two or three minutes, which fell back
again, giving place to steam. There was an eruption during
the night of September 17, and Mr. Cummings made another
ascent on the 19th, when he found the crater cleared of ashes
to the depth observed May 31 and with a small amount of
water in the bottom, from and through which steam and “fire”
were rising. The “fire” mentioned by Mr. Cummings must
have been red-hot fragments of rock. A short violent out-
burst occurred about 6 P. M. September 21, lasting about ten
minutes in its most vigorous stages.

La MontaGne PEeL&e, MARTINIQUE.

The destruction of human life overshadows every other con-
sideration, in popular estimation, at least, when one speaks of
the eruption of Mt. Pelée, Martinique, which took place May
8, 1902. The sweeping of between twenty-five and thirty
thousand human beings out of existence almost in a moment
presents a holocaust with but few parallels in the history of
the world. The present eruptions of Pelée and the Soufriere
will not, however, take first rank among those which have
torn these and other Caribbean volcanoes, but they are exten-
sive enough and are of such a character as to merit the study
they have been and are receiving.

Mt. Pelée, like La Soufriere, gave warnings of the approach-
ing catastrophe, but they were not heeded by the inhabitants
of Martinique. The waters of the Lac des Palmistes became
very noticeably warmer than usual several months before the
eruption took place; rumblings were heard and steam began
to issue from the old crater some weeks before dangerous
activity began ; a fortnight before the first great eruption took’
place the earthshocks were sufficiently strong to displace dishes
on the shelves of the house of Mr. Prentiss, the American
Consulin St. Pierre. The voleano became so threatening that
some uneasiness was felt by the people dwelling on the slopes
of the mountain, but the citizens generally were so deeply
interested in a political contest in which race prejudice was
playing an important and bitter part, that they paid little atten-
tion to the returning activity of the dangerous mountain on
and near which they lived, until it was too late for them to
escape with their lives.

332 EL. O. Hovey—FEruptions of 1902 of La Soufriére,

The area of devastation caused by the eruptions of Pelée
from May to July inclusive was less than that due to that
of La Soufriere (compare maps, fig. 1, p. 320, and fig, 2, below.

2

—_——-—: —-..

\ ‘
Guerin, Factory. il,
Riv. Blanche. RK

/
I
Ss .
No
Carbetss

June 14-28))

5 English Miices.

Fic. 2. Map or THE NORTHWESTERN PART OF THE ISLAND OF MARTI-
nigue, F, W. I.

The cross-lined area shows approximately the portion of the island devas-
tated by the eruptions of Mt. Pelée, from May 5 to July 6, 1902. The
tremendous outbursts of August 25, 28 and 30 and September 3 extended
the ruin greatly on the north, east and southeast sides of this area, so that
nearly all of the region represented here, with the exception of a few square
miles in the southeastern corner, now is included within the zone of destruc-
tion. The dot-and-dash lines indicate the principal routes traversed by the
author while on and near the volcano.

Plotting the area on the Admiralty chart as well as possible,
after inspecting all sides of the voleano, and measuring it with
a planimeter, I find that thirty-two square miles of Martinique
were laid waste by dust, lapilli, stones and ‘‘mud.” The area
is not as symmetrically disposed about the crater as is the case

St. Vincent, and Mt. Pelée, Martinique. 333

on St. Vincent, probably because the crater of Pelée is so
much lower on the southwest than on the other sides and the
great gash opening into the gorge of the Riviére Blanche,
together with the configuration of the neighboring “ mornes,”
or ridges, has given direction to all the violent explosions
which have occurred. Although the whole island has received
débris from some of the outbursts and dust has been scattered
over a wide area, the district over which the vegetation was
killed, at least temporarily, is included within a line beginning
at the sea coast, about midway between St. Pierre and Carbet,
though the palm trees along the coast at the base of the bluffs
were scorched as far as Carbet Point itself. Passing inland
about a mile, the line curves sharply to the north and east of
north to the Roxelane River, then goes northeastward along
this river and one of its tributaries, paralleling the main street
of Morne Rouge within a quarter of a mile, swings then to
the east of La Calebasse and rises somewhat on the northeastern
flanks of Pelée, apparently passing along the south side of Pain
de Sucre and then northwestward, leaving the island midway
between La Perle Rock and Cap St. Martin. Much of this
area is already springing into verdure again; the grass was
already very noticeable on the hill slopes encircling St. Pierre
by July 1, and green vegetation was to be seen even nearer
the source of destruction. When I first arrived at Martinique
(May 21) the line between the scorched and unscorched areas
was strikingly sharp, and was still very noticeable six or seven
weeks later. In many places the line of demarkation passed
through single trees, leaving one side scorched and brown,
while the other side remained as green as if no eruption had
occurred.*

The material ejected by Pelée during this series of eruptions
consists of dust in vast quantities,t fine and coarse lapilli,
breadcrust bombs (see fig. 8) of all sizes from one inch to
three feet and more across, and blocks of small and great size,
the cracked condition of which shows that they have been
highly heated. The freshly fallen ashes had a curious resem-
blance to snow, which gave a peculiar Alpine aspect to the

* The tremendous eruptions of August 25, 28 and 30 and September 3 have
extended the devastated area greatly on the northeast, east and southeast
sides of the area of destruction indicated on the accompanying map. It now
extends as far as Carbet, and includes Morne Rouge, but the reports thus far
received do not clearly show whether it reaches quite to the coast on the
north and northeast or not. Tho eruption of Mt. Pelée- now must nearly if
not quite equal that of La Soufriére in magnitude.

+ One hundred and twenty tons of dust and lapilli were removed from the
decks of the Roddam after her arrival in the harbor of Castries, St. Lucia,
according to the personal statement of one of the agents of the line to
which the steamer belonged.

3384 EF. O. Hovey—LHruptions of 1902 of La Soufriere,

mountain, and is noticeable in some of the photographs. No
stream of molten lava has issued yet from the voleano asa
feature of this eruption, though such flows were common in
the early history of Pelée, as they were in that of St. Vincent’s
Soufriére. The bread-crust bombs prove, however, that much
lava has been thrown out in the condition of half-melted

FiGurE 3.—‘‘ BREADCRUST’”’? Boms FROM Mt. PELEE.

Collected on the Séche-Blanche plateau, three miles from the crater. The
specimen is two feet two inches high.

masses. ‘These bombs usually are more or less pumiceous in
texture, and they show the “bread-crust”” surface much more
distinctly than do the more basic bombs of the Soufriere.
The largest of the bombs observed was one fifteen feet long
on the southeast slope of Morne Lacroix at an elevation of
3950 feet above the sea. The largest ejected block that we
saw was one on the surface of the mud-flow between the rivers

St. Vincent, and Mt. Pelée, Martinique. 335

Blanche and Seéche and not more than two hundred yards from
the sea coast. Its dimensions are about 22 feet high, 30 feet
long and 24 feet broad (see fig. 16), and it is of the light-gray -
andesitic lava forming one of the ancient lava beds near the
summit of the mountain. When I inspected this block on
June 25, | found it too hot on the surface to bear the hand
upon it long at a time; the great mass was cracked in several
directions, and steam and sulphurous gases were emanating
from the cracks. It seems certain that this enormous block
was thrown out of the crater in a highly heated condition
during the present eruption, but it may have reached the place
where it now is partly through the agency of the great mud-
flow on which it rests. Many other great bowlders, some of
which are of nearly half the dimensions of the one just
described, lie near by on this mud plain.

The area of distribution of the ejecta cannot be designated
accurately yet for lack of data. The U. S. collier Leonidas
received a quarter of an inch of dust on her deck from the
great outburst of June 6 when she was 102 miles west of
Martinique. It took the ship from 3 Pp. M. until nearly 6
o’clock to traverse the cloud of dust. This eruption began at
10.15 A. M. and was one of the heaviest of the whole series.
I was in Georgetown, St. Vincent, at the time and felt the
shock distinctly. From 3 o’clock onward that afternoon until
after sunset heavy clouds of dust from the Pelée eruption
passed over St. Vincent, much of it falling upon the island.
The top of the cloud of dust as it passed over the mountains
seemed to me to be about 6000 feet above the sea, so that the
last deposits must have been made far south of St. Vincent.
Kingstown is about 110 miles south of Mt. Pelée. The shocks
or detonations from some if not all of the great outbursts were
felt in St. Kitts and Trinidad, thongh not in some of the
intervening islands. |

Two illustrations of the force with which the bombs and
blocks strike may be permitted here. On the sea coast near
the Fort Villaret church in the portion of St. Pierre north of
the Roxelane River there was a large distillery in which there
were four big storage tanks constructed of quarter-inch boiler
iron plates riveted together. These tanks look as if they had
been through a bombardment by artillery, being full of irregu-
lar holes which vary in size from mere cracks at the bottom of
indentations to great rents 24, 30 and even 36 inches across,
while a strip several feet long was torn off from each of two.
The direction of impact was essentially the same in all instances,
namely, from the crater. The other illustration is found on the
southeastern flanks of Mt. Pelée, along the trail leading from

Am. Journ. Sc.—FourtTH Series, Vou. XIV, No. 83.—NOvVEMBER, 1902.

336 L. O. Hovey—Eruptions of 1902 of La Soufriere,

Morne Rouge to the summit, where numerous spoon-shaped
_ depressions occur in the rather loose soil of the mountain side,
especially between the elevations of 2400 and 3000 feet above
tide. These holes are of all sizes from 2 feet in diameter
upward, the largest one which I saw being 40 feet long, 25
feet wide and 5 feet deep, but the depth had evidently been
reduced by the sand which had been washed into it by recent
rains. The longer axis of this depression was N. 50° W.,
pointing directly at the crater, and the longer axes of all the
other holes observed were pointing toward the same center.
The deepest part was on the up-hill side. On the down-hill
side of each depression was found the cause of the phenomenon,
and it was a bomb or ejected block from the volcano, which
had struck the ground with a splash, throwing the earth in all
directions and usually bounding or rolling out of the hole
which it had made. Sometimes the blocks which did the
work were found intact, but more frequently they had burst
asunder after strikmg. All showed that they had come out of
the voleano in a highly heated condition. Such splashes as
these can be made experimentally on «a small scale in any bed
of stiff mud by means of well-directed stones.

Many stones must have fallen in St. Pierre, but they are so
mingled with the rubble stones from the walls of the ruined
buildings that usually they are not easily distinguishable there-
from. Great quantities of small, rounded fragments of yellow
pumice are to be found now amid the ruins, the fine gray dust
having been washed away to a considerable extent by the
copious rains which have fallen since the great eruptions.
Most of the pebbles of pumice were less than three inches in
diameter. They are evidently from the old tuff agglomer-
ate and must have been torn from the beds through which the
voleanie vents pass and from the interior of the old cone.
Stones fell all over the island in some of the eruptions.

Four ascents of Mt. Pelée, in the course of which the crater
rim was traversed from the great chasm on the southwest along
the southern and eastern edge more than two-thirds of the way
around the circle, and the remainder also of the rim was clearly
seen, enabled Curtis and me to form a reasonably definite idea
of the center of activity and what was going on therein. Twice
we followed the trail from Morne Rouge to the summit, which
led us for a considerable distance along the right (southern)
brink of the canyon of the Falaise River, and on the day inter-
vening between these ascents we examined the gorge of the
Falaise carefully from the point where the Morne Rouge trail
to the summit strikes it nearly to its junction with the Capot
River, a mile or more beyond the area of devastation. The
upper reaches of the gorge certainly present the scene of deso-

St. Vincent, and Mt. Pelée, Martinique. 337

lation so graphically described by George Kennan,* but the
“‘ Falaise crater” mentioned by him and by Professor Heilprint
and indicated on Hill’s map can hardly be a true crater. We
saw the same accumulations of volcanie ash in the gorge at an
elevation of 1800 to 2000 feet above the sea (aneroid reading)
that Heilprin and Kennan mention as forming a crater from
which mud-flows were hurled down the gorge to the sea, and
we saw steam issuing from them, but to us who had studied the
phenomena on St. Vincent it seemed perfectly evident that
the outbursts in the gorge of the Falaise were comparatively
feeble examples of secondary or superficial eruptions of the
same character as those which took place on such a grand scale
from the ash-beds of the Wallibou and Rabaka Dry Rivers.
This was the history of events in the Falaise and probably at
Basse Pointe, Macouba, Grande Riviere and other places; and
it was the history of some, but not all, of the mud-flows in the
Précheur, the Mare, the Blanche, the Séche and the Des
Peres Rivers. Since the eruptions began there have been
great floods in the Roxelane River, but it seems doubtful
whether or not this stream has carried any trne mud-flows
down its gorge.

The actual crater is apparently somewhat oval in shape,
with its longer axis stretching northeast and southwest. The
highest point of the rim is on the northeast side and is what is
left of the peak known as Morne Lacroix. By taking the
average between the readings of our two barometers, we deter-
mined its altitude to be 4200 feet above the sea.§ It consists
of ancient andesitic lava. Almost directly opposite this is the
lowest point of the crater, where the great gash formed by the
gorge of the Riviere Blanche occurs, The bed of andesite
forming what may be considered the rim of the crater on the
southeast side of the gash is 3350 feet above the tide, while the
real bottom of the gorge where it issues from the crater is five
or six hundred feet less in altitude. From this lava bed the
rim rises rapidly (80° to 35°) to about 3750 feet above tide (see
fig. 13), and then more gradually along the southern edge until
3950 feet is reached on the eastern edge. The northwest side
of the southwestern gash is formed by a pinnacle of ancient
lava which appears to be about 4000 feet above the sea, but
may be higher. From this point the rim drops somewhat
toward the north, but gradually rises again toward the east

* The Outlook, vol. Ixxi, pp. 778, 774, 26 July, 1902.

+ McClure’s Magazine, vol. xix, p. 363, August, 1902.
_ ¢{ National Geographic Magazine, vol. xiii, p. 260, July, 1902.

$The French engineers located at Martinique are reported to have deter-
mined (by triangulation ?) that Morne Lacroix had lost 150 feet during the
eruption, making its present altitude 4273 feet above tide.

338 E. O. Hovey—Eruptions of 1902 of La Soufriere,

until the point of rock on the northeast, already mentioned, is
reached again. This great crater is about half a mile across, an
estimate that is based upon the proportion which it bears to the
height of the mountain, looked at from the sea, and from the
fact that it took us twenty minutes to walk along the southern
third of the rim from our first cairn to the Riviere Blanche
gorge without stopping. The walking was not bad, consider-
ing the location of the route, and I should estimate the distance
traveled in this time at not less than half a mile.

The breadth of the rim varies from a mere knife-edge on the
south, north and northeast sides to a sloping plateau fifty
to one hundred yards wide on the eastern side. This plateau
is the site of the Lac des Palmistes, which was considered to be
the old crater lake of Mt. Pelée. Studying this plateau care-
fully, we saw that it sloped gently southward and eastward from
one side of a low divide running northeastward from Morne
Lacroix across to a high ridge which paralleled the northern
and northeastern sides of the crater. On the northwest side
of this divide, the altitude of which was 3950 feet above tide,
there is a shallow valley which rapidly changes into a gorge
discharging into the canyon of the Préecheur River. Heilprin’s
description® and his unpublished photographs show the existence
on the plateau of a small lake-basin not more than five or six
feet deep. He and his companion, E. E. Leadbeater, a New
York photographer, state, furthermore, that this plateau and
this portion of the crater rim were entirely or practically free
from ash and dust deposit. When Curtis and I visited the
spot, June 18 and 20, the surface was coated with a thick layer
(more than four feet deep in places) of dust and ashes, probably
from the great eruption of June 6. This materal had drifted
into depressions to such an extent that we saw no indication of
the existence of a lake-basin in this plateau. We had a per-
fectly clear and cloudless period when on the spot, and saw the
topography with distinctness. I cannot think that the plateau,
including the lake basin, ever has been a primary crater or
center of eruption, though at the time of my visits the ground
was hot, a scalding temperature being reached less than a foot
from the surface, and steam was issuing from numerous crev-
ices. This was the site of the Lac des Palmistes, but that lake
was not located in the great ancient crater of Mt. Pelée.t

* Op. cit., p. 360.

+The photograph of the Lac des Palmistes published by Dr. Emil Deckert
on page 425 of the Zeitschrift der Gesellchaft fir Erd-kunde zu Berlin
for 1902 shows that body of water as it appeared in the rainy season of 1898.
Deckert describes the lake as being but 2 meters deep and as lying in the
middle of a morass or swamp upon a bed of lava. He regards this asa
crater lake, though he mentions the fact that there is not and probably never
could have been any crater wall on the east and north [south] sides. The

St. Vincent, and Mt. Pelée, Martinique. 339

_ Judging from the account of the guide Romain, * from the
“ Notes relating to the history of the eruption of 1902,” as
translated in the Century Magazinet from the issue of Les
Colonies for May 7, from the description by Lafcadio
Hearnt{ and from my own observations while on and near the
mountain, the Etang Sec was the real crater lake of Mt.
Pelée corresponding, though dry.since the eruption of 1851,
to the crater lakes of La Soufriere, on St. Vincent, Mt.
Misery, on St. Kitts, and others. Its basin is said to have
contained water until the eruption of 1851 drained it. The
Etang Sec is stated to have been 700 meters (2300 feet)
above the sea; its plain was estimated to be about 300 meters
(986 feet) across, and the great circle surrounding it was judged
to be about 800 meters (2628 feet) in diameter at top. This
last estimate agrees closely with my estimate of the diameter
of the present crater at top. The walls of the ancient crater
must have risen almost precipitously from 1600 to 2100 feet
above the Ktang Sec, except on the southwest, where was
located the great gorge through which flowed the waters of
the Riviere Blanche, the sources of which were in the eastern
side of the ancient crater.§ Before the eruption which began
last April, the crater of Pelée, except for the size of the great
gash, must have been very much like the crater of St. Vin-
eent’s Soufriére and that of Mt. Misery, St. Kitts, as I saw it
July 8 on my way home, and probably those of the other
voleanic cones of the Lesser Antilles.

The whole interior of the crater was not seen entirely free
from steam at any one time, but enough was observed to
determine its character in its eastern, southern and western
portions and to infer the shape of the remainder. The crater,
like that of the Soufriére, is in the top of a broad, truncated

eastern end of the ‘‘Somma” ring of Pelée bounds the little plateau on the
north, but there is no cliff to the south; on the contrary, the plateau slopes
off to the south into the head ravines of the Falaise River. It seems as if
Deckert must have gotten his north and south points interchanged in his
description. The two smail craters of 1851 mentioned by Deckert (loc. cit.,
p. 426) are covered now, probably, by the inner cone, while the series
of little craters in the gorge below those of 1851 must have become covered
by the débris from the same cone, or have had the evidences of their
existence destroyed by the tornadic blasts of the present eruption.

*The Century Magazine, vol. lxiv, p. 623, August, 1902. ‘Translation from
Les Colonies of May 5.

+ The Century Magazine, vol. lxiv, p. 631, August, 1902.

{‘‘ Through a cloud-rift we can see another crater-lake twelve hundred
feet below, said to be five times larger than the Etang we have just left [the
Lac des Palmistes, near the. summit]; it is also of more irregular outline.
. . . . It occupies a more ancient crater and is very rarely visited ; the path
leading to it is difficult and dangerous,—a natural ladder of roots and lianas
over a series of precipices.’—Two Years in the French West Indies, by
Lafcadio Hearn, p. 288.

$See Romain, loc. cit.

340 EL. O. Hovey—FKruptions of 1902 of La Soufriere,

cone of ancient tuff agglomerate alternating with lava beds.
Some diametral enlargement has taken place, perhaps, during
this eruption, though not enough to change the sky-line of the
top to any great degree, except in the southwest side, where
the old gash has been greatly widened, and perhaps deepened.
The lower part of the outside of this old cone has an angle of
slope of 20°, while the upper part is as steep as 80°, according
to my determinations. Measurements of the inner slope gave
values of 40° to 65° in the portion carved out of the old
agolomerate, but the angles increased to 75°, 85°, and even
showed great overhanging blocks on the eastern side where

/
ete. rer

New

fragmental ;
\ ecne with Basin of
a crater in top. ——# Lac des Palm-
ee \, teers i istes on slope
: ~ : eed / of Morne
\ eae / Lacroix.
7 ‘ yh vs
os om NN ?
Zo \ y
Ae Ye
ay ; Saaatian TE Re N.S
hang See ant Us fia.

Fig. 2a. SECTION across Summit oF Mt. PrLke From S.W. To N.E.
JuLy 6, 1902.

Horizontal and vertical scales the same. Elevations estimated, except on
northeast side. The shape of the crater in the inner cone is entirely infer-
ential. The relation of the great southwestern gash to the fragmental cone
is indicated by the dotted line. The broken lines complete the profile of the
great ancient crater as it existed, probably, before the eruption of May 8,
1902. The fragmental cone was the scene of greatest activity, but there
seemed to be another important center of eruption in the northeastern por-
tion of the crater at the base of Morne Lacroix.

the old lava beds form the rim. In the western portion of the
crater rises a cone of fragmental material, consisting of dust,
ashes and large and small blocks and bombs. This cone is the
scene of the greatest activity in the crater and it grew materi-
ally in size between the day when I first saw it, May 21, and
July 6, when I got my last glimpse of it. It now entirely
covers the site of the Etang Sec and partly fills the old crater,
and probably more than compensates for the material torn and
undermined from the old walls and thrown out by the erup-
tive action of the voleano. A large proportion of the activity
of the volcano, aside from that of the great outbursts, has
gone into the building of this cone. The accompanying profile

St. Vincent, and Mt. Pelée, Martinique. 341

(p. 340) indicates the probable relationships between the inner
cone and the old crater.

The illustration, fig. 14, gives the sight I obtained of the
inner cone from the eastern side of the crater. At that time
its top must have been just about on a level with or, perhaps,
somewhat higher than the camera, which was 3950 feet above
tide, by aneroid reading. The photograph shows that there
was a depressed crater in the top of the inner cone. My
measurements of the angle of slope of the southern side of
this cone determined it to be 38° to 40°, but there were pre-
cipitous portions. The material which rolls and slides down the
southwest side of this cone continues directly into the gorge
of the Riviere Blanche. The steep-sided valley formed by the
inner cone and the inner slopes of the crater-rim forms a
continuation of the gorge of the Blanche and rises at a con-
siderable angle from the southwestern gash to the base of the
rocky precipice on the eastern side of the crater, where it may
be 800 feet in depth. The valley probably continues around
the northern side of the inner cone rising in a spiral, for it
appears at an elevation of at least 3600 feet on the western
side, between the rim of the crater and the cone on the north-
west side of the great gash. The new fragmental cone rises,
apparently, on the site of the new crater mentioned by
Romain, a conclusion which seems to be in agreement with
the account of the eruption of May 8 by M. Fernand Clere as
given by Kennan,* which is as follows: “ About eight o’clock,
with a rending, roaring sound, a great cloud of black smoke
appeared suddenly on the southwestern face of the volcano
near its summit, and rushed swiftly down in the direction of
St. Pierre, . . .” Before this outburst, M. Clere had been
observing the great column of vapor rising from the other
principal center of eruption, which is located in the valley
within the great crater at the base of the high point of rock
on the eastern edge (the remains of Morne: Lacroix). At
intervals columns of steam rise energetically from other parts
of the crater valley.t+

* The Outlook, vol. lxxi, p. 683, July 12, 1902.

+ Morne Lacroix is reported to have “disappeared altogether and the crater
to have extended greatly toward the east during the oreat outbursts which
occurred from August 25 to September 3. This may indicate that the vent
under the new inner cone above described has become partly clogged, and
that the main activity has shifted to the vent mentioned (see p. 340) as being
east of the Htang Sec at the base of the western face of Morne Lacroix. If
this has taken place, as seems highly probable from the reported destruction
of Morne Lacroix, the new inner cone acted as a dam in the great south-
western gash which played such an important part in the destruction of St.
Pierre, so that the last eruptions came from a centrally located vent (like ©
that of La Soufriére, St. Vincent) and the débris from the crater was dis-
tributed more symmetrically about the cone. The position of the great gash

342 E. O. Hovey—Eruptions of 1902 of La Soufriére,

The history of the present series of eruptions may be epito-
mized somewhat as follows: the gradually returning activity
of the voleano began to make itself very manifest in the latter
part of April, since visitors to the crater found warm water in
the basin of the Etang See on the 25th of that month, and the -
lake was deep. Columns of dust-laden steam rose from an
opening within the old crater on the east side of the Etang
See and from one on the west side of the same basin, and
cones rose about these openings. Water in large quantity
collected in the old lake basin, assisted, perhaps, by a dam
formed across the gorge by the ejecta from the western crater.
The water was heated by the action of volcanic forces. On
May 5 the heated waters of the crater broke through this dam
and rushed, as a deluge of mud and bowlders of all sizes, down
the gorge of the Riviere Blanche, and overwhelmed the
Guérin sugar factory, which was situated at the mouth of the
stream. On May 8 began the series of great explosions which
have sent steam, laden with sulphuronus gases, dust, ashes and
stones, again and again over the southwest slope of the
mountain with the violence of a tornado, several times reach-
ing to St. Pierre and beyond. The author would explain the
blasts in the same way as in the case of St. Vincent (see p.
326), but the great gash in the side of the crater of Pelée and
the position of the neighboring ridges concentrated the force
of the explosions in a certain direction and along a compar-
atively narrow zone—and the city of St. Pierre with its 26,000
inhabitants* and thousands of refugees lay in an amphitheatre,
a regular cul-de-sac, directly in the path of the blasts. :

There seems to be no crater or center of primary eruption
in the gorge of the Blanche below the great crater, or in the
gorge of the Seche,t but there has been much secondary

and neighboring cliffs with reference to the vent on the western side of the
Etang Sec, which, was the most active center of eruption in May, June and
July, directed the blasts of the earlier eruptions toward St. Pierre and away
from Morne Rouge. That directive factor having ceased to have force,
through the growth of the inner cone and the (apparent) shifting of the
center of activity to the eastern vent, Morne Rouge, a mile and a half nearer
the crater than the middle of St. Pierre, came far within the area of destruc-
tion and received the full fury of an eruption.

* Population in 1895, 25,382, according to the Century Atlas.

+ Prof. Angelo Heilprin has stated in his article in McClure’s Magazine
for August, 1902, and elsewhere, that eruptions have taken place from a
crater located in the gorge of the Riviére Blanche some distance below the
great crater. R. T. Hill has expressed the same idea in his extended article
in the National Geographic Magazine (vol. xiii, pp. 251, 261) for July,
1902, and speaks of this as the center from which came the blast that
destroyed St. Pierre. calling it the ‘‘Soufriére crater.” He has recorded the
matter on a map which was published on p. 260 of the National Geogra-
phic Magazine. Curtis was With Hill when the latter made the observa-
tions on which this map was based, and therefore knew the spot intended to
be represented. Curtis and I stood on the brink of the gorge of the Blanche
overlooking it on June 24, examined it again with field glasses from the rim

St. Vincent, and Mt. Pelée, Martinique. 343

action along the lower portion of their courses, and much
steam, with or without large quantities of dust, has been
thrown high into the air when water has reached the heated
interior of the vast beds of volcanic ash deposited there during
this eruption. Mud-flows and torrents have been very numer-
ous down the gorges of these streams and on the plateau
between them. Some of these flows have come directly from
the crater, especially in the case of the Blanche; others have
originated on the exterior slopes of the old cone, while others
have started in the heavy ash-beds in the gorges in the manner
already described in connection with the Wallibou River, St.
Vincent, and the Falaise River on the eastern side of Mt.
Pelée. The surface of one of the mud-flows on the plateau
between the Seche and the Blanche is shown in fig. 17. Sub-
sequent rain has washed the fine mud from the stones.

These streams of mud and stones present some characteris-
tics which distinguish them clearly from the surfaces of undis-
turbed ash-beds. The most striking of these is the existence
of curved folds or wrinkles transverse to the direction of flow
of the stream, the folds varying in size with the size of the
flow. The surface of an unmodified deposit of ash presents a
drifted appearance like that of a field of snow or of dry sand
on a sea beach, and the Richmond estate as illustrated in fig. 6
is a typical example. The plateau between the Des Peres and
Roxelane Rivers, on which was located the Fort Quarter of
St. Pierre, was covered with several feet of wind-drifted ash,
and it was nota mud-flow or a series of mud flows which
destroyed this portion of the city, as has been stated in several
publications.

In addition to the showers of dry dust and ashes, there fell
during the eruptions an immense amount of liquid mud which

of the crater, where we obtained a view directly through the gorge lengthwise,
and repeated the examination from the crater rim on June 26. We could
see no crater or center of eruption in the gorge of the Blanche below the
great crater, though there has been much secondary or superficial. eruption
of steam from the ash-beds along the gorge.

I cannot agree with the distribution of the ‘‘zones of devastation”
indicated on Hill’s map or with the location of ‘‘mud craters” as the origin
of the mud-flows of the Séche, the Blanche, the Falaise and other valleys.
Tt is well to separate the devastation into zones of ‘‘ annihilation’ and
** singeing,” in a general way, but the crater of the voleano should certainly
be included within the former instead of being placed outside of the latter,
as is done on Hill’s map. The real location of the singe line of the May
eruptions is nearly that of Hill’s ‘‘ ash line” and is indicated approximately
on my map (p. 532), where it is called the limit of devastation. The ‘‘ash
line’ should be placed at some undetermined distance far beyond the shores
of the island of Martinique. The existence of real ‘‘mud craters” on the
slopes of Mt. Pelée seems very improbable, for the reasons given on pages
337 and 338 and elsewhere in this article. The mud-torrents of the Grande
Riviere, which were among the heaviest of those experienced on the north
and northeast side of the island, are not indicated on Hill’s map.

344 L. O. Hovey—Eruptions of 1902 of La Soufriére,

had been formed within the eruption cloud through the con-
densation of its moisture. This mud formed a tenacious coat-
ing over everything with which it came in contact. That
drops of mud, too, formed in the air and fell as a feature of the
eruption is proven by the condition of the walls of the houses
in Précheur, on which I foand flattened spheroids of dried
mud which could have formed only in the manner indicated.
These flecks of mud were two, four and even six inches across,
where two or more had coalesced. They occurred mostly on
the northern and eastern walls of the houses. The testimony
of the people as to the occurrence of rain during the great
eruptions is conflicting, but the existence of this coating and
these drops of mud proves that much aérial condensation of
steam accompanied the outbursts.

During the latter part of our stay on the crater rim on June
24 the rain fell in torrents, and the deluge continued until we
reached the foot of the outer cone on our return journey, the
heaviest portion of the storm lasting for an hour or an hour
and a half. Here we found the fumaroles sending out more
steam than they did on our upward journey. When we crossed
the Séeche River, we found a foot and a half of yellow, muddy
water in place of the two or three inches which we had noticed
there in the morning. We had not climbed out of the lowest
gorge of the river before our attention was attracted by the
heavy eruption that was taking place from the crater, and that
was sending enormous clouds of dust-laden steam down the
gorge of the Blanche to a point below the so-called Soufriere
crater. Thunder-like noises nearer at hand had already made
themselves heard and in another minute a wall of hot water,
ten or fifteen feet high, swept with railroad speed over the place
where we had crossed the river, and rushed on tothe sea. The
roar of the torrent was like that of a train, and the water dash-
ing from side to side of the narrow gorge caused the ground
on which we were standing to tremble like a ship when her
propeller “races.” The water was thick and as black as ink
with its load of volcanic ash, and it transported with ease bowl-
ders five feet in diameter which it had excavated from its
banks. In many, if not most instances, these bowlders were the
ejecta of the present eruption. To the left a stream of thick,
yellowish mud was flowing down from the plateau of the
Seche-Blanche which we had left a quarter of an hour before
and was cascading into the Séche directly beneath us. Soon
the black torrent cut into the ash-beds along its banks suffi-
ciently to reach their still highly heated interiors and cause
columns of steam to shoot hundreds of feet into the air. The
steam columns carried great clouds of black and light-brown
voleanic sand scores of feet upward. The hot area of the

St. Vincent, and Mt. Pelée, Martinique. 345

plateau also was sending skyward great columns of steam, and
the whole formed a scene seldom witnessed, difficult to describe,
and never to be forgotten. The next day we measured the
gorge and found that the Seche had deepened its channel at
least ten feet in the loosely compacted recent ash during the
hour which the flood lasted. |

In this instance it seems evident that there was close con-
nection between the heavy rain, the eruption, and the black
torrent. Two explanations present themselves for considera-
tion: (a) the crater may have thrown a mass of accumulated
rainwater and ashes bodily over into the head canyon of the
Seche; (6) the rain which fell into the crater may have been
the exciting cause of the eruption, but the mud-torrent may
have been due to the soaking of the heavy coat of ashes on the
steep outer slope of the old cone at the head of the Seche until
the resulting fluid mass slid off from the comparatively hard
surface beneath and poured down the gorge of the river.
There was plenty of water-soaked mud and ashes on the upper
part of the mountain to supply the avalanche and some of it
was on the verge of fluidity at the time of our visit, hence the
latter explanation seems the more reasonable. The little tribu-
tary of the Seche which empties into it on the southeast side
close to our point of observation, did not show any correspond-
ing torrent, because it does not head on the side of the great
cone.

The mud-flows which have descended the Precheur River
canyon have had ample collecting ground in the “ Atrio del
caballo,” to use a Vesuvian term, on the north and northeast
sides of the great crater, where the fine dust settles in vast
quantities ready, when sufficient water has been added to it, to
descend through a narrow gorge into the valley of the Precheur.
When I was walking along the crater rim above the “ Atrio”
June 20, my footsteps started small mud-flows down the outer
cone, so liquid was the mud at that time. The ordinary action
of the voleano is to deposit dust of impalpable fineness on the
inner face of the crater rim. When this deposit becomes thick,
it is ready to be swept off by a copious rain and carried through
the great southwestern gash, out of the crater and down the
gorge of the Riviere Blanche as a mud-torrent or flow. There
does not seem to the writer to be any need of locating ‘ mud
eraters”’ at the heads of or along the line of the gorges which
have been the courses which these torrents of liquid mud have
followed to the sea.

Where the tuff agglomerate of the old (outer) cone had been
freed from its coat of ashes, especially in its lower portion,
1. e., from 1000 to 2000 feet above tide, it showed a smooth,
somewhat fluted surface, the soft bowlders having been planed

346 LF. O. Hovey—Lruptions of 1902 of La Soufriere,

off even with the matrix. The whole showed striations parallel
with the slope, so that the surface looked like the glaciated
rock surfaces so common in northern latitudes. The planing
and the striations seem to have been due to the scouring action
of the ash avalanches in this part of their course. They ceased
where the steep slope of the cone changed to the gentle slope
of the plateau, and thus gave opportunity for the material of
the avalanches to check its descent and pile up. The sides of
the radial gorges on the flanks of Mt. Pelée show approximately
horizontal striations. Near the stream-beds such striations
occur on both sides of the gorges, and are due to abrasion by
the sand and stones carried by the torrents. High above the
stream on the bluffs facing the crater there are similar stria-
tions, but these must have been made by sand-blast action dur-
ing the hnrricanes of dust-laden steam, resulting from the
explosions during the recent great eruptions. These striations
extend to the very tops of bluffs rising 500 feet and more above
the stream beds at their bases (see fig. 15).

Erosion seems not yet to have cut deeply into the old land-
surface since or as a feature of the eruptions, because here and
there all over the mountain side one can find undisturbed roots
and charred grass still in place. The shore line from Ste.
Marthe Point nearly to Cap St. Martin has been somewhat
altered since the eruptions began, some of the river deltas hav-
ing been built out by the new material brought down by stream
and torrent, while others have been cut back by the waves.
The most important example of the cutting back is near the
mouths of the Seche and the Blanche, where local land-slides
have assisted the sea in forming nearly vertical bluffs from ten
to thirty feet in height. These bluffs show sections of the old
and the new material now composing the plain. The little ash
island* which was formed near the mouth of the Riviere la
Mare between May 8 and 23, and which was visited on May
23 by Mr. Curtis and two companions, had been washed away
by June 14. The stone pavement laid on the beach of St.
Pierre was cut into in places, perhaps by the return waves
from the ocean accompanying the great outbursts.

The mud-flow which swept down the Grande Riviere
reached the village of the same name at 4 A. M., May 8, four
hours before the eruption occurred which destroyed St. Pierre.
Three other great mud-flows have traversed this river: on May
11, June 6 and June 22, though no great eruption of Pelée
took place on May 11 or June 22. The eruption of June 6
was one of the heaviest that occurred ; this time the mud-torrent
reached Grande Riviere village about an hour anda half before

* Mentioned by Hill; The Century Magazine, vol. lxiv, p. 773, September,
1902.

St. Vineent, and Mt. Pelée, Martinique. 347

the eruption took place.. The flood of May 8 was the most
violent and was three meters (about 10 feet) deep where the
valley of the river opens onto the sea coast, according to M.
Delsol Désiré, the mayor’s deputy of Grande Riviere. He
gave me the foregoing particulars in regard to these floods.
The fine mud of these flows entered the buildings on the banks
of the river as if it had been thick syrup. In one room that
we examined the line of highest level was even with the top
of an ordinary table, which would show that the mud was 30
inches deep in the room. At the time of our visit the deposit
was nearly dry and it showed a shrinkage of but eight inches
or 27 per cent. In another room the shrinkage was greater,
showing that the mud there was thinner when it flowed in.
Streams composed of such material as this would have great
power in the transportation of bowlders. The sizes of the
bowlders brought down by tne mud-torrents and deposited on
the flood plain of the stream above the village and in its old
channel may be inferred from figure 18. Some of those meas-
ured were eight feet across. The bowlders seem to be from
old deposits, since they have weathered surfaces. They show
fresh abrasion along edges and at corners, due to their recent
trip down the gorge. The mud is made up of gray material
from the present eruption, together with a large proportion of
yellow sand from the old beds through which the river runs.
The vast amount of material brought down by the torrents has
extended the delta plain fully five hundred feet into the sea
and has pushed out the shore-line for several hundred yards on
either side of the mouth of the river.

At Basse Pointe the history in regard to floods or torrents of
mud has been similar. The principal disasters occurred on
May 8 and 27, but the latter was the greater and most of the
destruction was wrought on that occasion. Here too the deltal
plain has encroached five hundred or six hundred feet on the
sea and the ocean currents have spread the surplus material as
a new beach for a long distance north and south of the mouth
of the river, destroying the little artificial harbor of the town.
Bowlders ten feet across were brought down and left in the
town by the floods, and a deposit of sand, gravel and bowlders
fifteen to eighteen feet deep rests upon the site of the old
market place, which was at the mouth of the river.

The ruins of the city of St. Pierre presented a very interest-
ing field of study, but mostly in the line of speculations as to
the cause or causes of the terrible destruction of human life.
The walls of the houses (see fig. 11) showed that one or
more blasts of tornadic violence had swept over the city and
that they came from the direction of the crater of Pelée, for
the east and west walls—transverse to the direction of the

348 #. O. Hovey—Lruptions of 1902 of La Soufriere,

crater from the city—were demolished more generally than
the north and south walls. The direction in which most of
the trees were felled indicates the same thing, but the trees
in the angle of Morne Mirail, which rises behind the middle
of the city, were thrown over at all angles progressively, show-
ing that a vortex was formed there. As is indicated by the
gradually decreasing degree of destruction in passing from the
northern to the southern part of the city, the blast diminished
in force as it progressed and expanded, but when it reached
Ste. Marthe Point it still had strength enough to throw the
statue of Notre Dame de la Garde from its pedestal. The
statue, which is of hollow iron ten or eleven feet high, now
lies on the edge of the bluff with its foot about fifty feet S.
10° W. from its original position on the pedestal, and directly
in line with the crater.* The guns in the Ste. Marthe and
Morne d’Orange batteries were thrown from their carriages in
the same direction. More than once when I| was on the rim of
the crater or on the west flanks of the mountain I saw great
clouds of dust-laden steam come out of the gash in the side
of the crater with sufficient force to descend the gorge of the
Riviere Blanche with great rapidity a full mile before rising
in columns. It was not difficult to imagine that, if this hap-
pened when the crater was sending a steam and dust column
only one or two thousand feet high, the action would be vastly
greater and even like a hurricane in violence when the crater
was in full eruption and was sending its ash-laden steam column
from five to seven miles into the air.t

It does not seem necessary to call in any forces new or
strange to the history of vulecanism to account for the phe-
nomena attending the eruption of Mt. Pelée, or the destruc-
tion of St. Pierre and its people. The “ flames” reported were
perhaps the incandescent stones and bombs flying through the
air; and these certainly would set fire to any combustible
material upon which they fell. The officers of the French
cable-repair ship Pouyer Quertier were eye-witnesses of the
eruption of May 8 and describe the cloud as being black when ~
it issued from the crater, but say that it became luminous as it
approached the coast.{ Several times at night during our stay
we saw the inner cone of the crater outlined and streaked with

* See Bul. Am. Mus. Nat. Hist., vol. xvi, pl. xlvi, fig. 2; and Nat. Geog.
Mag., vol. xiii, p. 200, for illustrations. ;

+ Lieut. B. B. McCormick, U. S. N., in command of the Potomac, was
on his vessel in the harbor of Fort de France May 20 and made measure-
ments of the angular distance to which the steam column rose during the
great outburst that morning. The column subtended an angle of about 30°
and the tug was thirteen and one-half nautical miles in a straight line from
the mountain.

t Heilprin, loc. cit., p. 367.

St. Vincent, and Mt. Pelée, Martinique. 349

lines of “fire” due to the rolling and sliding red-hot rocks and
Japilli, and this light was reflected from the steam clouds above
the cone. The existence of notable quantities of burning or
inflammable gases in the discharges from the volcano seems to
me to be as yet undemonstrated.

On two occasions, June 24 and 26, I went into the crater for
a short distance beside the southwestern gash and several times
was surrounded with heavy clouds of steam from within the
abyss. The steam, which was warm, but not hot, when it
reached me, contained much sulphur dioxide, SO,, and at
times some hydrogen sulphide, H,S, but I could not detect the
odor of any other gas. The sulphur gases made the atmos-
phere difficult to breathe, but the most uncomfortable sensation
was due to the irritation caused by the fine, angular dust get-
ting into the respiratory passages and the eyes. Such a mix-
ture, raised to a high temperature, and containing a large
amount of dust and a considerable percentage ot sulphur gases,
would be almost instantaneously fatal to life. It was a cloud
like this that rolled over and enveloped St. Pierre for several
-minutes about eight o’clock in the morning of May 8, and
must have caused most of the deaths. Some of the other
causes of death were, (1) blows from falling stones which had
been hurled out from the voleano, (2) crushing beneath falling
walls and various objects (one man was found with his back
broken by a sign which had fallen from over a store front), (3)
burns due to hot stones and dust, (4) burns caused by steam
alone and (5) by steam mingled with dust, (6) cremation in
burning buildings, (7) nervous shock, (8) suffocation from lack
of respirable air and, perhaps, (9) lightning. No autopsy was
made on any of the thousands of victims of the disaster on
Martinique, although men capable of performing such opera-
tions had the opportunity of making them within a very few
hours after the eruption ; hence there is no sure way of deter-
mining whether poisonous gases other than those mentioned
played any part in the destruction of life. Immanuel Lédée,
one of the survivors of the crew of the Roraima, told me
that when the mucous membrane of his mouth, throat and
nose sloughed off on account of the burning, it was found to
be full of the fine black (gray) dust. He was taken to the
hospital at Fort de France after the eruption. Samson Cil-
Barice,* the prisoner who is the sole survivor of the persons
within St. Pierre at the time of the eruption, told me in
Morne Rouge on June 18 that it was the hot dry “sand”
which sifted in through the window of his cell that caused
his terrible burns.

* This name is spelled very differently in the various accounts of the dis-
aster. The spelling here adopted is that given me by my interpreter.

3850 £. O. Hovey—FEruptions of La Soufriere, ete.

The term “stellar lightning” has been proposed by George
Kennan for the particular form of electrical discharge charac-
terizing the eruptions of Mt. Pelée. This expression, how-
ever, implies that the bolts shot out radially from a center in
all directions at the same instant, whereas the shafts flew out
successively in the different directions. Often they seemed to
come from centers, but this appearance probably was due to —
the foreshortening of the line along which the successive
flashes originated. The amount of electricity generated by the
friction of the ascending column and the moving clouds of
dust-laden steam against the surrounding atmosphere was very
great, but much of the discharge seemed to be comparatively
noiseless. At midnight of June 26 an eruption occurred which
sent up a steam column to a height estimated at 12,000 feet
above the top of the mountain. Much of this scintillating*
lightning played about the column and the “mushroom”
cloud above, but no sound of thunder could be heard from our
sloop, which was at anchor off Carbet, seven miles distant.
The same form of electrical discharge was observed in connec-
tion with the great outbursts of La Soufriére on St. Vincent.
The electrical phenomena attending the September eruptions of
Mt. Pelée are described as having been even more magnificent.
and terrifying than those observed in connection with the
earlier explosions.

American Museum of Natural History.
October 11, 1902.

* “ Qoruscating”’ is the excellent descriptive term applied by Dr. Jaggar
to these discharges.

Photograph by E. O. Hovey.
LA SOUFRIERE. SOUTHWESTERN PART OF CRATER-RIM.

Photograph by E. O. Hovey.
LA SOUFRIERE. SOUTHEASTERN PORTION OF CRATER. [Page 351.

‘INHONTA “LG “HLVISH GNOWHOINY AH], "MOSTEM “O ‘¢ AQ Ydvasojo4g

Photograph by E. O. Hovey.

LA SOUFRIERE. SECONDARY OUTBURST IN ASHBED OF WALLIBOU RIVER.

Dendritic erosion developing.

Photograph by E. O. Hovey.
La SOUFRIERE, MvUD-COVERED RIDGE AT 1900 FEET ALTITUDE.

Tot

[Page 393.

Photograph by E. VU. Hovey.
LA SOUFRIERE. ASH-FILLED GORGE OF THE RABAKA Dry RIVER.
New deposit is said to be 200 feet deep.

Photograph by E. O. Hovey.
La SOUFRIERE. WINDWARD SIDE. TREES FELLED AND CARVED BY VOLCANIC BLAST.
Page 304. |

Photograph by E. O. Hovey.
RUINS OF ST. PIERRE, MARTINIQUE. MrT. PELEE IN BACKGROUND.

12

Photograph by E. O. Hovey.
Ruins oF St. PIERRE, MARTINIQUE. VALLEY OF RIVInRE ROXELANE.

[Page 355.

18

Photograph by E. O. Hovey.
Mt. PELEE. SOUTHEASTERN PART OF OCRATER-RIM AT 8700 FEET ALTITUDE.

Photograph by E. O. Hovey.
Mr. PELEE. INNER CONE; FROM EASTERN EDGE OF

Page 356. |

CRATER AT 3950 FEET ALTITUDE.

Photograph by E. O. Hovey.
Mr. PELEE. BLUFF IN CENTER OF VIEW SHOWS SAND-BLAST ACTION OF VOLCANIC TORNADO.

Photograph by E. O. Hovey.
Mr. Pet&e. EsECTED BLOCK OF ANDESITIC LAVA ON SECHE-BLANCHE PLATEAU, 3 MILES
FROM CRATER.

Dimensions 30! x 24’ x 22’. [Page 307.

Photograph by E. O. Hovey.
Mt. PELEE. SURFACE OF ‘‘MUD-FLOW” ON SECHE-BLANCHE PLATEAU AT 800 FEET ALTITUDE,

18

Photograph by E. O. Hovey.

GRANDE RIVIERE. GREAT BOWLDERS BROUGHT DOWN BY MUD-TORRENTS OF May AnD JUNE, 1902.
Page 358. ]

Bumstead— Reflection of Electric Waves, ete. 359.

Art. XX XIV.—On the feflection of Electric Waves at the
Free End of a Parallel Wire System; by Henry A.
BUMSTEAD.

WHEN electrical waves are sent along two straight parallel
wires, and a series of standing waves is formed by reflection
from the unconnected ends of the wires, it is found that the
distance from the ends (where one would expect to find a loop
for electric force and a node for current) to the first node for
electric force is less than a quarter wave-length as determined
elsewhere along the wires; and the discrepancy always appears
whether the measurements are made by means of a resonator
or by the sliding bridge method of Lecher. This apparently
anomalous behavior excited considerable attention among the
earlier experimenters and writers upon this subject, and sev-
eral attempts at an explanation, by means of special assum p-
tions, have been made.* <A more careful consideration of the
question, however, shows that even with perfectly conducting
wires, and without any modification of Maxwell’s theory, or
any assumption of the escape of charge into the air, a displace-
ment of the current node beyond the end of the wires is to be
expected theoretically and that its magnitude should be of the
same order as that which is found experimentally.

The cause of the phenomenon is perhaps most easily seen
by using Mr. Heaviside’s ingenious device of considering elec-
trical resistance in the guiding wires as approximately equiva-
lent to a fictitious magnetic conductivity in the dielectric. Thus
if the vectors E and H are the electric and magnetic forces at
any point and C the total current, the fundamental equations
of the field, in the form used by Heaviside and Hertz, are

curl H = 470 (1)
—curl E = »H (2)
C is the sum of the displacement and conduction current in
the dielectric, so that

Ks
C= ey E+/E
T

where K and /7 are respectively the dielectric constant and
conductivity of the medium. Equation (1) thus becomes

| cull H = KE+47/E (1’)
If we assume that the medium possesses also a magnetic con-
ductivity, g, then (2) takes the entirely analogous form

—curl EB = vuH + 4rgH (2’)

* See Poincaré, Les Oscillations Electriques, arts. 110 and 125.

Am. Jour. Sct.—FourtTs SERIES, Vou.. XIV, No. 83.—NOVEMBER, 1902.

360 Bumstead—Reflection of Electric Waves at the

If now we are dealing with plane waves, propagated along
straight and perfectly conducting guides parallel to the axis of
2, then K and Hare in planes perpendicular to 2 and cut
each other orthogonally at all points. The actual form of the
lines of electric and magnetic force may, of course, vary greatly
with the shape and size of the guides, but, if we follow any
plane in the wave, the line integral of E from one guide to
the other is independent of the path, so long as we stay in the
plane; and, with this understanding, it may be called the dif-
ference of potential between the wires, V. In like manner,
the line integral of H about a closed circuit embracing one of
the wires is a constant, and equal to 4771 where I is the current
in the wire at this point. It should be observed that this eur-
rent has nothing directly to do with the current € in (1), but
is, from the present point of view, to be regarded merely as
the line integral of the magnetic force.

The advantage in using the scalars, V and I, instead of the
vectors, EK and H, in treating plane waves, is apparent, since
V and I vary only with the codrdinate along the wire, 2, and
the problem reduces to one dimension. Equations (1’) and (2’)
become, respectively,

dl ih
— 5 =8V + @V (3)
OO ee ,

where § and L are the capacity and inductance of the system,
G is the total electric conductance of the medium from one
wire to the other, and F is the total magnetic conductance of
the medium in circuits about the two wires following the lines
of magnetic force; all these being measured per unit length
along z.* 7

It is immediately apparent from these equations that the
rate of dissipation of energy in a wave plane, whose thickness
is dz, is GV°dz due to the electric conductivity, and 47FPdz
due to the magnetic conductivity ; and both take place through-
out the plane. 7

~The approximate equivalence of magnetic conductance in
the medium and electrical resistance in the guiding wires is
thus made evident. If we now have no magnetic condue-
tivity in the medium, but a resistance, $R per unit length, in
each wire, where R = 47F, we shall still have the same dissi-
pation per second in the plane, viz: RI’dz, and things will go
on approximately as before. In fact, if we use one of the old
theories, which ignore the action of the medium, to investi-

* See Heaviside, Electromagnetic Theory, vol. i, p. 884, where, however, a
different system of units is employed.

Free End of a Parallel Wire System. 361

gate the propagation of electrical disturbance along resisting
wires, we shall get exactly equations (8) and (4) with R in
place of 47F. But on Maxwell’s theory, as Heaviside has
pointed out, the correspondence is not exact; for, with the
resisting wires, we shall no longer have strictly plane waves,
as the electrical force is no longer normal to the wires; more-
over the dissipation of energy does not take place throughout
the plane but in the wires alone, and, unless the distance
between the wires is small in comparison with the wave length,
we must expect to find noticeable differences between the
approximate theory and the results of experiment.

It is in this lack of complete correspondence that the expla-
nation is to be found for the apparent discrepancy mentioned
at the beginning of this paper. To make the matter some-
what clearer, iet us consider first the conditions when the ends
are bridged. Jet the perfectly conducting wires end_nor-
mally upon a plate of infinite extent with perfect electrical con-
ductivity ; then, when any wave plane reaches it, the electric
force will be instantly annulled throughout the plane, but the
magnetic force will persist, and we must therefore have a
reflected sheet with its electrical force in the reyersed direc-
tion, and its magnetic force in the same direction, as in the
incident sheet. This can occur since the two sheets are mov-
ing in opposite directions. Exactly at the plate we shall have
a node of E and V, and a loop of H and I. If, instead of
using the infinite plate, we bridge with a straight, perfectly
conducting wire, we shall have nearly, but not quite, the same
result; for now the electrical force is not instantly annulled
throughout the whole plane, but only in the central portion ;
and the outlying tubes of electric force must travel inward
toward the bridge with the speed of light, thus generating
lines of magnetic force which surround the bridge and, in
general, making the reflection a complicated one. But the
most conspicuous effect will be a delay in the vanishing of the
electric force and a consequent apparent shift of the potential
node (or current loop) beyond the bridge. As, with circular
wires of equal section, one-half the whole number of unit
_ tubes in the plane lie within a circle whose diameter is the
distance between the wires, we should expect this shift to be
not far from one-half the distance between the wires. Asa
matter of fact, a displacement of the node beyond the termi-
nal bridge of about this amount is always observed, but it
excites no surprise, since even the most elementary considera-
tions make it natural to add half the length of the bridge to
the parallel wires, the middle of the bridge being, obviously, a
center of symmetry. It is, however, frequently assumed that
this correction would disappear if the bridge were a perfect

362 Bumstead—Reflection of Electric Waves, ete.

conductor ;* but the above considerations make it clear that
this is not the case.

But an exactly analogous thing happens also when there is
no bridge, when the simple theory would not cause it to be
expected. Thus, suppose the wires to terminate in an infinite
plate having no electrical conductivity but infinite magnetic
conductivity ; we now have reflection of the opposite kind to
that in the previous case, for now the magnetic force will
vanish while the electrical force will remain; and we shall
thus have a node for H and I and a loop for E and V. If,
instead of having the magnetically conducting plate, we cut
the wires at this point, we have nearly the same effect; for
infinite resistance in the wires is, as we have seen, approxi-
mately equivalent to infinite magnetic conductivity in the
medium. But the result is not quite the same, for the effect
of the termination of the wires is felt, at first, only in the parts
of the wave plane near the wire; it is equivalent, not to an
infinite plane with magnetic conductivity, but to a little ring
of infinite magnetic conductance, immediately surrounding
each wire. ‘These little rings will immediately annul the mag-
netic force near the wire, but, as in the case of the wire bridge,
the more distant tubes of force must come in with the speed of
light and there is an entirely analogous delay and displace-
ment of the apparent node beyond the ends of the wires. The
shifting of the node should be about the same as in the pre-
vious case, as, in fact, it is found to be experimentally ; if any-
thing, it should be slightly less, for the imitation of a ring of
perfect magnetic conductance by cutting the wire is probably
closer than the approximation to a perfectly conducting bridge
when a copper wire is put across the guides.

Yale University, New Haven, Conn.

* EK. g., Drude, Physik des Aethers, bottom of page 465.

G. H. Girty— Upper Permian in Western Texas. 363

Art. XXXV.—The Upper Permian in Western Texas ;* by
GEORGE H. GIRTY.

DurRinG the field season of 1901, as a member of a party
under the direction of Mr. R. T. Hill, 1 examined a very inter-
esting series of Carboniferous strata in western Texas. The
route traversed was in part the same, though in an opposite
direction, as that pursued by the expedition under Captain
Pope in 1855, of which G. G. Shumard was a member. Our
point of departure was El Paso, and we passed eastward,
approximately along the thirty-second parallel, as far as the
headwaters of Delaware Creek, in the Trans-Pecos region.
_ There, finding that the old trail down Delaware was impractica-
ble, we turned southward, meeting the railroad at Toya.

The region of special interest was the Guadalupe Mountains,
where Shumard collected a number of species, subsequently
described by his brother, B. F. Shumard, which indicated the
existence at this point of a peculiar Carboniferous fauna, essen-
tially different from anything known elsewhere in North
America. Shumard gives the following section of the Guada-
lupe Mountains :t+

1, Upper or white limestone -_.---.--------- 1,000 feet.
2. Dark-colored, thinly laminated and foliated
MEMERRGING 1 lu a crete evil i ds eel. a2 50-100
3. Yellow quartzose sandstone._-...--------.--. 1,200-1,500
4. Black, thin-bedded limestone ._-_.---.---- 500

The geologic series is finely exposed, and the order of super-
position of the beds is obvious. The constituent formations
are of rather unusual thickness and uniformity of composition.
In fact, though we were able to make only barometric measure-
ments, and though only the upper limestone was exposed in a
continuous section, I believe that the thicknesses assigned by
Shumard are too small. We measured 1,700 to 1,800 feet of
the upper limestone, while the middle member, a yellow sand-
stone, must contain from 2,000 to 2,500 feet. Of the black
limestone, which forms the lowest member of the section,
probably not more than 500 feet are exposed, but the base was
not seen.

Fossils were obtained from all these formations, but it is
especially of those from the upper one that I desire to speak.
The rock is a white limestone, sometimes stained yellowish or
reddish in the lower portion, where it is also quite siliceous.

* Published by permission of the Director of the U. S. Geological Survey.
+ Acad. Sci. St. Louis, Trans., vol. 1, 1860, p. 280.

364 G. H. Girty— Upper Permian in Western Tewas.

Horizons which are dolomitic also occur, and one of these is a
pisolite. The highest beds seen are exposed on the summit of
El Capitan, and contain P’usulina elongata Shumard in great
profusion. The main fossiliferous horizon, however, was found
about 1,000 feet below. Near this level crystalline calcite
occurs in great abundance and in cleavage blocks of considera-
ble size. The locality was difficult of access, and could only
be reached by hard climbing, so that the collections obtained
were less complete than might be wished. The above remarks
relate to the “ white Permian,” which is peculiarly massive and
shows little evidence of bedding. The hundred feet or more
of brownish or gray limestone immediately below it, however,
which represent Shumard’s “dark Permian,” are obviously
stratified. From the “white Permian,” chiefly at the locality -
mentioned, were obtained upward of 75 species, some of the
more interesting of which I will mention. The fauna is very
varied, and includes representatives from nearly all the inverte-
brate groups, though the brachiopods and pelecypods are most
numerous in species and individuals.

Of the Protozoa we have Fusulina elongata Shumard, a
species unique so far as I am aware, which probably attains a
length of nearly 2 inches.

Sponges are remarkably abundant, and belong chiefly to the
Calcispongie. The orders Lycones and Pharetrones are repre-
sented by a number of genera and species, most of which are
probably new. In addition the genera Mammillopora and
Bothroconis have been identified.

Though a species of Lophophyllum was found, coelenterates
are rare, except for the group of Hydrocoralline similar to
those described by Waagen from India. Both in these forms
and those representing the Calcispongie the rock is crystalline
and altered in such a way that the outlines of structures are
partially obscured. The original difficulties attendant upon the
study of these groups are therefore very much enhanced.
Though of so widely different zoological affinities, such is the
preservation in which they appear that in a number of cases I
am not sure that some of the supposed Hydrocoralline are not
calcareous sponges, and I even entertain similar doubts with
regard to certain of Waagen’s representatives of that group. —

Of the echinoderms nothing appears in the collection, nor of
the helminths.

The Bryozoa, except for a single group, are mostly few and ~
of a character alien to our known Carboniferous faunas. They
consist of Fenestella sp., Acanthocladia? sp., a Gontocladia
near G@. indica Waagen, and a peculiar group of fistuliporoid
Bryozoa much more like Mesozoic than Paleozoic types, which

G. A. Girty— Upper Permian in Western Texas. 365

are more abundant in the “dark Permian” than in the white
limestone whose fauna I am attempting to summarize.

The brachiopod fauna is abundant, and while of an undoubted
Carboniferous facies, presents one which is apparently later and
certainly different from that of our usual Carboniferous rocks.
The following species were found:

Streptorhynchus near pelargonatus Schl.
Derbya n. sp.

Derbya sp.

Derbya Bennetti?

Meekella n. sp.

Chonetes n. sp. near trapezoidalis Waagen
Productus semireticulatus var.

Productus pileolus Shum. ?

Productus cora d’Orb. ?

Productus near Wallacianus Derby (2 sp.)
Spirifer Mexicanus Shum.

Spirifer n. sp.

Squamularia? Guadalupensis Shum.
Martinia n. sp. near elongata Waagen
Ambocoelia planiconvexa ?

Spiriferina Billingsi Shum.

Spiriferina sp.

Spiriferina n. sp.

Seminula n. sp.

Hustedia? papillata Shum.

Pugnax n. sp.

Puenax Swallowiana Shum. |
Camarotoechia indentata Shum.
Camarotoechia sp.

Dielasma near truncatum Waagen (2 sp.)
Dielasma ?? sp.

Dielasma ? n. sp.

Hemiptychina near H. sparsiplicata Waagen
Spirigerella ? sp.

Lyttonia sp.

Richthofenia (several sp. ?).

Besides the forms listed are several new genera and species
of terebratuloid shells, which, even in the comparatively few
specimens in our collections, present a really remarkable
differentiation. Many of the genera cited above are new to
North America. Among these may be mentioned Lyttonza,
Lichthofenia, Spirigerella, and Hemiptychina, though in one
case, at least, the generic reference is questionable. Many of
the genera common to our other North American faunas are
represented in part or wholly by species of pecaliar types not
found in them. An orbicular Spiriferina with short hinge
line is one of these.

366 G. Hl. Girty— Upper Permian in Western Texas.

Of the pelecypods the following species were collected:

Aviculopecten (Pterinopecten ?) sp.
Aviculopecten ? sp.

Streblopteria (3 sp.)

Entolium ? (3 sp.)

Lima n. sp.

Euchondria? sp.

Camptonectes ? (2 sp.)

Avicula n. sp.

Myalina? near M. squamosa Sow.
Myoconcha n. sp.

Modiola? sp.

Edmondia? sp.

Cypricardinia? sp.

Macrodon ? sp.

Axinus securus Shum. ?

The aspect of this fauna also is quite varied. The shells are
as a rule small and preserved as internal casts, so that it is
dificult in most cases to ascertain the characters of really
generic value. There is a general absence of the types which
characterize the Upper Carboniferous or even the Permian of
the Mississippi Valley. The development of the pectinoid and
mytiloid groups is rather striking. The genus Myoconcha, of
which I seem to have virtual representatives, has not before
been recognized in North America, and Camptonectes is new

among our Carboniferous faunas. The latter is abundantly

represented by individuals, and has every external appearance
of belonging in the Mesozoic genus.

Gastropods are rare and in size generally minute. So far
only the tollowing have been found:

Pleurotomaria? group of carbonaria (2 sp.)
Pleurotomaria ? (n. gen. ?) sp.

Worthenia? sp.

Euconispira ? sp.

Orthonema ? sp.

Holopella ? sp.

But little need be said of the gastropod fauna, in which there
is nothing striking, except the absence of our familiar Carbon-
iferous species.

Of the cephalopods, at a horizon where novel and interesting
ammonoid types might be expected to occur, only an imperfect
and undetermined species of Zemnocheilus has been found.

No Crustacea have yet been noted, except Phzllipsia peran-

nulata Shumard, which occurs in abundance in the “dark —

ev rer oe

es

G. H. Girty— Upper Permian in Western Texas. 367

Permian” below, but which is in this instance doubtfully
identified.

From the “dark Permian,” some of it in place, but most
from float pieces probably belonging to that horizon, were
obtained the following species:

Fusulina elongata Shum.
Sponges (of several types)
Lophophyllum (2 sp.)
Striatopora ? sp.

Acanthocladia sp.

Fistuliporoid Bryozoa of Mesozoic types
Crania ? sp.

Streptorkynchus ? sp.

Derbya sp.

Chonetes Permiana Shum.
Chonetes n. sp.

Productus Popei Shum.
Productus Popei var.

Productus semireticulatus var.
Productus, semireticulatus type (5 sp.)
Productus cf. Norwoodi Swal.
Spirifer sulcifer Shum. ?
Squamularia ? Guadalupensis Shum. ?
Spiriferina Billingsi Shum.
Hustedia ? papillata Shum.
Hustedia ? (2 sp.)

Pugnax ? bisulcata Shum.
Pugnax ? n. sp.

Pugnax Swallowiana Shum.
Camarotoechia indentata Shum.
Camarotoechia sp.

Dielasma perinflatum Shum. ?
Dielasma cf. truncatum Waagen
Aviculopecten sp.

Myalina ? near squamosa Sow.
Euomphalus n. sp.

Bairdia sp.

Phillipsia perannulata Shum.

This fauna requires little comment. Its relation to that of
the “ white Permian,” nearly a thousand feet above, is obvious,
yet each has an individual facies. In this fauna Phillipsia
perannulata is abundant, and also the curious small fistuli-
poroids to which attention has already been directed. The
Producti are for the most part of the semzreticulatus type,
small, and characterized by strong curvature and a deep
median sinus.

368 G. A. Girty— Upper Permian in Western Texas.

The fauna of the yellow sandstone I will only hastily refer
to. It contains /usulina elongate Shumard and some brachio-
pods, chiefly Producti of somewhat different types from those
above. The major portion of the fauna consists of pelecypods,
many of which are different from those of the limestone, and
are new species.

The black limestone at the base of the section is sparingly
fossiliferous. A number of species were obtained, however,
about 18 in all. This fauna is related to those above, espe-
cially to that of the “white Permian,” and it also includes a
species of Ltzchthofenia.

These faunas are clearly pre-Mesozoic, though containing
certain obvious Mesozoic relations. They are also different
from any American faunas yet known, and are more especially
related to the late Paleozoic faunas of Asia. Among related
faunas may be mentioned that of China, described by Richtho-
fen,* that of India, described by Waagen,f that of the Carnic
Alps, described by Schellwien,{ and, samewhat more remotely,
that of the Palermo province, described by Gemmellaro,§ and
the Permian of England, described by King.|

Not only are these faunas very different from any known in
America elsewhere, but they give evidence of being later in
geologic time. For this reason I propose to give them a
regional name which shall be employed in a force similar to
Mississippian and Pennsylvanian. For this none more appro-
priate than one derived from the locality where they were first
discovered can be found, and the term Guadalupian is sug-
gested. The stratigraphic limits of the Guadalupian period
will have to be determined on intrinsic evidence. At present
it seems to include the whole section at the southern end of
the Guadalupe Mountains, but the central fauna will be that
of the “ white” and “ dark Permian” as described by Shumard.

I plan to resume field work in this region during the season
of 1902, and hope to make important additions to the Guad-
alupian fauna, and to discover its relations with those which
preceded and those which followed it. It is my purpose to
more fully study and to describe this interesting fauna at an
early opportunity.

Washington, D. C.

* China, Richthofen, vol. iv, Berlin, 1883.

+Geol. Surv. India, Paleontologia Indica, Salt Range Fossils, vol. i, 1887.
acai aaa der Trog-Kofelschichten, K. K. Geol. Reichsanstalt, Bd. 16,

§$ La Fauna dei Calcari con Fusulina, Molluscoidea, Palermo, 1898-99.
| Mon. Perm. Foss. England, Paleontograph. Soc., 1850.

ee ee ee

Gooch and Stookey— Reduction of Vanadic Acid, ete. 369

Arr. XXX VI.—The Reduction of Vanadic Acid by the Action
of Llydrochloric Acid, by ¥. A. GoocH and L. B. STooKEy.

[Contributions from the Kent Chemical Laboratory of Yale University—CXII.]

BunsEN* and Mourt have proposed to estimate vanadic
acid by the action of concentrated aqueous hydrochloric acid,
thus reducing it from the condition of oxidation represented
by the pentoxide to that of the tetroxide, in accordance with
the eqnation

V,0,+2HCl = V,0,+H,0+Cl,,

and determining iodometrically the chlorine evolved.

In an attempt by Czudnowiczt to make use of this reaction
by collecting in standard arsenious acid the chlorine evolved
from the boiling mixture of hydrochloric acid and vanadice acid,
and titrating the excess of the arsenious acid by iodine,
accordant results were not obtained. On the other hand,
Gibbs§ was able to determine with a fair degree of accuracy
the small amounts of vanadium pentoxide found in the vana-
dio-tungstates and other complex combinations, by boiling
with strong hydrochloric acid, collecting in potassium iodide
the chlorine evolved, titrating the freed iodine with sodium
thiosnlphate, and calculating from the amount of iodine thus
found the vanadium pentoxide corresponding to a change of
condition from V,O, to V,O,.

Milch| was unable to substantiate the work of Gibbs, and,
obtaining in four experiments, upon approximately 0°2 grm.,
05 grm., and 1:0 grm. of vanadium pentoxide, amounts of
-chlorme varying with the quantity of material handled and
never exceeding one-fourth of the amount demanded by theory,
was led to suppose that the degree of reduction was not con-
stant, being larger in proportion as the quantity of pentoxide
used was smaller. Milch concluded that while accordant
results might be obtained by working with approximately
similar amounts of material, the process afforded no ground for
the calculation of the actual amount of pentoxide present.

Rosenheim4 also tested the process, working with pure van-
adates and pure vanadinm pentoxide. Concentrated hydro-
chloric acid, of sp. gr. 1:19, was employed and care was taken
to dry the distillation flask before introducing the materials,
because even slight dilution of the acid lessened the evolution

* Ann. Chem. (Liebig), xcvi, 265. { Titrirmethode v. Auil., 314.
iAnn. Phys., exx; 17, § Proc. Amer. Acad., x, 200.
|| Inaug. Dissert., Berlin, 1887, 10.

“| Inaug. Dissert., Berlin, 1888; Ann. Chem. (Liebig), ccli, 197.

370 Gooch and Stookey-—Reduction of Vanadic Acid

of chlorine in marked degree. Three stages were noted in the
action of hydrochloric acid upon vanadic acid; the vanadie
acid first dissolved to a solution of a deep brown color without
perceptible evolution of chlorine ; upon warming, the solution
suddenly evolved chlorine and took on a deep green color;
thereafter the evolution of chlorine became weaker, and the
solution, giving off a small amount of chlorine on strong boil-
ing, assumed a clear blue color. After this, further heating
effected no change. It is recommended to raise the tempera-
ture gradually in order that the acid may not boil over into the
receiver, and to keep this receiver cool so that iodine may not
be volatilized by steam from the flask during the prolonged
heating. In the experiments with ammonium vanadate, in
which the heating was interrupted after the first violent evolu-
tion of chlorine, the iodine found in the receiver indicated,
according to Rosenheim’s figures, about 50 per cent of the
vanadium pentoxide present ; in three experiments longer con-
tinued the indications advanced about 6 per cent; and in three
experiments continued presumably to the end—to the appear-
ance of this blue color—the amount of pentoxide indicated
were 61°77 per cent, 63°71 per cent, 61°39 per cent of the
quantities actually taken, and further heating with another
receiver until almost all the hydrochloric acid had been volatil-
ized, resulted in no further progress in the reaction. Experi-
ments with potassium vanadate, and with vanadium pentoxide
resulted similarly. Rosenheim draws the inference that a
titrimetric method based upon a theory of reduction from V,O,
to V,O, by the action of hydrochloric acid, must lead to false
conclusions, and suggests that his results lend probability to
the supposition that an oxide of the formula (V,O,),V,O, is
formed, although the actual reductions observed by him are
not quantitatively in accord with the formation of such a body
—which implies a reduction covering about 66 per cent of the
interval between V,O, and V,O,. Rosenheim coneludes that
the whole method is to be rejected in quantitative estimations.
An inspection of Rosenheim’s tables, however, shows that the
standards of the solutions of thiosulphate and iodine have
been interchanged either in the printed record or in the eal-
culation of results. In the latter event the actual indications of
the process would be about 15 per cent higher than Rosen-
heim’s figures, but still approximately 20 per cent below the
theory of the reduction from V,O, to V,O,.

In Holverscheit’s* hands the process gave results most irregu-
lar and unsatisfactory. Ammonium vanadate was the material
taken for the experiments, and the residue obtained by treat-

* Inaug. Dissert., Berlin, 1890.

by the Action of Hydrochlorie Acid. 371

ing this salt with potassium hydroxide and evaporating to
expel ammonia, was boiled in a Bunsen distillation apparatus
with concentrated hydrochloric acid. The solution, at first
yellow, became green on long boiling. .In some eases ebullition
was kept up for 15 minutes. In working with amounts of
material represented by 0:0585 grm., or 01171 grm. of V,O,
Holverscheit obtained from about 28 per cent to 39 per cent,
or in the average about a third of the theoretical reduction,
and, while noting the fact that by working with wholly con-
eentrated (“‘ganz. konz. HCl”) Rosenheim doubled this yield,
ventures the opinion that it is scarcely conceivable that Gibbs
obtained by boiling vanadic acid with hydrochloric acid and
estimating the chlorine liberated, results even approximating
the truth.

The green color of Holverscheit’s residual solution shows, of
course, an incomplete reduction—the consequence of the use
of an acid which, as intimated, was not of full strength. Ros-

enheim, however, got the blue residual solutions characteristic
of the condition of oxidation corresponding to V,O,. Why
the deficiency in the chlorine indicated in the receiver in Ros-
enheim’s experiments was so marked is not obvious, unless it
be supposed that chlorine was lost from the apparatus in the
sudden evolution which takes place when action begins. Our
experience, moreover, shows that the liberation of chlorine
begins as soon as the hydrochloric acid and vanadie acid come
into contact. In our experiments upon the process, therefore,
we have used a form of apparatus, shown in the accompanying
figure, such that no chlorine can be evolved before the appa-
ratus is connected and ready. For the retort we have nsed a
Voit flask upon the inlet tube of which was sealed a stoppered
funnel, while the outlet tube was sealed to a Drexel wash-
bottle and this in turn was joined to Will and Varrentrapp
bulbs. The Drexel bottle and bulbs used as receiver and trap

372 Gooch and Stookey—Reduction of Vanadic Acid

were charged with water containing 3 grm. of potassium
iodide ; the vanadate was introduced into the dry flask and the
apparatus was adjusted. Hydrochloric acid was put in the
stoppered separating funnel and, when all was ready, allowed
to run into the flask. Connection was made between the
funnel and a carbon dioxide generator, so that by passing ear-
bon dioxide through the apparatus all danger of regurgitation
might be avoided. Sometimes in the course of our work, too,
a branched connection with the funnel tube was so arranged
that either hydrochloric acid gas or carbon dioxide or both
might, at pleasure, be sent into the apparatus. We have used
in the experiments to be detailed ammonium vanadate either
purchased in pure condition or made by precipitation with
alcohol, or prepared from the oxychloride obtained by acting
with chloroform upon ignited vanadium pentoxide. In each
case the content in vanadium pentoxide was determined by
ignition and confirmed by Holverscheit’s* bromide process.

In Table I A are gathered results obtained by the use of
the ordinary pure hydrochloric acid of the laboratory of sp. gr.
1:17; while in B of the same table are recorded results got by
using acid of sp. gr. 1°20, specially prepared by saturation of
the former acid kept cold by ice, with hydrochlorie acid gas.

In these experiments, in which the precautions described to
catch all the chlorine evolved were taken, the actual registra-
tion of chlorine amounts in the average to 92°62 per cent of
the theory when the ordinary pure acid of the laboratory of
sp. gr. 1:17 was used, and to 94°46 per cent when the specially
concentrated acid of sp. gr. 1:20 was employed. While the
error of these determinations would amount to a considerable
figure when large amounts of vanadic acid are to be deter-
mined, it is obvious that a correction of, let us say, 6 per cent
or 8 per cent upon the small proportions of vanadium pen-
toxide present in many of the complex salts to which Gibbs
applied this method, would not change the significance of the
analysis. Indeed, reference to the original shows that the
formule deduced by Gibbs would in many cases apply as well
to the corrected results as to the original figures. Thus it
appears that, so far as concerns the special cases in which Gibbs
used this method of analysis, the conclusion of Milch that the
process affords no ground for the calculation of the amounts
of pentoxide present, and Holverscheit’s difficulty in suppos-
ing that results were even approximately true, prove to be
unfounded, while Rosenheim’s supposititious oxide disappears.

In the light of these results it is of interest to note the effect
of a repetition of this treatment with strong acid of the residue

*Tnaug. Dissert., p. 48.

373

Aci

TG

of Hydrochlor

~on O

by the Act

99 99 op)

9 ) aan) ”
9 ” ”

9 99 99

peBinjzVs-at UOYM UMOAIG f oNnTG 0} pol[log

”
LO[O9O ONG 0} po[lod

” ”

” ”

99 99

JoJOO onjq 0} soynuIM ¢ pallog
‘JU9T}VIT} FO STIVJOd

G&L 9200-0
90-4 F000
CO.V 1600.0
G0-P T&00-0
Lé-P 6600 0
I¥-9 6700-0

ad
90-G 9110-0
[Ph OL10-0
F8-L 0900-0
19-9 1¢00-0
G&-L 9900-0
80-0T 600-0
‘4ue0 Log “WI
“OFA FO SULLE} UL

LOITG

vr

‘[ W1av,

60L0-0
1140-0
7810-0
FEL0-0
ZEL0-0
91L0-0

6414-0

C616-0
040-0
FL L0-0
6040-0
8480-0

m1i3
“punof
g O2A

910-0
¢940-0
G9L0-0
G940-0
G910-0
G940-0

[666-0
2662-0
G9L0-0
G910-0
G9L0-0
1860-0

A UOM Re}
‘MOLye[NOTeo
fq °OFA

9

99

>y)

by)

9

Ou ato cds

9)

oy)

oy)

by)

9)

99

9)

9)

99

93

29

>)

99

99

9

GG
GG
iss
O&
O€
O&

GG

06

OV
OF
OF

LL-T ‘18 ‘ds 0g

eUld
“TAye4y
10H

OO00T-0O
OOOT-0
OOOT-0
OOOT-O
OOOT-0
OOOT-O

000€-0
0008-0
0O00T-0
OO0T-0
OOOT-O
0090-0

UIs
"Ua ye4y
"OATHN

Acid

(ZG

of Vanad

don O

Gooch and Stookey—Leduct

3874

"m90Ly tf
“09° A F000-0 0} JuayeAinbo ag Avs AgTy
YA 10 [OH pozerzuoouoy yA on[q onpised
” g
” ¢:
SoyNUlU OZ po[log
“O'A 1000-0 07 4U9]
-vAInba 1g OAV IGY YAM poyvoi, onpiser
9) 99
” ”
99 99
ssoudip 0} poyeiodeaq
sayNUIUL GT poplog
TOT pee
-U90U0D YIM poejvor}, UoYyAM UMOIG oONpIset
3) 9

3) 9).
ssoudip 07 poyerodeaq
‘qUOTAYvEIY FO S[IejJod

‘ong +

41000-0
46200-0
¥£900-0

40100-0
Udy
te Ba

6900-0

¢L00-0

‘ulIs
“LOLLA

‘TI aTav T,

9)

”

6910-0

9?

9

6910-0
"TIS

“peye[noleo

207A

“UMOIg x

9)
{sex
Oe oS: ms OF

” ” tL
” oy) xh
” ” xh

OG: a0 eas OF

” ” G
” ” *G
OMT PLease sec0T
‘Ua ye}
IOH

OOOT-0

OOOT-0

OOOT-0

"UIs
"Ma yey
°OA°HN

by the Action of Hydrochloric Acid. Y 8%

of the first action weakened by boiling. Plainly two methods
of bringing about the re-treatment with strong acid are open:
either the weak acid remaining after the boiling process may
be wholly removed by evaporation and replaced by concen-
trated acid, or it may be cooled and re-saturated with gaseous
acid. Table II contains the record of results obtained in both
these ways.

These experiments show very clearly that several successive
treatments with strongest hydrochloric acid, consisting either
in the addition of the concentrated aqueous acid to the dry
residue in successive portions, or in passing gaseous acid through
the residual liquid cooled in ice-water, finally bring the residue
to a condition of reduction in which the recharging of the
liquid residue with gaseous acid produces no brown or green
color but a clear blue. When this condition is reached the
Holverscheit procedure produces no appreciable further reduc-
tion. The holding of the blue color when the solution cooled
in ice-water is thoroughly charged with the gaseous acid, appears

to be the best evidence of the complete reduction of V,O, to
_V,O,; it is not sufficient that the boiled solution, in which the
acid has been weakened, should be blue.

In the several manipulations of the repeated processes there
is likely to be some mechanical loss of chlorine. For this
reason, the direct titration of the residues by potassium per-
manganate in presence of a manganous salt was combined, as a
control, with the determinations of the iodine set free by
chlorine in the distillate, as shown in the experiments of
Mable Til... -.

In these experiments we have again evidence that it is pos-
sible to effect the reduction of vanadic acid to within a few
per cent of the amount present by a single treatment with con-
centrated hydrochloric acid, and that the amount of the reduc-
tion may be determined by titrating the residue with potassium
permanganate. It is plain, therefore, that Gibbs’ application
of this mode of reduction and titration to the determination
of the small proportions of vanadium pentoxide found in many
of the complex salts studied by him is also justified.

The experiments show also that when the action of hydro-
chlorie acid is sufficiently continued, ordinarily large amounts
of vanadic acid may be completely reduced to the condition of
the tetroxide; and the method of reduction is of special
advantage in those cases which eall for titration of the tetrox-
ide by permanganate, since in such cases the use of Holver-
scheit’s admirable method of reduction by hydrobromic acid is
precluded.

Am. Jour. Scl.—FourtTH SERIES, Vout. XIV, No. 83.—NOVEMBER, 1902.

376

NH,VO;
erm.

0°1000
0°1000
0°1000
0°1000

0:1000

0°3000

0°3000

0°3000

Gooch and Stookey—Reduction of Vanadic Acid, ete.

Aq. HCl
em?,
25 sp. gr. 1°20
95 ce ce
25 66 (7%

30 sp. gr. 1°20
Gast

25 sp. gr. 1°20
Gast
Gast

25 sp. gr. 1°20
Gast
Gas{

25 sp. gr. 1°20
Gast

25 sp. gr. 1°17
Gast
Gast
Gast
Gast

V2.0;
calcu-
lated.
erm.
0°0765
0°0765
0°0765

0°0765

---+-

0°0765

TABLE III.

V205
found by

chlorine in

distillate.
germ.

\=-s-+

00711
0°0734

0°0714
0°0746
0:0765

0°2259
0°2288

0°2207
0°2244
0°2258
0°2269
0°2281

* Residue after boiling was blue.
+ Residue blue after boiling was still blue when re-saturated.
t Residue was cooled in ice-water and re-saturated with gaseous HCl.

Error.
erm.

e-~-=
----

0°0036
0°0007

0°0088
0°0051
0°0037
0°0026
0°0014

V205-

found by —

KMn0O, as
residue.

Error.

grm.
0:0025 —
00029 —
0°0021—

00007 —

0:'0000

0:0011—

0°0007 +

a

John Wesley Powell. 377

JOHN WESLEY POWELL,

Founder and Director of the Bureau of Ethnology of the
Smithsonian Institution, for thirteen years Director of the
United States Geological Survey, died at Haven, Me., Sep-
tember 23d, in his sixty-ninth year.

Althongh well known locally for his seientifie work and
enthusiasm before the civil war, and acquiring military repn-
tation in that war, he first came prominently before the
scientific world and the wider general public by his daring
exploration of the great Colorado Canyon in 1869, and since
that date has been a conspicuous personage among American
scientists for his zeal for science, his eminent administrative
ability—shown in the organization and management of geo-
graphical and geological surveys and scientific bureaus—his
broad grasp of scientific questions, his varied activities in the
promotion of research in several branches of science, and by a
charming personality.

He was born March 24th, 1834, at Mount Morris, then a
small village in the Genesee Valley of Western New York.
His parents were English, his father a Methodist clergyman
who came to this country but a short time before the birth of
his son. The requirements of his profession caused many
changes of home, and the family moved to Ohio in his early
childhood ; eight years afterwards to Wisconsin; and again,
when the boy was fifteen years old, to Illinois, which was young
Powell’s home until the breaking out of the civil war, in his
twenty-seventh year. |

The boy was an ardent lover of nature and the migratory
home of the family during the days of his youth and early
manhood gave him unusual facilities to see outdoor nature
under many aspects, but the conditions of his environment were
very unfavorable for obtaining a college or university training.
He was fond of roaming, a keen observer, and in his studies
was from the first strongly attracted to the natural sciences,
especially such of them as could be pursued out of doors. He
studied botany and geology and used every opportunity to
learn these and the kindred sciences. He was fora while in the
Illinois College at Jacksonville, later in Wheaton College and
still later in Oberlin College in Ohio. Unable to attend any
of these continuously, he alternated between teaching school
and studying when and where the opportunity occurred. To
him there was no continuing curriculum or studies available and,
looking to an academic degree, he studied as the opportunities
offered, now while teaching in some country or village school,
then as a temporary student in some college, or,while roaming.

378 John Wesley Powell.

He made excursions and collected specimens which found
their way into the museums of the several colleges and soci-
eties with which he had: been connected. Some of these
excursions are noteworthy. He journeyed to St. Paul on the
Mississippi and across Wisconsin to Mackinaw. In 1856 he
descended the Mississippi alone in a row-boat from the falls of
St. Anthony to its mouth, making collections on the way. In
1857 he rowed the whole length of the Ohio River, from
Pittsburg to its mouth, and in the fall of that year studied the
geology and mineralogy of the Iron Mountain Region in Mis-
souri. In 1858 he made a trip down the Lllinois River to its
mouth, and up the Des Moines River, returning, as usual, laden
with specimens.

Meanwhile he had become a member of several local
scientific societies as well as colleges. These institutions had
given him much encouragement and some facilities in the pros-
ecution of journeys, but this encouragement was moral rather
than pecuniary and the necessary funds for the excursions—
explorations we may call them—he was obliged to earn for,
himself by teaching during a portion of each year.

All this brought him into acquaintance with a great variety
of people, scientific and otherwise, and the experience was rich
in incident and adventure. I have some most pleasing recol-
lections of the charming way in which he recounted some of
these experiences to his intimate friends; of the enthusiasm
and humor with which the stories were told, and the touches
of philosophy with which they were embellished.

Such was the great university where he was educated.
What a training it was for his future career! It often required
as delicate tact, as careful diplomacy, as ingenious planning
and skillful management, as enthusiastic argument and as per-
sistent effort to carry out his plans to success, as it did later to
deal with politicians in and out of congress and to successfully
carry out the great works with which his name will be asso-
ciated so long as science shall be studied. Those row-boat
excursions on the gentler currents of the Ohio, Mississippi, and
Illinois, were the forerunners of the daring one through the
madly rushing waters in the great canyon, and the plans of
the great Geographical and Geological Survey of the nation
had their elements in these earlier trips.

His school changed when the great rebellion broke out. He
enlisted as a private in the army and rose through the successive
steps of lieutenant, captain, and major, in which office he
lost his right arm at the battle of Shiloh. As soon as he
recovered he returned to his post and continued in the service
until the very end. He was made lieutenant-colonel and in
the last days of the war received the commission of colonel,

John Wesley Powell. 379

which he declined, not wishing to enter the profession of war,
but the military title of major clung to him through life.
His war experiences may here be considered as a sort of post-
graduate study, following the gentler training of previous
years, a schooling, both as private soldier and commander of a
regiment, a training in the management of men in both field
and office, and for bolder exploration.

The war over, he refused a lucrative civil position in his own
town, as he already had a higher military one, and accepted a
much less remunerative position as Professor of Geology, and
Curator of the Museum of the Illinois Wesleyan University at
Bloomington, which was followed later by a similar position in
the Illinois Normal University.

In 1867, Professor Powell visited the Rocky Mountains of
Colorado, taking with him his class in geology, for the double
purpose of exploration and research, and for the instruction of
his students in field work. He was practically the pioneer in
the actual and practical introduction of extensive expeditions
with students as a part of their college training for future field
work, a phase of college instruction since so extensively prac-
ticed and which has been so rich in results.

Major Powell, on this excursion, became interested in the
Colorado Canyon and its surroundings. For a century or more
vague rumors of this region and its wonders had reached the
outer world: stories of its awful and mysterious chasms, abso-
lutely impassable and entirely preventing passage over the
region. The stories became much more numerous and the
information more definite after 1850, when the gold-seekers
attempted to reach California; but curiously little was accu-
rately known more than that the waters of the Rocky Moun-
tains, from as far north as the 48d parallel, found their way
through an awful canyon hundreds of miles in length to the
borders of California and thence to the Gulf. It was reputed
to be many thousands of feet deep. One and another had
been on its brink here and there; that was about all, except
the disappearance of luckless travelers who had by accident,
got into it at certain points.

In 1868 Major Powell organized a little party of movufntain-
eers and others and explored a portion of the region, studied
the problem, resolved on the exploration of the canyon, and
finally went into winter quarters on the White River. From
there he made further reconnaissances and other preparations
for the bold work he had planned.

The transcontinental railroad, then under construction, had
progressed far enough to bring in such supplies and appliances
as were not otherwise obtainable, and in the early spring of
1869 the expedition left its winter quarters on White River

380 John Wesley Powell.

and proceeded to where the Union Pacific Railroad crosses the
Green River. Four small row-boats had been built in Chicago
for the specially dangerous voyage and transported on the still
unfinished railway to this place.

The small party consisted of but ten men. Their make-up
is noteworthy. Of his nine companions, Major Powell
describes four as having served in the army—in the war
recently over; three were described as ‘“‘ hunters and trappers ”—
‘Indian fighters” is incidentally mentioned ; one as “a pen-
sive young man;”’ and one was an Englishman “looking for a
glorious trip.”

With rations sufficient to last ten months, the little fleet
started on its perilous trip on May 24th. The departure is
briefly chronicled in the narrative. The people at the cross-
ing turn out to see them start, he tells us, and “ we raise our

little flag, push the boats from shore, and the swift current —

carries us down.” They disappear from the outside world

and emerge from the mouth of the Grand Canyon August 29,

and the next day arrive at the mouth of the Virgin River.

But not all of them. Only a few days earlier, the dangers
of the passage becoming even greater than before, and a place
occurring where it was thought they could get out of the
canyon, ahead of them rapids or falls that seemed more dan-
gerous than any before encountered, the dark abyss beyond
visible but a short distance, three of the men resolved to leave
the party while they could. They took out with them duplicate
notes, that the results of the trip might not be lost with the
party. They made their escape, and along the river below
watched for fragments and traces of their abandoned com-
panions.

Of the voyage itself and its brilliant success, of the results
that grew from it, of the adventures and experiences encoun-
tered, I need not speak further. The subject forms a brilliant
chapter in the annals of exploration and adventure in the
interests of science.

Never was a bolder voyage planned and executed. I know
of no equal in the annals of exploration and navigation.
Whilé comparisons between this and polar exploration are
difficult, yet, there were in this features of the possibilities
which seemed such eminent probabilities of disaster, and the
dangers were of such a kind, as to deter the attempt. The
disastrous end of the expedition, and the manner in which it
would come about, seemed so plain that several of our enter-
prising newspapers published more or less minute accounts of
its sad end; all of the party but one being lost was the most
common plan of the tales; the nature of the dangers were such
that one of the party had to be saved or no story written, ex-

ee oe |

ee ee

x

John Wesley Powell. 381

cept the disappearance of the party at an unknown time and
place.

: Some years later, while smoking an after-dinner cigar with
some of his friends, he gave his reasons for his faith in under-
taking it. I told him that for some years previous to his
famous trip, | had been much interested in that canyon and
had picked up all the rumors and information pertaining to it
that I could, and being in Colorado while he was making the
trip was intensely anxious as to his fate, for I thought it was a
mad scheme; the canyon was a long and vastly deep one, cut
mostly in strata lying relatively level, that owing to unequal
hardness the erosion created waterfalls ; that I had been reared
in central New York where such waterfalls were especially
numerous, cited Niagara and various other examples elsewhere ;
that this long and deepest canyon in the world was mostly in
such rocks; that he embarked on the river at over 6000 feet
elevation, to emerge some 500 or more miles below at nearly the
sea level, the river having an average fall of ten or fifteen feet
per mile and I had assumed that there must be great falls, and
that the explorer must approach them from above.

He answered in substance, ‘“ Have you never seen the river?
Jt is the muddiest river you ever saw. I was confident that I
would find no considerable falls. Rapids I expected, of course,
but not falls. I was convinced that the canyon was old
enough, and the muddy water swift enough and gritty enough
to have worn down all the falls to mere rapids. I entered the
canyon with confidence that I would have no high falls to stop
us, although there might be bad rapids, and [ believed that we
might overcome them in some way,—and we did.”

The next year he induced Congress to establish a geological
and topographical survey of the Colorado River and its tribu-
taries; it was placed under his direction and on it he was en-
gaged much of the following ten years.

Incident to this, he became interested in the study of the
arid regions and the problem of their improvement, also the
impounding of the floods of the western rivers for the double
purpose of controlling the floods and using the water for irri-
gation. The present hydrographic survey of the country is
the outcome of his interest in this matter.

Between 1865 and 1875 many surveys in the western country
were established, acting independently of each other, often in
competition as well as rivalry, but not mutually helpful, and
working under different departments of government.

Major Powell took an early.and active part in the efforts that
came up for a more satisfactory adjustment of these and their
unification under a more rational system of operation. After
much agitation, discussion and opposition, Congress finally, in

382 | John Wesley Powell.

March, 1879, discontinued the separate surveys and established
the United States Geological and Geographical Survey, under
the Department of. the Interior, and Clarence King was
appointed Director.

Beginning with his first visit to the Rocky Mountains, Minar
Powell began ethnological and anthropological studies ‘of the
American Indians, for the Smithsonian Institution.

Ten years later, in 1876, Professor Henry, the then secre-
tary, placed this along with other accumulated material per-
taining to this subject in his hands, for arrangement and
publication. The next year, his first volume of “ Contribu-
tions to- North American Ethnology” was published by the
Geological Survey. Later followed five more of these quarto
volumes under the same auspices.

With the establishment by Congress of the Geological Survey,
the Bureau of Anthropology was also established in the Smith-
sonian Institution and Major Powell made Director, an office
he retained twenty-three years and until his death, and the
annual volumes of ‘Contributions’ have continued in the
same general form.

Mr. King resigned the directorship of the Geological Survey

in 1881, and Major Powell was appointed his successor, retain-
ing however the direction of the Ethnological Bureau, ‘and for
thirteen years he administered both offices, and both institutions
were greatly widened in their work and improved in their
methods under his administration.

In 1894 he resigned the office of Director of the Geological
Survey and since then has devoted himself more to other work,
ethnological, anthropological, psychological and philosophical.

Major Powell was endowed with an eminently philosophical
mind, had great administrative ability, was rich in suggestions
and fertile in originating and planning, in devising new work
and methods and in improving old ones; had a personality of
great force, persuasive in inducing men to ‘do, and he inspired the
confidence of those with whom he held official or social rela-
tions. He was a powerful advocate of reform in laws affecting
the permanent welfare of the West and was for many years
one of the most conspicuous personages in the scientific corps
under the government. He was a member of the National
Academy of Sciences and of other societies and clubs, and
several colleges and universities conferred academic degrees
upon him.

Major Powell was a faithful and genial friend, and his most
interesting individuality made him many friends. He died of
apoplexy. His wife and a daughter survive him.

Wm. H. BREWER.

New Haven, Conn.

Chemistry and Physies. 383

SCIENTIFIC INTELLIGENCE.

I. CHEMISTRY AND PHYSICS.

1. The Combination of Hydrogen and Oxygen.—It is well known
that many chemical reactions do not take place in the absence of
moisture, but heretofore it has not been shown that the combina-
tion of hydrogen and oxygen could be prevented in this manner.
By using a new method for preparing the gases, the electrolysis
of barium hydroxide solution, H. BREREToN Baker has succeeded
in obtaining the mixture sufficiently pure and dry to withstand
heating to redness without combination. Even when a coil of
silver wire was heated to its melting point (over 1000°) in contact
with the detonating gas it did not cause combination, but red-hot
platinum wire and electric sparks produced explosions. The
ordinary ignition-point of these gases is 600°. The interesting
fact was noticed that when the tubes in which the experiments
were made were subjected to only two days’ drying in the pres-
ence of phosphoric oxide, instead of a period of ten days, which
was employed for thorough drying, water was formed by heat-
ing and combination took place slowly without explosion. This
is a result which is entirely different from what takes place,
either with the moist or the perfectly dried gases. This beha-
vior is explained by the hypothesis that without an electrolyte no
chemical action is possible, and that since the water formed by
the union of very pure gases is itself very pure, it is not suffi-
ciently electrolytic to produce an explosion. During this inves-
tigation the important observation was made that hydrogen and
oxygen combine slowly when exposed to direct sunlight if they
are moist, but do not combine under this influence when they
are perfectly dry. The exposure to sunlight took place outside a
south window for four months, and the contraction of the moist
mixture amounted to one twenty-third of its volume. On account
of this result care was taken to dry the gases in darkness in pre-
paring them for the experiments which have been described.—
Jour. Chem. Soc., 1xxi, 400. snd Wr

2. Arsenic in the Animal Organism.—Several investigators,
especially Gautier, have recently arrived at the conclusion that
minute quantities of arsenic exist normally in the animal organ-
ism, a fact which had been previously denied. The same conclu-
sion has now been reached by Brerrranp after a series of very
careful experiments. The human thyroid gland, thymus, skin,
and other organs, have been found to contain arsenic, but Ber-
trand has avoided experiments with human tissues, because it is
impossible to be sure that the individuals have not been subjected
to treatment with arsenical medicines, or have not been otherwise
contaminated with compounds of arsenic. He has also not at-
tempted experiments with horses, because these animals are
sometimes treated with arsenious acid. Investigations were

384 Scientific Intelligence.

therefore carried out with the thyroid glands of calves and swine,
then with the bristles of the latter animal, the horns of beeves,
the feathers of geese, the hair and claws of dogs, etc. It was soon
found that the horny and hairy products contained even more
arsenic than the thyroid glands. The horns of oxen gave arsenic
in the relatively enormous amount of 5™S per kilo. Traces of
arsenic were found in the hoofs, hide, and liver of a calf one
month old, while larger quantities were evidently present in the
corresponding parts of a heifer 18 months of age, although the
horns of the latter contained less of it than those of more mature
animals. It appears, consequently, that arsenic accumulates with
the age of the animal. As a convincing proof of the normal
occurrence of arsenic in an animal which had not breathed air
contaminated by industrial fumes, the element was found in the
thyroid glands of a seal which was captured in the vicinity of
Spitzbergen.— Bulletin, xxvii, 847. H. L. W.

3. Atomic Weight of Radium. —By fractional crystallization of
the greater part of the radiferous barium at her disposal, Mux.
CuriE has succeeded in obtaining about a decigram of radium
chloride which appeared to be perfectly pure. Atomic weight
determinations were made with this material with the result that
225 was found to be the atomic weight of the element, assuming
that it is bivalent. Radium, therefore, appears to be a higher
homologue of barium, with a place in the periodic table under
barium and in a line with thorium and uranium. It is stated
that pure anhydrous radium chloride is: spontaneously luminous.
— Comptes Rendus, cxxv, 161. H. L. W.

4. Double and Triple Thiocyanates.—A number of these com-
pounds has been prepared by We ts and his pupils. Most of
the double salts contained czsium as the alkali-metal, while ferric,
lead, mercuric, cuprous, silver, thallous, magnesium, zine, ¢al-
cium and strontium compounds were obtained. It appears that
the double thiocyanates are not generally formed in as great
variety as the chlorides, bromides and iodides, but most of those
obtained corresponded to known double halides. The calcium
and strontium double thiocyanates, 2CsSCN:Ca(SCN),-3H,0.
and 2CsSCN‘Sr(SCN), .4H,O are of considerable interest, because
double salts containing these alkali-earth metals as the negative
constituent are very rare.

The triple salts show great variety in composition. Several of
them crystallize remarkably well, and can be recrystallized from
water. Fourteen of these compounds were obtained, belonging
to seven different types, as follows:

. One salt, ; CsAgZn(SCN),-H,O
II. One salt, . Cs,AgZn(SCN),
Ill. Five salts, analogous to Cs,Ag,Ca(SCN ),:2 H,O
IV. Four salts, analogous to Cs wAg, Ba(SCN).
V. One salt, i K Ag, Ba(SCN), ‘H, O
Wily One salt, ; : CsAg,Zn, (SCN),
VII. One salt, : ; CsAg.Zn,(SCN),

eo Ce

Chemistry and Physics. 385

The other salts belonging to type III have magnesium, man-
ganese and nickel in place of calcium, and in one instance cu- »
prous copper in place of silver. The replacements of barium by
strontium and silver by copper give the other members of type IV.
The four members of the latter type are isomorphous and resemble
apophyllite in form.—Amer. Chem. Jour., xxvill, 245. H. L. W.
5. The Determination of Thallium in the Thallous State.—To
make this determination THomas adds auric bromide solution in
excess to the warm thallous solution, then keeps the liquid warm
for eight or ten hours, and finally filters off the metallic gold
which is produced and weighs it. Three molecules of the thal-
lous salt precipitate two atoms of gold. The test-analyses given
are very satisfactory. The method will evidently be useful in
analyzing substances containing both thallous and thallic com-
pounds.— Bull. Soc. Chim., xxvii, 470. H. L. W.
6. Chemisches Praktikum, von Dr. A. WoLFRuM. 12mo, pp.
562. Leipsic, 1902 (Engelmann).—This volume is the first
part, dealing with analytical operations, of a guide in laboratory
work. The course is designed ona practical basis with the object
of giving the student an idea of factory operations, and of en-
couraging the study of technical chemistry. This feature will
doubtless appear still more conspicuously in the second part, not
yet published, which will deal with chemical preparations. The
volume under consideration covers a wide field of analytical
chemistry in a very compact form. It embraces a course in
quantitative analysis for the common metals and acid radicals, as
well as the detection of the more important groups in organic
compounds and the application of these tests to a large number
of natural and artificial organic products. Quantitative analysis
is presented in all its branches; not only the gravimetric and
volumetric methods for determining inorganic substances are
included, but also the quantitative determination of organic
groups, gas analysis, molecular weight determination, toxicologi-
cal analysis, and many methods of technical analysis. The
methods appear to be well chosen, so that the book will be use-
ful, not only to students pursuing courses in laboratory work,
but also to analytical and technical chemists. Heb W:
Vacuum Thermo Element.—PrtER LEeBEpEw discusses the
ae caaces of enclosing a thermo element in rarified air. Kundt
and Marbury have shown that a body loses heat less rapidly in a
vacuum, and this fact doubtless lies at the root of the superior
efficiency of a bolometer or thermo element in a vacuum. The
author employs a thermo element of platinum constantin, diam-
eter of the wire d= 0°025™". The sensibility steadily increases
and reaches its highest point at a pressure of 0°01™™. It is
believed that this method will allow measurements of electrical
waves, which hitherto could not be made.—Ann. der Physik, No.
9, 1902. pp. 209-213. J. T.
8. Electromotive Force of Ozone.—In the course of an investiga-
tion on the possibility of converting the heat of deozonization

386 Scientific Intelligence.

into electrical work, A. Branp determined the electromotive force
of ozone. The electromotive force of the OQ, and the O, electrode
of a Grove gas cell was measured against a normal mercury ele-
ment Ce covered with mercuric sulphide), in a normal sul-
phuric acid( 3H,SO, g. to the liter). It was found that the electro-
motive force with increasing proportion of ozone approached a
limit which at 17° was 0°919; at 0°, 0°950 volt.—Anmn.. der
Physik, No. 10, 1902, pp. 468-474, Joa

9. Influence of Electrification of the Air on Electric Sparks.—
The passage of electricity through a gas is such a complicated
phenomenon and is influenced by so many conditions that it is
dificult to frame a comprehensive theory of it. Ernst LecHEr
shows that a hitherto unknown condition results from electrifying
the air or gas through which the electric spark is discharged.
For this experiment two sources of electrification are employed.
An induction coil produces the spark and a Holtz machine serves
to electrify the dielectric. Many interesting phenomena result

from this arrangement. The author believes that the phenomena

are due to an ionization of the discharge space.—Ann. der Physik,
No. 10, 1902, pp. 442-451. 3. T.
10. Onanew Reaction between Electrostatic Tubes and Insulators.
—M. W. ve Nicoxarzve believes that certain attractions and
repulsions due to an electrostatic field set up in the process of
electr olysis form a new phenomenon and confirm Professor Poynt-
ing’s theory of the action of tubes of force.— Phil. Mag., No. 19,
July, 1902, pp. 133-138. Ji
ddd Magnetic Detector for electrical waves.—In a communica-
tion to the Royal. Society, of London, Marconr describes his
magnetic detector. Ona core of thin iron wires is wound a coil
consisting of one or two layers of insulated copper wire, and over
this and separated from it by insulating material is wound ‘a
second and longer coil. The ends of the inner coil are connected
to earth and the aerial conductor ; and the ends of the outer coil
to a telephone. The iron core is magnetized by a permanent
magnet at one end, which is rotated by clockwork so as to cause
a continual slow change in the magnetization. The magnetization
lays behind the magnetic force, owing to hysteresis ; but when a
high frequency current passes through the inner winding there is
a decrease in the hysteresis. A sudden variation in the magnetiza-
tion of the iron results, and this results in inducing a current in
the coil connected to the telephone. ‘This receiver is more sensi-
tive than the coherer. Experiments have been carried out between
points 135 miles apart. This detector seems to be Rutherford’s
magnetic detector.—Wature, July 31, 1902, p. 334. ba
12. Radiations from Radioactive Substances.—Professor Ruru-
ERFORD and Miss Brooks conclude from extended observations
that radioactivity is a very complicated phenomenon. Both
uranium and radium emit negatively charged particles with high
velocities, similar in all respects to cathode rays. Uranium,
radium and thorium emit rays which are not deflected by the

Ss

Chemistry and Physics. 387

magnetic field and are absorbed by gases and thin layers of metal.
These rays differ from one another in penetrating power. The
emanations from thorium and radium differ greatly in their rates
of decay of radiating power. The presence of emanations gives
rise to a complicated phenomenon of ‘excited ” radioactivity.
Elster and Geitel have recently shown that a negatively charged
wire in the open air, free from all possible contamination of radio-
active substances, becomes strongly radioactive. This activity
decays at a different rate from that due to thorium and radium
and is of greater penetrating power.— Phil. Mag., No. 19, July,
1902, pp. 1-28. B os
13. Induced Radioactivity in Air.—lIt has been shown by Elster
and Geitel that a strongly electrified wire becomes after several
hours strongly radioactive. The authors conclude that the atmos-
phere contains some radioactive substance which is attracted to
the wire, and that this emanation is like that from thorium.
Professor J. J. THomson, however, comes to the conclusion that
it 1s not necessary to make this assumption and believes that
negatively electrified surfaces may become radioactive without
the deposition upon them of substances possessing radioactive
properties. He found that a negatively electrified wire placed in
a closed vessel, the enclosed air or gas being exposed to Rontgen
rays, became radioactive. When the enclosed gas had been
bubbled through air the effect was very large. Professor Thom-
son thinks that the ionizing power of the wire may be due to a
kind of polarization which produces an electric field which makes
the wire into a cathode emitting cathode rays of feeble penetrat-
ing power, which ionize the gas in the neighborhood of the wire.
— Phil. Mag., No. 21, 1902, pp. 352-367. B Peat
14. Deviable Rays of Radioactive Substances.—Professor Rutu-
ERFORD and Mr. Grier have discovered that uranium, thorium
and radium all emit both deviable and non-deviable rays. They
differ from polonium, which does not emit deviable rays. Uranium
gives out more deviable rays than radium or thorium. The authors
believe that most of the deviable rays are given out by a second-
ary product, produced by a disintegration of the uranium or
thorium atom or molecule. These secondary products differ from
the uranium or thorium in chemical properties.— Phil. Mag., No.
21, Sept., 1902, pp. 315-330. 3. T.
15. Removal of Negative Electricity from the air by falling
drops of water.—A. Scumauss shows that there is a possibility
that each rain drop passing through higher layers of air rich in
ions draws with it negative ions and conveys them to the earth.
—Ann. der Physik, No. 9, 1902, pp. 224-237. 5 A
16. Lhe Elements of Experimental Phonetics ; by Epwarp |
WHEELER SCRIPTURE. Pp. xvi+627. Charles Scribner’s Sons.
Yale Bicentennial Publications.—Valuable contributions have
been made by Professor Scripture to the analysis of speech
sounds by means of his ingenious apparatus for enlarging and
tracing the curves recorded by the graphophone, and his experi-

388 Scientific Intelligence.

ments are fully described and a large number of curves given in
a series of plates at the end of the volume. The careful analysis
of these curves leads to a number of most interesting and impor-
tant results, and enables the author to arrive at a definite conclu-
sion upon several controverted points, such as the action of the
vocal cords, the theory of vowel sounds, the melody and rhythm
of speech, the duration of long and short vowels, etc. In regard
to the action of the vocal cords, the opinion is strongly main-
tained that they do not vibrate like stretched strings or mem-
branes, but, at least in the chest register, act by alternately open-
ing and closing the orifice between them, somewhat (one would
judge) in the manner of a striking reed blown from the wrong
side ; the mechanics of such a vibration is not altogether obvi-
ous, but, putting aside this difficulty, the hypothesis makes the
cord tone consist of a series of puffs of air similar to those pro-
duced by a siren, and more or less explosive in character. Many
of the curves appear to support this view. One of the conse-
quences, however, which the author draws from this theory, and
insists upon repeatedly (pp. 41, 94, 97, 263, ete.), cannot be
accepted : it is that the cord tone, since it consists of such a series
of explosive puffs, cannot therefore be resolved into simple har-
monic (sinusoidal) constituents by the use of Fourier’s Theorem.
As a matter of fact, a series of explosive increments and decre-
ments of the air pressure can be expressed by a Fourier expan-
sion, even when the rise and fall are assumed to take place really
instantaneously, that is, in a mathematically infinitesimal time,
and for however small a fraction of the interval between two
puffs the increased pressure endures; if the puffs are all alike
and occur at equal intervals of time, then the lowest term in the
Fourier series will have a frequency equal to the number of puffs
in a second ; if they are not all similar the series will begin with
a term of lower frequency, but in any physical case there can be
no doubt that the expansion is possible, and any argument based
upon the supposed non-harmonic character of the cord-tone will
be fallacious. The convenience and utility of such an expansion
in any given case is another question. In the discussion of the
vowel sounds, some confusion is introduced by this erroneous
assumption; but much experimental evidence is adduced in sup-
port of the theory of Willis, according to which the resonance
tones characteristic of each vowel are fixed in pitch and have no
necessary relation to the cord tone, and against Helmholtz’s
theory that the resonance tones are always harmonies of the cord
tone.

Taken as a whole, Dr. Scripture’s work is distinctly interesting
as well as instructive ; it is written with spirit and enthusiasm,
the illustrations are admirable, and the descriptions of apparatus
and of experimental methods are so clear and detailed that it
might well serve as a laboratory manual. To the student of
phonetics it will of course be indispensable on account of the
wealth of scientific material which it contains. H, A. B.

Geology and Mineralogy. 389

II. GEroLoGy AND MINERALOGY.

1. United States Geological Survey.—The following publica-
tions have recently been issued.

Twenty-First ANNUAL Report, Pr. III. 644 pp., 68 pl., 104
figs.—This volume contains papers on general geology, ore
deposits and the Philippines. Professor Hosss’s report on the
Newark System of Pomperaug Valley, Conn. (reviewed in this
Journal, vol. xiii, p. 70), occupies pp. 19-160. ‘The Laccoliths
of the Black Hills” by Professor Jacear (pp. 163-303) is a
description of the interesting intrusions of rhyolite and phonolite
in western South Dakota and eastern Wyoming. The laccoliths
are so numerous and the amount of erosion so varied that all
parts of atypical intrusion—conduit, basement contact, wedge,
flank, crest—are exposed in different parts of the region. These
intrusions are assigned to the Eocene and are an effect, not the
cause, of the great fractures and dome uplift of the Black Hills.
“The fractures reached downward to a zone where molten rock
was under pressure. The liquid shot upward into every ramifi-
cation of the fracture system.” Ernest Howe has made an
extremely interesting series of laboratory investigations (pp.
291-303) imitating the processes of laccolith intrusion and con-
sequent deformation of the invaded beds. Professor Van HiszE
contributes a paper (pp. 313-434) on the “Iron Deposits of the
Lake Superior Region” in which both the geologic and economic
aspects of the area are discussed. The Arkansas Bauxite and the
Tennessee white Phosphate deposits are described by C. W.
Hayes, pp. 435-487. “The Report on the Geology of the Phil-
ippine Islands” (pp. 493-625) is by G. F. Becker. It contains
general descriptions of the rock types and formations, the min-
eral resources, a bibliography and a paper by K. Martin on Ter-
tiary fossils.

TweEntTy-First ANNUAL Report, Pr. IV. 741 pp., 156 pl.,
329 figs. Hyprocrapuy ; by F. H. Newetr, Hydrographer in
charge. The tabulation and interpretation of stream measure-
ments for 1899 has added much data which is essential if intelli-
gent use is to be made of the water resources of the country.
Work has been carried on in all parts of the country, and the
descriptions (pp. 45-488) are so given as to render the facts
directly useful to interested people. A special paper on the
Geology and Water Resources of the Black Hills region (pp. 489-
599) is written by N. H. Darron. (Reviewed in this Journal,
vol. xiii, p. 68.) “The High Plains and their Utilization,” by
Wiarp D. Jounson (pp. 609-741) is an interesting study of
the topography, climate and history of the region including parts
of Texas, Kansas, Nebraska, New Mexico, and Colorado. The
causes of agricultural failure and the conditions necessary to suc-
cess are discussed in the light of recently acquired knowledge of
soil character, streams and underground water.

MonocrarH XLI.—Glacial Formations and Drainage features

390 Scientific Lntelligence.

of the Erie and Ohio basins; by Frank Leveretr. (Held for
review.)

Buuuetins. No. 179.—Bibliography and Catalogue of the Fos-
sil Vertebrata of North America ; by OxriveR Perry Hay. Pp.
1-868. Washington, 1902.—The author has rendered an immense
service to his fellow paleontologists in completing this list of
all the fossil vertebrates described, up to the end of 1900, from
all that part of the North American continent lying north of
Mexico. The list of systematic names recorded reaches nearly
10,000 ; and the number of titles referred to in the bibliography
is about 4600. The author estimates that there are over 40,000
citations recorded. The method of arrangement of the statistics
and the mode of citations have greatly condensed the mass of
material. The specific names are arranged according to their syste-
matic order in the catalogue. A separate key is given of this
catalogue, and under each scientific name in the catalogue are
given all the citations recorded, but abbreviated to author’s name,
date, and a letter (A, B, C, etc.) standing for the separate papers
of each author published during one year. In the Bibliography
these letters mark the several papers whose titles are given in
full. By this means the size of the volume is greatly reduced
without omitting the important details called for by the user. A
complete index of all the systematic names cited occupies the
last 72 pages. The author deserves the gratitude of all paleon-
tologists who may use this model catalogue of fossil vertebrates.

H. 8. W.

No, 182.—A report on the Economic Geology of the Silverton
Quadrangle, Colorado; by F. L. Ransome. 258 pp., 16 pls., 23
figs. This is a complete and detailed report on the economic
geology, of the ore deposits and mining operations of a quad-
rangle situated in the midst of the San Juan Mountains in south-
western Colorado. The first part of the report consists of a
general description and discussion of the ore deposits and includes
a chapter on the general geology of the region by Whitman
Cross, while the second part consists of detailed description of
special areas and individual mines. The area is covered with a
complex of volcanic rocks consisting of a thick series of tufts,
agglomerates and lava flows. The ores occur in these rocks ;
either as lodes or as nearly vertical stocks of almost solid ore.
The chief ore minerals carrying silver are galena, tetrahedrite,
enargite, stromeyerite, bornite and chalcopyrite. Gold is present
in the ore as native gold almost entirely, no important occur-
rences of tellurides being known. The ores were deposited from
acid solutions for the most part in preéxisting openings and some
of the deposits show the effects of descending surface waters in
the secondary enrichment of their upper portions. Ww. E. F.

No. 188.—Bibliography of North American Geology, Paleon-
tology, Petrology and Mineralogy for 1892-1900 inclusive ; by
F. B. Weeks, 717 pp. This bulletin is a combination of bulletins
Nos. 130, 135, 146, 149, 156, 162, 172 with the bibliography of

Geology and Mineralogy. 391

the literature for 1900. The index (bulletin 189) is to be used in
connection with this bibliography.

No. 189.—Index to North American Geology, Paleontology,
Petrology and Mineralogy for 1892-1900 inclusive; by F
WEEKS. 337 pp. The indexes contained in bulletins Nos. 130,
135, 146, 149, 156, 162, 172 are here combined. See above No. 188.

ies 190.—A Gazetteer of Texas; by Henry GANNETT. 162
PP. Mr. Gannett
ae a brief description of the topography, climate, character
of the population, ete.

No. 191.—North American Geologic Formation Names : bibli-
ography, synonymy, and distribution; by F. B. Wrexs. 448
pp-—This volume is a valuable addition to the scientific registers
of facts grown too bulky for memory. The author does not
claim that every name that should appear in such a list is actually
needed, but every name reported in a list of some five hundred
official and serial publications which are given is mentioned.
The author has made the list of special value to investigators by
presenting the facts exactly as they were presented by the author
of the name; thus throwing upon the user of the register the
decision as to which name is entitled to first consideration as the
proper name of any particular formation. When it is stated that
the bulletin contains 448 pages, the indispensable nature of such
a register, for the avoidance of further duplication of names, is
apparent. We are pleased to see also that it is proposed to con-
tinue this list in the annual Bibliography and Index of North
American Geology, which has become a feature of the Bulletins
of the United States Geological Survey. H. S. W.

No. 192.—A Gazetteer of Cuba ;.by HENRY GANNETT. 113 pp.

No. 193.—Geological Relations and Distribution of Platinum
and associated metals; by J. F. Kemp. 91 pp., 6 pls., 9 figs.

No. 194.—Northwest Boundary of Texas ; by Marcus Baker.
50 pp., 1 pl., 5 figs.—An account of the surveys of western Texas
from 1859 to 1900 is given and the present status of the boundary
is discussed. ‘Our knowledge of the location of the west
boundary of the panhandle is very imperfect,” and Mr. Baker
advises re-mapping the area.

Water Supply AND IRRIGATION PAPERS Nos. 57 (ALABAMA—
Montana) AND 61 (NEeBRASKA-—W yomineG).—Preliminary List of
Deep Borings in the United States; by N. H. Dartron.—A list
of wells over 400 feet in depth has been compiled for each State
—arranged alphabetically and accompanied by references to the
literature.

2. Lowa Geological Survey ; by SamuEt Carvin, State Geol-
ogist. Vol. xii, Annual Report 1901, with accompanying papers.
—Detailed geologic maps have been completed for fifty-three
counties in Iowa, seven counties having been surveyed during
poor: Uhe report on the Mineral Production in Iowa for the year
(by S. W. Beyer) shows a substantial increase in coal and gyp-
sum, but a decided falling off in the production of zinc. The

Am. Jour. Sci1.—FourtH SERIES, VoL. XIV, No. 83.—NOVEMBER, 1902.
27

~

392 Scientific Intelligence.

Geology of Webster County, by Frank A, Wilder, contains (pp.
138-224) an exhaustive study of the lowa gypsum deposits, as
well as those of other American and of foreign localities. The
Geology of Henry County is written by T. E. Savage. Cherokee
and Buena Vista Counties have been mapped by Thomas H. |
Macbride, who presents new facts regarding the limit of the
Wisconsin drift and its effect on topography. The Geology of
Jefferson County is by J. A. Udden, and that. of Wapello by A.
G. Leonard, Assistant State Geologist. Vol. xu is fully up to
the high standard of previous volumes of the Iowa survey, espe-
cially as regards contributions to stratigraphy and glaciology.

3. The Evolution of the northern part of the Lowlands of
Southeastern Missouri ; by C. F. Mansur. University of Mis-
souri Studies, vol. i, No. 3, 63 pp., 7 pls.—Southeastern Missouri
consists of belts of lowlands enclosing belts and isolated areas of
upland. As the result of field study (involving the making of a
topographic and a geologic map), Professor Marbut finds the
explanation of these belts as follows: The Mississippi River
flowed west of (the present) Crowley Ridge, then between
Crowley and Benton ridges and finally through Benton Ridge at
Grays Point. These successive channels were abandoned in
favor of small tributaries of the Ohio; “the Mississippi has been
captured by the Ohio twice in succession.”

The author’s former paper on this region (Bos. Soc. Nat’! Hist.,
vol. xxvi, pp. 478-488) described Crowley Ridge as a cuesta, and this
erroneous interpretation shows clearly the uncertain character of
conclusions based on map study unaccompanied by field work.
As Professor Marbut says: ‘‘T’o have determined the whole suc-
cession of events and the processes operating to produce them,
would have been probably impossible by the methods of library
study alone” (p. 39).

4. Geology of the Potomac Group of the Middle Atlantic
Slope; by W. B. Cuarxk and A. Brszpins. Bull. Geol. Soc. Am.,
vol. xiii, pp. 187-214, 7 pls.—A narrow belt of sand and clays,
mostly unconsolidated and often ferruginous, form the basal
element of the Atlantic Coastal plain. These sediments—called
collectively the Potomac Group—are described in detail as regards
their composition, fossil contents and economic importance. The
relationship of the beds is given as follows: |

Raritan
Patapsco

§ Arundel
( Patuxent

Cretaceous
Jurassic (?)

In regard to the conflicting evidence of paleontology and paleo-
botany (see this Journal, vol. ii, p. 433, Dec. 1896), the authors
think it essential to suspend final decision until more exhaustive
investigations of the faunas and floras has been made throughout
the entire Coastal region. |

Geology and Mineralogy. 393

5. Pleistocene Geology of Western New York ; by H. L., Fatr- »
CHILD. Report of progress for 1900. 20th Ann. Rept. N. Y.
State Geologist, pp. 105-139 ; pls. 9-41.—Professor Fairchild has
made a special study of the Iroquois shore line between Rich-
land and Watertown, and finds that the warping of the Ontario
basin has occurred mostly since the extinction of glacial Lake
Iroquois, and that “the rate of deformation has been much
greater than the present rate.” The territory between Syracuse
and Oneida was studied with reference to the higher channels cut
by overflow of glacial waters.

6. Michigan Geological Survey ; AtrrEeD C. Lane, State
Geologist. Annual Report, 1901, pp. 1-304, figs. 1-7, plates i-xv.
—Besides the ordinary reports of progress for the year this vol-
ume contains a paper by B. E. Livingston on the Distribution of
Plant Societies of Kent County, on the lines introduced by Dr.
H. C. Cowles. Mr. F. B. Taylor presents an excellent graphic
representation of the nature and distribution of surface deposits
of Lapeer County (Plate VI). A detailed analysis of the local
faunules of the Traverse (Devonian) formation of the northern
part of the State is given by A. W. Grabau : the State Geologist
_ publishes a geological map (scale 1 in.= 56 miles) revised up to
1902: an illustrated paper on Wave Cutting on the west shore of
Lake Huron is contributed by C. H. Gordon. H. S. W.

7. On Vertebrates of the Mid-Cretaceous of the Northwest
Territory ; by Henry Farrrietp Ossorn and Lawrence M.
LamBe. Geol. Survey of Canada. Robert Bell (acting) Director.
Contributions to Canadian Paleontology, vol. iii (quarto). 81 pp.,
22 pl. Ottawa, 1902.—Osborn and Lambe have contributed a
second part to the quarto memoirs on Canadian fossil vertebrates,
begun by the volume on the Cypress Hills species contributed by
the late E. D. Cope, in 1891. Professor Osborn contributes the
first paper on the Distinctive features of the mid-cretaceous fauna,
in which he reaches the conclusion “that the Belly River fauna is
more ancient in character both as to the older types of animals
which it contains and as to the stages of evolution among ani-
mals which are also represented in the Laramie.” (p. 21.)

L. M. Lambe contributes the second descriptive paper :—Vew
genera and species from the Belly River series (mid-cretaceous),
in which are described thirty-four species, of which sixteen are
new. H.S. W.

8. Queneau on Size of Grain in Igneous Rocks ;* by ALFRED ©
C. Lane. (Communicated.)—In this paper (a thesis for the mas-
ter’s degree) the author takes up the theory of grain proposed by
myself in part 1 of vol. vi of the reports of the Geological Sur-
vey of Michigan, and successfully applies the same to the “ Pali-
sade” intrusive near New York, and, with very interesting re-

* Size of grain in Igneous Rocks in Relation to the Distance from the
Cooling Wall, by Augustus L. Queneau. School of Mines Quarterly, January
1902, pp. 181-195, Contributions from the Geological Department of Colum-
bia University, vol. ix, No. 80. (This paper was briefly noticed on p. 70.)

394 Scientific Intelligence.

sults, to the minette dike of Franklin Furnace, New Jersey. The
mathematical foundation has also been worked out for him anew
by Prof. Woodward, with accordant results, but whereas I treat
the case where the country rock is appreciably heated up, he
takes only the case in which the walls of the cooling body are
kept at a constant temperature.

Unfortunately we have taken different initial temperatures,
and otherwise the two methods differ so that the curves and
numerical tables cannot be directly compared. The curves of
falling temperature given in my paper are referred to the time,
whereas in Queneau’s paper the abscissas are laid off proportional
to the square root of the time. The early stages of cooling are
given, therefore, in more detail, but his remark at the bottom of
page 188 as to the slopes of all the curves being equal, is errone-
ous. It applies to my curves, but not to his.* Queneau’s results
upon the minette of the Franklin Furnace are especially note-
worthy; both biotite and apatite vary in size according to the
distance from the margin. This shows clearly that these min-
erals are mainly not pre-intrusive. Moreover, the size of the
apatite increases. clear to the center, while the biotite has a zone
of nearly uniform grain occupying nearly the middle two-thirds
of the dike. The following inferences of theoretical interest,
which, however, Mr. Queneau does not mention, may be drawn:

The apatite was formed early, much before the biotite, and
probably before the dike had lost even one-tenth of its initial
temperature. The dike therefore must have been Jargely viscous
when the apatite formed, and the broken condition of the apatite
mentioned by Queneau may well be attributed to a strain in the
otherwise viscous magma.

The biotite was formed later, and supposing, as seems likely
from the figure, that the breadth of the zone that has a markedly
less grain than that at the center is about one-eighth of the width
of the dike, we must infer that it was formed when the dike
retained but little more than half its initial excess of tempera-
ture.

A small dike of diabase, a meter wide, shows in its feldspars
that the temperature of its injection was not much above that of
the formation of feldspar, at any rate.

When he discusses the Palisade trap sheet, however, Mr. Que-
neau neglects to. mention that he does not carry his section clear
across, nor does he state the thickness of the sheet, nor say

* Some other minor mistakes in our papers are: on page 187, one column
should be headed x=°85c, and line five above should begin with 95c. I be-
lieve there are also some numerical mistakes in the body of the tables. The
last column of the table on page 188 is erroneous, for instance. Also on
page 191 the units for « are not in square but linear millimeters. In my
table, when m=7 if g=°92, the entry should be 74265, and if q is ‘91 the
entry should be *73000. In my equation 11 an n has crept in that does not
belong there, and on page 117, line 2, for (w) read (2w). Also, in the plate,
the decimal point for the horizontal unit is wrongly placed,—it should be
log = —01.

Geology and Mineralogy. 395

whether he continued the section until the grain appears uni-
form. From Kiimmel’s report for the New Jersey Geological
Survey, 1897, p. 62, we learn that at one of the points—Fort
Lee—the sheet is 950 feet thick, of which Quenean’s section cov-
ered only about 100 feet. Of the grain, Ktimmel says that while
on the whole the trap is coarse-grained, near the contact it is
fine-grained or occasionally slightly glassy, but this rapidly in-
creases in size within a few feet of this shale, where coarsest
tin tabular crystals of feldspar occur, two or three-eighths of an
inch in diameter. The King’s Point trap near Weehawken he
estimates from 700 to 875 feet thick, and again Queneau can only
cover the margin of varying grain (3449, or 105 feet).

From his plates and figures it appears that, so far as he meas-
ured, the feldspars still steadily and uniformly increase in size,
but the rate of increase of the King’s Point augite drops deci-
dedly between 23 and 34 meters. Thus the breadth of the zone
of augite which was formed before the center cooled is not far
from 100 feet, or about one-eighth of the total breadth. This
would imply that the consolidation of the augite took place un-
der conditions of temperature, etc., a little nearer those of the
initial magma than those of the country rock, and that the feld-
spar was earlier formed. We may also infer from his equations
of grain (as shown by investigations of mine as yet unpublished)
that there is a contact zone of noticeable thickness in which the
temperature-was appreciably increased, and that while his data
are consistent with Kiimmel’s observations, the grain of the
augite at the center is not more than 5™™, being probably about
3™™—q result which it would be interesting to verify. Had we
the thickness noted, and that of the contact zone if convenient,
and a section representing the central belt, we could have a check
on the theory, as well as on the accuracy of the observations ;
for the estimation of the grain is a difficult matter, and I doubt
if an accuracy of 10 per cent can be obtained, though the diver-
gence of Queneau’s observations from his curves is no fair test, for
he has done himself injustice, since the curve of linear grain of a
great intrusive sheet with a contact zone, like the Palisades, ought
not theoretically to be a straight line, but is enclosed by three
straight lines as external tangents, which it will usually follow
very closely, having only short easement curves near the points
of tangency.

Queneau’s statement that the zone of varying grain will vary
inversely as the temperature of the country rock, is not mathe-
matically accurate, and is probably intended merely as a reword-
ing of what I say on page 111, that the hotter the country rock
the less pronounced will be the zone of finer grain. But I may
say that if we have the curve of grain located well enough to
determine the three tangents just mentioned, we may infer the
ratio of the temperatures of injection and consolidation, the
thickness of the intruded sheet and the breadth of the contact
zone appreciably affected, as well as a constant involving the

396 Scientific Intelligence.

diffusivity of the rock and the tendency of the particular mineral
to crystallize. So that we may often give a shrewd guess at the
size of the sheet from observations on the grain only at one side.
I hope to take up this matter soon.

Geological Survey of Michigan, Lansing, Mich.

9. Les Roches alcalines caractérisant la Province petrograph-
ique d’ Ampasindava; by A. Lacroix. (Nouv. Archives du
muséum d’ histoire naturelle. 4th Ser., Tome I, pp. 152, pl. 10,
4°, Paris 1902, Fascicule I.)—This work is based chiefly on speci-
mens and observations collected by M. Villiaume, a colonial
official. It is divided into three parts, the first being a descerip-
tion of the geology and rocks of Nosy Komba, an island off the
northwest coast of Madagascar between the island of Nosy Be
and the bay of Ampasindava.* In this portion are described
gabbros, nepheline syenites of various types, including dark basic
varieties passing into the essexites, ‘They are accompanied by
acidic dike rocks, and there is described, in addition, the phe-
nomena of the contact of these intrusive masses, both exomorphic
and endomorphic. The second part relates to the occurrence and
petrography of igneous rocks from portions of the mainland in
this part of Madagascar and from Nosy Be. In this are described
granites with soda amphiboles, pulaskite, nordmarkite and laur-
vikite types of syenites, essexites, quartz bostonites, camptonites,
phonolites, trachytes, monchiquites, ijolite, tinguaite, etc., ete.

The petrography of these rocks is given with the care and de-
tail for which the author’s works are so well known. A number
of analyses enhances the value of the memoir, which adds greatly
to our knowledge of this little known region. The appearance
of Fascicule II, containing chapter ili, giving the general con-
clusions and resumé, will be awaited with interest by petrog-
raphers. L, Vii.

10. Ueber mariupolit, ein extremes Gilied der EHlaeolithsyenite ;
von J. Morozewrcz (Tscher. Min. u. Petro. Mitt. Band xxi,
s. 238, 1902).—On the north coast of the sea of Azov in Russia
lies an area of crystalline rocks which is being investigated by
the author. In the district of Mariupol a portion of this area,
between gneiss and granite and comprising about 10-12 square
kilometers, has been found to consist of eleolite syenite and pyrox-
enite. The investigation of this former rock shows it to consist
of 73 per cent albite, 14 per cent of nephelite, 7°6 per cent of
aegirite, 4 per cent lepidomelane and 1°6 per cent zircon. The
mass analysis gave the following results :

SiOz ZrO. Al,Os; Fe2.0; FeO MnO MgO CaO K.O Na.O H,O Sum.
62°53 1:08 18°72 38:26 0°34 0:16 0°08 0:54 0-79 11°77 0°68 = 99-95

In this the relation K,0: Na,O::1:24. The texture is variable,
sometimes coarse, sometimes fine, and sometimes porphyritic.
The feldspathic components are set through with aegirite needles.

* (Stieler’s Hand Atlas gives Ampassandava. )

Geology and Mineralogy. 397

Separate analyses of the mineral components are also given. To
this rock of the syenite family composed almost solely of soda-
bearing constituents, the author gives the name of Mariupolite.
L. Vieibe
11. Dahamit, ein neues Ganggestein aus der Gefolgschaft des
Alkaligranit ; von A. Prexixan. (Denkschriften der mat.-nat.
wiss. classe der K. Akad. der. Wiss. Wien, 1902. Bd. lxxi.)—In
an article dealing with the petrography of a rock collection from
the islands of Socotra, Abd el Kuri and Semha the author de-
scribes a dike rock of a chocolate brown color, compact texture,
containing red phenocrysts of thin tabular feldspar. The micro-
scopic study and the chemical analysis shows the rock to consist of
6°8 per cent riebeckite, 43°8 per cent albite, 2°8 per cent anorthite,
12°2 orthoclase, 31°5 per cent quartz. The chemical analysis by
E. Ludwig gave

$i0, Al,0, FeO; FeO MgO CaO Na,O K.O H,O Sum.
74:02 13:56 1:93 1:09 0:28 0:56 5°80 2°06 1:05 = 100-30

The phenocrysts are albite or albite:oligoclase, containing the
small amount of anorthite. The other components are mixed in
the groundmass. ‘The author suggests the name from the locality,
Dahamis, and offers the fact that it differs from grorudite (quartz
tinguaite) in containing riebeckite instead of aegirite, as a reason
for the new name. He overlooks entirely, however, the fact that
Osann (Tscher. Min. Mitt., vol 15, p. 485, 1895) has already
described a dike rock consisting of alkali feldspars, quartz and
riebeckite under the name of paisanite, the analysis of ‘which also
agrees in essential particulars with that above. The rock should
be regarded as a variety of paisanite. Be Vuk:

12. Mineral Resources of South Dakota ; by C. C. O’Harra |
and J. E. Topp. South Dakota Geol. Survey Bull. No. 3, pp.
130; pl. 31.—The mineral wealth of the Black Hills has already
been described by Professor O’Harra (this Journal, vol. xiii, p.
474). Professor Todd, State Geologist, gives a report (pp. 81-
130) on the mineral Building Materials, Fuels and Waters of
South Dakota.

13. Note on anew occurrence of Native Arsenic; by NEvit
Norton Evans. (Communicated.)—During the past summer
native arsenic was discovered in a calcite vein cutting the nephe-
line syenite at Montreal, Canada, by Mr. Edward Ardley, Museum
Assistant at the Peter Redpath Museum, McGill University. The
mineral appears to be pure arsenic, not more than traces of other
elements having as yet been found in it. The vein has a maxi-
mum thickness, so far as has been seen, of two inches, and has
already yielded about fifty pounds of the arsenic. A complete
analysis, together with a detailed study of the occurrence, is
being made by the writer, and will be published shortly.

McGill University, Montreal.

14. The Inverness Earthquake of September 18th, 1901, and
the Carlisle Earthquake of July, 1901; by C. Davison. Quart.

398 Scientific Intelligence.

Jour. Geol. Soc., vol. lviii, pp. 371-397, 2 pls.—The recent Inver-
ness earthquake gives additional proof of the existence of a fault
line reaching from Loch Ness to the sea and this region “ appears
to be in a stage of more rapid development than any other in the
British Islands” (p. 396). The earthquakes of 1816-18, 1888,
1890, and 1900 were along the same fault.

The Carlisle earthquake has led to the recognition of a deep-
seated fault (running N. 5° E ) in a region where there is no sur-
face sign of such a structure.

MISCELLANEOUS SCIENTIFIC INTELLIGENCE.

1. Beitrdge zur chemischen Physiologie und Pathologie, her-
ausgegeben von FRanz Hormeister. II Band. 1902 (Braun-
schweig, F. Vieweg und Sohn. 15 M.).—This completed volume,
of which the earlier numbers have already been reviewed in this ~
Journal, brings evidence of the same high standard which char-
acterized the first one. The Beitriage have now won a place in
the permanent literature of physiological chemistry. Particular
attention may be directed to the recent contributions on the
products of proteolytic digestion and the studies on various pro-
teids of animal and vegetable origin. L. B. M.

2. Les Dirigibles ; by M. H. Anpr&. Pp. ii+346. Paris:
Ch. Béranger, editeur. 1902.—This is a well arranged and use-
ful hand-book of the theory and practice of aerial navigation as —
it is known at present. As indicated by the title, the larger part
of the book is occupied with dirigible balloons, and much inter-
esting and valuable information “has been collected upon the
various conditions of this difficult problem ; the laws of the resist-
ance of the air, so far as they are known, are applied to the
determination of the proper shape and dimensions of the balloon
itself, of the propellers, rudders, etc., and some consideration is
given to the motors which may be employed ; the very important
question of stability also receives attention. The book concludes
with a critical account of the various experiments which have
been made with dirigible balloons from the celebrated attempts
of Gifford in 1852 and 1855 to the recent achievements of Santos-
Dumont. H. A. B.

38. An Introduction to Physical Geography ; by GrovE Kar.
GitpeERT and ALBerT Perry Bricuam. D. Appleton & Co.
370 pages, 263 illustrations.—Gilbert and Brigham’s Physical
Geography is thoroughly scientific, up to date and attractive. It
is a good introduction to out-of- door science of any sort.

3

ff

|

Am. Jour. Sci., Vol. XIV, 1902. Plate IX.

Fic. 1.—BOWLDER TRAIN UPON THE SLOPE OF SHERMAN HILL

The view looks north-northwest along the train toward the crown of the hill.

wo

Fic. 2.—SouTHERN END OF THE TRAIN OF BasaLT BLOCKS FROM SHERMAN HILL.

The view looks north-northwest along the train, Sherman Hill itself, distant a mile
and a half, appearing between the two largest blocks.

THE

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES.}]

Art. XXX VII.—An Instance of the Action of the Ice-sheet
upon Slender Projecting Lock Masses ;* by WILLIAM
HERBERT Hopss. (With Plate IX.)

THE present paper will discuss what seems clearly to be a
case of degradational action by the ice mantle exerting its
pressure in a lateral direction against steep and slender masses
of projecting rock. The conditions are in some respects pecu-
liar to the locality, which lies in an island-like area of Newark
traps and arkoses within protecting walls of the crystalline
schists—the Pomperaug Valley area of Connecticut. The pre-
servation here of the Newark rocks from erosion, which has
elsewhere so generally removed them, is a direct result of their
deformation by jointing and their depression in the form of a
composite crustal block along the marginal joint planes as
fault walls.t This depression has been irregular in so far as
the blocks of the smaller orders of magnitude within the larger
composite block have been depressed by different amounts so
as to produce a mosaic of small prismatic blocks. The present
positions of basalt and arkose within the area are also dependent
upon the high position of the main basalt flow in the local
Newark series, and upon the low easterly dip (420°) of the
beds and flows previous to their dislocation. Their positions
are further affected by the fact that the displacement by fault-
ing was distributive largely within a marginal zone. As a
result of all these conditions the dense resistant basalt now
occupies the central portion of the depressed area and is bor-
dered by the softer shales and arkoses.

Could the area have been visited subsequent to its deforma-
tion by faulting but previous to its degradation through the

* Read before the Wisconsin Academy of Sciences, Arts, and Letters, at
Milwaukee, December, 1901.

+The Newark System of the Pomperaug Valley. 2ist Ann. Rept. U.S.
Geol. Survey, Pt. III, pp. 1-160.

Am. Jour. Scl.—FourTH SERIES, Vou. XIV, No. 84.—DEcEMBER, 1902.

400 W. Hf. Hobbs—Instance of the Action of the

agencies of subzerial and of ice erosion, we may believe that it
would have presented an irregular surface not unlike that of a
mosaic from which a local area of the back had become dis-
placed and the overlying blocks allowed to slide down by
small amounts while still restrained by their friction upon their
neighbors. The effect of subzerial erosion has been to etch
out the marginal areas of soft sandstone and leave the basalt
prisms of the central area in strong relief like the image of a
cameo. The basalt itself discloses no marks of the suberial
erosion, for the reasons: first, that it is intensely resistant; and,
second, that its area is so small (six miles in length by two
miles in greatest breadth) that no streams of any power have
been developed upon it. It is not, however, to be assumed
that no considerable degradational action has occurred within
the area of the basalt masses of the valley, for the three upper
members of the Newark series found in the Connecticut Val-
ley area, which begins less than a score of miles to the east, are
missing from the Pomperaug Valley series, and were doubtless
removed by subeerial erosion, while large thicknesses of the
surrounding schists were being carried away.

Fic. 1. Schematic profile of basalt ridges of the Pomperaug Valley.
Black, basalt; black spotted with white, amygdaloidal basalt ; white, shale ;
white with black circles, conglomerate (where stippled, baked zone of con-
tact) ; stippled area, drift and alluvium.

The work of the ice within the valley is revealed in the pro-
files of the basalt ridges. These ridges have generally fault
scarps on their western and northern sides (which face in the
direction from which the ice moved) and gentle slopes to the
eastward and southward, conforming to the dip of the beds and
flows of which they are composed. In these general outlines
the action of the ice is not disclosed, but the caps of all the
ridges seem to have been removed by an appreciable fraction
of their height. This is brought out in the schematic figure 1
and in the author’s report above cited.* That this degrada-
tional action by the ice is localized largely at the crests of
ridges is also shown by the texture of the rock found at the
crests when compared with that upon the flanks of the southern
ridges. Dense and massive at the crest in correspondence
with the lower beds of the flow, it is amygdaloidal and vesicu-
lar upon the southeastern flanks, where it doubtless represents
upper layers of the flow.

* Loe. eit.; pl. Ves A.

Ice-sheet upon Slender Projecting Rock Masses. 401

Some clue to the manner of decapitation of the basalt ridges
may be found in two trains of basalt blocks, one of which
is represented in Plate IX. This train, which is the one
of the greater interest, proceeds from the ridge known as
Sherman Huill,* the extreme southwestern elevation of the

2

g
ty ®

en
KEY PE Ke

Fic. 2. Sketch map of the Pomperaug Valley showing the distribution of
the terrace deposits and of basalt trains. W., Woodbury; S., Southbury ;
S. B., South Britain; W.O., White

H., Hotchkissville; P.. Pomperaug ;
Oaks; S. H., Sherman Hill; R. H., Rattlesnake Hill; C. R., Castle Rock.

The stippled area represents roughly the distribution of the terrace deposits,
and the lines of larger spots the trains of basalt blocks; A., area of terrace

excavated for railroad fill.
* Sometimes considered a part of Rattlesnake Hill.

4.02 W. H. Hobbs—Instance of the Action of the

basalt within the valley. _The position and direction of this
train is brought out in fig. 2, which is a sketch map of the
Pomperaug Valley. Upon the south-southeastern flank of
Sherman Hill the train begins as a collection of irregular
blocks of basalt, occupying a belt about one hundred feet in
width and extending from the present low tower-like crest of
the hill to the corner of the highway, where is the edge of the
terrace floor of the valley and where the train appears to end
abruptly. Within this portion of the train are a dozen or
more blocks six to eight feet in diameter, and very many
smaller ones. Fig. 1, Pl. LX is a view looking up the train
from a point near the road corner toward the summit of Sher-
man Hill.

Lost in the terrace deposits of the valley the train is again
picked up upon the south side of the valley so soon as the
slope begins to rise above the level of the terrace. This local-
ity is near and west of the road from South Britain over
Georges Hill (see fig. 2) and just above the great fill upon the
railroad. The sand and gravel for this fill was obtained from
the area immediately north where indicated upon the map
(A of fig. 2). Mr. Henry M. Campbell of South Britain —
pointed ont to the writer a place where three large blocks of
the basalt, one of them six feet or more in diameter, were
unearthed and removed by blasting during the excavation of
this section of terrace. The location would fall within the
line of the train from Sherman Hill.

The blocks now in evidence south of the railroad fill are of
special interest because of their location in the line of the train
and because of their unusual size. Fig. 2, Pl. LX, which looks
northwestward along the train, includes two of the largest.
The hill from which they were separated is visible in the dis-
tance (more than a mile and a half away) between their tops.
There are several blocks of this size (15 to 25 ft. in largest
dimensions), some partially buried in earth along with many
smaller blocks, but all located within a belt less than a hun-
dred feet in width. The train does not appear to extend south-
ward beyond this point, but no attempt has been made to
follow it farther. The direction from Sherman Hill of these
great blocks of basalt, which form a landmark in the valley, is
S. 28° to 29° E., a value near that of the average movement of
the ice over the higher points in the vicinity.

The conditions seem here to be best explained by assuming
that from the pre-glacial surface produced by subeerial erosion
the faulted prisms of basalt projected in pinnacles above the
softer sandstones, in part opposing to the ice stout walls and in
part comparatively slender walls and towers. Sherman Hill,
as indicated by the geological map in the report here cited,

Ice-sheet upon Slender Projecting Rock Masses. 4038

unlike practically all of the other ridges of the valley, opposed
its long face and, therefore, its weakest direction to the ice
mass, which was directed against its thin vertical walls. Owing
in part to dip and in part tothe manner of faulting, its western
end (the present peak) projected above its general wall-like
mass. This part also was unprotected by the higher mass of
Rattlesnake Hill. In this way, it is believed, the great blocks
which are now found over a mile and a half away and con-
nected by the train with this elevation were separated from
their parent mass.

Castle Rock, a prismatic block of nearly equal basal dimen-
sions and now having a sheer cliff upon its northern face
nearly one hundred feet in height, is located in the north-
eastern part of the area.* Like the west end of Sherman Hill
its position left it unprotected by other masses from the north-
west, where a considerable valley allowed the ice to sag well
below the crests of the elevations surrounding the valley. The
row of basalt blocks located southeast of White Oaks is
believed to have been derived from this elevation much as the
other train has been from Sherman Hill. The slender nature
of the prism of Castle Rock, its considerable elevation above
the general level, and its exposure upon the northwestern
edge of the area all favor such an hypothesis. The direction
of this train (S. 13° E.) would indicate that the long valley east
of the basalt masses has modified the direction of movement of
the lower portions of the ice mantle.

That the ice found a depression in the Pomperaug Valley,
into which its substance sagged, is indicated by the present
altitudes of the higher basalt ridges which have been decapi-
tated, in comparison with those of the gneiss hills of the vicin-
ity (the basalt ridges are some 200 to 300 feet the lower).
It must be assumed that erosion in post-glacial time has been
more effective upon the gneiss than upon the basalt, though
this would doubtless be modified by the soft arkose members
above the main basalt.

To ascribe these trains to the separation of blocks from the
lee side of ledges due to the frictional action of overriding, is
to leave unexplained their large size and the distance to which
some have been carried, more especially, however, the restric-
tion of the trains to ledges which the geological study shows
to have been slender in form and with deep valleys northwest
of them into which the ice could sink and act in a horizontal
direction against their walls.

Professor Chamberlin has informed me that he has observed
in Greenland an undoubted instance of such degradational
action of the ice from lateral pressure upon rock walls.

University of Wisconsin, Madison, Wisconsin.

Soe. cit, fie: 42, pe bit,

404 Koenig—New Species Melanochaleite and Keweenawite.

Art. XXXVIII.—On the New Species Melanochalcite and

Keweenawite. With Notes on some other known species ;
by GEorGE A. KOENIG.

1. Melanechaleite.

For the material of this investigation 1 am indebted to
Captain James Hoatson, of Calumet. I received it early in
January of 1902, and informed that gentleman—about January
17th—that I had much reason to distinguish the black mineral
as a new mineral species, for which I proposed the name
melanochalcite (from pwédas and yadxes).

Occurrence.—The material comes from the exploration shaft
of the Calumet and Arizona Copper Mining Company, near
Bisbee, Arizona. I am informed by Captain Hoatson that
these specimens represent the character of the ore at a depth ©
of 800 feet. All the mines in that district exhibit oxidized
ores to a very considerable depth. Sulphides are rare. I had
examined and assayed a number of samples for Captain Hoat-
son from time to time as the sinking of the shaft proceeded;
all oxide ores, and of exceptional richness. Nosamples assayed
under 10 per cent copper. Cuprite was always preponderating,
sometimes mixed with much hematite; say, for instance, 30 per
cent of the former to 70 percent of the latter, as an extreme.
Malachite and chrysocolla appeared sparingly ; noazurite; whilst
at the Copper Queen Mine, of the same belt, the silicate and the
carbonates are in the foreground. The material under consid-
eration in this paper differs from that of the upper shaft.
The material before me presents hard spheroidal nodules,
cemented together by a soft, brown-red, clayey material, easily
removable. The nodules’ nucleus is formed by granular
cuprite, with occasional druses, the latter lined with octa-
hedral crystals. This kernel is surrounded by a zone of
pitchy-black mineral, a few millimeters in thickness. Upon
this follows a banded green zone of chrysocolla and malachite.
Thereupon follows white, or transparent, quartz. Within the
quartz are smaller cuprite kernels, each with its aureole of
black and green. Here the black material is thicker, but less
pure asa whole. The purest substance is always thin, lying
close to the cuprite. It passes into deep olive green; then
light olive-green into the pure green of chrysocolla, or mala-
chite. Thus the fracture-surface of a nodule is of striking
beauty. The one before me, which served as model for the
description, has an average diameter of 120 millimeters. The
kernel is not centric, and rather oblong than circular. The
black, green and white parts are massed chiefly on one side.

Koenig—New Species Melanochaleite and Keweenawite. 405

This material reminded me of the German “ Kupferpecherz,”
which has been declared by eminent authorities as a mineral
mixture of various bodies—chiefly chrysocolla and limonite.
Dana places it under chrysocolla. JI have analyzed such
material coming from the Old Dominion Mine, Arizona, of a
brown-black color. It was more the esthetic beauty of the
present material which induced me to enter into its study,
than the expectation of finding new facts. Absence of definite
erystalline form is ever apt to call up a prejudice against the
homogeneity or chemical integrity of bodies. In the present
instance, additional reluctance was caused by thinness of the
black zone. But since there was plenty of material, the ques-
tion of obtaining a sufficient quantity of satisfactory substance
could be answered by care and patience.

Physical Examination.—l picked out successively, and
from different nodules, three portions of material—A, B, C.
The sequence of the letters denotes the degree of scrutiny
employed. “A” was intended for preliminary work, “B”
for chemical orientation, and ‘‘C” for the final trial. "At the
beginning the chief care was the rejection of either red or
green particles. But it was quickly discerned that among the
dark material there was a lustrous and a dull portion; a banded
and a bandless part. High lusterand absence of band-strueture
go together; dullness and banding are yoked. Sample “A”
was tainted somewhat by red, by green and by dull particles.
Sample “B” was only contaminated by dull parts; but “C”
was picked over several times, and contained only brilliantly-
black material. This was time-consuming. It will be seen,
however, from the analyses that the blemishes were rare even
in “A” and “Bb.” The black mineral is fairly hard—about 4;
but it is exceedingly brittle. The cause of this brittleness
lies, probably, in numerous microscopic fissures. As support
for this opinion, I mention the long time required for the ceas-
ing of air bubbles to rise, when the mineral was placed under
water (specific gravity determination). And also the failure to
obtain a thin plate by grinding; the plate going to small
pieces long before translucence was reached. When the min-
eral is eround i in a mortar it shows the disagreeable property
of © “smearing.” The fine powder is coffee-brown in color.
Some of the finest dust from sample “ C,” which had stuck to
the mortar, was brushed upon a glass slide, imbedded in
Canada balsam and examined under the microscope. It proved
to be translucent, letting through yellow-brown light, one par-
ticle exactly as the other. The mineralogic uniformity and
singleness of this material (“C”) is undoubtable. In polar-
ized light these dust particles behaved like an amorphous,
or isotropic, body. The specific gravity was found to be

406 Koenig—New Species Melanochalcite and Keweenawite.

— = 4-141 at 21 C. (material “C”). . This number is
rather under than above the true weight, because I wished to
avoid boiling, in order not to interfere with the water-per-
centage of the material. To even up matters, the water used
was saturated with air at 21 C. for both weighings.

Chemical Hxamination.—B. B. In closed tube the sub-
stance looses water and CO, and the coffee-brown powder turns
to brown black. With the fluxes only copper reaction, except
with microcosmic¢ salt, when a fine skeleton of SiO, appears in
a blue glass. The mineral is readily decomposed by HOl of
all concentrations, even when in coarse fragments. If such a
fragment be placed in 3 per cent HCl, and one observes with a
pocket lens, one sees the margins of the splinter turn white. The
white zone widens steadily, until nothing but a white mass
remains, which occupies the entire space of the original black
substance. And yet the percentage of SiO, is not quite 10.
While this dissolution is progressing ove sees a steady spray
arise of minute bubbles of CO,, very different from the big air
bubbles arising from the microscopic capillary fissures. I con-
sider this behavior as the essential foundation of my hypothesis
regarding the constitution of the molecules of this mineral—
to wit: SO, and CO, are simultaneously liberated, whilst
CuO and H,O pass into the solution. The resultant SiO, is
what I would eall semigelatinous—not colorless and transpar-
ent, but white and translucent. It is readily dissolved by the
water solution of NaHO. There is no euprous chloride formed.
The copper is altogether cupric.

Material ‘“‘ A.’—preliminary analysis. Weight of substance
taken, 0°5361 gr. |

0:0671 loss by ignition.
00532 SiO,

00010 Fe,O,

0-4188 Cu,S

The sulphide was dissolved and electrolyzed, yielding 0°3277
Cu=0°4107 CuO, to which must be added 0:00015 CuO,
obtained from the electrolyte by H,S. Total, 0°4122 CuO. |

In percentage : 3

Gud. = 76:72
SiO, = 9-91
TGQ) facrerse

CO, == Ra?
Ke:0.. = 0219

Koenig—New Species Melanochalcite and Keweenawite 407

Material “B.” 0°5658 gr. The tenacity for water at in-
creasing temperature was tested by exposing the powder, for
two honrs each, at the following temperatures :

0°0092 loss at 88C.
0°0142 va Hoe:
0°0082 s5 160 C.
0°0128 eG 210.C.

0°0444
0°0451 loss at red heat.

0°0895 total loss.

CO, was not separately determined in this sample; but from
the determinations in “O,” when H,O and CO, are nearly
alike, the inference may be drawn that all the water is expelled .
at 210 ©.

0°0500 SiO, ; 0°3415 Cu+0-0082 CuO from electrolyte by
HS, 0°0008 Fe,0,.

CuO = 76°46
SiO, = Soo
ORE uae,
H,O ( == 4°20
Fe,O, =) 0°14
99°63

Material ““C” — This having been proved irreproachable
material, the analysis was made with great care. OO, and
water were determined as follows: 0°5260 gr. of the fine
powder was placed in a porcelain boat and the latter heated
in a combustion tube to redness, whilst a slow current of per-
fectly dried and purified air passed through the tube. An
“U” calcium chloride tube received the water and a Geissler
Potash bulb received the CO,. The increase of weight in the
former was 0:0407 gr.; in the latter, 0:0378 gr., a total of
0-0785 gram. At the same time the weight in the boat
decreased by 0:0766 gram. Hence there is here an error of
0-0019 gr., which may be evenly distributed between H,O and
CO,, or may even be neglected altogether without affecting
the result sensibly. The copper was precipitated electrolyti-
cally, and in this case H,S found no residue of copper in the
electrolyte. The latter being thereupon rendered ammoniacal,
a flocculent precipitate fell, which, after filtering and ignition,
showed by its color that it did not consist of iron oxide alone.
It was found to be ZnO+Fe,O,. I had overlooked the zine in
the other analyses. |

408 Koenig—New Species Melanochalcite and Keweenawite.

The percentages are:
Molecules.

CuO, 117688 2 79 =] OESSs
SiO, = 780: 60 =0-12914
9

CO, = 7:17:44 = 0°16295 to eager
H,O = 771 :18 = 0-42833
ZnO = 0°41 : 81 = 0:00506
BO” = OVE

100-04

This unusual, and surely novel composition, may give cause
for several inter pretations. Roughly speaking, it might be an
intimate mixture of copper carbonate, copper silicate and
copper hydrate; but it may also be interpreted as the basic
salt of an ortho-silico-carbonie acid H,(Si,C)O,, in which Si
and C may replace each other within certain limits. A scrutiny
of the three analyses shows a practically identical percentage
of copper oxide. The percentages of SiO, and of the volatiles
vary. As to SiO,, we have 9-91 (“A”); S8ai("ai ee
(“C”), a variation of about 1 per cent. In “A” we have
12°52 volatiles; in “ B,” 14:2; in “C,” 14°88. These figures
are evidently not due to accident, nor due to mechanical mix-
ture. Unfortunately their significance was recognized only
after it was too late for a determination of CO, in “A” and
““B,” for there was nothing, or too little, of the materials left.
Notwithstanding this serious experimental deficiency, it seems
to me that the logical comparison forces the admission of
strength in my hypothesis. Admitted the existence of a com-
plex H,(Si, C)O,.H,O (and I can see no chemical reasons
against it) in which hydrogen is wholly replaced by Cu, and
H,O partly by CuO and in which, moreover, Si and C are
atomically interchangeable: but so that the percentage of Ou
be constant, then a change in the percentage of Si,0, must
influence both the percentages of CO, and H,O. Now, com-
paring “A” and “C,” we find in the former 2°11 SiO, more,
and 2°35 volatiles less than in the latter. 2°11 SiO, are equiv-
alent to 1°53 CO,, leaving 0°82 as that portion of the volatiles
attributable to water, thus conforming with the hypothesis.

If the composition of melanochalcite be interpreted accord-
ing to this plausible hypothesis, the figures of analysis “C”
take the following shape.

Molecules Si,O, = 0°12914
CO. 220° 16295

0°29209

These 0°29209 molecules require 2x 0°29209=0°58418 mole-
ecules of CuO.

Koenig—New Species Melanochalecite and Keweenawite. 409 -

There are disposable 0°96583 CuO+0-00506 ZnO, a total of
0-97089 molecules of basic oxides. Deducting from this total
the requirements of the silico-carbonate, there will be left
0-97089—0°58418 = 0°38671 (CuO,ZnO} to constitute with the
water, the hydroxide. But we have 042833 molecules of water ;
hence there is a surplus of the latter of 0°42833—0°38671=
0°03762. This surplus, which amounts to 0-749 per cent, must
be declared present as hygroscopic water. It is not determin-
able by experiment, since the substance lost at 87 C.—1°63
per cent, and since it is a well-known fact that boilmg water
converts the hydroxide Cu(HO), into the oxide CuO.

Putting together these figures, we get :

Copper silico-carbonate 0°87627 : copper hydroxide
0°77342 = 1°000: 0°882.

The formula of melanochalcite is:
Cu,(Si,C)O, - Cu(HO),

Paragenesis.—In regard to the forming of this mineral sub-
stance, one may conceive of at least two modes :

(1). Cuprite crystallizes first from a solution of copper ear-
bonate, with the codperation of a deoxidizing agent, and pro-
duces centers or nuclei. At a later period an oxygenated
aqueous solution of silicon and carbon meta-acids invests these
nuclei. In presence of an overwhelming basic substratum, the
meta condition of the acid changes into the ortho state, and
the melanochalcite molecules are formed rapidly, falling out
in amorphous aggregates. As the crust increases in thickness
the basic substratuim’s influence decreases and we find mixtures
of melanochalcite aggregates with those of chrysocolla and
malachite, because the ortho reverts to the meta state. The
complex molecule Cu,(Si,C)O, is no longer possible; and,
shortly after, we have clear alternating bands of chrysocolla
and malachite.

(2). One can conceive of an aqueous solution holding from
the start all the constituents, excepting only oxygen, and from
which the least soluble constituents, eee CUTS), will fall out
first. But considering the great bulk of the latter compared
with that of melanochalcite, chrysocolla and malachite, this
view would seem less simple than at first. Both are purely
hypothetical ; no experimental facts are known to me outside
of the formation of copper sulphides and arsenides. A large field
of investigation is open here. The relative absence of native
copper in these Arizona cuprite ores is surprising, because these
ores are evidently neither metamorphic nor pseudomorphic ;
but appear to be automorphic, the same as the native cope
ores of Lake Superior.

410 Koenig—New Species Melanochaleite and Keweenawite.

2, Heweenawite.

Occurrence.—In April 1901, driving the fifth level at the
Mohawk Mine, Keweenaw Co., Michigan, southwestward from
shaft No. 1 to shaft No. 2, a narrow cross vein was cut through,
carrying domeykite and a reddish-metallic mineral having the
color and general appearance of massive niccolite. Superin-
tendent Fred A. Smith sent me some of the material. Being
very busy at the time, I did not investigate this matter until
June, after my vacation had begun. The examination revealed
then my wrong, first impression, inasmuch as the substance
showed copper, nickel, arsenic in the ratio of 2:1. Early in
July I visited the mine and informed Superintendent Smith
that I proposed the name Aeweenawite for this undoubtedly
new mineral species. I went under ground for the purpose of
gaining some knowledge of the paragenesis of the arsenides for
which this mine has now become famous. The large domey-
kite-emohawkite vein crosses the amygdaloid copper-bearing
bed near the No. 1 shaft at an angle of nearly 45°. The
arsenides in large shining masses sit mostly against the reddish
amygdaloid, without the intervention of any selvage. There
is little parallelism among the several minerals (the least soluble
species forming the bordering fringe), which suggests the notion
of the fissure having been filled by crystallization from a stand-
ing mother liquor. In the absence of parallelism, the mother
liquor must have been in motion. The stoping operations on
this vein between the second and third levels tend to show
the vein as a system of flat lenses. One of them has been
stoped out with a rich yield. I found the new ‘vein to be
about 1,300 feet from the large one just described, near shaft
No. 2. The vein comes into the level very flat, almost parallel
with the strike of the bed. It is thin, with a maximum width
of 6 inches. The general character is very like that of the
large vein; to wit, absence of parallelism. Sometimes calcite
sits against the amygdaloid, sometimes quartz, or again either
domeykite or keweenawite. The conditions under which the
filling out of the vein took place must have been alike in both
veins. I intended to analyze a large sample of the amygdaloid
adjacent to the vein in order to learn whether the arsenic per-
vades the rock in minute quantities, to find proof or disproof
of a lateral leaching. Thus far I have not done this, though
the intention still exists. This is one of the reasons why I kept
back the publication of the present notice.

Physical Properties.—I have observed the mineral only as a
massive, very fine granular aggregate which is very brittle and
shows flat conchoidal fracture. Hardness about 4, similar to
that of domeykite. The color on the fresh fracture is pale

Koenig—New Species Melanochaleite and Keweenawite. 411

pinkish brown. On exposure tarnishes to a deeper brown
red; but is far more constant towards the atmospheroids than
domeykite or mohawkite. I have now some specimens before
me which have been exposed to the laboratory gases for over
a year and still show the characteristic color, but slightly
deepened. Specific gravity at 20° C.= 7-681 (with 3:140
grams of the mineral).

Chemical Characters.— B. B. melts easily to a globule.
Gives vapors of arsenic. On continued blowing in oxidizing
flame, a metallic globule is obtained. But if a borax bead be
placed along the globule from the first, while the O. F. is act-
ing, then a blue glass is obtained ; later a brown glass, and
ultimately a green glass, showing successively cobalt, nickel,
copper. The qualitative behavior is thus identical with
mohawkité. Dissolves in concentrated HNO, and even in
HNO, (Specific gravity 1-2).

A first preliminary quantitative determination gave Ou=
38°5; (Ni+ Co) = 17°98; quartz = 4°52; no iron—the dif-
ference must be arsenic. This was very different from
mohawkite.

A second analysis of the same sample, carefully made with
O, 500 grams, gave

00249 insoluble

0°3760 Mg, As,O,

0-2449 Cu,S |

0:0945 Ni + Co (by hydrogen)
0°0047 Co (by hydrogen).

From this follow the percentages.

Quartz = 4°98
As — eae 5 a “== 0493
Cu en ae ae = G2 t
a = a : 58-6 = o-ag5 (9944
Fe = trace
99°96

Thus the ratio obtains As : (Cu,Ni,Co) = 1,000: 1,915 and
again Ou : (Ni+ Co)=2:1 nearly.

It was. assumed by me that this ratio between the metals
would probably not be constant; but instead to a certain limit
a replacement of each by the others, the same as was assumed
for the mohawkite and has been fully proven by numerous
analyses, I have made since.

Keweenawite = (Cu,Ni,Co), As.

412 Koenig—New Species Melanochalcite and Keneenawite.

This result was communicated to Superintendent Smith at my
visit to the mine on July 6, 1901. The material had come
from the sixth level. I collected material at the vein crossing
on the fifth level. In appearance it was identical with the one
just described. The analysis made with 0°5 gram samples gave
the following data:

0°3379 Cu,S

0'0487 Ni

0:0047 Co

0°3560 Mg, As,O,

0°0039 quartz.

The percentages are:
Cu 53°96 :63 = 0°856

A,
—
atta ea
fe)
~J
i

34°18 Aoi oi pees (A
Quartz 0°78

99°60

The ratio (Cu,Ni,Co): As = 2:27:1, is not so close to 2:1
as that of the first material, but still sufficiently so. The
material, though freer from quartz, was not so faultless through-
out as the first. There appeared certain pseudocleavages
along which a thin olive-green film could be seen. Just to
what extent this condition has to do with preponderance of the ~
metals over the arsenic, [ am not prepared to say. But on the
other hand, this analysis shows unmistakably the replacement
of copper by nickel, and vice versa.

A third analysis was made with another sample, the arsenic
not determined.

Quartz = 060
Cu == 40°72, 3: 63 (= 06646
ae a se : 58-6 = 0346 ( 999?
Fe = trace
(diff) As = 88°42 os. Was 0°515
100°00

(Cu, Ni, Co)inats = T1282 4

x
In my paper on mohawkite (this Journal, December, 1900)
I showed how an arsenide Cu,As forms very easily when the
vapors of arsenic act upon copper at red heat. This artificial
Cu,As shows the color and crystalline structure of chalcocite.
The difference in color in keweenawite must, therefore, be
owing to the nickel.

Koenig—New Species Melanochaleite and Keweenawite. 413

I give the specific gravity (1. ¢.) as 7°71, the calculated specific
gravity as 7°75, whilst the keweenawite gives the gravity 7-681,
which is very close indeed. Nickel and cobalt having nearly
the same specific gravities and not much different from copper,
this slight difference is readily accounted for.

38. Mohawkite and Domeykite.

During the past year I have made several more analyses of
these two minerals from the Mohawk mine. The material
was in part collected by me on the spot or sent to me by Mr.
Fred Smith, to whom I wish to express my indebtedness.

The colors of these minerals seem alike to those of the
fabled chameleon, and the identification by the eye is made
very difficult, in fact impossible.

a. The mineral appears in form of large masses in calcite
with a color equaling that of chalcopyrite; altogether unlike
any previously seen domeykite or mohawkite. Mr. Smith

- states that it had this yellow color when it came up the shaft..

It is exceedingly brittle, crumbling between fingers into blue
and purple fragments. B. B. shows trace of cobalt and trace
of antimony.

Cu = 70°56

As = 29°50
It is therefore domeykite.

6. Large iron gray mass, looks like arsenopyrite. B. B. gives.

strong cobalt reaction.

Cu = 67°86
Ni + Co =. 3°32
AS =e nO

99°28

The mineral proves to be mohawkite ; but the sum of cobalt
and nickel is only one-third of that in my original mohawkite.
c. Mr. Knight, one of my students, picked up a specimen at
the Mohawk mine, which seemed to be peculiar. It is tough,
but not so much as what I have called mohawk-whitneyite,
gray color and fine granular structure somewhat like algodo-
nite in appearance. 7
B. B. reaction for antimony, no cobalt, no nickel.
Spec. gr. 8°378 — 8°364 (with 1°747 grams)
Cut= S02. :. G3 = 1281
Ais == VOT ts ==)0'255
Sb = 0re# 120'=" 0-007
100°68
Cur (Acs, a9 2 eo

414 Koenig—New Species Melanochalecite and Keweenawite.

On the face of this result the establishment of a new species
would seem indicated. But as I have expressed myself (1. ¢.),
the higher ratios of these copper arsenides are either down-
right mixtures (grain very fine) or they pertain of the nature
of alloys and hence all ratios are possible. In order to throw
a little more light on this subject, a long sliver was knocked
off the type specimen. On this sliver the mixture of two
substances was discernible. The sliver was broken into three
fragments and in each the copper was determined.

a (0'5063 gram) gives Cu,S = 0°4836 per cent Cu = 76:2
8(0'500 gram) “ So sa OUIASY ps Cu.= 75°76
y (0506 gram) “ cs eS ODS. kn as Cu = 80°10

These differences would seem to bear out my statement. A.
straight ratio marks an accident rather than a necessity. In
speaking of mixtures like these, it may perhaps be well to use
the term mohawk-algodonite, as distinguished from true
algodonite.

d. Mr. Rogers, another of my students, collected a specimen
which did not quite correspond to anything I had seen before.
It was much tougher than the preceding specimen. A frag-
ment weighing 0391 gram was dissolved and. the copper
determined. Arsenic by difference. Nothing insoluble.

Cu,S = 0:4208, Cu = 0°3361 = 85°9
As = o= dard
: 100°0
Ratio 7°25: 1
It must be classified as mohawk-whitneyitte.
e. Genuine Algodonite from the Champion Mine, on the
South Copper Range.
I am indebted to Dr. L. L. Hubbard, the Superintendent,
for the specimen. It presents a piece weighing about 400
grams, through which a drill-hole passes. The specimen was
encountered on a cross vein, very narrow, similar to the oceur-
rence at the Mohawk mine. The metallic mineral is inti-
mately mixed with calcite. The fresh fracture is steel grey.
Several fragments were digested with dilute HCl. The residue
carefully washed and dried weighed 0°1968 grams. This quan-
tity was dissolved in HNO, and boiled nearly to dryness in
order to bring all the arsenic to the form H,AsO,. Only a
trace of white insoluble material (quartz) was observed.
00683 Mg,As,O, = 0:0328 As
0:2063 CuO = 01645 Cu
Hence Cu= 83°53

As = 16°55). @u: As=5:9678

100°18

Koenig—New Species Melanochaleite and Keweenawite. 415

Ff. Pulveriform chalcocite from the Champion Mine.

This very interesting material was also encountered in a
small cross vein and recognized by Dr. L. L. Hubbard as some-
thing peculiar. Neither of us thought of chalcocite. It soils -
the paper or fingers like soft pyrolusite or graphite. An
aggregate of small crystals, the largest less than 0°5™™ to micro-
scopic individuals.

The dust appears under the microscope as made up of single
hexagonal plates, or groups of plates, easily disengaged and
reduced to individuals, perfectly opaque. I have since observed
this same substance both at the Champion and Mass mine,
dusting over groups of calcite crystals, either loosely or firmly
imbedded in the calcite, causing the latter to look gray or
black. It becomes bluish by tarnish. Bb. B. reacts for copper
and sulphur only. The record of the quantitative analysis is
lost; but I remember distinctly that the percentage of copper
was close to 79.

g. Nodular nuggets found in the coarse material from the
mortar at the stamp mill of the Baltic mine were singled out
by Mr. William Vivian as whitneyite. If seen by themselves
they look much like rounded native copper; but seen along-
side of the latter a difference becomes noticeable even to
the untrained eye. Compared with one another, on filed or
ground surface, slighter differences appear. Held in a vise, a
strong blow with a cold-chisel will break the whole in two.
Native copper will not do this. The fracture is hackly; there
are small geodes of caicite. Calcite is visible between the
imperfect crystals of gray color. But the lens also reveals
minute black globular bodies. These latter I sneceeded in
identifying as chalcocite. The material for analysis was
extracted with diluted HCl. After extraction material
weighed 0-4003 gram.

It gave: Insolubles = 0°0081; Cu,S = 0°4575; Mg, As,O,=
0:0555
pelence: .Cu = 91°33 : 62 = 1-450
es: = OnGOs 75 == 0:0888
Quartz =" 2°20

100°13 Cu : As = 16'22:1

h. From Captain James Hoatson, of Calumet, I received a
specimen showing quartz and what looks like native copper, at
first sight. The supposed native copper is similar in color to
the nodules just described; also similar in hardness and rela-
tive toughness. A filed surface shows a very uniform, dense
texture, with a decided yellowish, brassy, color; but etched

Am, Jour. Scl.—Fourts Series, Vou. XIV, No. 84.—DECEMBER, 1902.
29

416 Koenig—New Species Melanochalcite and Keweenawite.

with HNO,, the heterogeneous nature becomes visible even to
the naked eye.
Two fragments gave :

Cu = 92°78 Ou = 93°96 1:°4755 Cu: As = 18°8:1
As = 5°85 As = 5°74 .0°0765 Cua: As =19°3:1

98°63 99°70

Two other fragments were detached close to the one whose
composition has just been given.

They gave: 3
Cu = 96°2 and 94°5
Ag == 38 5'5 (by difference)

—— =

100°0 100°0

Here the ratios are Cu: As=30°1:1 and 25°5:1. The figures
show plainly that definite proportions cannot be looked for; ~
but they indicate, as I look at it, that we must assume here
conditions similar to those of pig iron, or of alloys. Further
studies may give better light, especially in the direction of
etching. In speaking of such alloys as these (g, 2), I ven-
ture to propose the word “ Semi- Whitneyite,” to be used like
Mohawk-Whitneyite, or Mohawk-Algodonite, meaning neither
a species nor a variety; in the sense of a rock rather than of a
mineral.

Chemical Laboratory, Michigan College of Mines,
Houghton, Mich.

T. Holm--Studies in the Cyperacee. 417

Art. XXXIX.—Studies in the Cyperacee ; by THEO. Hou.
XVII. Segregates of Carex Tolmiet Boott. (With figures
in the text.) :

SIMILAR to the prevalent distribution in the north the genus
Carex, as represented in the Rocky Mountains of Colorado,
does. not attain its greatest development until the timber-line
is reached, where conditions exist which are favorable to the
development of Carices. From the timber-line with the low,
open thickets of willows (S. glauwcops) accompanying the water-
courses, to the alpine slopes with the snow-banks, we meet with
no small number of species over a relatively limited area.
Those localities correspond, in many respects, with those of
the north, and it is quite natural that we often find ourselves
confronted with species that are, also, known to occur even in
the polar-regions. Several of these alpine and subalpine
Carices have been well recognized as being identical with
certain northern, arctic and circumpolar, species, and we need
only refer to such as: C. nardina, incurva, festiva, rupestris,
atrata, misandra, etc., while others appear as indigenous only
to the Rocky Mountains, e.g. Carex jilifolia, Engelmannii,
elynoides, variabilis, ete.

The mode of variation in alpine Carices corresponds well
with that of northern species, indeed we have exactly the same
variations exhibited by several of these, e. g. C. alpina, atrata,
rigida, misandra, etc. And Carex variabilis, at least the
plant from the Sphagnum-swamps at higher elevations in the
Spruce-zone, varies so much, that we have observed forms that
are absolutely analogous with the subarctic 0. anguillata,
hyperborea, stans and the various forms of C. rigida, even if
they may not be looked upon as specifically distinct. We have
collected a number of such forms of C. variabilis growing
almost side by side, and possessing the same structure and
shape of utriculus, besides the same color of squamee, but vary-
ing in the length of bracts, peduncles and spikes, characters
that are very prominent in C. anguzllata, stans and hyper-
borea ; in these species, however, the utricle offers such charac-
ters as seem to afford good evidence of their specific validity,
as they were formerly established by Salomon Drejer, the most
eritical of Caricographers.

However, the ability to distinguish and recognize several of
these species from the alpine regions depends to a large extent
upon observations in the field as well as a thorough acquaintance
with their homologues in other countries. And one of the
species, which we “intend to describe in the present paper,
Carex scopulorum, offers an excellent example of analogous

418 T. Holm—Studies in the Cyperacee.

variation, which might have resulted in the establishing of
several species, if it were not that we had the opportunity of
studying the plant in its natural surroundings. It is a species
that has, so far, been entirely misunderstood, and, strange to
say, 1t has at various times been confounded with such species
as C. Tolmiei, rigida, vulgaris and pulla,—to none of which
it bears any close resemblance.. It seems as if the mode of
growth, the color, shape and direction of the utricle in this
species has not heretofore been very critically examined. And
the morphological structure of the rhizome with its floral and
leafy shoots, besides the mere color of utriculus, its outline and
direction, are characters of no small importance for distinguish-
ing critical species, and are often much more constant than the -
number of stigmas, of spikes, distribution of sexes, ete.

Thus is the ramification of the rhizome in Carex either
mono- or sympodial, the former being rare, but, as we have
mentioned in a previously published paper,* it is characteristic
of a number of alliés of C. laxiflora, digitata, ete., to which it
appears as absolutely constant, while the latter, the sympodial,
is to be observed in most of the other Carices. The cespitose
and the stoloniferous growth is nearly always constant in the
respective species, and such variations as have been noticed in
a few cases, for instance in C. vulgaris, where a ceespitose rhi-
zome changes into one that is stoloniferous, it appears to
depend merely on the character of the soil, and is, therefore,
of less importance.

Finally we may mention an additional character derived from
the foliar organs at the base of the flower-bearing stem, which
was first described by Elias Fries,t and which consists in these
basal leaves being provided with large, assimilating blades or
being merely scale-like with rudimentary or non-developed
blades, at the time of the flowering. He gave the name “ Phy]-
lopodic ” to the former, and ‘ Aphyllopodic” to the latter.
Besides the species with monopodial ramification, which are all
“aphyllopodic,’ not a few species of those with sympodia
belong to this same category. Such phyllo- and aphyllo-podic
species are readily to be distinguished from each other, inas-
much as the character appears constant; low forms of the
aphyllopodic Carex macrocheta exhibit in this wise an entirely
different aspect from the phyllopodic C. ustulata, with which
such dwartfish forms have often been confounded, as well as
the aphyllopodic C..cwspitosa may be distinguished at once
from the phyllopodic C. vulgaris and its allies. It is a feature
which deserves much attention, but seems, so far, to have been

* This Journal, vol. i, 1896, p. 348.

+ Fries, Elias: Synopsis Caricum distigmaticarum, spicis sexu distinctis,
in Scandinavia lectarum. (Botaniska Notiser, 1843, p. 97.)

T. Holm—Studies in the Cyperacec. 419

overlooked or neglected by descriptive authors; as a matter of
fact it constitutes one of the principal differences between C.
Tolmiet Boott and our C. scopulorum, and the confusion
would have been avoided had this particular point been duly
considered.

In respect to the utricle, the peculiar dull, brown color
noticeable in C. sty/osa from Greenland and Alaska seems very
rare, but is, nevertheless, also common to C. scopulorum, and
constitutes, thus, an excellent distinction, and appears to be
constant. A very prominent character may, also, be derived
from the outline and direction of this same organ, the utricle,
but it is very important to study this from life or from mate-
rial preserved in alcohol. The utricle in C. scopulorum is
turgid with the beak abruptly bent to the horizontal, a feature
that makes the plant at once distinct from any of those to
which it was formerly referred. But in regard to the other
characters, such as the relative development of the spikes, their
length, thickness, the peduncles, distribution of sexes, etc.,
several of these prove less constant and are, as we remember,
often the principal foundation for the establishment of varie-
ties, viz: zsogyna (C. dioica), epigejos (C. aquatilis), simpli-
ctor (C. paniculata), ramosa (CU. teretiuscula), acrogyna (C.
Pseudocyperus), remotifiora (C. vulpina), pendula (C. stricta),
spinosa (CU. hirta), anomala (C. acuta), ete. |

It, thus, appears as if there were a number of reliable points
to be observed and by which one might determine obscure or
imperfectly known species of the genus Carex, so as to place
them in the sections where they naturally belong; but it is,
nevertheless, a very difficult task to draw such distinction in so
large a genus as that of Carex, when the supposed nearest
related species are not at hand, but must be studied from the
diagnosis alone. And in regard to Carex Tolmiei Boott it
would seem absolutely fruitless to gather any knowledge about
this particular plant from the latest descriptions published in
this country, where the species is said, for instance, ‘‘ to repre-
sent C. vulgaris Fr. on the western side of our continent,”
where the true C. vulgaris Fr. is amply represented and occurs
with much the same variable habit as in the old world!
Moreover the diagnosis of C. Zolmiez as reproduced in recent
systematic works is so as to make it “specifically distinct”
from the one established by Boott; in other words, the mate-
rial has not been correctly identified, nor has the original diag-
nosis been carefully compared. Very little comfort is gained
from consulting the larger herbaria, where we found, for
instance, C. atrata, Parryana, Raynoldsii, macrocheta, ete.,
identified as Boott’s C. Tolmiez.

Under such circumstances one feels obliged to consult the

420, T. Holm—Studies in the Cyperacee.

original diagnosis as a last resource, and it is, also, the safest
method. We were in the present instance under the impres-
sion that our Carex from Colorado did belong to C. Tolmiez,
judging from the very numerous references in literature and
herbaria to the same plant from various parts of the Rocky
Mountains, but our suspicion became aroused when comparing
the diagnosis as written by Boott and reproduced by recent
authors. Boott’s comprehensive work is, of course, a great
help, but it is, as may be said of all other monographs of large
genera, far from infallible, and the writer is greatly indebted
to Mr. C. B. Clarke of Kew for his kindness in loaning us one
of Boott’s own specimens of C. Zolmiei, and for calling our atten-
tion to certain points in the figured details, which are less cor-
rect. We have, thus, had the opportunity to compare one of
Boott’s specimens with his diagnosis, and have succeeded in
acquiring a fuller knowledge of the plant than otherwise
obtainable.

Most of the other species with which C. Zolmzez has been
confused were already familiar to us; the remaining are evi-
dently undescribed and will be treated as such, where sufficient
material becomes accessible. But before we present the diag-
nosis of some of these segregates, we deem it necessary to
give a-few data concerning some external characters of the
true C. Tolmiet Boott, by which it may be readily distinguished
from its supposed allies.

The species C. Zolmiez of Boott is aphyllopodic, with the
rhizome densely matted and with short stolons covered by a
mass of fibers from the dissolved scale-like leaves, and the
sheaths of the basal, proper ones, by Boott himself cor-
rectly described as: ‘ Rhizoma horizontaliter repens, fibris
lanatis.” The relative large number of almost contiguous pis-
tillate spikes is a striking feature, also the long and slender,
setaceous peduncles of the lower spikes. The number of stig-
mas is constantly three, and the utricle is, as it appears, mostly
two-nerved, granular above and often purplish spotted with
a short emarginate or, sometimes, slightly bidentate beak.
Among the specimens which we have examined of C. Yolmiez
were several which answered very well to Boott’s diagnosis of
C. nigella: “A C. Tolmied differt spicis paucioribus, masculis
2-3, perigynio bidentato majore, squamis lanceolatis mucro-
natis.” However, we observed among typical CO. Tolmiei
utricles with bidentate beaks, and several with the scales
mucronate. It, thus, appears as if C. nigella Boott may not
be specifically distinct from C. Zolmzec Boott, and, moreover,
Mr. Ciarke has informed us that he has examined the very
specimens in herb. Boott collected by Toulmie and labelled
“nigella,” and that he feels most inclined to consider them

T. Holm—Studies in the Cyperacee. 421

identical with C. Tolmiei, even that they were collected from
‘“‘the same tuft” as this. It must, also, be borne in mind that

the republication of C. nigella (Ill. genus Carex, p- 194) was
done after the death of Boott, and Mr. Clarke, therefore, thinks
that Boott himself would have reduced it.

429 T. Holm—Studies in the Cyperacee.

But whatever C. nigella may be, a distinct species, a variety
. or identical with C. Tolmzez, the latter itself cannot in aceord-
ance with Boott’s diagnosis be confounded with any of the
other species of this section, the MZelananthe Drej., among
which we prefer to place it rather than among the Micro-
rhynche Drej. Some authors, among which Boott himself,
did think that the affinity should be looked for among the
Microrhynche Drej., and he compared it with C. rigida Good.,
but was, nevertheless, well aware of the fact, that even if C.
Tolmiet “has much the aspect and in some respects the habit
of C. rigida, it differs from this in the number of spikes, the
three stigmas, triquetrous achenium, etc.” Following the sug-
gestion of Boott, Professor Bailey places the species among
the Microrhynche, in which he includes both Drejer’s Zora-
stachye and Melananthe ; thus the section becomes very unlike ©
that proposed by Drejer, and must be credited to Professor
Bailey alone; the plant becomes, thus, associated with such
species as C. vulgaris, stricta, acuta, glauca, salina, ete., a
classification too unnatural to be acceptable. But then it is
not so strange to understand how a number of very remote
and very distinct species of no immediate affinity have been
plunged together into one: “ C. Tolmzez,” the characters of
which have gradually vanished in the systematic treatises of
the genus in this country. And some of the plants which in
this way have been united with C. Zolmiei are the species
which we intend to describe in the following pages as inde-
pendent species, and not by any means closely related to the
original C. Tolmiet of Boott. These segregates of C. Tolmies
are;
Carex scopulorum Holm (figs. 1-6).
(C. Tolmiet Bail. non Boott var. subsessilis Bail. ex parte.)

Roots thick, very hairy ; rhizome stoloniferous, densely cov-
ered with persisting (not fibrillose) scale-like leaves; culm from
10 to 40™ in height, erect, rather coarse, triangular, more or
less scabrous, the base surrounded by green leaves (phyllo-
podic) ; leaves shorter than the culm, relatively broad and flat,
scabrous along the margins, otherwise glabrous; spikes very
variable in number, from two to seven, the upper ones mostly
contiguous: the terminal mostly purely staminate, sessile or
short-peduncled, clavate, or sometimes androgynous or, though
seldom, gyneecandrous,* the scales (fig. 4) oblong-lanceolate,;

* As to the distribution of sexes and number of spikes I find in 65 speci-
mens of Carex scopulorum from various localities in Wyoming, Montana,
and Colorado :

42 specimens with the terminal spike purely staminate.
17 an a & ‘¢ androgynous (staminate above).
6 ey x ce ‘¢ gyneecandrous (pistillate above).

T. Holin—Studies in the Cyperacee. . _ 423

obtuse, black or déep brown with thin, membranaceous mar-
gins, the midrib pale, not excurrent; lateral spikes mostly
three, thick, erect or spreading, when crowded near each other
(fig. 3), purely pistillate or the upper ones androgynous (fig. 2),
at maturity almost squarrose, the upper sessile, the lower borne
on short, stout peduncles; bracts not sheathing, with black
auricles, only the lower ones leaf-like and about as long as the
axillary spike; scale of pistillate spike (fig. 5) ovate, acute,
black with pale, not excurrent midrib; utriculus (fig. 6)
minutely granular above, mostly longer than the scale, thin in
texture, shortly stipitate, turgid, with a short entire beak,
abruptly bent to the horizontal, dull brown, often purplish
spotted above, two-nerved; caryopsis sessile, light brown,
roundish in outline with both faces slightly convex ; stigmas
two, base of style persisting.

We found this species very abundant in the region of Clear
Creek Carfion (Colo.), also near Leadville (Colo.); it grows in
thickets of willows along creeks at au elevation of between
3,600 and 3,900 met.*

Carex prionophylla Holm (figs. 7-11).
(Carex Tolmiei Bail. non Boott var.)

Roots thick, very hairy; rhizome densely cespitose with
numerous persisting, very scabrous, reddish brown scale-like
leaves; culms aphyllopodic, until 45°" in height, erect, stiff,
triangular, very scabrous and leafy to abont the middle ; leaves
of sterile shoots as long as the culm, with long, tubular sheaths,
as those of the culm, very scabrous along the margins besides
on both faces of the blade and on the sheaths, the blade flat ;
spikes mostly five, contiguous, or the lowest one in some dis-
tance from the others; the terminal staminate, short, sessile ;
the scales (fig. 9) elliptical, black with pale, not excurrent mid-

24 specimens, some of the lateral spikes androgynous.

41 _ all the lateral spikes purely pistillate.
de Specimens with in all 3 lateral spikes.
a ee ee 9 ims ee
ie 66 ce 4 4 a4
3 ee ce 1 ce iad
I a4 ce 5 ce ee
at ee oe 6 66 ee

* Specimens examined: Wyoming: Head of Big Goose Creek, Big Horn
Mts. (Fr. Tweedy); Sierra Madre Mts., Carbon County. Battle Lake Mt.,
9,500 ft., scarce on a wet, shaded north-slope (A. Nelson) ; Little Goose
Creek (A. Nelson). Montana: Old Hollowtop, near Pony Mt. (P. A. Rydberg
and H. A. Bessey). Colorado: Marshall Pass, 12,000 ft. abundant in wet.
places, covering extensive areas (C. F. Baker); Silverplume (P. A. Rydberg) ;
Headwaters of Clear Creek and the alpine ridges lying east of Middle Park
(C. C. Parry); abundant along Quail Creek near Stevens’ mine, in low
thickets of Salix giauwcops ; Headwaters of Clear Creek ; in swamps on Gray’s
peak ; very frequent in swamps on Mt. Massive near Leadville (the author).

494 _ I. Holm—Studies in the Cyperacec.

rib ; pistillate spikes very short, about $™ in length, sessile, the
lowest one subtended by a sheathless leaf-like bract, reaching
to the base of the staminate spike, the other bracts scale-like,
bristle-pointed ; -scales (fig. 10) ovate-acuminate, black with
excurrent pale midrib; utricle mostly shorter and broader than
the scale (fig. 11), rhombic-oval, compressed, stipitate, granular,
brownish green with purplish spots, membranaceous, with a
short, straight, emarginate beak, obscurely two-nerved ; stig-
mas two with the style enclosed ; caryopsis immature.

Habitat: Northern Idaho. Region of the Ceur d’Alene
Mountains; near mountain streams; Divide between St. Joe
and Clearwater River. Alt. 1,510". July 10, 1895; collected
by John B. Leiberg.

Carex gymnoclada Holm (figs. 12-14).
(C. Tolmiez Bail. non Boott var. angusta Bail. ex parte.)

Roots thick, densely hairy ; rhizome stoloniferous with shin-
ing, reddish brown persisting, scale-like leaves; culm from ¥8
to 44° in height, erect, slender, triangular, scabrous, leafless,
phyllopodic ; leaves as long as the culm, narrow, flat, scabrous
along the margins; spikes few, mostly three, 1™ in length,
situated near the apex of the culm, but not contiguous; the
terminal staminate or, sometimes, androgynous, short-pedunceled ;
the scales (fig. 13) somewhat spathulate-oblong, deep brown
with pale, not excurrent midrib; lateral spikes pistillate, mostly
two, erect, sessile or nearly so; bracts not sheathing, with
_ black auricles, the lower leaf-like and reaching to the staminate
spike; scales of pistillate spike (fig. 14) elongated oval, black
with pale, not excurrent midrib, shorter and narrower than the
utricle; the utricle minutely granular, especially above, with a
few prickle-like projections near the beak and around the ori-
fice of this, membranaceons, rhombic-oval, stipitate, compressed,
with a short, entire beak, yellowish green, obscurely two-
nerved ; caryopsis sessile, light brown, oblong; stigmas two,
style flexuous within the utricle.

Habitat: Eastern Oregon, bogs of Hurricane Creek, 6,000
ft. alt. Aug. 28, 1900, collected by Wm. OC. Cusick.

In considering these segregates of Carex Tolmiei Boott, C.
scopulorum is at once distinguished by being phyllopodic, by
its turgid utricle, the two stigmas, etce.; C. prionophylla differs
from C. Tolmiei especially by its profuse, scabrous covering,
its small spikes and two stigmas, and C. gymnoclada by being
phyllopodic, besides by the spikes not being contiguous, by
the structure of utriculus and the two stigmas. Carex scopu-
lorum may be placed among the JZicrorhynche Dre}j., where
it is somewhat anomalous, however, on account of the turgid

_ LT. Holm—Studies in the Cyperacee. 425

utricle ; C. prionophylla shows some attinity to C. cespitose L.
of the same section, and (. gymnoclada may be placed with
this near C. vulgaris Fr.

Brookland, D. C.

426 Ford—Chemical Composition of Dumortierite.

Art. XL.—On the Chemical Composition of Dumortierite ;
by W. E. Forp.

Introduction.—Dumortierite was first discovered near Beau-
nan, France, by Gonnard, who recognized it as a new mineral
and named it after the paleontologist, Eugene Dumortier.*
The mineral occurs at the original locality rather sparingly as
fine grains or needles enclosed in pegmatite, associated with a
gneiss. Sufficient material, however, was obtained for an
analysis which was made by Damourt while the optical prop-
erties of the mineral were studied by Bertrand.t Dumorti-
erite was next found near Harlem, Manhattan Island, New
York, where it occurs in a pegmatoid portion of a biotite
gneiss. ‘The mineral from this locality was first thought to be
indicolite, the blue variety of tourmaline. An analysis, how-
ever by Riggs, anda study of the optical properties by Diller
proved that it was not tourmaline, but presumably a new min-
eral. Later it was shown by E. S. Dana to be identical with
the dumortierite of the French locality.| A little later the
mineral was discovered at Clip, Yuma County, Arizona, and
two analyses of material from this last locality, together with
one of the Harlem material, were made by Whitfield.4]

All of the five analyses thus far published show considerable
variation in the composition of the mineral and none of them
yield a satisfactory formula. The original analysis by Damour
showed the mineral to be essentially an aluminium silicate con-
taining a small amount of water. The presence of boron evi-
dently was not suspected. The analysis, it should be stated,
was made on a limited quantity of material, only 0°41 gr. being
used. The analysis by Riggs of the Harlem dumortierite
showed the presence of B,O, and also a considerable ‘amount of
alkalies. The presence of alkalies has given rise to some ques-
tion as to the purity of the material analyzed, while Whittield
ascribes the B,O, to the possible presence of tourmaline. The
amount of B,O, found by Riggs was about 4 per cent ; hence,
since tourmaline contains only about 10 per cent of B,O,, it
would be necessary, according to Whitfield’s hypothesis, to
assume the presence of over 40 per cent of tourmaline in the
material analyzed, a supposition quite untenable, as indicated
by the proportions of the other constituents found by Riggs.

The analyses by Whitfield of dumortierite from Harlem and
Clip are discordant, since only a trace of B,O, is reported in
the Harlem mineral, while varying amounts, 4°94 and 2°62, are
indicated in the two analyses of material from Clip. Of the
Harlem material only 0:217 gr. was available for analysis and a

* Bull. Soc. Min., iv, 2, 1881. + Ibid., iv, 6, 1881.

t Ibid., iv, 9, 1881. $ This Journal (3), xxxiv, 406, 1887.

|| Ibid. (3), xxxvii, 216, 1889. “| Ibid., xxxvii, 216, 1889.

Ford—Chemical Composition of Dumortierite. 427

direct determination of B,O, was not made. It is stated,
however, that but the smallest trace of B,O, was observed. It
may here be noted that it is not especially easy to get a satis-
factory qualitative test for boron in dumortierite, and without
a quantitative determination one might easily be led to a wholly
misleading conception as to the amount of B,O, present;
hence it cannot be assumed that the absence of notable quan-
tities of boron in the Harlem mineral has been satisfactorily
proved. It has been found, as will be shown in the course of
this article, that the dumortierite from Harlem does contain
boron, and it also occurs in notable quantities in the mineral
from the original locality in France, as indicated by a distinct
qualitative test made by fusing on platinum wire with potas-
sium bisulphate and fluorspar; hence, boron seems to be an
unfailing constituent of dumortierite.

All analyses thus far made agree in indicating that dumor-
tierite is a very basic silicate of aluminium, yielding a small
loss of weight on ignition, presumably water, but they show
marked variations as regards boron. The present investiga-
tion was undertaken, therefore, with the idea of making a
special feature of the boron determination and of establishing,
if possible, the exact chemical composition of the species.

Material for Analysis.—It was possible to obtain material
from three independent sources, Clip, Arizona; Harlem, New
York; and San Diego Co., California. From the first locality
material was obtained from specimens in the Brush collection,
and also from an abundant supply of the mineral sent by Mr.
Geo. L. English of New York. The cumortierite from Clip
occurs as small columnar aggregates imbedded in a matrix of
granular quartz, associated with a little magnetite and cyanite.
Pure material was separated by pulverizing a large quantity of
the rock and treating, first with potassinm mercuric iodide in
order to separate the quartz, then by suspension in barium
mercuric iodide so as to obtain the mineral in condition of
greatest possible purity. After separation in this manner the
dumortierite was picked over carefully by hand and then
allowed to stand for several days in hydrofluoric acid in order
to remove any remaining traces of quartz which might be
attached to the mineral. It may be stated that hydrofluoric acid
in the cold has practically no action on crystals of dumortierite.

_Material from Harlem was obtained from specimens in the
Brush collection and rendered pure by the means. described
above. The material from San Diego Co., California, was
obtained from some specimens which were sent to Prof. S. L.
Penfield by Mr. E. Schernikow of New York. The general
locality of the occurrence of this type of dumortierite has been
confirmed by Mr. W. Tassin of the U. S. National Museum in
Washington, who kindly sent a specimen for comparison with

428 Ford—Chemical Composition of Dumortierite.

the material that was analyzed. The exact locality from which
the mineral came could not, however, be learned. This mate-
rial shows a marked difference in appearance from that of
dumortierite from other known localities, as it has a pro-
nounced lavender color rather than blue. The pleochronism
of this type is analogous to other dumortierite, but ranges from
deep lavender to colorless instead of from deep blue to color-
less. The optical orientation of the crystals is the same as in
the case of the dumortierite from Clip. It occurs in masses
of considerable size, of radiating columnar structure, associ-
ated with granular quartz and a light colored mica from which
it was easily separated by means of the heavy solutions, and
thus obtained in a pure condition. |

Method of Analysis—The method of analysis presents no
new features and need be only briefly discussed. The silica,
alumina, and ferric oxide were determined as usual. Boron
was determined by the Gooch method* using the precautions
suggested by Pentield and Sperry, recorded in their paper on
the Composition of Howlhtet and by Penfield and Foote in
their paper on the Composition of Tourmaline.t It was found
that water could not be determined by simple loss on ignition
and, accordingly, the method suggested by Penfield§ of igniting
with a weighed quantity of lime, was employed. The loss on
ignition in this case was found to be less than when the min-
eral was ignited alone. A somewhat different method of
analysis had to be adopted in the case of the dumortierite
from Harlem on account of the small amount of material
available, and it is fully realized that the results are not to be
looked upon as being as exact as those of the other analyses.
The analysis was made on only 0°6 gr. of the mineral, which it
was necessary to divide into two portions. In one of these,
water was determined in the manner described above, and in
the other boric oxide and the remaining constituents, silica,
alumina and ferric oxide.

The results of the analyses are as follows:

Analysis No. I. Dumortierite from Clip, Arizona.
Specific gravity close to 3°319.

Average. Ratios.
ION wi Be 30°00 29°66 29300 29°86 "497
ALO} oS 6320 63°74 63°76 63°56 617 X%6 =a
FerGioxie °23 "23 "23 “001 X6 ‘006
Boyne 5°47 5°06 5°26 1: > COF0RS "420
HO) ase 1°45 1°38 1°41 “070X2 *140
100°32 4°268
4°268 : 497 = 60: 6°99
* Amer. Chem. Jour., ix, 23. + This Journal (3), xxxiv, 220, 1887.
t Ibid. (4), vii, 97, 1899. § Ibid. (3), xxxii, 109, 1886.

Lord—Chemical Composition of Dumortierite. 429

Analysis No. II. Dumortierite from San Diego Co., Cal.
Specific gravity between 3°226 and 3°43.

Ratios.
2 ee 30°58 509
2) ae 61°83 600 X6 = 3°600
Bete. *36 002 X6 “O12
B,O, ra 2 5°93 085 <6 510
H,0 ee ee 2°14 cL Se2 "238
100°84 _ 4°350

4°350 : 509 = 60: 7°02

_ Analysis No. III. Dumortierite from Harlem, N. Y,
Specific gravity between 3°211 and 3°302.

Ratios.
lee ees > 2 31°24 520
Pee os. 2 (L926 596 X6 = 3'576
eer i Fo ED "0006 X 6 004
Peele 6714 ‘088 x6 528
120) Pee ee 2°09 116 <2 "232

4°340
4-340, =°520 = 60+-7°19

Discussion of Analyses.—It would seem that the chief
problem in arriving at a formula for dumortierite would be
the determination of the roles played by boric oxide and
water, but the analyses indicate that these constituents are
present in such variable proportions as not to yield definite
ratios with the silica. In the first two analyses the ratios of
SiO,: B,O, are respectively 7:1 and 6:1, with percentages of
H,O increasing with increase of B,Q,. Tt may be assumed,
therefore, that the boron in the mineral has not the nature of
an acid element, occurring in a definite proportion, but must
be considered rather as having basic qualities, replacing alu-
minium as an isomorphous constituent. Moreover, there is no
simple relation between the H,O and the other oxides, which
fact, taken‘ with that of the small amount of water present
and the difficulty with which it is driven off from the min-
eral, has led to the assumption that it, too, plays the part of a
basic constituent in the composition of the ‘mineral. That
water does at times play such a réle has been abundantly
proven. The results of the analyses were, therefore, treated
as follows: The ratio of each of the basic oxides was multi-
plied by a number which represented its equivalent in hydro-
gen, and in this way an expression for the acid from which
dumortierite is derived was obtained. Thus treated sag
No. I gives the ratio of SiO,: H = 6:99:60, No. II, 7:02: 60

430 FHord—Chemical Composition of Dumortierite.

and No. III, 7:19:60, the ratios of all three being very close
to 7:60. Analysis No. II] shows the greatest variation from
this ratio but the method of analysis which had to be employed
in this case, as explained above, would account for a high ratio
of the SiO,. The ratio as given leads to an acid having the
empirical formula, HS. 5. ‘and, on the assumption that the
hydrogens are wholly replaced by aluminium, the formula
becomes Al,,8i,0O,, or developed as a basic orthosilicate,
(A10),,A1,(Si0,),.. In the foregoing formula it must be under-
stood that boron and a little hydrogen replace a portion of the
aluminium. In Analysis I, in which the ratio of B,O,: SiO,=
1:7, the formula approximates very closely to [ AlO],,A1,B,
(SiO,),.. The theoretical percentages penta oe to this
formula are as follows:

SiON gist PM ey ome
BIO Sous. Ua agere eb
BeOican Ase wis Caan 4:97

100°00

It is of interest to note here a new occurrence of dumer-
tierite. During the preparation of this article a blue mineral
was sent to Prof. E. 8. Dana for determination by Mr. R. M.
Brereton of Woodstock, Oregon, and which was examined by
the present writer. It proved to be dumortierite of a some-
what new habit. The mineral occurs in small spherules, about
1™= in diameter, imbedded in a_ light-colored, fine-grained
siliceous gangue. Each spherule when broken shows a radi-
ated fibrous structure, and the mineral has a beautiful blue
color. When the gangue is pure white a polished specimen
of this material gives a very pleasing effect. This dumor-
tierite shows the pleochroism characteristic of the blue variety.
It was scarcely possible to obtain material sufficiently pure for
analysis, but a decisive qualitative test for boron and approxi-
mate determinations of the other constituents left no doubt as
to the identity of the mineral. The locality of this occurrence
of dumortierite as given by Mr. Brereton is on the headwaters
of the North Fork of the Washougal River in Skamania
County, Washington.

The writer wishes to express his obligations to Messrs.
English and Schermikow for the material which they gave for
analysis and to Prof. S. L. Penfield for his constant advice and
assistance.

Sheffield Laboratory of Mineralogy, Yale University, New Haven.

J. S. Hmerson—Some Characteristics of Kau. 431

Art. XLI.—Some Characteristics of Kau ;* by J.S. EMERSON.

Or the Hawaiian group of islands, Hawaii is the largest,
the most lofty and the most recent. The fires of its voleanoes
are still burning, and from time to time new material is poured
out over the surface of a land which is still in process of forma-~
tion. Here is the place to study Nature in her workshop and
see a world in making. But it is not of Hawaii as a whole
that we propose to speak. It is made up of districts with
characteristics marked and distinct from one another.

The Hilo-Hamakua-E. Kohala district, occupying the north-
east side of the island, from Hilo Bay to Upolu Point, is the
land of gulches and streams of water, of disintegrated rock ©
and deep heavy soil.

The W. Kohala-Kona-W. Kau district, occupying the west
side of the island from Upolu Point to South Cape, is the land
of slowly disintegrating rock, almost without gulches and
running streams. The extreme richness of its coffee lands is
due to the fact that new soil is ever being formed, as wanted,
_ from the loose mass of aa rocks, which contains every element
needed for plant growth.

Puna is the district where frequent showers fall upon loose
stone and sink out of sight, without forming gullies and
streams, and where vegetation thrives without soil. Here the
cocoanut tree grows in immense forests as nowhere else in this
group.

Kau is unique, a district by itself; “Aaw ka maka lepo”’;
Kau the dusty. That portion extending from South Cape to
and including Kapapala Ranch, a distance of some thirty
miles, contains all the land of any value for grazing or agri-
culture. Briefly described, it is a wide expanse of compara-
tively modern lava, forming a floor of rock upon which is
spread a superficial covering of fine, light, reddish or yellowish
dust of varying depth. In places the ephemeral mountain
torrents have washed away this unstable soil and revealed the
bed rock beneath, which has scarcely yet begun to disintegrate
and ally itself with earth. This superficial coating of dust is
distinct from the rocky stratum on which it rests, as the dust
which the cleanly housewife removes is from the floor which
she has swept. And such dust as that of Kau is found in
quantity in no other district of these islands. In the olden
days the natives of this section, when engaging in the popular
pastime of lehe kawa, jumping from a high bank, often sub-
stituted a bath of this dust for the usual pool of water. An

* Read before the Social Science Association of Honolulu, Oct. 14, 1895..
Am. JOUR. SCI.—FouRTH SERIES, VOL. XIV, No. 84.—DecEmBER, 1902. ©

432 J. S. Emerson—Some Characteristics of Kau.

ordinary dirt road, unless properly covered with stone, is quickly
worn down by constant travel. The wheels of loaded teams
sink into the soil, which in time is blown away by the winds,
or washed out by the heavy rains which occasionally fall.
From Waiohinu to Hilea the cane fields are mostly on the
hills, where the soil is of great depth and remarkably free
from rocks or stones, so that almost anywhere a crow bar or
walking stick can readily be thrust down to its full length.
On the low lands, however, the soil is usually shallow, with the
bed rock frequently cropping out, or may be wanting altogether.
The Hawaiian Agricultural Co. are more fortunate in the lay
of their plantation, and have an extensive tract of about four

thousand acres of rich cane land, not perched on isolated hills,

but occupying two large valleys with the gentle slopes between
them. These fields are somewhat scattered about, with unpro-
ductive areas between them, but as a whole they are far more
compact and accessible than the fields from Waiohinu to Hilea.

Kau is not a well-watered district. It often suffers from
prolonged drought, when the dry winds parch and destroy whole
fields of young cane and raise clouds of dust from the newly
plowed lands. When at length the long wished for rain comes, it
sometimes pours down in such torrents as to cause great destruc-
tion of property. The light soil soon dries up again, so that
the irregularity of the rains is the more keenly felt, while the
cane stalks register the fact in the varying and irregular length
and diameter of their joints. During these periods of drought
great care has to be exercised to prevent any one from smok-
ing in the fields or making a fire in the grass, for when a fire is
once started, it is sometimes extremely difficult to prevent it
from spreading under ground. The roots and other vegetable
matter in the soil are consumed, while the mineral basis of the
soil itself is so light and spongy as to allow enough air to enter
to support combustion many inches below the surface. At
times, when such a fire was supposed to be extinguished, it has
burrowed its way unobserved, to show its presence at some
other point where the surface would cave in and the ground
itself seemed on fire. Mr. W. E. Rowell states that “one
peculiarity of the soil on the Pahala plantation is the entire
absence of any clay or anything adhesive in its composition,
so that it does not stick to the shoes, however wet it may be.”

At various points through the district it is possible to exam-
ine the record of the rocks and to see the character of the suc-
cessive strata resting one upon another. It would appear from °
such observations that most of the strata represent ordinary
lava flows, with but thin separating sheets of ash or earth.
But Providence has most considerately anticipated the indas-
trial wants of man at the present time, and above all these ©

J. 8. Emerson—Some Characteristics of Kau. 433

successive layers of rock has added this crowning layer of dust,
which makes it possible to plow and planta district whose
formation is so recent. How different it might have been if
this superficial layer had been aa@ or pahoehoe rock, as in Kona,
requiring another ten thousand years to distintegate it suffi-
ciently for farming purposes.

Between the South Point and the road from Kahuku Ranch
to Waiohinu is a beautiful sketch of smooth grassy pasture.
At various points near the Cape the writer has measured the
depth of the soil to the bed rock, and recorded it as having an
average thickness of about ten feet, separated into two layers
of nearly equal thickness by a thin layer of whitish earth per-
haps half an inch thick. One branch of the flow of 1868
traversed a portion of this plain from the vicinity of the
Kahuku Ranch house half way to the Kaalualu landing.
Flows of an earlier date have also covered other portions with
a horrid mass of rock. But it is ontside of the purpose of
this paper to describe the various lava flows which cover much
of the territory below the cane lands and the Kapapala Ranch.

The woods above the plantations are difficult, if not posi-
tively dangerous, to traverse on horseback, so much so that
Mr. Julian Monsarrat, who knows the country well, pronounced
the carrying of supplies throngh these woods to his workmen
engaged in building a fence along their upper edge, as impracti-
cable. ‘The soft treacherous mnd gives no foothold for man or
beast. Consequently the supplies were carried a long distance
around.

The Kau hills, Iholena, Puu Enuhe, Makanao and Kapuna,
bear a most striking family resemblance to each other, so
that the profile of one will serve pretty well for all. Cap-
tain Dutton in his Hawaiian Volcanoes speaks of them as
“mere remnants of a large alluvial formation which was origi-
nally continuous,” and the valleys which separate them as “ val-
leys of erosion.” In speaking of these hills he uses the follow-
ing language: “The more we see of this country the more
will the evidences accumulate that these buttes are silent wit-
nesses of an extensive upheaval of this part of the island at
an epoch not very remote.” Again he says, “It is difficult to
estimate with precision the amount of elevation attested by
these terraces, but there are evidences still legible of several
of them—one of them 1,200 to 1,400 feet high, another
about 2,800, and perhaps, though more doubtful, a third at
3,400 feet.” (See Dutton’s Hawaiian Volcanoes, page 98.)

This theory, though advanced by a recognized authority on
the geology of portions of the western United States, lacks the
support of facts. It has not a single ledge of coral, a bed of
shells, or a vestige of marine life of any sort on which to rest.

434 J. S. Emerson—Some Characteristics of Kau.

Until such proof of marine origin is found, we cannot accept
the theory of upheaval as proven or even as a working hypoth-
esis. ‘To speak of the Kau dust as an “alluvial formation ”
only removes the question of its origin one step further back.
Whence came it? The process of disintegration has scarcely
begun in much of the Kau bed rock. The most casual study
of its soil shows that it is not decomposed rock. It is totally
unlike the soil of Hilo or Kona, and certainly was not washed
down from the aa and pahoehoe of upper Mauna Loa.

Evidently it is a formation similar to that which covers a
large portion of the slopes of Vesuvius and overwhelmed the
ancient city of Pompeii. This volcanic ash is the evidence of
a series of explosive eruptions in Kan on the grandest scale,
probably far surpassing anything of a similar nature of which
we have any evidence in this group. We must then look to
the medium of the atmosphere rather than to the action of
water to explain the remarkable distribution of this dust which
has been sifted down on the ridges, as well as in the hoilows,
in a manner far more regular and uniform than could have
been accomplished by the action of running water.

To locate the source or sources of this aérial eruption is a
problem not yet fully solved, but some light may be thrown
upon it. We need hardly look to Mokuaweoweo, the summit
crater of Mauna Loa, as asource. Tor had such an explosion
occurred there, the dust would also have been earried in the
direction of Kona. But asa fact, neither in Kona nor on any
side of the great dome of Mauna Loa is this dust found in
quantity, until we reach the wood belt of Kau. Kilauea, the
only remaining active voleano, may be looked to as a possible
source with slightly greater probability. But between the
district covered by the Kau dust and Kilauea lies the Kau
Desert with an area of “many miles” of a totally different
formation, while nothing is to be seen of the dust anywhere
around the voleano. The writer has examined fissures in the
Kau Desert to a depth of. perhaps thirty feet and failed to see
any of the dust. If then Kilauea were the real source of the
dust eruption, the evidences of it have been covered most com-
pletely by these later eruptions. |

The supposition is barely possible, but extremely improbable.
Another possible source in this enumeration should not be over-
looked. It is the vicinity of Puu o Keokeo, greatest of all the
offsprings of Mauna Loa, with an elevation of 6870 feet; it
appears to those who sail around South Point as another grand
mountain, the rival of its mighty parent. It is located on the
great fissure, or rift, in Mauna Loa, extending from Pohakn
Hanalei to the vicinity of South Cape. This fissure, probably
as old as the mountain itself, lies in the direction of the line

J. S. Emerson—Some Characteristics of Kau. 435

joining Mauna Kea and Mauna Loa, the same as the line of
the axis of Mokuaweoweo produced. The flows of 1868 and
of 18587 seem to have escaped from the central shaft through
this lateral rift in the mountain, and to have come to the sur-
face at points on this line of least resistance. It is an axis of
great volcanic activity, which has built up for itself the
immense ridge of Mauna Loa which divides the geological
district of Kau from that of Kona. About Puu o Keokeo as
a center there is a large area cevered with pumice and gravel,
quite similar to the so called Kau Desert, southwest of Kilauea.
Manifestly it has been produced by explosive eruptions on a
large scale. But as in the case of the vicinity of Kilauea, so
here the peculiar Kau dust is altogether wanting. Further
there is no sign of any such dust deposit at any distance, on
the Kona side, of this area covered with pumice, while the
entire portion of Kau covered with dust is at a distance of
many miles in the opposite direction.

The conclusion seems forced upon us that we must look for
the origin of this great eruption within the limits of the dis-
trict covered by it. With this idea guiding us, we naturally
look first to the immediate vicinity of the hills already men-
tioned, where the peculiar Kau soil attains its maximum depth.
One cannot but be struck with the suggestion that Pun Iki
marks a point on the upper rim of a vast extinct crater extend-
ing south to Kaiholena, possibly even to the great whaleback
ridge on which Kapuna Trig. Station is situated, while the
eastern rim, now largely washed away, would be marked by the
hills Makanao, Puu Enuhe and Kaumaikeohu with the steep
side hill just to the east of it. Everything seems to point to
this locality as the source of the stupendous explosion, or series
of explosions, which has rescued Kau from being a waste of
unproductive rock and transformed it to so large an extent into
a land of pastures and plantations.

If we admit that this is indeed the source of these eruptions,
the whole problem of the distribution of the dust is greatly
simplified. As would naturally be expected, the ejected matter
has been deposited on all sides, but the action of the trade
wind has carried the finer particles to a much greater distance
to the south and southwest than in the opposite direction. So
that we find the country about South Cape covered to a depth
of ten feet, as before stated, with the finest dust, without any
admixture of coarse material. The whole district west of
Kaaluala landing has been tilted up on its western edge, along
the line of the great fissure already alluded to, forming an
extensive fault from the sea to the Government road just above
Col. Norris’ residence. The visible deposit of dust breaks off
abruptly at the edge of this precipice. But according to native.

436 J. S. Emerson—Some Characteristics of Kau.

tradition the whole plane of Pakini at its base, on the Kona
side, was formerly a rich field, cultivated with sugar cane and
sweet potatoes. Since that time, however, various flows, among
which are those of 1868 and 1887, have converted this garden
into an uninhabited waste of aa and pahoehoe.

If, however, we examined the earth in the vicinity of the
supposed center of eruption, we find a considerable admixture in
places of stones, for example the so called Mud Flow of 1868
covered several hundred acres of good land to a depth of from
ten to thirty feet with a heavy, red, clayey earth, abounding in
stones, wholly unfit for cultivation, and producing a grass of
such inferior character that even the cattle and horses shun it.
This mud flow was simply a land slide. Kaapao pali was full
of water and the great earthquake of April, 1868 loosened the
superficial mass of earth and lubricated the rocky bed on
which it rested. Gravity acting on a steeply inclined plane
did the rest. Mr. Walton, the energetic and wide-awake man-
ager of the Pahala plantation, has used his wits to good advan-
tage in the search for water on the Kaapao pali and neighboring
hillsides, and during the past few months has found enough, as
I am informed, to irrigate five hundred acres of cane.

There are no traditions, so far as the writer can learn,
relating to these great explosive eruptions. The only oceur-
rence of the sort of which we have any historical information
took place in 1790 at the volcano of Kilanea, which destroyed,
as Dibble tells us, about 400 of the warriors of Keona’s army,
one entire division. The sand, ashes, pumice and stones
ejected at this time cover the country about the volcano for
miles, and have been fully described by Dana and others.

The extensive area about Pun o Keokeo already alluded to
is covered with material very similar to that of the Kau
Desert. The “ Alanui Umi,” built early in the 16th century
by King Umi, traverses this mountain desert from north to
south. This road was made necessary by Umi’s occupation
with his court and warriors of the barren waste between
Mauna Loa and Hualalai. When the political and military
necessity for Umi’s occupation of this strategic position ceased,
he sought a more agreeable home, and spent the last years of
his life at Kaawaloa, by the sea. The Umi road seems to have
been little used since those days, save by the race of bird
catchers and perhaps by the sandal-wood cutters of a later day.

Its very existence, however, is scarcely known to most of the
dwellers on Hawaii. This ancient road is chiefly interesting to
us at this moment as an evidence that no great explosive erup-
tion has probably taken place in that portion of Hawaii for
the past 350 years. Itis pretty much in the condition in which
Umi left it.

J. S. Emerson—Some Characteristics of Kau. 437

Some years since I had occasion to ride overa portion of this
road, which was two or three feet wide and is still readily fol-
lowed. So long as the mule kept to the old path he made
good progress, but when he deviated but a few feet on either
side, he sank down to his girth in the sand and pumice and
floundered helplessly. It was most instructive to follow this
path to the great natural amphitheatre on the southern slope of
Puu o Keokeo, where the famous cock fights used to draw
immense crowds to witness one of the great national games of
Hawaii. The cock-pits or rather pens still stand, probably as
Umi left them three and a half centuriesago. Had there been
a shower of ashes or pumice from the vicinity of Puu o Keo-
keo during this interval, the old road and these cock-pits would
have disappeared forever beneath the sands of the desert.

Since the above was written, Dr. A. B. Lyons of Oahu Col-
lege has kindly furnished the following statement: ‘“ I have had
occasion to examine the soil from some of the cane fields at
Pahala. They were remarkable in several particulars. They
contain a large proportion of organic matter, and yet could not
be called peaty. They seemed rather sandy. They contain
almost no clay. The mineral matter consists in fact of vol-
eanic sand rich in olivine and very little decomposed—very
similar in many respects to the sandy or gravelly soil of
Punahou, which is made from recent voleanic sand. The
abundance of lime in the soil confirmed also this view of its
probable origin. Hawaiian soils composed of decomposed lava
contain very little lime.”

For the sake of clearness, a brief recapitulation and summing
up of the argument with reference to the origin of the pecu-
liar Kau soil may properly conclude this paper.

A district equal in extent to one-half the area of the Island
of Oahu, 300 square miles, is covered with a soil quite unlike
that of any adjoining district, and totally distinct from the
very recent voleanic bed rock on which much of it rests. The
entire absence of a single ledge of coral, bed of shells, or other
positive evidence of marine formation, and the frequent occur-
rence of caves and caverns which remain unfilled with silt,
together with the very porous character of the whole formation,
diseredits the theory “of an extensive upheaval of this part of
the island at” any “epoch not very remote.” On the contrary,
this formation originated on dry land and has not been sub-
merged. If this soil were alluvial it would show stratification.
Instead of that it is blanketed like newly fallen snow upon the
uneven contour of hill, plain and ridge. |

If this formation were deposited at tide level, it would be
interpenetrated by marine growths, animal and vegetable. But
the evidence that such growth exists is conspicuously wanting.

438 J. S. Emerson—Some Charucteristics of Kau.

But, finally, no better evidence as to the origin of this minute
sand can be produced than its appearance under the mierc-
scope. Its mechanical features are sharp and broken as of
volcanic sand, wholly unlike the rounded and worn features of
beach sand. In short, the positive testimony of chemical analy-
sis and microscopic examination shows that the mineral portion
of this soil is voleanic sand, not the wash from any higher level
of decomposed rock. This necessitates the theory of explosive
eruptions on a scale of magnitude proportional to the extent of
territory covered, which ‘would have been greater but for the
ocean which abruptly terminates it along the entire coast from
Punaluu to South Cape, and a precipice on the southwest,
beyond which all trace of this soil has been effectually covered
up by flows of lava. Further, three great centers of volcanic
activity, viz., Kilanea, Mokuaweoweo and the vicinity of Puuo
Keokeo to the northeast, north and west respectively of the
district in question, have covered extensive areas about them
with other formations, so that it is impossible to locate the
original limits of this unique formation in those directions.
At the same time, the evidence is very strong, if not conclusive,
that neither of these three volcanic centers was its source,

while every consideration points to the great crater-like area

below Puu Iki.

This locality is so little known and so difficult of exploration
that we are unable to point out the exact location of the center
of explosion. The unstable character of the material ejected
would tend in a measure to cover up its source, which from the
nature of the case could not be as sharply defined as that of a
flow of aa or pahoehoe, whose birthplace is marked with solid
rock.

As to the time when these remarkable explosions occurred,
it may be observed that the only event of this character in
these islands, of which we have any definite information, took
place at Kilauea a little over one hundred years ago, as already
stated. From the fact that the old cock-pits near Puu o
Keokeo and the ancient road leading to them remain as they
must have been left by King Umi 350 yearsago, it seems quite
evident that no explosive eruptions on a large scale have taken
place during that interval either from the summit crater of
Mauna Loa or from the great lateral rift which has been so
active in building up the immense ridge which marks its
southwest slope. The conclusion therefore is evident that the
eruption which produced the Kau dust under discussion was
earlier than the beginning of the sixteenth century. Though
we may not locate the time of these eruptions as definitely
as the place, yet the fact that they form the last of a long

EE a ee

J. S. Emerson —Some Characteristics of Kau. 439

series of distinct strata, each of which represents a consider-
able interval of time, shows that they must have occurred
at a period very recent in the growth of the island. On the
other hand, the entire absence of any tradition relating to them
and the occurrence in several places of aa and possibly pahoehoe
flows from Mauna Loa of limited extent and of uncertain age
superimposed on this formation, make it probable that the stu-
pendous convulsions of Nature which gave birth to this crown-
ing feature of the district and prepared it for the support of
man in an advanced stage of civilization, occurred many cen-
turies ago, probably before the advent to these shores of its
first Polynesian inhabitants.

440 Phelps—Tirtrimetric Estimation of Nitric Acid.

Art. XLIIT.—The Titrimetric Estimation of Nitric Acid ;
by I. K. PHetps.

[Contributions from the Kent Chemical Laboratory of Yale University—No. CXIII.}

In the methods for the quantitative estimation of nitric acid
which depend upon its reduction with a ferrous salt and the
determination of the amount of oxidation produced, serupu-
lous care is necessary that the atmosphere in contact with the
ferrous salt while the nitrogen dioxide is present shall be
free from oxygen. This fact was recognized by Fresenius,*
who modified the original process of Pelouzet by filling the
flask with carbon dioxide or hydrogen at the outset. HEdert
used carbon dioxide similarly. Holland’s method,§ roughly
described, consists in boiling, until the air is expelled, the solu-
tion of the nitrate in a flask provided with a doubly bent exit
tube, rubber-jointed and fitted with a pinch cock, then admit-
ting through the tube as the flask cools a mixture of ferrous
salt and strong hydrochloric acid, heating the mixture on a
water bath, and, finally, titrating the resulting ferric salt with
stannous chloride.

All of these methods give results which are higher than the
theoretically expected—those which use carbon dioxide on
account of the oxygen invariably present in the gas as ordi-
narily produced in the laboratory, and Holland’s process pos-
sibly because of the slow leakage through the rubber connections
during the long heating, or because the nitrogen dioxide, which
is not driven out completely from the solution of the iron salts,
acts somewhat with the atmospheric oxygen during the titra-
tion.

Another series of methods has depended upon the deter-
mination of the nitrogen dioxide evolved, usually by direct
measurement as nitrogen dioxide. Among ‘these may be men-
tioned the methods of Schloessing, | Reichart,{ Schulze and
Wulfert,** Tiemann,t+ Wildt and Scheibe,t{ Warington,$$
Boehmer, | | Kratschmer, Wilfarth,*** Morse and Linn.,t++
Berger,tt{ and Roberts.§§$ In general, these methods give
results lower than demanded by the theory—in many cases, on

* Ann., cvi, 217; Zeit. anal. Chem., i, 32.
+ Ann. de Chim. Phys. [3], xx, 129. t Zeit. anal. Chem., xvi. 267.

§ Chem. News, xvii, 219. | Ann. de Chim. Phys. [8], xl, 479.
“| Zeit. anal. Chem., ix, 26. ** Chid.. ix, 400

++ Anleitung zur Untersuchung von Wasser von W. Kubel. Zweite Auflage
von F, Tiemann, 55.

tt Zeit. anal. Chem., xxiii, 151.

S$ Jour. Chem. Soc., , xxxviil, 468; xli, 345.

|||| Zeit. anal. Chem., "xxii, 20). “[4] Ibid., xxvi, 608.

eet Dbid:. aos, 411, ttt Amer. Chem. Jour. , viii, 274.

ttt Chem. Zeit., xix, 305. S85 This Journal, xlvi, 126 (1893).

Phelps—Titrimetric Estimation of Nitric Acid. 441

account of the solubility of the nitrogen dioxide in the ferrous
solution used as a reducer and, in other cases, on account of
the solubility of that gas in the solution of sodium hydroxide
over which it is measured. Further there is the effect of
oxygen upon the nitrogen dioxide; if present in the measur-
ing burette in the proportions found in air, it will not give
any appreciable effect in moderate amounts, as was observed
by Roberts; but if it is present while the gas is in con-
tact with the ferrous solution, the nitrous or nitric acid formed
will be taken up by the solution, and, should the oxygen be
introduced continuously, as an impurity in carbon dioxide
furnished by a Kipp generator, the iron salts will never be free
from oxidized nitrogen. If, however, the oxygen is present
in some other proportion of dilution than that of the air—as,
for example, in the proportion in which it dissolves in aqueous
solutions—it will produce an error no matter whether it comes
in contact with the gases of the ferrous solution or those of the
collecting burette: in the example taken, the proportion of
oxygen being greater than that in the air, the error produced
is one of deficiency. The effect of oxygen has been recog-
nized before. For example, in his process of determining
nitrates by reducing them with ferrous salt and hydrochloric
acid in an atmosphere of carbon dioxide (furnished by the
action of hydrochloric acid on marble), boiling the ferrous solu-
tion completely to dryness, collecting the gases over mercury,
absorbing the carbon dioxide by sodium hydroxide solution
and the nitrogen dioxide by repeated treatment with a con-
eentrated solution of ferrous salt, Warington found that,
when the carbon dioxide was made as free from air as possible
by using marble boiled in water and hydrochloric acid also
boiled and adding a little cuprous chloride to the acid, the
results were decidedly better.

The estimation of the amount of oxidation of a ferrous salt
may be accomplished with a simpler apparatus than that required
when the nitrogen dioxide is measured, and the purpose of this
article is to put on record such a method, following most nearly
the procedure of Holland. The apparatus used consisted
of a 250 flask closed with a rubber stopper carrying in
two perforations the inlet and exit tubes. A stoppered funnel
of 50°* capacity with its tube constricted at its lower end was
used as the inlet tube; and a glass tube of ‘8° internal diam-
eter, enlarged just above the stopper to a small bulb (to pre-
vent mechanical loss of the solid contents of the flask during
the boiling) and bent twice at right angles, served as the exit
tube.

The analysis was made as follows: A solution of ferrous
sulphate, mildly acidified with sulphuric acid, was made of

442 Phelps—Titrimetric Estimation of Nitric Acid.

about one-fifth normal strength and standardized against stand-
ard decinormal arsenious acid solution by treating portions of
it, measured from a burette (and weighed as a check on the
burette readings) with an excess of decinormal iodine solution,
adding about 3 grms. of Rochelle salt in solution, neutralizing
with acid potassium. carbonate and adding in succession 15°%™
of a saturated solution of acid potassium carbonate, starch
paste and then standard arsenious acid solution to the bleaching
of the starch blue, and, finally, titrating to color with the
iodine solution. For the determination of nitric acid, the pure
potassium nitrate of commerce was used. In the smaller
amounts, the nitrate was taken in a solution of known strength
measured from a burette and in the larger amounts as the dry
salt. Where the solution of nitrate was used, it was measured
into the flask, the stem of the separating funnel being com-
pletely filled with water, and the exit tube of the flask reach-
ing to the surface of mercury which was placed in a test tube to
the depth of about three centimeters. The nitrate solution
was then boiled to small volume, an amount of the standardized
ferrous sulphate solution known to be in excess introduced into
the separating funnel, the exit tube plunged a centimeter or
two deep into the mercury (which is readily accomplished by
changing the position of the flask on the wire gauze provided
that the gauze is depressed well at the center and the flask is
set well up on the higher part at the beginning of the opera-
tion), and then the flame withdrawn until diminution of pres-
sure sufficient to draw the ferrous solution into the flask is
made evident by the rise of the mercury in the exit tube. By
applying and. withdrawing the flame and by regulating the
rate of inflow of the solution, the ferrous salt may be intro-
duced without admitting air and the funnel washed carefully
with an amount of concentrated hydrochloric acid nearly
enough to equal the total volume of the liquid in the flask.
After the pressure has been restored in the apparatus by heat-
ing the flask, the exit tube is again raised to the surface of the
mercury and the solution in the flask boiled to a volume of
10-15". The excess of acid is then nearly neutralized with
sodium carbonate solution, the carbon dioxide evolved assisting
in maintaining the pressure in the apparatus so that the con-
densed liquid in the test tube which may contain oxidized
nitrogen dioxide is not returned to the iron solution ; the flask
is cooled and the ferrous salt remaining determined by iodine
as previously described or by a standard solution of potassium
permanganate. in case permanganate is used, the contents of
the flask are diluted with 600° of water, 2-3 grms. of crystal-
lized manganous chloride,* and titrated to color with potassium

* Gooch and Peters, this Journal, vii, 461 (1899).

Phelps—Titrimetrie Estimation of Nitric Acid. 448

permanganate. In the experiments where the dry salt was
used, the air was expelled from the apparatus by boiling 10°"
of water to small volume in the flask, arranged as previously
described, the ferrous sulphate solution introduced and concen-
trated by boiling to a volume of about 20°™* so that the amount
of acid used may not be so large; then, the nitrate, dissolved
in a small amount of water, was allowed to flow in and was
finally washed in, as before, with an amount of concentrated
hydrochloric acid, which approximates in volume that of the
liquid contents of the flask.

Experiments numbered I-VIII, inclusive, of the following
table show the results of a series of experiments made as
described above.

Oxygen value Oxygen value

KNO;

of ferrous salt of ferrous salt Error on
taken. taken. found. oxygen.
erm. erm. erm. erm.

"070500 0°01823 0°00621 0°00015 +
II 0°0500 0°01865 0°00681 0°00003 —
VE 0:0500 0°01954 0:00768 0°00000 +
IV 0°1000 0°'02881 0:00507 0°00001 +
VV 0°1000 0°02822 0°00441 0:00008 +
VI 0:2500 0°06453 0°00512 0:00009 +
VII 0:5000 0°138394 0°01524 0°00005 +
VIII 05000 0712210 0:00340 000005 +
IX 0:0500 0°01720 0°00747 0:002)4—
~X, 0-0500 0701550 0700389 000026 —
XI 070525 0°01550 0°00318 0°00014—
XII 0:1000 0°02765 0°00432 0:00040 —

Experiments numbered [X and X were made to determine
whether the long boiling in the previous experiments is actually
necessary ; for in the work by Fresenius as well as that of Eder,
directions are given merely to boil the hydrochloric acid
solution of the iron salt with the nitrate until the color of
the dark compound of the nitrogen dioxide with the ferrous
salt wholly vanishes and is replaced by the clear color of the
ferric salt, which means an active boiling of not longer than
five minutes in the experiments recorded in the table. Experi-
ment IX was made exactly like the similar ones above it in
_ the table, except that the boiling was interrupted after the
dark colored ferrous compound with the nitrogen dioxide was
completely broken up; and experiment X similarly, except that
the boiling was continued for five minutes after the complete
disappearance of the dark color. Obviously the period of boil-
ing allowed by Fresenius and Eder is not enough, but since
the main error incidental to these processes is an error of excess
due to oxygen present, as already stated, it will readily appear

444 Phelps—Titrimetric Estimation of Nitric Acid.

how this second error, which is one of deficiency and hence
correcting the first error, may have been overlooked. |

Ferrous ammonium sulphate has been recommended by
Austin and Chamberlain* and by Rosat as more convenient,
stable, and sensitive than the crystalline ferrous sulphate. It
would seem possible that nitric acid might act as an oxidizing
agent on the ammonium salt, and certainly nitrous acid (if it
were produced, as might be, as an intermediate product in the
reduction of nitric acid to nitrogen dioxide) would act accord-
ing to the equation,

NH,Cl4+ HNO,=HCl1+2H,0+N.,,.

Such an action either of the nitric acid or nitrous acid would
produce anerror of deficiency. Experiments Xl and XII were
made to test this point, being made exactly like those num-
bered I-VIII, except that 1 gram of erystallized ammonium
sulphate was added with the ferrous salt. The results show
slight but appreciable losses.

The concentration of the hydrochloric acid in this operation
does not allow of much diminution under these conditions,
where immediate reduction of the nitric acid without any
volatilization is a necessity,—a fact which has been recognized
before by Roberts and obviated under the conditions there
used.

Thus it would appear—

First. That the method outlined above is capable of yield-
ing very accurate results, affording as-it does an easy and com-
plete means of shutting out oxygen from the nitrogen dioxide
while that gas is in contact with the ferrous salt.

Second. That prolonging the boiling only until the dark
colored compound of nitrogen dioxide with ferrous salt is
broken up, results in the incomplete reduction of the nitric
acid. |

Third. That ammonium salts must be absent, if the highest
accuracy is desired.

*Amer. Chem. Jour., v, 209. + Gazz. chim. ital., xv, 295.

ee

Lt. M. Brown—Clays of the. Boston Basin. 445

Art. XLIL.—TZhe Clays of the Boston Basin ;* by ROBERT
MarsHALL Brown.

THE problem of the correlation of the clays of New England
has never been solved. Numerous papers have been published,
in which various clay beds have been described and in part dis-
cussed. Some of the writers have ventured to correlate the
clays of neighboring sections. No one has yet undertaken the
solution of the entire problem. Not enough light has been cast
upon the subject for a general discussion. The isolation of
clay, beds in the same neighborhood with no evidence to show
whether they are parts of the same deposit or disconnected
deposits of the same time or unrelated beds, offers little of
interest. Similarity of clay itself cannot be taken asgood data
for correlation, unless by chance some unique peculiarity of
composition is discovered. Ordinarily, the zest to further search
is furnished the student of clays by the constant excavations
that are made in the streets for the laying of pipes, the opening
of new clay pits, the extension of the old, and the removal of
the cover of the clay for various purposes. All new exposures
which furnish interesting evidence in any direction should be
recorded. By the accumulation of such evidence, it is believed
that in time a correlation of the clays of New England and
beyond may be safely attempted. For this reason, the writer
of this article desires to place on record some of the results of
this field study on the clays about Boston, with a few observa-
tions that arose during the process of the investigations.

Field Work.—Near Chelsea Street, between the cities of
Everett and Chelsea, Massachusetts, extensive excavations,
during the fall of 1901, revealed the clays. A hill, drumlin-
like in appearance, of no great height, and half a mile long,
with a trend a little south of east, parallel to the drumlins
of the neighborhood, was in part removed. An almost com-
plete cross section of the hill was exposed. The core of the hill
is an igneous rock (fig. 1). The rock does not extend to the
southern slope, however. Here a clay and sand interior was
found. Both clay and rock was capped by a thin layer of till.
The relation of the clay to the rock was unknown, as neither
the quarrying of the rock to the north nor the removal of the
elays to the south showed the contact. The first appearance of
the south end of the hill in the process of removal of the
material showed an anticline in the clay. The stratification of
the beds was emphasized by the interlamination of thin layers

* This paper is a part of a thesis presented in the course on Glacial
Geology at Harvard College, under Prof. J. B. Woodworth.

446

RL. M. Brown—Clays of the Boston Basin.

A. ='Clay;

Figure 1.—Chelsea Street, between Chelsea and Everett, Mass.
Dotted lines in clay represent sand.

5. .> Talus, C. = Bock, D. = far:

: Gi
= 50.0 DOS LAGOS. G GOD: SD

_—

—
—

Figure.2.—Chelsea Street, between Chelsea and Everett, Mass. A. = Sand,
Dotted lines in clay represent sand.

Bo = Clay, ‘C= Tis,

—

2S SS SSS SS = A Ms Oh. WAGES
A. = Clay,

Figure 3.—Chelsea Street, between Chelsea and Everett, Mass.
Dotted lines in clay represent interstratified sand.

5: = Sand, oo tat

R. M. Brown-—Clays of the Boston Basin. 447

of sand in the lower portions of the clay and by dark bands,
probably of vegetable matter, in the upper portions. Below
the clay was a bed of sand, and above a thin layer of till which
ran out towards the south (fig. 2). The disappearance of the
till was due probably to the removal of the material in a pre-
vious excavation further to the west. Before a second visit to
the pit was made enough material had been removed to yield a
better section (fig. 3). The bed of sand at the bottom of the
clay (shown in fig. 2) proved to be a stratum of sand 14 inches
thick, underlain by more clay. The sand was of fine texture,

Figure 4.—Cross-section of end of Drumlin in Revere, Mass. A. = Clay,
B. = Sand and Clay, C. = Till.

similar to our beach sands, and homogeneous throughout. A
single small slate pebble was the only break in the uniformity
of the sand bed. At the same time enough of the talus had
been removed to show the sand bed extending to the edge of
the till and there cut off before the deposit of the till. To the
north the clay was much more crumpled, and the interstratified
sands were, as a general rule, thicker than the layers above the
thick sand bed. The till averaged 18 inches in thickness, was
made up of coarse and fine material, and extended over the
entire section of the hill remaining. The slope of the hill was
very gradual and the question of the slipping of the till over
the clays was negatived by the rolling surface of the clay as well
as by the slope.

In seeking for other instances of the same thing, a section
across the trend of a drumlin at the junction of Shirley Avenue
and Nahant Avenue in Revere, Mass., was found. The drum-
lin had the northwest—southeast position of the drumlins of
the district. The core of the drumlin was clay (fig. 4).
Above the more solid bottom clay was a thickness of about
three feet, made up of alternating layers of sand and clay, the
sand increasing upwards. Over the sand was 14 feet of till.

Am. Jour. Sc1.—FourtH Series, Vou. XIV, No. 84.—DrEcEmBER, 1902.
; 31

448 ft. M. Brown—Clays of the Boston Basin.

The till contained bowlders of varying sizes; the largest over a
foot in diameter. Thickly intermixed as a part of a till was
a great deal of clay. The clay itself, below the layers of sand,
was homogeneous and contained no foreign matter. The bot-
tom of the clay was seen in one instance only, and the clay
rested in that case on gravels. This exposure was at a distance
from the two mentioned above.

Clays in Literature.—An early mention of the clays was by —
Edward Hitcheock.* His conclusion stated that nearly all of
our clays are in the Tertiary formation. At the same time he
is frank in admitting that he has no evidence bearing on the
matter, and that his result was reached solely by analogy. The
European plastic clays resting on the chalk was the basis of his
reasoning. In the Geology of New Hampshire many cases are
instanced by Uphamt where clays overlie till. This is especi-
ally true of the Winnipiseogee Lake beds. His explanation is
quoted: ‘“‘ The ice-sheet probably remained in a high mountain-
like mass over these lakes after it had disappeared on each
side from the basin of Ossipee Lake and from the lower part
of the Pomigewasset valley. As the melting continued, the
drainage over this area was frequently obstructed because the
ice-sheet retreated from the lines of watershed towards the
middle of these hydrographic basins. The water seems then
to have melted large open spaces beneath the ice near its mar-
gin in which beds of clay and sand were deposited. This
would occur at the various heights and in the situation where
these beds are found, and the till which overlies them is shown
by its material to be that which was contained in the ice-sheet
and fell upon the surface when its melting was completed.
We thus see how these deposits came to be spread over the
slopes of the hills, thinly covered by large boulders and till.
The frequent accumulation of such deposits in other parts of
the state was prevented by unobstructed drainage from the
melting ice. This modified drift over-laid by till does not
therefore appear to bear testimony to a warm interglacial
period, or even to any retreat and subsequent.advance of the
ce:

Emersont reports clay resting upon till in the “ Northamp-
ton Lake” deposits. In a few localities he observes till over
clay, and, furthermore, explains the contortions of the clay beds
by ice moving over them. In the succession of events a minor
advance (being the third) of the ice is considered to explain
these phenomena. Shaler§ publishes a section at Weewocket,

* Report on Geology, Mineralogy, Botany and Zoology of Massachusetts,
1888, 36.

+ Geology of New Hampshire, iii, 136.

t Monograph 29 U. 8. G.S., p. 697. § Bulletin 53 U. S. G. S., p. 16.

ft. M. Brown— Clays of the Boston Basin. 449

Nantucket, in which is shown a till deposit over clay. Pro-
fessor Shaler at another time,* in summing up the facts of the
clays, suggests that they are reconcilable with the supposition
that there were several clay-making periods, each marking a
retreat of the ice, followed by a readvance. The clays about
Boston were studied by Marbut and Woodworth.t Sections of
drumlins in Somerville have beeu published by them showing |
the clay beneath the till. Professor Woodworth,t in the corre-
lation of the deposits along the coast of New England, places
the Mystic clays in the second of three glacial epochs.

Discussion of Lrelation of Clays to the Glacial Period.—
The points that must be explained are as follows:

1. The position of the clays beneath the drumlins in Chelsea
and Revere. 2. The considerable amount of clay that makes
up a constituent part of the till of the drumlins. 3. The dis-
tortion of the clay-beds.

Only one conclusion can be reached in regard to the first of
these. The clays must have been deposited before the advance
of the ice which formed the drumlins. The clayey composition
of the drumlins in the neighborhood points to the same con-
clusion. In a pit recently opened in the flank of a drumlin
situated near Hyde Park, Massachusetts, enough clay was found
on the floor to give an aspect of a clay bed. This clay had
washed out of the exposed till during the storms, and collected
at the bottom of the pit. A little digging exposed a few
inches of clay only, and further examination explained the
origin of the clay. The per cent of clayey material appears to
be too large to be produced by the natural process of grinding,
during or just previous to the formation of the drumlins. It
seems just, under the circumstances, to presuppose a reservoir
of clay from which the drumlins drew their supply. Local
advances of the ice have been offered as an explanation for the
position of the clay under the drumlins as well as the large
percentage of clay in their composition. Emerson considered
such an event in the “ Northampton Lake” area. While a
local advance of the ice after the last general retreat has been
proven along the line of the chain of lakes—Fresh Pond, Spy
Pond, the Mystic Lakes, ete.,—it is not consistent to extend this
local advance over an area large enough to account for the
drumlins of Revere and Chelsea. The local advance would
then grow to a general invasion. _

The distortion of the clay-beds offers some additional testi-
mony. No explanation has been given for the folding of the
clays that has been more satisfactory than the thrust of the

*17th Ann. Rep. U.S. G. S., p. 969.

417th Ann. Rep. U.S. G. S., p. 989.
+17th Ann. Rep. U.S. G.§S., p. 987.

450 R. M. Brown—Clays of the Boston Basin.

advancing ice. The agent of this work has also been con-
sidered a local advance, but again the widespread occurrences
of these folds indicate the general event rather than a limited
one. :

Where the surface has been planed off, as in fig. 3, the ice
was supplied with its clay, which it deposited in the drumlins
of the harbor and the harbor vicinity. The beds must have
been developed previous to the ice advance. Upham’s ex-
planation, cited above (dated 1883), cannot be considered as
applicable for the Boston district, as he demanded a valley slope
by means of which the water was held against the side of the
glacier so that the warmer water might undermine the edges
of the ice. It is not easy to conceive how clays of any amount
could be deposited under the ice by such a supposititious theory.
There seems to be evidence strong enough to support the belief
that some of the clays of the Boston Basin were deposited before
the last general advance of the ice; are therefore inter-glacial
in origin.

New Bedford, Mass.

Barbour and Fisher-——Caleite-Sand Crystal. 451

Art. XLIV.— A New Form of Caleite-Sand Crystal; by
Erwin H. Barsour and Cassius A. FISHER.

UNTIL quite recently the knowledge of calcite-sand crystals
was confined to a very few occurrences, that of the well-known
“Fontainebleau limestone” being the most important ;* here
the erystals are rhombohedral in form (—2R). <A few years
since sand-calcite crystals of different type from those of
Fontainebleau were found on Devil Hill, in the Indian Reser-
vation, Washington Co., South Dakota. During the winter
the writers have added still another locality, partly in western
Nebraska and partly in Wyoming, which yields a third form

Fig. 1. Pen drawing of a scalenohedron with rhombohedral terminations
illustrating the theoretical form of the sand-lime crystal shown in figure 2.

Fic. 2. Photographic reproduction of a sand-lime crystal, nearly natural
size, the type found at Devil Hill, Indian Reservation, Washington County,
South Dakota.

of sand crystals, namely combinations of acute and obtuse
rhombohedrons, constituting a group similar to and quite as
imteresting as those from Devil Hill, though smaller in size.

Crystals from Devil Hill, South Dakota.

The sand-ealcites from Devil Hill, which is the base of the
Arikaree sand, vary from single to doubie, triple, quadruple
and multiple crystals, up through clusters and concretions to
solid sand-calcite rock, which though solid and compact reveals
on fracture, and especially on weatliered sections, the indis-
tinct hexagonal outlines of its component crystals.

In point of size they vary from a quarter of an inch (6™”) to

*Lassone, Mem. d l’acad. royal, Paris, 1775, p. 65. Haiiy, Traite de
Mineralogie, vol. 1, p. 424, 1882. Seealso Dana, Syst. Min., p. 266.

452 Barbour and Fisher—Calcite-Sand Crystal.

those exceeding 15 inches (28™) The average crystal, of
which there are myriads to be dug out of loose sand, has a
length of two and one-half to three inches (6 to gem), (See
figs. 1 and 2, which are somewhat reduced.) They are barrel-
shaped with rounded ends, and have been described by Penfield
and Ford* as hexagonal pyramids, but to the writers a study of
the form of the faces and a measurement of the angles seems
to show that they are a combination of scalenohedron and
rhombohedron.

Analyses.—Analyses of four crystals were made by Wood-
ruff and Warner, who determined the average percentage of
sand to be 63°81, leaving 36°19 soluble matter, as shown in
detail in the accompanying table.

Per cent of

Weight of Per cent soluble

crystal. sand. matter.

Crystal ING ee i se Fe 1297572" 63°07 36°93
me “NEagplt One, Se ris 80°368" 63°55 36°45

fe Bole 2 gece 33°408T 64°22 35°78

Re > arene 2 OL BS OSGae" 64°40 35°60

Average of the above -._----------- 63°81 36°19 ©
Large comeretions.. -- 2... =. 22. Specs 38°12
Sand Clee aingGhew. 0 Ue 63°43 36°57

Average of four radiate sand-lime con-

cretions, Sioux County, Nebraska.- 58°89 4111

Crystals from Goshen Hole Region, Wyoming.

The new crystals, which also consist of sand cemented by
calcite, are very similar in color, texture, and in general appear-
ance to those from Washington County, South Dakota, differ
from them in being a combination of acute and obtuse "rhom-
bohedrons. They show greater uniformity of size, varying
but little from one and one-half inthes (40™) in. length, by
seven-eighths inch (21™™) in thickness (figs. 3, 4). Apparertly
they occur much more sparingly than the others, although the
field remains to be fully explored.

They show the same tendency to become doubled, clustered,
concretionary, and massive to such an extent that descriptions
already written of the one kind serve well for the other. If
the crystals happen to be particularly small, the size of the
sand grains is relatively so large that the exact outline of the

* Penfield and Ford, Siliceous Calcites from the Bad Lands, Washington
County, South Dakota, this Journal, vol. ix, 1900, pp. 352-4, 1 plate, 4
figures.

See also Barbour, Sand-Crystals and their relation to certain concre-
tionary forms, presented before the Geol. Soc. of America, Dec. 27, 1900,
printed in Bulletin of the Society, vol. xii, pp. 165-172, pls. 13 to 18, April
16, 1901.

Barbour and Fisher—Calcite-Sand Crystal. 453

crystal is obscured, and yet to the collector in the field, at
least, the identity is plain. The less obvious forms occur in
great masses and over wide areas, and pass for concretionary
sand. Their extent may be judged better from the fact that
they may be traced from the Indian Reservation of South
Dakota to northwestern and western Nebraska and eastern
Wyoming, as far west, at least, as Bates Hole, which is prac-
tically the extent of the Arikaree formation. Without a knowl-
edge of the actual sand crystals these obscure forms would
ordinarily be unrecognizable. In the case of the new forms as
in the case of the older ones they are the result of erystalliza-

Fic. 3. Drawing of a crystal showing combination of acute and obtuse
rhombohedrons, the theoretical form of the actual crystal shown in figure 4.

Fie. 4. Sand-lime crystal, natural size, type found west of Mitchell, in
the Chadron sandstone near the North Platte River, on the Wyoming-
Nebraska line.
tion in a sand bed saturated with water bearing calcium carbo-
nate in solution. The lime in erystallizing out followed some
one of the forms common to calcite and in the process cemented
the sand, together making a sand-calcite crystal. The erystal
proper is calcite. The sand may be viewed as an accident
incident to the crystallization of lime in a sand bed. A chem-
ical analysis shows that those from western Nebraska do not
differ essentially from those of South Dakota, as shown by
analyses of Mr. Willis Warner (Univ. Nebr. 1901).

Analysis of Sand Crystals from Goshen Hole Region.

Per cent Per cent of lime
of sand. and soluble matter.
Baer ebystal = 022 ieee 63°63 36°37
pare GConeretion. 2. 12 2 62°54 37°46

Analyses of the above in full as furnished by Mr. Warner
are as follows:

Crystal: silica, 49°32 ; phosphorus, ‘011; iron, Fe,O,+Al,0,,
14°21; lime, CaCO,, 33°27 ; magnesia, MeCO,, 3-14; “undeter-
mined, probably manganese, "049.

454 Barbour and Fisher—Caleite Sand Crystal.

Concretion : silica, 47°94; phosphorus, ‘01 ; iron, Fe,O,+ Al,O,,
14°52; lime, CaCO,, 34:24; magnesia, MeCO,, 3:25; undeter-
mined, probably manganese, ‘04. .

Distribution and Geologic Range of Sand- Calcite Crystals.

The distribution of sand crystals is very wide though they
are little known and seldom recognized from the fact that the
more obscure forms pass for concretionary sand and receive no
attention. ‘They have been personally observed at and around
Devil Hill, in the Indian Reservation of South Dakota; in
Sioux County, Nebraska; in the North Platte region in
Nebraska, at Goshen Hole, "Bates Hole, and inter mediate points .
in Wyoming.

The larger crystals present no such disparity between the
component sand and gravel grains and the crystal itself as do
the lesser ones, hence their forms are defined and perfectly
apparent. Such forms are restricted and very local in area.
Those at Devil Hill occur in enormous quantities along about
100 yards of exposure. The total extent of the bed, which is
but a mere remnant of the original, is scarcely one acre.
Those from the region of the North Platte river, at or near
the Nebraska line, seem to be even more local, and are very
scattered in number. The same seems to hold true of those in
the Goshen Hole country, which may be classed with the North
Platte region, both producing the same crystalline forms shown
in figure 4, though differing in horizon. There are but the
three above named localities known in this country which pro-
duce these crystals. In France, at Fontainebleau, the crystals
are similar in physical properties to those from the Great
Plains though differing crystallographically, inasmuch as they
are simple unmodified rhombohedrons.

In vertical range these forms occur in greatest numbers in
the Arikaree, though found as low as the Chadron sands at the
base of the Oligocene. In the basal sands of the Laramie of
Wyoming occur numerous examples of the obscure or concre-
tionary type.

Locality of sand-calcites.
Mica ( Devil Hill, S. D.
Tertiary (Arikaree) Sion Coe ae
Goshen Hole, Wyoming
| Bates Hole, Wyoming

Oligocene
(Chadron sand) Mitchell region, Nebr.

South of Buffalo, Wyoming
Cretaceous Laramie Obscure crystals or concre-

tionary sand

University of Nebraska, Lincoln, Nebr.

Chemistry and Physics. 455

SCIENTIFIC INTELLIGENCE.

I. CHEMISTRY AND PHysIcs.

1. Transformation of Carbon into Diamond.—It was sbown
by Pepys, in 1815, that an iron wire heated to redness in contact
with diamond was converted into steel. Dr. ALBERT Lupwic now
claims that under great pressure in an atmosphere of hydrogen
this reaction is reversed, and also that carbon fused in the electric
are in absence of iron, under the same conditions of pressure, is
likewise converted into diamond. In carrying out the experi-
ments, pressures as high as 3100 atmospheres were used. A spiral
of iron wire was embedded in powdered retort-carbon and heated
by an electric current in a highly compressed atmosphere of
hydrogen. In afew moments the resistance, which was slight at
first on account of the conductivity of the carbon, rose to that of
the iron spiral. This indicated that the carbon in contact with
the spiral had become non-conducting, and careful examination
showed that on some of the pieces of carbon were br illiant, minute
crystals, possessing the hardness, specific gravity, and refraction
of the diamond. The greater part of the carbon contained in the
iron spiral had also been converted into diamond. The crystals
had the characteristic irregular surface shown by the diamonds
_ produced by Moissan by the sudden cooling of molten iron. To
produce diamonds without the use of iron, it was found necessary
to use much higher temperatures and actually fuse the carbon
under very high pressure. Thus fused it was shown to be a non-
conductor, and hence was considered to be molten diamond.
Under the pressure employed this fusion took place very easily
in the electric arc, and there were thus obtained spherical masses,
of the size of peas, having the great hardness and crystalline
structure of carbonado. It is the intention of the author to
develop the process for the commercial manufacture of diamonds.
— Chemik. Zeitg., xxv, 979. H. L. W.

2. An Lodometric Titration of Thiocyanie Acid. —It has
been shown by Rupp and Schied that sulphocyanides are oxidized
in sodium bicarbonate solution, according to the equation,

HSCN +41,+4H,O=H,S0,+7HI+CNI,
but starch cannot be used as an indicator in applying this reaction
on account of the presence of cyanogen iodide ; and, moreover, the
yellow color of the latter compound makes the method applicable
only to small quantities where the disappearance of the iodine
color is used. I[t has been found by Meinecke, however, that in
acid solutions the following reaction takes place: CNI+ HI=I,+
HCN. Tuite has now used the facts which have been stated for
devising a very convenient method for making this titration.
The thiocyanate is first treated with an excess of iodine solution
in the presence of sodium bicarbonate, enough water being used
to dissolve the latter. The reaction is complete in four hours at

456 Sctentijic Intelligence.

ordinary temperature. Then an excess of hydrochloric acid is.
added, and the excess of iodine is determined at once by means of

sodium thiosulphate solution, with the use of starch as an indica-
tor. The final result corresponds to the equation,
HSCN +-3],+4H,O=H,SO,+6HI+HOCN.

The experimental results obtained were extremely accurate,
and the author advises the standardization of volumetric thio-
cyanate solutions by this method.— Berichte, xxxv, 2766. H.L. w.

3. Lodine Pentafluoride.—This compound was prepared, pro-
bably in an impure condition, by Gore in 1871 and by Maclvor
in 1875, by the action of iodine upon silver fluoride. Morssaw
has now prepared it by allowing fluorine gas to act upon iodine.
Combination takes place with evolution of heat and the product
is a colorless liquid boiling without decomposition at 97°, and
solidifying at 8°. In the solid state it resembles camphor. The
liquid fumes in the air, and when poured into water it is decom-
posed without violence into hydrofluoric and iodic acids accord-
ing to the equation,

' 21K, +5H,O = 1,0,+10HF.
The fluoride possesses great chemical activity ; most elementary
bodies decompose it, and it enters into reactions with a great
number of compound substances. When its vapor is heated to
the neighborhood of 500° it decomposes, showing the color of
iodine vapor.— Comptes Rendus, cxxxv, 563. H. L. W.

4, Separation of Manganese from Magnesium, Zine and
Aluminum.—To effect these separations DieTricnw and HassEn
precipitate the manganese as higher oxide in dilute sulphuric or
nitric acid solution by means of ammonium persulphate. For
quantities of manganese corresponding usually to -1¢ MnO or
less they used 5°¢ of dilute sulphuric acid (1:10) and 15—20° of
a ten per cent solution of ammonium persulphate. In the separa-
tion from magnesium the total volume of the liquid was 150-200°,
while in the other cases, where the quantities of zine and alu-
minum oxides present were about the same as that of the man-
ganese Oxide, a volume of 400-500°° was used. The precipita-
tions were made in solutions heated upon the water-bath and
frequently stirred, and the heating was continued for about two
hours in order to decompose the excess of ammonium persulphate.
The manganese precipitates, after filtering and washing, were
ignited directly, and Mn,O, was weighed. The test analyses
gave very satisfactory results. It seems probable that this
method will find useful application in mineral and rock analyses,
in which, by the usual methods, the separation especially of man-
ganese and aluminum presents some difficulties.— Berichte, xxxv,
32662" 2% B.D we

5. Plasticity and Adhesiveness of Glass at Ordinary Tem-
peratures.—That glass possesses a certain amount of plasticity
when not heated has been known for a long time, but it has not
been shown definitely that this plasticity is sufficient to produce
adhesion similar to the welding of metals. Piccarp has now

eee

ES ee ee eee

Chemistry and Physics. 457

made experiments which indicate that the thin cracks extending
from ‘05 to :1™™ below the scratch made by the glass-cutting
diamond may be healed to a considerable extent in from one to
three days by gentle pressure. The experiments were conducted
by comparing the loads required to fracture freshly scratched
pieces of glass and those which had been slightly bent for several
days in such a way as to press together the ruptured surfaces.
— Berichte, xxxiv, 3635. BEE Welt 4

6. General Principles of Physical Science; by Artuur A. ~
Noyes. 8vo, pp. 172. New York, 1902 (Henry Holt & Co.).—
The purpose of this volume is to present the principles and laws
of physics and chemistry which lie at the basis of the modern
science of theoretical chemistry. The treatment is non-mathemati-
cal toa greatextent. Thetwo main chapters of the book treat respec-
tively of the general principles relating to matter and to energy.

The book forms the first part of a work on the general princi-
ples of chemistry, which has had to be discontinued. It is to be
hoped that Professor Noyes will continue the work in the future.

H. W. F.

7. Lehrbuch der Allgemeinen Chemie ; by Dr. Wi~HE Lm Ost-
WALD. 2 vols. Leipzig, 1902 (Engelmann).—The last section of
the second part of volume ii has just appeared. A third part is
to appear in the future. This work is by far the most compre-
hensive which has yet appeared in the field of physical chemistry
and probably comes as near treating the whole subject as it is
possible to do. It is to be regretted that volume i and the first
part of volume 1i are already out of print. H. W. F.

8. The Spectra of Hydrogen and Reversed Lines in the
Spectra of Gases ; communicated by Joun TrowpripGr.—In a
previous paper* I described the spectra produced by powerful
condenser discharges through Geissler tubes filled with hydrogen.
A fairly continuous spectrum was obtained between the HH lines
and the red end of the spectrum which was traversed by reversed
lines. In that paper I expressed the hope of being able to obtain
quartz tubes. This hope has been realized. Through the kindness
of manufacturerst working under the direction and according to
the method of Professor Shenstone of Clifton College, England, I
have obtained suitable tubes, and the results given by such tubes
are so remarkable that they seem worthy of a preliminary paper.

The tubes are eight centimeters in length ; with a capillary
four centimeters in length and about two millimeters in diam-
eter. On account of the difficulty of inserting platinum termi-
nals in quartz, I had the ends of the tubes ground smooth ; and
the glass blower of the laboratory prepared glass bulbs in which
suitable electrodes were inserted. These bulbs were luted to the
ends of the quartz tubes. In certain cases where metal plates were
luted directly to the ends of the quartz tubes I employed silicate
of soda as a luting agent ; and after this had hardened I applied
on the outside of the joint a hard preparation of pitch and shellac.

The glass bulbs were covered with other bulbs which allowed

* This Journal, vol. xiv, p. 1. + Baird and Tatlock, London.

458 Scientific Intelligence.

a current of water to circulate from the upper end of the tube
to the lower. The great heat, however, was excited in the capil-
lary of the tube. Quartz prepared by the method of Professor
Shenstone possesses the property of resisting changes of tem-
perature in a remarkable manner. One of these quartz tubes can
be heated to a white heat and plunged into water without crack-
ing. Such tubes, therefore, are very valuable for the experiments
I have been conducting on gases at high temperature.

They also possess the great advantage over end-on tubes of
glass provided with quartz window, that the capillary can be
placed close to the slit of the spectroscope ; thus giving a very
intense light and a broad spectrum. Moreover, the quartz is not
melted by the intense heat. A photograph of gaseous spectra
can be obtained with a single discharge and a very narrow slit ;
with tubes filled with hydrogen excited by a difference of poten-
tial of twenty thousand volts, condenser ‘3 microfarads, an
extremely intense light is obtained. This light is dazzling white
with a bluish cast. It has more than three times the actinic
effect of the same quantity of electricity discharged between
magnesium terminals. Viewed with a strait vision spectroscope,
the spectrum appears continuous, and even photography fails to
reveal bright lines between the HH lines and the red end of the
spectrum. In the region, however, beyond the limit set by the
absorption of the glass Geissler tubes there are both bright lines
and dark lines. The principal reversed lines are at wave lengths
2889°70 ; 2549°89; 2528°60; 2524°29 ; 2519°3 5; 2516°21.

These lines correspond with the lines of silicon volatilized by
the spark in air. It seems that we have in this phenomenon
another instance of selective solarization mentioned in my pre-
vious paper. The strongest metallic lines or gaseous lines are not
those which show the strongest reversal. For instance, the cal-
cium line at approximately 4227 is strongly reversed, while the
stronger 3968, 3933 do not show a reversal, except with much
stronger and longer continued discharges.

A careful inspection of the negatives shows that the reversals
of the metallic lines occur when they fall on bright gaseous
lines or bands. In the same way a bright gaseous line falling on
a continuous spectrum can show a similar reversal. We can
express this in symbolic language as follows: let A represent the
intensity of the line and B the amount of the previous action of
light on the photographic plate, then the reversal appears to be
proportional to AB.

It seems probable that there are similar reversed lines running
through the solar spectrum and I hope to detect them.

This investigation shows that the presence of dark lines in the
spectra of stars does not imply necessarily the presence of revers-
ing layers of a colder state of the gases; for such reversal may
arise from photographic action on the plates which are used.
Moreover a gas may show a continuous spectrum to the eye, or
even when photography is employed with glass tubes and glass
lenses; while with quartz tubes such as I have employed a large

Chemastry and Physics. — 459

region in the ultra violet is shown to be traversed by both dark —

and bright lines and bands.

Jefferson Physical Laboratory, Harvard University.

9. A New Holtz Machine.—The influence electrical machine in
many respects has been found to excel the induction coil for the
excitation of X-ray tubes. H. WommeEtsporr describes a new
electrical machine which he calls the condenser machine. It con-
sists of many discs similar to those now used in the Holtz
machine. ‘There is a large condenser action between the glass
plates which are provided with suitable sectors. We have thus
a row of condensers; one set of plates or coated layer remaining
fixed while the intervening plates revolve.

A small model, 30™ high, 28° long and with a breadth of 22,
affords a larger quantity of electricity than the largest Holtz
machine, and this new machine promises to be of the oreatest use
in the excitation of X- -Tay tubes. — Ann. der Physik, No. 11, 1902,
pp. 651-659. Lega,

10. Hlectrical Conductibility of Metals and their Vupors.—
Hon. R. J. Strutt finds that (1) Mercury vapor is an insulator,
while liquid mercury is a conductor. Since the liquid and satu-
rated vapor are indistinguishable above the critical temperature,
one or both of these must undergo a remarkable change of elec-
trical properties as that temperature is approached.

(2) Attempts to predict the critical temperature of mercury
seem to lead to results altogether inconsistent with one another.

(3) Attempts to observe the critical temperature of mercury
and arsenic in quartz tubes have failed. In both cases experi-
ment proves that the critical temperature lies above a dull yellow
heat.

(4) Up to a full red heat the conductivity of saturated mer-
cury vapor remains of quite a different order of magnitude from
that of the liquid, the latter being ten million (10’) times as great
as the former. But on the other hand, the conductivity of the
saturated vapor is immensely greater than that of the vapor at
atmospheric pressure. For the former was found to have a
resistance of 10’ times that of the liquid, the latter more than
410" that of the liquid. Thus the vapor at atmospheric pres-
sure has a resistance about 4X10’ times that of the saturated
vapor, both at a full red heat. It need scarcely be said that this
ratio is of quite a different order from the ratio of the densities
of those vapors. It seems likely that as the critical temperature
is approached the vapor begins to conduct freely, while the
liquid changes its electrical character to a much less extent.

(5) The conductivity of saturated arsenic vapor at a bright red
heat is of the same order as that of mercury, and obeys Ohm’s
law, at all events up to an electromotive intensity of more than
100 volts per em.— Phil. Mag., Nov., 1902, pp. 596-605. J. 7.

11. Lonization of Nuclei Produced by Violent Agitation of
Dilute Solutions ; communicated by C. Barus.—Discharging the
nuclei produced by violent agitation of dilute solutions, in a steady
stream at once into a tubular condenser, the following deflections

460 Scientific Intelligence.

of the electrometer (with very light needle) were observed at
intervals of half a minute: Charge+, 14:4, 10°5, 6°5, 3-4; time
rate 7°5. Charge—, 21°5, 20°0, 18:2, 16°2, 14°2, 12°2; time rate
4°0. Charge+, 18°3, 14:5, 10°4, 6°7, 3°3, °9; time rate 7°6. The
surprising result is thus obtained that the electric current is con-
stant while the initial charge of the inner coating of the con
denser (the outer being earthed), at nearly 20 volts, gradually
quite vanishes.

In explanation it may be assumed either that the number of
ionized nuclei at constant ionic velocity varies inversely as the
potential difference, or far more simply, it seems to me, that the
velocity of the nuclei is independent of the potential gradient,
each nucleus retaining its own specific velocity in the presence or
the absence of an electric field, while the number of nuclei is
appreciably constant—the point of view taken in my earlier work
(Smithsonian Contributions, 1901). The difference of current for
positive and negative charges follows from the known excess of
negative nuclei. The initial ionization is of the same order as
that of the phosphorus emanation.

12. Handbuch der Spectroscopie ; by H. Kayser. Vol. 11, 698
pp., 4 tables, 57 figs. (Leipzig, S. Hirzel) —The appearance of
the second volume of this work gives further assurance of the
thorough treatment of the whole subject of spectroscopy which
we may expect from this eminent worker in the field. In the
present volume, the first chapter is devoted to emission and
absorption, with the history, development and proofs of Kir-
choff’s law. The chapter on radiation of solids furnishes an
exhaustive discussion of the various laws which have been formu-
lated connecting the emission with the absolute temperature, the
distribution of the energy in the spectrum, and the attempts to
apply the results in the measurement of temperature. Under
radiation of gases, the sources of energy and the production of
ether vibrations in general are discussed. This is followed by a
detailed treatment of the spectra of compounds and of the differ-
ent spectra given by the same substance under varying con-
ditions. The succeeding chapters deal, respectively, with the
influence of pressure, temperature and the nature of the electric
discharge upon spectra; the appearance of spectral lines, their
broadening, and reversal ; Doppler’s principle and its applications,
in the discussion of which the author has been assisted by Dr. H.
Konen ; the general relations which have been discovered among
the lines of the spectra of individual elements and of those af
Mendeleeff’s groups. The concluding chapter brings up to date
the work which has been done upon the vibrations of light in the
magnetic field. Prof. C. Runge has assisted in the development
of the Zeeman-effect in accordance with the modern ionic theory.

Evidently no effort has been spared in making the references
to the literature of every subject discussed as complete as possi-
ble up to 1901. The volume is replete with suggested problems
of research, to stimulate which is one of the purposes which the
author hopes the book may serve. D, Ans

Geology and Natural History. 461

II. Guotocy anp Narurat History.

1. United States Geological Survey.—The following publica-
tions have been received :

Minera Resources oF THE UNITED STATES FOR THE CAL-
ENDAR YEAR 1901; by Davin T. Day. 973 pp.—For the second
time the annual mineral production exceeds $1,000,000,000. The
important metals except iron and zinc have decreased in output
and value. 1,408 ounces of platinum were produced in 1901
compared with 400 in 1900. Building materials, clays and abra-
sions show greatly increased production. For the first time in
the United States arsenious oxide was manufactured at Seattle,
and rutile was produced on a large scale—in Nelson Co., Va.

Buuietin No. 195.—Structural details in the Green Mountain
Region and in Eastern New York; by T. Netson Date. 10 pp.,
4 pls., 8 figs. Some interesting examples of structures in meta-
morphic rocks are described and well illustrated. These details
have been collected since the publication of Professor Dale’s
previous papers on this region (16th Ann. Rept., Pt. I, pp. 549-
570, 19th Ann. Rept., Pt. ILI, pp. 199-217).

Bu.uetTin No. 203.—Bibliography and Index of North Ameri-
can Geology, Paleontology, Petrology and Mineralogy for 1901 ;
by F. B. WrEExs. 144 pp.

2. Geological Survey of Kansas—Special Report on Mineral
Waters ; by EK. H. 8S. Batrzy. Vol. vii, pp. 25-333, 38 pls.—
Kansas is well supplied with marketable mineral waters. Mr.
Bailey describes and gives the analyses of some 87 springs and
wells separated into groups according to their mineral content.
_ He also discusses the medicinal value of the various waters and
the industrial uses of the brines. Dr. W. R. Crane contributes
a chapter (pp. 323-333) on the Geological Distribution of mineral
springs and wells.

}
‘

3. A Quantitative Chemico-mineralogical Classification and '

Nomenclature of Igneous Rocks ; by Wuirman Cross, J. P. Ipp-
ines, L. V. Pirsson and H. 8. Wasurneaton (Jour. of Geol., vol.
x, No. 6, pp. 555-690, 1902).—In this work is presented an
entirely new system for the classification and nomenclature of
igneous rocks, a result which the authors believe is to be best
attained by the codperation of several workers agreeing on funda-
‘mental principles. Originally the late Prof. G. H. Williams was
also one: of the collaborators. :

The authors state that after many attempts to modify and use
existing systems this was perceived to be impossible if one should
evolve at the same time a comprehensive and rational one based
on recent investigations and knowledge. Thus through repeated
trials the present system was gradually worked out, and with its
evolution has gone hand in hand the calculation of thousands of
analyses by which it has been tested and its formation in part
controlled. It is a chemico-mineralogical system based on its
own principles, and as its concepts of rocks are in great bart new
it demands a new momicnelatuie,

462 Scientific Intelligence.

What the authors propose is as follows: All igneous rocks are
classified on the basis of their chemical composition ; all rocks
having like chemical composition are grouped together. |

The definition of the chemical composition of a rock and of a
unit of classification is expressed in terms of certain minerals
capable of crystallizing from a magma of a given chemical com-
position, and the expression is guantitative. For this purpose the
rock-making minerals are divided into two groups, the one of the
more siliceous alkali-and calcic-aluminous ones, the other of the
ferro-magnesian minerals. The first group is called mnemonically
the salic group; the second one, the femic group. From this
category the aluminous augites and amphiboles and the micas are
excluded for reasons given.

To completely classify a rock by this system its chemical com-
position must be known by chemical analysis or approximately
so by physical or microscopic optical methods indicated by the
authors. Rocks once determined become types by which similar
rocks may be approximately classified. ,

Since a given magma may crystallize into quite different min-
eral combinations according to the different conditions attending
its solidification, it is necessary to select a certain set of salic and
femic minerals as uniform standards of comparison, These are
the ones ordinarily formed, but aluminous augite and hornblende
- and micas are excluded. In practice, the molecular composition
of a rock obtained from its chemical composition (determined as
mentioned above) is computed into amounts of these standard
minerals and its place in the system is then easily determined.

The standard mineral composition of a rock is called its norm,
and this may be quite different from its actual mineral composi-
tion or mode. Methods are given for obtaining the latter and
indicating its relation to the former.

On the relative proportions of these two groups of standard
minerals the rocks are divided into five Classes, accordingly as
one or the other of these two groups alone constitutes the norm
or is extremely abundant ; whether one or the other is dominant ;
or whether the two are present in about egual proportions. These
Classes are then divided into Orders on the relative proportions
of the minerals forming the predominant group in each case and
in the middle group on the relative proportion of the salic min-
erals. So in the preponderantly salic classes the orders are based
on the relative proportions of quartz, feldspars and feldspathoids.

The Orders are divided into Jtangs on the chemical character
of the basic oxides in the minerals in the preponderant group in
each case; thus if these were feldspathic, as to whether they are
alkalic, alkali-calcie or calcic. The lowest division or grad obtains
only in the three intermediate classes and results from the con-
sideration of the relative amounts of the minerals forming the
subordinate group in each case. Where necessary there are sub-
classes, suborders, subrangs and subgrads.

Texture is considered of minor importance and is taken into
account after the chemical and mineral composition.

Geology and Natural History. 463

Nomenclature. ‘The system demands a new nomenclature and
this has been provided for according to a definite system. As
proposed it consists of three parts, substantive names for the mag-
matic units, implying the chemical composition and the norm ;
then two sets of adjective terms to qualify these nouns, one
referring to the mode and the other to the texture.

The magmatic name consists of a root, geographical in all cases
except for the names of the five classes, and of a suffix. The suf-
fixes are chosen so that they vary in a definite way with the
division of the system to which the magmatic name belongs.
Thus for Class, Order, Rang and Grad, the letters n, 7, s and ¢ in
alphabetical order are used with the vowel a, giving in English
ane, are, ase, ate. For subclass, suborder, etc., the vowel is 0,
giving one, ore, ose and ote.

In the geographical roots, so far as possible those in present
use are retained, advantage being taken of their connotation as
to magmatic character.

The authors propose a nomenclature for field use based on
purely megascopic characters.

The work concludes with a discussion of methods of calculat-
ing mineral composition from chemical composition and the
reverse, and presents tables to aid such calculations. 1. V. P.

4, Petrography and Geology of the Monzoni region in Tyrol.
—The exhaustive study of this area, so long known and so much
investigated as to be perhaps rightly called a classic one, has
recently received a new impetus, apparently from the suggestive
work of Brogger (ii Die Eruptionsfolge der triadischen Eruptiv-
gesteine bei Predazzo in Siidtirol). A number of shorter articles
by various authors have appeared describing new types of rocks
or various phases of its geology. Quite recently, however, more
important memoirs on the geology and petrography of the region,
the one by Rompere (Geologisch—petrographische Studien im
Gebiete von Predazzo 1 and i, (Sitzber. d. K. preuss. Akad. d.
Wiss. Berlin, 1902, phys. mat. class. 675), the other by DoELTER
(Tscher. Min. and Petro. Mitt., vol. xxi) have been published.
The various authors have not entirely agreed among themselves
regarding various points and this has led to a stimulation of in-
terest in their work. As a result many new analyses and descrip-
tions of important rock types have been given. It is shown that
the forecast of RoszeNsBuscu on theoretical grounds, regarding the
existence of nephelinitic and other alkaline types, was correct, and
in general a considerable amount of new information, of value
to systematic petrography, has been developed by these excellent
studies. hia Vo

5. Gesteinskunde fiir Techniker, Bergingenieure etc., von F.
RinNE. (8°, 206 pp., 4 pls., 235 figs. Geb. Jinecke.) Hanover, 1901.
—In this excellent and well written little volume is given a general
survey of the subject of petrology from the modern standpoint,
with especial reference to its use by beginners and those desiring
some knowledge of the subject for technical purposes. It is also
a good introduction to the science for workers in other fields in

Am. Jour. Sci.—Fourts SERIES, Vou. XIV, No. 84.—DxEcremBer, 1902.
32

464 Scientific Intelligence.

geology. Some optical mineralogy is introduced and the use of
the microscope indicated. A further advantage is that a con-
siderable amount of physical geology, in so far as it relates to the
subject, is brought in. <A great help will be found in the numer-
ous well executed half-tones of photographs, especially those of
rock sections showing texture, etc. A more careful selection of
more modern and better executed analyses of rocks and more of
them, would have been of advantage. The volume is well printed
and in convenient form. Lv. Be
, 6. The Terlingua Quicksilver Deposits, Brewster County, Texas.

The University of Texas Mineral Survey, Bull. No. 4, pp. 74.—
These deposits are of quite recent development and are as yet
only worked in a small way by means of open pits. They are
located in the center of an extensive area of Cretaceous sedimen-
tary rocks, which have been much faulted and into which igneous
rocks have been intruded. The ore deposits are found within
the sedimentary rocks and usually at a short distance from some
igneous intrusion. The chief ore mineral is cinnabar, with which
is associated some iron oxides and large amounts of calcite. The
ore occurs in vein formations of a number of different types, none
of which are very wide and most of which vary considerably in
their richness. It is thought that the veins and their ore contents
will be found to continue in depth. The ore is treated near the
mines in furnaces similar to those in use in California. The total
output of the district up to the end of 1901, is estimated at 3,700
flasks of mercury, of which 3,000 flasks were produced during
the year 1901. W. EL F.

7. Notes on New Minerals.—BavuMuavERITE is a sulpbarsenite
of lead having the formula 4PbS:3As,8;,. The analysis made by
H. Jackson gave the following results :

Found. Calculated.
ote ee Ce 48°86 48°75
aka OEY ease 24°39 24°61
IAS a a: uray os tele 2 Sai See 26°64
99°67 100°00

The specific gravity was determined as 5°329. It is monoclinic
in its crystallization, having the axial relations,
a:b:¢ = 1:136817 :1:0°947163 ; B = 82° 422’.

The crystals closely resemble those of dufrenoysite and jordanite
in appearance. They may be distinguished from dufrenoysite
by the marked obligue development of the zone 100—001, and
from jordanite by the absence of twin striations and by the color
of the streak. The mineral is described by R. H. Sotty and
named in honor of Dr. H. Baumhauer, Professor of Mineralogy
in the University of Freiburg.—Min. Mag., xii, 60.

A new analysis of material from the original locality has been
made by G. T. Prior and shows KitiricKENITE to be identical
with the earlier described geocronite. The important difference
between the new analysis and the original analysis is in the find-

Geology and Natural History. 465

ing of arsenic. No determination of arsenic was made in the
first analysis and its presence was evidently not suspected. Prior’s
results are as follows: )

Pb, 68°49; Sb, 9:13; As, 4:59; S, 17°20; total, 99°41.
These results yield the formula 5PbS(Sb, As),S,, which is the
accepted formula for geocronite.

An analysis of MrersitE by Prior gave the following results :
Atomic ratios.

ee eee 38°17 "353 or 4

pe. hoe Bd 090 © J

a ees 56°58 "446 “5
100°39

which yield the formula, 4AqI. Cul. The specific gravity of the
mineral was determined as 5°640.
MarsHite was also analyzed by Prior and gave the following
results :
Atomic ratios.

Seo ey eee 32°35 513

5 Se eee 119 ‘O11

eat SS 65°85 520
99°39

which yield the formula Cul. Specific gravity = 5°590.—Min.
Mag., xiii, 60. 3

CooLeaRrpiTE,—This mineral was discovered by A. Carnot at
- Kalgoorlie. It is steel-gray to yellow-gray in color, is practically
devoid of cleavage and breaks with a conchoidal fracture. Its
formula is (Au,Ag,Hg),Te, and it is, therefore, related to colo-
radoite (Hg,Te,) and melonite (Ni,Te,). The following are
Carnot’s analyses:

| I. IL. EE.

iT (eas 25:15 at ea AS. 37°06
oe Se i 16°65 13°60 4°71
5 TE open ea 3°10 3°70 3°70
0 eee ae “10 “a5 °88
er trace trace 90
1G 2 3 ee 56°55 53°70 5a lB:
“> NS cal Srere 20 ES) E20
1 99°75 99°15 99°58

— Geol. Surv. W. Australia, Bull. No. 6, p. 18.

CoLORADOITE occurs freely and in large pure masses at Kal-
goorlie. A new analysis of pure material is of interest: Hg,
50°40; Au, trace; Ag, 12; Te, 49°48; total, 100-00. Specific
gravity, 9°21.

The formula for coloradoite calculated from these figures is Hg,
Te,, requiring : Hg, 51°7 per cent.; Te 48°3 per cent. The pre-
viously accepted formula based upon analysis of small and impure
specimens was HgTe, requiring: Hg, 61°6 per cent.; Te, 38°4
per cent.— Geol. Sur. W. Australia, Bull. No. 6, p. 27.

466 Scientific Intelligence.

STIBIOTANTALITE, a tantalo-niobate of antimony, Sb(Ta,Nb)O4
is found in the stream works at Greenbushes. The exact crystal-
line form of the mineral has not been determined for it occurs in
water-worn pebbles with a very smooth bright surface. It is
brittle, with a subconchoidal to granular or occasionally fibrous
fracture. Its hardness is 5 to 5°5, and specific gravity 6°4 to 7°4.
Its luster is adamantine to resinous ; color, various shades of yel-
low and brown, also gray. It is subtranslucent to opaque. Its
composition is shown by the following analysis by Goyder: Sb,O,,
40°23 ; Bi,O,, 82; NiO, -08; Ta,O,, 51:13 ; Nb,O,, 7°56 ; total,
99°82.

A hydrated variety of this mineral occurs also, resembling it
in color, but having a rough surface. Its formula would appear
to be 2Sb(Ta,Nb)O4:7H,O. Before the blowpipe stibiotantalite
is practically infusible and colors the flame greenish-gray. It is
reduced to metallic antimony by fusion with potassium cyanide.
In the closed tube the anhydrous mineral gives no sublimate ; the
hydrous, a sublimate of water. After mixing with sulphur it
gives in the closed tube a sublimate which is black when hot and
brownish red on cooling. It is soluble in hydrofluoric acid ; this
solution on adding a little potassium fluoride and evaporating
somewhat, deposits on cooling a felt-like mass of colorless crystals
of potassium fluotantalite. If some of the solution in hydro-
fluoric acid be poured into a platinum dish and a piece of pure
zinc be dropped into it, a black stain immediately develops on
the plate. Stibiotantalite is decomposed by fusion with potas-
sium bisulphate.— Geol. Sur. W. Australia, Bull. No. 6, p. 42.

HIsTRIXITE, an apparently new sulphide of antimony and bis-
muth, is described by W. F. Pretrerp. It occurs in radiating
groups of crystals which are orthorhombic, with acute but indis-
tinct terminations and are striated longitudinally. It is slightly
sectile with a hardness of about 2. Luster eminently metallic.
Color and streak, steel-gray. The mean of two analyses is :—
S, 23°53; Bi, 56°00; Sb, 9°70; Cu, 6°49; Fe, 5°31; total, 101-03.
which gives for the formula, 7Bi,S,,2Sb,8,,5CuFeS,.—John Vail,
Gov. Printer, Tasmania.

PETTERDITE, a new oxychloride of lead. This mineral, which
was found at Zeehan, Tasmania, was described by Mr. W. H.
TweELveTrees. It occurs in thin hexagonal plates of a white
color. Fracture is irregular, luster dull, hardness 1°5 to 2, specific
gravity 7:16. The analysis made by Mr. O. E. White of Hobart
is as follows :—PbO, 74:04; As,O,, 2°60; P,O,, 2°10 ; Sb,O,, -50;
Cl, 20:00; total, 99:24.—John Vail, Gov. Printer, Tasmania. w.. F.

8. Gold in Meteorites—At the September meeting of the
Royal Society of New South Wales, Prof. LiversipeE exhibited
under the microscope particles of a malleable yellow metal, insol-
uble in nitric acid, which have all the appearance of gold obtained
from certain Australian and European meteorites (siderolites).
The presence of gold in meteorites bears upon the presence of
gold in “meteoric” dusts, and it is also of great interest in con-
nection with the presence of gold upon the earth and in sea-water,

Geology and Natural History. 467

inasmuch as meteorites and the dust of meteorites are constantly
falling upon the earth, to the extent of probably many million
tons a year. Further information upon the question of the pres-
ence of gold in meteorites is promised in a subsequent paper.

9. Triassic Ichthyopterygia from California and Nevada ; by
Joun C. Merriam. Bulletin of the Dept. of Geol., Univ. of
Cal., vol. ii, No. 4, pp. 63-108, pls. 5-18.—A valuable contribu-
tion to Paleontology, based chiefly upon saurian material from
the upper Triassic of Shasta County, California, collected in 1891
by field parties from Stanford and California universities. |

The great value of the monograph under review lies in the
carefully executed plates and in the accurate measurement and
description of the bones of a hitherto little known group of
extinct reptiles. Collectively, the several specimens described
afford an excellent generic definition of Shastasaurus, the anterior
part of the skull and the distal ends of the paddles being practi-
cally the only portions of the skeleton which now remain unknown.
Prof. Merriam has met with the great, though by no means rare,
good fortune to find in his collected material nearly as many
species of Shastasaurus as individuals, seven examples offering
more or less complete descriptions of five new forms, S. perrini,
osmonti, alexandra, careyi, and altispinus, although none of the
new acquisitions have been identified with the type species S.
pacificus, which was described by him in 1895 (this Journal,
vol. iv, p. 56). The type species is shown to differ from S. per-
rint in regard to certain posterior dorsal vertebrez and the pubis.
None of the other new species, however, are represented by these
parts, and therefore cannot well be differentiated from the type.
The author wisely states: “ At the present time we cannot deter-
mine definitely the relations of pacificus to the better known
Species, and it is not impossible that when the other parts of the
skeleton of those forms are better known it will be found that
some one of them should be included in pacificus.” In view of
this lack of comparison and the urgent need of simplicity, the
doubt naturally arises whether the purpose of classification would
not have been better served by uniting, provisionally at least, one
of the new forms with the type species.

Accompanying the discussion of Shastasaurus is a careful
revision of Leidy’s genus Cymbaspondylus from the middle Tri-
assic of Nevada, which is shown to be closely related to Shasta-
saurus. The family Shastasauride is proposed to embrace these
two genera, distinguished as they are from the Mixosauride,
Ichthyosauridex, and Baptanodontide, by the peculiar articulation
of their dorsal ribs, the form of the pelvic girdle, and their long-
spined chevron bones. In a postscript, Prof. Merriam states that
a good collection of saurian material has just been obtained from
the middle Triassic of Nevada. When these specimens have been
properly worked out, another contribution is promised by the
author. This will be awaited with great interest, and it is confi-
dently hoped that further investigation will yield as satisfactory
results as the monograph before us. G. F. E.

468 Scientific Intelligence.

10. Zoological Results based on material from New Britain,
New Guinea, Loyalty Islands and elsewhere, collected during
the years 1895, 1896, 1897; by Arraur Wittey. Part VI
(August, 1902), pp. 691-826, 8 pls., 33 figs—Dr. Willey is to be
congratulated upon the completion*® of a valuable series of zoolog-
ical papers. The present volume contains the results of a study
of the development of the Pearly Nautilus. The interesting “ per-
sonal narrative” (pp. 691-734) is illustrated by photographs of
the natives who carry on the Nautilus-fisheries on the different
islands, and who came in contact with Mr. Willey during his
visits to the ar chipelago (1894 and 1897) in search for the eggs of
species of Nautilus. The “special contribution” (pp. 736-826)
is prefaced by a complete review of the bibliography of nautilus
since the first description with figures of the external characters
by Rumphius, 1705, and of the animal by Bennett, 1831. Many
new and interesting facts are mentioned. It was found by obser-
vation that the color markings on both the shell and animal
“exerted a protective influence”; that the “wart—like gibbos-
ities” on the upper surface of the hood form a shield for pro-
tecting the aperture of the shell when the animal is in retraction,
aud not a foot for locomotion. The studies of the anatomical
characters show the relation of the Vautilus to the dibranchs, but
with reference to the other cephalopods opinions will differ. It
is, however, one whose external shell is probably primitive as
compared with the internal shell of Spirula. No definite opinion
is given as to whether the many similarities in anatomy between
the Nautilus and diotocard prosobranchs (Haliotis, Fissurella
and Pleurotomaria) are of the nature of affinity or convergence,
but an interesting figure (15) is given showing a shell with a
median groove and shell-slit similar to Pleurotomaria. XK. J. B.

III. MiIsceELLANEOUS SCIENTIFIC INTELLIGENCE.

1. National Academy of Sciences—The following is a list of
papers read before the National Academy at the meeting in
Baltimore, Nov. 11-12.

S. L. PenrreLp: A possible explanation of the difficult solubility of cer-
tain compounds containing fluorine and hydroxyl.

GrorGE E. Hate: The spectra of stars of Secchi’s fourth type.

T. C. MENDENHALL: Biographical memoir of Henry A. Rowland.

W. K. Brooxs: The embryology of Salpa cordiformis.

CASWELL GRAVE: The occurrence of reef corals near Beaufort, N. C.

D. H. Tennent: The Trematode parasites of the oyster.

H. N. Morse: The preparation of cells for the measurement of osmotic
pressure.

R. W. Woop: A substance with remarkable optical properties, and screens
transparent only to ultra-violet light.

J. B. WHITEHEAD: On displacement currents.

L. A. Parsons: On the spectrum of hydrogen.

Lewis Boss: A new system of positions for standard stars, with notes
relative to its bearing upon sidereal astronomy.

H. F. Ossorn: Complete skeleton and restoration of the Cretaceous fish
Portheus molossus Cope. A new small dinosaur from the Jurassic or Como

* See this Journal, vii, 79, 322; viii, 398; x, 89; xi, 330.

Miscellaneous Intelligence. 469

beds of Wyoming, apparently a bird-catcher. New or little-known elephants
and mastodons of North America.

A. Acassiz: On elevated Oceanic Islands in the Pacific.

2. American Association for the Advancement of Science.—
The next meeting will be held in Washington, Dec. 29 to Jan. 3,
inclusive. ‘This is the first time the Association has met in the
winter. |

3. Smithsonian Institution, Annual Report for 1901; by S.
P. Lanewey, Secretary. Pp. lvii+ 739, 71 pls.—The last report
of the Smithsonian Institution shows that increased appropria-
tions means increased volume of important work. Fifty articles
by well known specialists tell of the latest progress in the princi-
pal branches of knowledge. .

4. United States National Museum. Bull. 51,168 pp. List or
PusiicaTIons 1875-1900, by Ranpo.tpu I. Geare.—The National
Museum has issued a complete list of its publications, arranged
both asa chronological record and as a subject and title index.

5. British Museum — Catalogue of the Collection of Birds’
Figgs ; by Evcene W. Oares. Vol. ii, 400 pp., 15 colored plates.
—Descriptions are given of 15,000 specimens of eggs belonging
to the Carinate. Ten orders and 726 species are included in
the list. :

6. United States Coast and Geodetic Survey ; O. H. Tirrman,
Superintendent :

ANNUAL Report 1901. 423 pp., 46 illustrations, 4 maps in
pocket.—The superintendent of the Coast Survey reports the
inauguration of the survey of the Philippine coast line and the
establishment of a suboffice for the publication of charts and
notices to mariners. Parties were also at work in Porto Rico,
the Hawaiian Islands and Alaska. Perhaps the most import-
ant field work now in hand is the measurement of the 98th
meridian. Mr. A. L. Batpwin reports on the Measurement of
Nine Bases along the 98th Meridian (pp. 241-302); and Mr. J.
F. Hayrorp, inspector of geodetic work, describes the present
state of the triangulation along this meridian in Kansas and
Nebraska. The measured are will cover 23° of latitude within
the United States; the Mexican government is expected to extend
it 9° southward, and it is possible to extend it far northward.

THe Eastern Opsriquze Arc oF THE UNITED STATES AND
OscULATING SPHEROID ; by Cuas. A. Scuotr. Special Publication
No. 7, 394 pp., 38 illustrations, 2 maps in pocket.—The second long
are measured by the Coast Survey extends from Calais, Maine, to
New Orleans, La. The line is 2612°3 kilometers in length,
covering 23° 30’ 57”, and is unigue in that it is the first one which
utilizes on a grand scale a measurement oblique to the meridian.
The work was begun in 1833 and the field work finished in 1898.
The Eastern Oblique Are intersects the Transcontinental Are of
the 39th parallel in Maryland and Virginia. With the measure-
ments along the 98th meridian well under way and two long ares
already completed, the Coast Survey may well feel proud of its
contributions to geodesy.

470 Scientific Intelligence.

7. United States Naval Observatory, Captain C. H. Davis, U.
S. N. Superintendent. Publications, Second Series, Vol. II. Zonz
OBSERVATION WITH THE N1INE-1NcH Transit CrrcLE, 1894-1901;
by Aaron N. SKINNER, assisted by Franx B. Lirrett and THeo.
I. Kine. pp. xxviii, 525. is volume embodies the results of
observations made, in accordance with the suggestion of Dr. A,
Auwers of the Astronomische Gesellschaft, on stars to the ninth
magnitude, inclusive, between the parallels, —13° 50’ and —18°
ae

8. Bureau of American Ethnology.—The following volumes
have recently come to hand:

NINETEENTH ANNUAL Report; by J. W. Powett, Director.
Pt. L, pp. xcii+548. 79 pls., 48 figs. An exhaustive report on
the Myths of the Cherokee Indians has been prepared by James
Mooney, (pp. 3-548).

Buiietin No. 26—Kartuiamet Texts: by Franz Boas. 261

p., 1 pl.

i 9. Ostwald’s Klassiker der Exakten Wissenschaften. Leipsig,
1902 (Wilhelm Engelmann).—The following are recent additions
to this valuable series :

Nr.-129. Allgemeine Methode partielle Differential gleichungen zu inte-
griren ; von Johann Friedrick Pfaff. Pp. 84.

Nr. 130. Pangeometrie ; von N. J. Lobatschefskij (Kasan 1856). Pp. 95.

Nr. 131. Experimental-Untersuchungen uber Elektricitét (xiv und xv
Reihe) ; von Michael Faraday (London 1838). Pp. 48.

Nr. 132. Uber die Continuitat der gas formigen und fltissigen Zustande
der Materie, und Uber den gas férmigen Zustand der Materie; von Thomas
Andrews (London 1869). Pp. 81.

Nr. 193. 3. Ranberhsl Abhandlungen zur Bahnbestimmung der Come-
ten, Pp. 148.

10. Hodgkins Gold Medal Awarded.—The second Hodgkins
Gold Medal has been awarded -to J. J. Thomson of Cambridge,
England, “for his investigations on the conductivity of gases,
especially on the gases that compose the atmospheric air.” The
committee of award consisted of Mr. Richard Rathbun, Chair-

n; Doctor A. Graham Bell, for Electricity ; Doctor Ira Rem-.
sen, for Chemistry ; Doctor Charles D. Walcott, for Geology ;
Professor E. C. Pickering, for Astronomy ; Doctor Theodore N.
Gill, for Biology ; Professor Cleveland Abbe, for Meteorology ;
Mr. William H. Holmes, for Anthropology, and Mr. S. W. Strat-
ton, for Physics.

OBITUARY.

Dr. OaprEN Nicnoxas Roop, Professor of Physics in Columbia
University, died November 12 at the age of 71. (A sketch of
the life and work of Professor Rood will appear in this Journal
for January.)

Dr. Rosert C. Kenzie, for forty years Professor of Chemistry
at the Michigan Agricultural College, died November 7, at the
age of 79 years.

Dr. Rozert Rusenson, Director of the Central Meteorolog-
ical Institute of Sweden, died Oct. 14, aged 73.

Joun Hari GLapstox¢, “one of the founders of physical chem-
istry,” has died in London at the age of 75.

Pmpax FO. VOLUME XILV.*

A

Academy of Sciences,
meeting at Baltimore, 468.
Aerodynamics, experiments
Langley, 318: André, 398.
Alpen, die, im Hiszeitalter, Penck and

Briickner, 315.
André, M. H., Les Dirigibles, 398.
Association, American, meeting at
Pittsburg, 166; meeting at Wash-
ington, 469.
— British, meeting at Belfast, 318.
Austin, M., double ammonium phos-
phates in analysis, 156.

in,

B

Bacubirito, the great meteorite of
Sinaloa, Mexico, Ward, 316.

Barbour, E. H., calcite-sand crystal,
451

Barus, C., velocity and structure of
the nucleus, 220; ionization of
nuclei, 459.

Beitrage zur chemischen Physiologie

und Pathologie, Hofmeister, 69,
398.

Bermuda, ascidians of, Van Name,
74

Black Hills, laccoliths of, Jaggar, 389.
Blake, J. C., estimation of bromic
acid, 285.

BOTANY.

Birches, relationships of some Amer-
ican and Old World, Fernald,
167.

Cyperacez, studies in the, Holm,
No. xvi, 97; xvii, 417.

Plant growth, relation to ioniza-
tion, Plowman, 129.

Bottger, W., Grundriss der quali-

tativen Analyse, 69.

Brewer, W. H., obituary notice of|_

J. W. Powell, 377.

British Museum, catalogue of birds’
eggs, Oates, 469.

Brown, R. M., clays of Boston Basin,
445,

Bumstead, H. A., reflection of elec-
tric waves, 309.

National,

€

California, Cretaceous in, 33.

— Quaternary history, Hershey, 71.

— and Nevada, Triassic Ichthyop-
terygia from, Merriam, 467.

Cathode rays, effect of ultra-violet
light on, Lenard, 67.

— laws of velocity of, Gehrcke, 67.

Chemie, Lehrbuch der Allgemeinen,
Ostwald, 457.

Chemische Physiologie und Pathol-
ogie, Hofmeister, 69, 398.

Chemisches Praktikum, Wolfrum,
385.

Chemistry, Elementary, Clarke and
Dennis, 305.

— First Book of Qualitative, Prescott
and Sullivan, 65.

CHEMISTRY.

Ammonium phosphates in analysis,
Austin, 156.

Ammonium vanadate, precipitation,
Gooch and Gilbert, 205.

Arsenic in animal organism, Ber-
trand, 3838.

Bismuth, radio-active, Marckwald,
503.

—1lead sulphide-halides, Ducatte, 64.

Bromic acid, estimation of, Gooch
and Blake, 285.

Carbon, transformation into dia-

mond, Ludwig, 4595.
compounds, solubilities

Speyers, 298.

— dioxide, action on the borates of
barium, Jones, 49.

Chemical combinations, probable
source of heat, Richards, 64.

Copper sulphate, action upon iron
meteorites, Farrington, 33.

Glass, plasticity and adhesiveness
of, Piccard, 456.

of

b

Helium and liquid hydrogen,
Dewar, 305.

Hydrogen, liquid, and helium,
Dewar, 305.

— and oxygen, combination, Baker,
383 ; vapor-pressures of liquid,
Travers, Senter and Jaquerod,
311.

* This Index contains the general heads, BOTANY, CHEMISTRY (incl. chem. physics), GEOLOGY,,
MINERALS, OBITUARY, Rocks, and under each the titles of Articles referring thereto are

mentioned.
*

472

CHEMISTRY.

Iodine pentafiuoride, Moissan, 456.

Lithium silicide, Moissan, 64.

Manganese, separation from magne-
sium, etc., Dietrich and Hassel,
456.

Matter, heatless condition of, Brink-
worth and Martin, 304.

Nitric acid, titrimetric estimation,
Phelps, 440.

Oxygen and hydrogen, vapor-pres-
sures of liquid, Travers, Senter
and Jaquerod, 311 ; combination,
Baker, 383.

Radium, atomic weight, Curie, 384.

Rocks in the Nile cataracts, black
color of, Lortet and Hagouneng,
64.

Silver chabazite and silver analcite,
Steiger, 31.

Thallium in the thallous
Thomas, 385.

Thermochemical constant, Clarke,
303.

Thiocyanates,
Welis, 584.

Thiocyauic acid, iodometric titra-
tion, Thiel, 455.

Uranium, atomic weight, Richards
and Merigold, 65.

Vanadie acid, reduction of, Gooch
and Stookey, 369.

Clarke, F. W., Elementary Chem-

istry, 305.

Coleman, A. P., nepheline and other

syenites in Ontario, 147.

state,

double and triple,

Colorado, Silverton quadrangle,
economic geology of,. Ransome,
390.

Connecticut, Barkhamsted, essonite
and corundum at, Emerson, 234.

— extent of Newark System, Hobbs,
nae

— ice sheet in Pomperaug valley,
Hobbs, 399.

— Tungsten mine at Trumbull,
Hobbs, 72.
Cornwall, H. B., greenockite on

calcite from Joplin, Missouri, 7.
Crook, Z. E., electromagnetic alter-
nating currents, 133.

Cross, W., Chemico-mineralogical
Classification of Igneous Rocks,
461.

Crystallography, solution of prob-
lems in, Renfield, 249.

Cumings, E. R., variation in the
fossil brachiopod, Platystrophia
lynx, 9

INDEX

D

Davis, W. M., terraces of the West-
field River, Mass., 77; Elementary
Physical Geography, 318.

Dennis, L. W., Elementary Chem-
istry, 305.

Dewar, J., liquid hydrogen and
helium, 805.

Dirigibles, Les, André, 398.

Dresser, J. A., contribution to the
geology of Quebec, 48.

Drude, P., Theory of Optics, 68.

E

Earthquake of 1901 at Inverness, and
at Carlisle, Davison, 397.

Electric conductibility of metals,
Strutt, 459.

— discharges through gases, role of

self-induction in, Eginitis and de

Gramont, 68.
displacement, magnetic effect,

Whitehead, 109. 4

— resistance of glass, quartz, etc.,
Rood, 161.

— sparks, influence of electrification

of the air on, Lecher, 386.
waves, interference, tubes for,

Becker, 67.

— — magnetic detector for, Marconi,
386.

— — reflection of, Bumstead, 359.

Electricity, removal of negative from
the air by rain, Schmauss, 387.

Electrolytes, free ions in aqueous
solutions of, Olsen, 237.

Electromagnetic alternating
rents, Crook, 138.

Electromotive force of ozone, Brand,
386.

Electrostatic tubes, de Nicolaiéve,
386

cur-

Emerson, B. K., corundum and a
graphitic essonite, 234.

Emerson, J. S., characteristics of
Kau, 481.

Ethnology, Bureau of American,
annual report 1901, 470.

Evans, N. N., new occurrence of
native arsenic, 397.

F

Farrington, O. C., action of copper
sulphate upon iron meteorites. 38.

Fernald, M. L., relationships of some
American and Old World birches,
167.

Fisher, C. A., calcite-sand crystal,
451.

INDEX.

Ford, W. E., chemical composition

of dumortierite, 426.
Forest Reserves, U. S., Gannett, 70.

G

Gesteinskunde fir Techniker, etc.,
Rinné, 463.

Gibbs, J. W., Elementary Principles
in Statistical Mechanics, etc., 69.
Gilbert, R. D., precipitation of am-

monium vanadate, 205.
Girty, G. H., upper Permian in west-
_ ern Texas, 363.
Glass, plasticity and adhesiveness of,
Piccard, 456.

GEOLOGICAL REPORTS AND
SURVEYS.
Iowa, 1901, Calvin, 391.
Kansas, 461.
Louisiana, 314.
Michigan, Lane, 393.

United States, 21st annual report,
70, 389, 461.

GEOLOGY.

Alkaline rocks from the region of |

Ampasindava, Lacroix, 396.

Alpen, die, im Hiszeitalter, Penck
and Brickner, 315.

Carboniferous ferns, Sellards, 195.

Clays of the Boston Basin, Brown,
445.

Cretaceous outliers in California,
Hershey, 33.

— turtles, Wieland, 95.

Crossotheca and Myriotheca, fronds
of, Sellards, 195.

Dikes in Syracuse, New York, erup-
tive, Schneider, 24; petrography
of, Smyth, 26.

Economic geology, Silverton quad-
rangle, Col., Ransome, 390.

Eocene Mammalia in the Marsh col-
lection, studies of, Wortman,
17.

Eparchean interval, Lawson, 71.

Eruption of Mt. Pelee, 312, 319;
peculiar character of, Verrill, 72.

Eskers, origin of, Crosby, 316.

Fossil brachiopod, Platystrophia
lynx, variation in, Cumings and
Mauck, 9.

— plant from Illinois, Sellards,
203.

— vertebrata of North America, bib-
liography and catalogue, Hay,
390.

|

473

Geologic formation namesof North
America, Weeks, 391.

Geology of Quebec, contribution to,
Dresser, 43.

Ice-sheet, action of, on projecting
rock masses, Hobbs, 399.

Idiophyllum rotundifolium, valid-
ity of, Sellards, 203.

Igneous rocks, chemico-mineralog-
ical classification, Cross, Iddings,
Pirsson and Washington, 461.

— — size of grain in, Queneau, 70;
Lane on Queneau, 393.

Laccoliths of the Black Hills, Jag-
gar, 389.

Lowlands of southeastern Missouri,
evolution of, Marbut, 392.

Mammalia, Eocene in Marsh col-
lection, 17.

Monzoni region in the Tyrol, petrog-
raphy and geology, Romberg and
Doelter, 463.

Newark system, former extent of,
Hobbs, 71.

Niagara limestones of Indiana, Kin-
dle, 221.

Permian, upper, in western Texas,
Girty, 363.

Platystrophia lynx, variation in the
fossil brachiopod, Cumings and
Mauck, 9.

Pleistocene geology of western New
York, Fairchild, 393.

Plissements et dislocations de
Vécorce terrestre en Grece,
Negris, 71.

Potomac group of the middle At-
lantic slope, geology, Clark and
Bibbins, 392.

Quaternary history of
California, Hershey, 71.

Rocks in the Nile cataracts, black
color of, 64.

Terraces of the Westfield River,
Mass., Davis, 77.

Triassic _Ichthyopterygia from Cal-
ifornia and Nevada, Merriam,
467.

Vertebrates of the Northwest Ter-
ritory, Osborn and Lambe, 393.
Voleanic eruptions of 1902 on St.
Vincent and Martinique, 72, 312;

Hovey, 319.

Volcanoes, Hawaiian, Kau, Emer-
son, 431.

Gooch, F. A., precipitation of am-
monium vanadate, 205; estimation
of bromic acid, 2850; reduction of
vanadic acid, 369.

Grundriss der qualitativen Analyse,
Bottger, 65.

southern

474

H

Havana, study of yellow fever in,
Gargas, 79.

Hawaii, the volcano of Kau, Emer-
son, 481.

Hershey, O. H., Cretaceous outliers
in California, 33.

Hillebrand, W. F., additions to the
Alunite-Jarosite group, 211.

Hobbs, W. H., action of ice-sheet on
projecting rock masses, 399.

Hodgkins gold medal awarded, 470.

Hofmeister, F., Beitrige zur chem-

__ischen Physiologie, 69, 398.

Holm, T., studies in the Cyperacee,
No. xvi, 57; xvii, 417.

Holtz machine, new, Wommelsdorf,
4959.

Hovey, E. O., eruptions of 1902 on
St. Vincent and Martinique, 319.
Hydrogen, spectra of, Trowbridge,

457.

I

Iddings, J. P., Chemico-mineralogi-
cal Classification of Igneous Rocks,
461.

Indiana, Niagara limestones of, Kin-
dle, 231.

International Quarterly, 318.

Iowa geological survey, 1901, 391.

J

Jones, L. C., action of carbon diox-
ide on the borates of barium, 49.

K

Kansas geological survey, 461.

Kau, characteristics of, Emerson, 481.

Kayser, Hk, Handbuch der Spectro-
scopie, 460.

Kindle, E. M., Niagara limestones of
Indiana, 221.

Koenig, S A., new species melano-

chalcite and keweenawite, 404.

Lc

Lacroix, A., \iinéralogie de la France
et ses Colonies, 70.

Lane, A. C., Queneau on size of
grain in igneous rocks, 393.

Langley, S. P., Experiments in Aero-
dynamics, 318.

La Soufriére, St. Vincent, and Mt.
Pelée, Martinique, eruptions of,
Hovey, 319.

Light, electrical effect of, Lenard,
67.

.Louisiana geological survey, 314.

INDEX.

M

Maldive and Laccadive Archipela-
goes, Gardner, 74.

Marsh collection, studies of the
Eocene mammalia, Wortman, 17.

Martinique, 72, 312.

— and St. Vincent, eruptions of 1902,
Hovey, 319.

Massachusetts, Boston Basin, clays
of, Brown, 445.

— Westfield River, terraces of, Wis
Mauck, A. V., variation in the fossil
brachiopod, Platystrophia lynx, 9.
Measurement of 98th meridian, 469.
Mechanics, Elementary Principles in

Statistical, Gibbs, 69.

Metals, electrical conductibility of,
Strutt, 459.

— reflection power for ultra-violet
and ultra-red rays, Hagen and
Rubens, 66. ee

Meteorite, iron, from Sinaloa, Mex-
ico, Ward, 316

Meteorites, action of copper sulphate
upon, Farrington, 38; gold in,
Liversidge, 466.

Mexico, Sinaloa, meteorites from, 316.

a geological survey, Lane,
93.

Mineral resources of So.
O’Harra and Todd, 897. . .

— — of the United States, 1901,
Day, 461.

— waters, Bailey, 461.

Dakota,

'Minéralogie de la France et de ses

Colonies, Lacroix, 70.

MINERALS.

Alunite, Colorado, 216. Analcite,
dl. Arsenic, native, 807.

Baumhauerite, "464,

Calcite-sand crystal, new form, 451.
Chabazite, 31. Coloradoite, 465.
Coolgardite, 465. Corundum,
234.

Dumortierite, chemical composition,
426.

Essonite, graphitic, 234.

Garnet, 234. Gold in meteorites,
466. Greencckite on calcite,
Joplin, Missouri, 7.

Histrixite, 466. Hussakite, 316.

Jarosite, 216.

Keweenawite, Michigan, 410. Kill-
rickerite, 464.

Marshite, 465. Melanochalcite,
Arizona, 404. _Melilite, Syracuse,
INFN Gr Mercury-deposits in
Texas, "464. Miersite, 465.

Natrojarosite, Nevada, 211.

INDEX.

MINERALS.

Petterdite, 466.
New Mexico, 213.

Stibiotantalite, 466.
Xenovtime, 316.

Missouri, evolution of the lowlands
of southeastern, Marbut, 392.

Mt. Pelée, 312, 319; peculiar char-
acter of the eruption of May 8th,
Verrill, 72.

Plumbojarosite,

N

New York, Syracuse, eruptive dikes
in, Schneider, 24; petrography of
dikes, Smyth, 26.

— Western, pleistocene geology of,
Fairchild, 393.

Nile cataracts, black color of rocks
in, Lortet and Hugouneng, 64.

Noyes, A. A., General Principles of
Physical Science, 457.

Nucleus, velocity and_ structure,
Barus, 225; ionization of, Barus,

459.

O
OBITUARY.

Gladstone, J. H., 470.
Kedzie, R. C., 470.
Powell, J. W., 377.
Riva, Dr. Carlo, 166.
Rood, O. N., 470.
Rubenson, R., 470.
Virchow, Rudolph, 318.
Olsen, J., free ions in aqueous solu-
tions of electrolytes, 237.
Ontario, syenites in,Coleman, 147.
Optics, theory of, Drude, 68.
Ostwald, W., Lehrbuch der Allge-
meinen Chemie, 457.
Ostwald’s Klassiker der Exakten
Wissenschaften, 76, 470.
pe” electromotive force of, Brand,
86.

P

Penfield, S. L., solution of problems
in crystallography, 249.

Phelps, I. K., titrimetric estimation
of nitric acid, 440.

Phonetics, Elements of Experimen-
tal, Scripture, 387.

Physical Geography, Gilbert and
ao 398 ; Klementary, Davis,

8.

— Science, General Principles of,
Noyes, 457.
Pirsson, L. V., Chemico-mineralogi-
ce Classification of Igneous Rocks,
61,

475

Plowman, A. B., relation of plant
growth to ionization, 129.

Porto Rico, actinians of, Duerden,
74,

Powell, John Wesley, obituary no-
tice, Brewer, 377.

Prescott, Qualitative Chemistry, 65.

Q

Quebec, petrographical contribution
to the geology. of, Dresser, 43.

Queneau on size of grain in igneous
rocks, 70; Lane, 398.

R

Radio-active bismuth, Marckwald,
303.

— substances, Rutherford and Grier,
387; radiations, Rutherford and
Brooks, 326.

Radioactivity in air, Thompson, 387.

Rinné, F., Gesteinskunde fiir Tech-
nicker, etc., 463.

ROCKS:

Dahamite, Pelikan, 397.
Eleolite syenite, Morozewicz, 396.
Mariupolite, Morozewiez, 396.
Nepheline syenites, Ontario, 147.
Syenites, Ontario, 147.

Rood, O. N., electrical resistance of
glass, quartz, etc., 161.

Ss

St. Vincent and Martinique, erup-
tions of 1902, 72, 312 ; Hovey, 319.

Scientia, 76.

Scientific Literature,
Catalogue, 317.

Schneider, P. F., eruptive dikes in
Syracuse, New York, 24.

Scripture, E. W., Elements of Ex-
perimental Phonetics, 387.

Self-induction, role of in discharges
of electricity through gases, Hgin-
itis and de Gramont, 68.

Sellards, E. H., fronds of Crosso-
theca and Myriotheca, 195 ; validity
of Idiophyllum rotundifolium, 203.

Smithsonian Institution, annual re-
port, 1901, 469; publications of,
76.

Smyth, C. H. Jr., petrography of
dikes in Syracuse, N. Y., 26.

South Dakota, mineral resources,
O’Harra and Todd, 397.

Spark discharge from metallic poles
in water, Wilsing, 67.

International

4776

Spectra from the dissociation of

water vapor, Trowbridge, 1.

ee hydrogen and gases, Trowbridge,

Spectroscopie, Handbuch der, Kay-
ser, 460.

Spectrum of Novae, cause of pecu-
liarities in, Wilsing, 67.

Speyers, C. L., solubilities of some
carbon compounds, 293.

Steiger, G., silver chabazite and
silver analcite, 31.

Stookey, L. B., reduction of vanadic
acid, 369.

Sullivan, Qualitative Chemistry, 65.

SO
Terlingua quicksilver deposits,
Texas, 464.
Texas, northwest boundary of,
Baker, 391.

— upper Permian in western, Girty,
369.

— Terlinqua quicksilver deposits, 464.

Thermo element, vacuum, Lebedew,
389.

Trowbridge, J., spectra from the
dissociation of water vapor, 1;
spectra of hydrogen and reversed
lines in the spectra of gases, 457.

Tungsten mine, at Trumbull, Conn.,
Hobbs, 72.

Turtles, Cretaceous, Wieland, 95.

Tyrol, petrography and geology of
the Monzoni region, Romberg and
Doelter, 463.

U

Ultra-red rays, reflection power of
metals for, Hagen and Rubens, 66.

Ultra-violet rays, reflection power of
metals for, Hagen and Rubens, 66.

United States Coast Survey, annual
report, 1901, 469.

— — Geol. Survey,21st annual report,
70, 389, 461, Bulletins, 390, 391, 461 ;
mineral resources, 461.

INDEX.

United States National Museum, list
of publications, 469.

— — Naval Observatory, 470.

— — Weather Bureau, 76.

V

Verrill, A. E., peculiar character of
the eruption of Mt. Pelee, 72.

Volcanic eruptions on St. Vincent
and Martinique, 72, 312, 319.

W

Washington, H. S., Chemico-min-
eralogical Classification of Igneous
Rocks, 461.

Whitehead, J. B., Jr., magnetic
effect of electric displacement, 109.

Ween G. R., Cretaceous turtles,

Wolfrum, A., Chemisches Prakti-
kum, 385

Wortman, J. L., Eocene Mammalia.
in the Marsh collection, studies of,
17.

Zz

Zoological Results from New Britain,.
New Guinea, etc., Willey, 468.

ZOOLOGY.
Actinians of Porto Rico, Duerden,,.

Ascidians of ‘Bermuda, Van Name,

Brachiopod, embryology of, Conk-.
tim, Vo:

Corals of Maldive and Laceadive
Archipelagoes, Gardner, 74.

Nautilus from New Britain, etc.,
Willey, 468.

Stegomyia fasciata infected by yel-
low fever in Havana, study of,.
Gargas, 75.

oi stv. gl JULY, 1902.

Established by BENJAMIN SILLIMAN in 1818,

oY See
i a we

s AMERICAN
}JOURNAL OF SCIENCE.

Epitor: EDWARD S. DANA.

ASSOCIATE EDITORS

| Prorzssors GEO. L. GOODALE, JOHN TROWBRIDGE, [J
| W.G. FARLOW anp WM. M. DAVIS, or CamsripcE, a

«,

- e ok : ;
~ ‘ ; | 7 a q < ~
4 2 ee ede my (tae ee hs Op eh ie aie} © r)
Ea) ere 2 is aa e's teed. pee ee i pe te aad eau A¢
Ste ae oe
gale ise i eee AEN rete a

per or)

ChE

}| Proressors A. E. VERRILL, HENRY S. WILLIAMS anp a
| L. V. PIRSSON, or New Haven.

Proressor GEORGE F. BARKER, or PHILADELPHIA, , _
ProFressor JOSEPH S. AMES, or Battimorg, a
Mr. J. S. DILLER, or WasHINGTON. ay

FOURTH SERIES. ee
VOL. XIV—[WHOLE NUMBER, CLXIYV.] ae x
: No. 79.—JULY, 1902. |

WITH PLATES I-III.

é dk 5 naa J : LE din § ~ “4,8 t/t! = f wt loti Pi. tae
i e oes He F e-8y ate Aa Slik vs er rae ‘toa a9 i % ar. re eke Aa, oe
een SIRT ee Wied :

NEW HAVEN, CONNECTICUT.

ee eee

1902.

m= A.
oar ce
0b
RES wa
Ate ee : Ny

Bs _ THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 125 TEMPLE STREET,

‘ublished monthly. Six dollars per year, in advance. - $6.40
Union. Remittances should be made either by money
: Has bank checks Ceseeer by on New York banks).

Zs OPES! i the, /)

ders, registered %

JUL 3) 1902/

SAPPHIRE BLUE

of the true shade is to be found in the popularly named “ False Sapphire”
or CYANITE from the far-famed St. Gothard. A trip by our collector
and considerable work done for us yielded some superb specimens. This
locality has been known for over half a century, but like many others in
Switzerland, is quite unworkable, save during two months of the year, and
rarely visited even then. Thus the specimens are not new—just vastly supe-
rior to those in the large museums, all of which have the early specimens, but
are fast replacing them with the later ones,

Months of expert work were devoted to the careful removal of the Para-
gonite matrix, exposing the transparent blue crystals associated with lustrous
dark brown Staurolites, often in parallel and penetrating habits. This is
mentioned in some of the mineralogies, which likewise add ‘‘ rarely termi-
nated.” Yet we have perfectly terminated Cyanites six inches long, pene-
trating the length of a Staurolite. This peculiarity, together with the
contrasting blues and browns standing out on the pearly background, affords
one of the most striking combinations to be seen in any collection.

The stock of really fine specimens is limited and rapidly diminishing. 50e.
to $15.00.

Detached Crystals, 25c. per dozen to $1.00 each.

Terminated Crystals, 50c, to $2.00 each,

OTHER SWISS MINERALS.

We secured by exchange the few duplicates left in one of the oldest of
European collections. Among them were the following historical things of
high value: Hisenrose, Sphene, Octahedral Rose Fluor, Axinite, Apatite,
Green Fluor, Smoky Quartz, ete., ete.

EDUCATIONAL COLLECTIONS.

For 26 years we have supplied mining schools, universities, colleges and
secondary schools throughout the world with mineralogical material, Dur-
ing that period the quality of our elementary and advanced collections has
steadily improved, so that to-day the highest grade of study specimens are
offered at unprecedentedly low prices. An inspection of our Laboratory
List will show that European minerals are sold not simply below American
prices, but often at lower rates than prevail in Europe. The wide connec-
tions of our European house alone permit this economy to the consumer,
our prices being the same on both sides of the Atlantic. If in Paris this
summer favor us with a call—15 minutes from the Opera Quarter.

Illustrated Collection Catalog Free,

The Largest and Most Complete Stock of Scientific and Educational Minerals
in the World, Highest Awards at Nine Expositions.

FOOTE dCi 250.4200 Gee

FORMERLY DR. A. B, FOOTE,

PHILADELPHIA, PARIS,
1817 Arch Street, 24 Rue du Champ de Mars.

Dr. Cyrus Adler, | 2 RR aot ce aie aes si:

Librarian U. S, Nat. Mase. ad ok sees A A 7 Sea ie ae dee re deechs

Be AGB, 1908.

i a Established by BENJAMIN SILLIMAN in 1818.

AMERICAN 3
JOURNAL OF SCIENCE.

Epitorn: EDWARD S. DANA.

[s : ASSOCIATE EDITORS
| Proressors GEO. L. GOODALE, JOHN TROWBRIDGE, ae
_ W. G, FARLOW anp WM. M. DAVIS, or Camprince, a

_ Proressors A. E. VERRILL, HENRY S. WILLIAMS anv |[
Ll: V-- PERSSON; oF NEw HAVEN, — aa.

PROFESSOR GEORGE. F. BARKER, OF PHILADELPHIA, =
Proressor JOSEPH S. AMES, oF BALTIMORE, na =

Mr. J. S. DILLER, or Wasuincrton. :
FOURTH SERIES. : i

‘VOL. XIV-[WHOLE NUMBER, CLXIV] - i
No. 80.—AUGUST, 1902. ee

i: WITH PLATE IV. bere

NEW HAVEN, CONNECTICUT.

| a its A 5 Se

t AGS 2 ey en

: ye hd. =
THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 125 TEMPLE STREET. is
ss

lished monthly. Six dollars per year, in advance. $6.40 to countries in the |
Union. Remittances should be made either by money orders, registered
or bank checks spent on New York banks). !

SAPPHIRE BLI

of the true shade is to be found in the popularly named “‘ False S phi
or CYANITE from the far-famed St. Gothard. <A trip by our collec
and considerable work done for us yielded some superb specimens. yA
locality has been known for over half a century, but like many others
Switzerland, is quite unworkable, save during two months of the~ year,
rarely visited even then. Thus the specimens are not new—just vastly supe-
rior to those in the large museums, all of which have the early speceieres bat ;
are fast replacing them with the later ones, = “pe

Months of expert work were devoted to the careful removal of the: Para-: a) 3a
gonite mairix, exposing the transparent blue crystals associated with lustrous —
dark brown Staurolites, often in parallel and penetrating habits. This is €
mentioned in some of the mineralogies, which likewise add ‘‘rarely termi- ee
nated.” Yet we have perfectly terminated Cyanites six inches long, pene- _
trating the length of a Staurolite. This peculiarity, together with the é .
contrasting blues and browns standing out on the pearly background, affords
one of the most striking combinations to be seen in any collection. iy

The stock of really fine specimens is limited and rapidly pp 50c.
to $15.00.

Detached Crystals, 25c. per dozen to $1.00 each.

Terminated Crystals, 50c. to $2.00 each.

OTHER SWISS MINERALS.

We secured by exchange the few duplicates left in one of the oldest of
European collections. Among them were the following historical things of
high value: Kisenrose, Sphene, Octahedral Rose Fluor, Axinite, Apatite, —
Green Fluor, Smoky Quartz, ete., etc.

EDUCATIONAL COLLECTIONS.

For 26 years we have supplied mining schools, universities, colleges and
secondary schools throughout the worid with mineralogical material. Dur-
ing that period the quality of our elementary and advanced collections has s
steadily improved, so that to-day the highest grade of study specimens are_
offered at unprecedentedly low prices. An inspection of our Laboratory —
List will show that European minerals are sold not simply below American
prices, but often at lower rates than prevailin Europe. The wide connec-
tions of our European house alone permit this economy to the consumer, ~
our prices being the same on both sides of the Atlantic. If in Paris this 3
summer favor us with a call—15 minutes from the Opera Quarter.

illustrated Collection Catalog Free.

The Largest and Most Complete Stock of Scientific and Educational Minerals — ‘
in the World. Highest Awards at Nine Expositions. >

Foote Miwa A cee

FORMERLY DR. A. E. FOOTE,

PHILADELPHIA, PARIS,
1317 Arch Street. ; 24 Rue du Champ de Mars.

: ~ Sys
? “” ey Fes
7 te NEO
fit ai ee a
in = Cur’ < ri 4;
ia Mes

by BENJAMIN a: ae

THE. ae ee 2

" AMERICAN

Eprror: EDWARD S. DANA.

- ASSOCIATE EDITORS

rEssORS Been VERRILL, HENRY ‘SS. WILLIAMS anp
PV. PIRSSON, oF New Haven,

~ PRovEssor JOSEPH oF AMES, OF BALTIMORE,
Mr. J. S. DILLER, or Wasurncron.

oe FOURTH SERIES.
_ Vor. XIV—[WHOLE NUMBER, CLXIV.]

No. 81.—SEPTEMBER, 1902.

WITH PLATES Y-VII.

-

NEW HAVEN, CONNECTICUT.
EO. > Shi

H ; TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 125 TEMPLE STREET.-

Se URhie Six dollars per year, in advance. $6.40 to countries in the
.. Remittances should be_made either by money orders, registered —

“ES sureteceby on 1 N ew York Hse)

URNAL OF SCIENCE.

SS aS ear re
’

a a eee ok eee ae ee, a ae oe ,. nee ed Ae ee

: Cs A ised Pe ie eee
, i 4 , ‘ . r

“ ‘ ) é +3

characteristic habits. Breislakite, Hauynite.

Ten months were spent eae our collector in visiting the es
tions and several of the most important localities. A large numb
specimens were thus secured, which cannot be duplicated eve:
tempting exchanges that can be personally offered. But few good miner a
now coming from the mines of the peninsula, and the high values plac
examples of the old tinds are steadily rising. The following are all good
examples of their kind, extra choice crystallizations being Coie we

: - PIEDMONT -
Piedmontite, V iolan, Diopside, Idocrase, large Hessonite Bi of
fine quality.
BAVENO

(At lower prices than foreign collectors pay at the quarries.) a

Bayenite, a very rare new species, Babingtonite, rare, Orthoclase in Ba
and Manebach twins, Fluorite.
CARRERA

Limpid Rock Crystal on Marble.

: PORRETTA
A large assortment of the well known cavernous Quartz Crystals, many sho
moving bubbles, Calcite in neat groups of primitive rhombs. —

ELBA :
A large lot of Tourmaline crystals including excellent terminations,
green and white. A few vari-colored. Cookeite, Ilvaite, Pollucite. We maintain:
the old scale of prices on this consignment in spite of a great advance in Tal

BELLISIO SOLFARE ey

Sulphur in brilliant transparent crystals of different types from the Site
and of finer quality. Selenite, perfectly limpid and flawless crystals of unrivalled
lustre, offering the most highly prized specimens of this mineral known. ee 3
with the Sulphur and isolated. A remarkably beautiful but rare occnrrenee- :

CAPO DI BOVE ae
Nephelite in groups of sharply defined hexagons. Gismondite ayeie 0!

MONTE SOMMA :
“What's in a name” is answered for the mineralogist in the eee is.
Of some of them we obtained but one specimen, and at ‘the best got but a sad

from which Prof. Scacchi made the descriptions: _

Granuline Neochrysolite Melanothallite I
Neociano Nocerite Sodalite, well eer :
Euchlorite Aphthitalite, fine crystals Thermonatrite ce
Humboldtilite Belonesite with Humite
Cuspidine Cryphiolite Hydrodolomite

Meionite, large xls Séméline Hydrccyanite

Erythrocalcite Facellite, fine group of :
Dolerophanite crystals | a

Illustrated Collection Catalog Free.

The Largest and Most Complete Stock of Scientific and Educational Mine
in the World. gas pare Awards at Nine Expositions.

FORMERLY DR. A. E. FOOTE,

PHILADELPHIA, ee - PARIS{S 22
1317 Arch Street. 24 Rue du Champ de Mar .

‘Cyrus Adler,
Librarian U. S. Nat. Museum.

OCTOBER, 1902.

Established by BENJAMIN SILLIMAN in 1818.

THE

AMERICAN

Epitor: EDWARD S. DANA.
are

ASSOCIATE EDITORS -

EV. PIRSSON, or New Haven,

PROFESSOR GEORGE F. BARKER, or PHILADELPHIA,
ProFEssor JOSEPH S. AMES, or BALtTImore,
Mr. J. S. DILLER, or WASHINGTON.

FOURTH SERIES.

VOL. XIV—[WHOLE NUMBER, CLXIV.]

No. 82.—OCTOBER, 1902.

NEW HAVEN, CONNECTICUT.

/ ZS. , OF, "
THE TUTTLE, MOREHOUSE & TAYLOR CO.,, ay S, 125 TEMPLE STREET.

Pee hlished monthly. Six ALTIUS Asace $6.40 to countries in the
Postal Union. Remittances should ffiade either by money eects registered
tter or bank checks (preferably on New York banks).

(oes eee _—
re * Fg as ons, oe
a any i
my Lael
ey a

ITALIAN MINERAI

MANY IMPORTANT RARITIES. bac

Ten months were spent by our collector in visiting the principal mineral coll
tions and several of the most important localities. A large number of histori
specimens were thus secured, which cannot be duplicated even by the m
tempting exchanges that can be personally offered. But few good minerals ar
now coming from the mines of the peninsula, and the high values placed upon
examples of the old finds are steadily rising. The following are all good typica
examples of their kind, extra choice crystallizations being osucerne noted. ms

PIEDMONT
Piedmontite, Violan, Diopside, Idocrase, large Hessonite crystals of unusually 2
fine quality. lc hae
BAVENO ’

(At lower prices than foreign collectors pay at the quarries.)

Bavenite, a very rare new species, Babingtonite, rare, Orthoclase in Baveno
and Manebach twins, Fluorite. 3
CARRERA

cea Rock Crystal on Marble.

PORRETTA . et
A large assortment of the well known cavernous Quartz Crystals, many showing
ets bubbles, Calcite in neat groups of primitive rhombs.

ELBA . eee
A large lot of Tourmaline crystals including excellent terminations. Red, ~
green and white. A few vari-colored. Cookeite, Ilvaite, Pollucite. Wemaintain ~~
the old scale of prices on this consignment in spite of a great advance in Italy.

BELLISIO SOLFARE . fon

Sulphur in brilliant transparent crystals of different types from the Sicilian

and of finer quality. Selenite, perfectly limpid and flawless crystals of unrivalled

lustre, offering the most highly prized specimens of this mineral known. Grouped
with the Sulphur and isolated. A remarkably beautiful but rare occurrence. —

CAPO DI BOVE < : Po
Nephelite in groups of sharply defined hexagons. Gismondite crystals of
characteristic habits. Breislakite, Hauynite. . ; ¥

MONTE SOMMA

‘What's in a name” is answered for the mineralogist in the following list.
Of some of them we obtained but one specimen, and at the best got but a sadly
limited supply. All are old and quite’a number are of the original meagre fee so oe
from which Prof. Scacchi made the descriptions :— :

Granuline ; Neochrysolite Melanothallite

Neociano Nocerite Sodalite, well crystallized —
Euchlorite Aphthitalite, fine crystals Thermonatrite
Humboldtilite Belonesite with Humite

Cuspidine Cryphiolite | Hydrodolomite

Meionite, large xls Séméline Hydrceyanite
Erythrocalcite Facellite, fine group of

Dolerophanite crystals

Illustrated Collection Catalog Free.

The Largest and Most Complete Stock of Scientific and Educational Minerals
in the World. reas tae Vg Awards at Nine Expositions.

a ES NENT eA Coes

FORMERLY DR. A. E. FOOTE,

PHILADELPHIA, PARIS,
1317 Arch Street. : 24 Rue du Champ de Mars.

# y T U Ss Aaler,
4 ian U. S. Nat. Museum.

s AMERICAN
URNAL OF SCIENCE.

_ Eprror: EDWARD S. DANA.

ASSOCIATE EDITORS
oe GEO. L. GOODALE, JOHN TROWBRIDGE,

L. V. PIRSSON, oF NEw fies

- Proressor GEORGE F. BARKER, oF PHILADELPHIA,
_ PROFESSOR JOSEPH S. AMES, or Battimorg,
Mr. J. S. DILLER, or WasutncTon.

FOURTH SERIES.

No. 83.—NOVEMBER, 1902.

WITH PLATE VIII.

NEW HAVEN, CONNECTICUT. 0/9
1909. . Shona Huse’

“ete ~*~

blished monthly. Six dollars per year, in advance. $6.40 to countries in the
Union. Remittances should be made either by money orders, registered
co Pails checks Se on New York erste

WYER LO
“

Several recent’ shipments aoe from ae oe of A

cabinet specimens. Small study specimens SS 15c. 6 40e. z

Hematite in hexagonal = New: and rare.

ae examples, but of different erystalline habit.

Rose Apophyllite in handsome groups. Beautifat white and co. pS a
erystallizations. Also detached crystals. pase

aS Hyalite. Clear masses of pearly lustre in twisted and botryoi
& Attractive and typical. | ae

Bais: : Fire Opal in large pieces. ogre i=
sie : Octahedral Fluor showing interesting modifications. — = ee
| Quartz Crystals containing moving bubbles. Saas
Amethyst in groups of unrivaled richness and depth of oglege

“be Stilbite. Delicate cream-colored groups. Crystals symmetrical sind wells
) | defined. ea

offered at sinpiecod sakeaie oh low prices. An inspection of our Laborato
tag List of minerals for analysis, will show that European minerals are sold me
simply below American prices, but often at lower rates than prevail —
Europe. The wide connections of our European house alone permit 1]
economy to the consumer, our prices being the same on both sides ery |
Atlantic. >

Illustrated Collection abi Free.

Foote MINERAL co.

FORMERLY DR. A. E. FOOTE,

PHILADELPHIA, oan F
1317 Arch Street. 24 Rue du Champ de Mars.

_ Cyrus Adler,

Librarian U. S. Aa

THE

AMHRICAN

Epror : EDWARD S. DANA.

7 Sr | ASSOCIATE EDITORS
BSSORS GEO. L. GOODALE, JOHN TROWBRIDGE,

FOURTH SERIES.
VoL, XIV—[WHOLE NUMBER, CLXIV.].
No. 84.—DECEMBER, 1902.

ee WITH PLATE IX.

re
NEW HAVEN, CONNECTICUT. /7% 3:

1902.

hed monthly. Six dollars per year, in advance. $6.40 to mere in the
mn. Remittances should be made either by money orders, registered
es checks. Se a on New RES banks).

————————

oe =

steadily improved, so that to-day the highest grade of study specimens are ~

Several recent shipments direct from the ‘‘ Andreasberg of America”?

+8

afford a large choice of the more recent-finds of this famous district. Prices
-average lower than formerly, generally 50c. to $3.00 or $4.00 for the best sy
cabinet specimens. Small study specimens at 15c. to 40e.

Hematite in hexagonal prisms. New and rare. Small but marvel-
ously sharp and brilliant, coating a volcanic rock suggesting the Vesuvian
examples, but of different crystalline habit.

Rose Apophyllite in handsome groups. Beautiful white and colorless
erystallizations. Also detached crystals,

Hyalite. Clear masses of pearly lustre in twisted and botryoidal forms. |
Attractive and typical.

Fire Opal in large pieces. Fe
Octahedral Fluor showing interesting modifications.

Quartz Crystals containing moving bubbles. - :
Amethyst in groups of unrivaled richness and depth of color; 45558 , Re za

Stilbite. Delicate cream-colored groups. Crystals symmetrical and well .
defined. f st

STUDY SPECIMENS —

For 26 years we have supplied mining schools, universities, colleges aie
secondary schools throughout the world with mineralogical material. . “Dur-
ing that period the quality of our elementary and. adyanced- collections: has —

offered at unprecedentedly low prices. An imspection of our Laboratory ie
List of minerals for analysis will show that European minerals are sold not ~
simply below American prices, but often at lower rates than prevail in 3
Europe. The wide connections of our European house alone permit this
economy to the consumer, our Pass being the same on both sides of the
Atlantic.

Illustrated Collection Catalog Free.

The Largest and Most Complete Stock of Scientific and Educational Minerals. t
in the World. Highest Awards at Nine Expositions.

Foote MINERAL CO... a

FORMERLY DR. A. E. FOOTE,

PHILADELPHIA, PARIS,
1317 Arch Street. 24 Rue du Champ de Mars.

et

313 a

(H )

NIA

ii
3 90

ARIE

WAMU
88 01298 5651

80