ee en

ae Se
a

De De thath- Him he ata hn Dol ee me
pte ita tan he a tentn ita otha

oe Nah Ata Ba Fig em nahn a ha
re een wen

pee
ee ee
pt om nba ey Bon

THE

AMERICAN
JOURNAL OF SCIENCE.

Epirorn: EDWARD S. DANA.

ASSOCIATE EDITORS

Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE,
W. G. FARLOW anp WM. M. DAVIS, or Camprince,

Proressors ADDISON E. VERRILL, HORACE L. WELLS,
L. V. PIRSSON anp H. E. GREGORY, or New Haven,

Proressor GEORGE F. BARKER, or PHinapELpHia,
Proressor H™NRY S. WILLIAMS, or Irwaca,
Proressor JOSEPH S. AMES, or Bautimore,
Mr. J. S. DILLER, or Wasuinaton.

FOURTH SERIES

VOL. XXIV—[WHOLE NUMBER, CLXXIV.]

WITH TWO PLATES.

NEW HAVEN, CONNECTICUT.
LOO?

/9309

‘

THE TUTTLE, MOREHOUSE & T
NEW HAVEN

CONTENTS TC’ VOLUME XXIV.

Number 189.

Page
Arr, I.—Cnrrent Theories of Slaty Cleavage; by G. F.
1 BR BaO cop NU ede ca ft 2s) IRS 2) MA a ae ae eI i
II.—Origin and Definition of the Geologic Term Laramie
by A. (CLONE IUNG CG] s Gav Lady AU ANAM RE NCIS 8 2 UR aa eB WEES hae At 18
IIL. ee Method for the Determination of the Hardness of
Miimenals:s: soi, al Zi Iie jn aie Ee A a Beek 23
JAG Migr ceaus Stratigraphy of the Santa Clara Valley
Region in California ; by RopEric CRANDALL -------- 33

V.—Notes on the Habits and External Characters of the

Solenodon of San Domingo (Solenodon paradoxus) ; by

JAN VEN YON op elt) 6) CpG Pan 2 en cr et 55
VI.—Mississippian (lower Carboniferous) Formations in the

Rio Grande Valley, New Mexico; by C. H. Gorpon __ 58
VIU.—Iodometric Determination of Copper; by F. A. Goocu

SOGOU SMe ide Weal ONG Aye CI OS Sens ae a G0 pm UES GD 65
VIUI.—Strength and Elasticity of Spider Thread; by J. R.
IBSEN INI esr LO ARERR Sea ica ENS UN Dh I 75

SCIENTIFIC INTELLIGENCE.

Chemistry and Physics—New Intermediate Product in Thorium, Haun :
Interference of Fluorides with the Precipitation of Alumina, HinricHson,
79.—Maenetic Compounds of Manganese with Boron, Antimony and Phos-
phorus, WADEKIND, 80.—New Variety of Chromium, JasSONNEIX : Memoir
and Scientific Correspondence of the late Sir George Gabriel Stokes, 81.-—
Elektrische Fernphotographie und Ahnliches, A. Korn, 82.

Geology—United States Geological Survey Publications, 82.—Wisconsin
Geological and Natural History Survey, 83.— Geological Survey of Ala-
bama, EH. A. SmirH: Geological Survey of Western Australia, Mt. Margaret
Goldfield, C. G. Gipson, 84.—Geology of the Marysville Mining District,
Montana, J. BARRELL, 8).—Invertebrate Paleontology of the Upper Per-
mian Red Beds of Oklahoma and the Panhandle of Texas, J. W. BEEDE:
Stromatoporoids of the Guelph Formation in Ontario, W. A. Parks, 86.

Miscellaneous Scientific Intelligence.—Bulletin of the Bureau of Standards,
S. W. Srrarron: Carnegie Institution of Washington, 87.—Field Museum
of Natural History, 88.—Twenty-fourth Annual Report of the Bureau of
American Ethnology to the Secretary of the Smithsonian Institution, 1902-
1905, W. H. Houmes, 89.—Bulletin of the Imperial Harthquake Inves-
tigation Committee: Studies in Plant Chemistry and Literary Papers,
H. A. Micuant, 90.—Beitriige zur chemischen Physiologie, fF. HOFMEISTER:
Common Bacterial Infections of the Digestive Tract and the Intoxications
arising from them, C. A. Herter: Handbook of American Indians North
of Mexico, F. W. Hoper: University Text-Book of Botany, D. H. Camp-
BELL, 91.—Birds of the Chicago Area, F. M. WoopRrurrF: Seventh Inter-
national Zoological Congress: Centenary of the Geological Society of
ce Neues Jahrbuch fir Mineralogie, Geologie, und Paliiontologie,

2.

1v CONTENTS.

Number 140.

Page
Art. IX.—Radio-Activity of Thorium Salts; by B. B.
BOLT WOOD i)... 2203) i eee See 93
X.—Wave-lengths and Structural Relation of Certain Bands
in the Spectrum of Nitrogen; by E. E. Lawron._---- 101
XI.—Tertiary Peneplain of the Plateau District, and Adja-
cent Country, in Arizona and New Mexico; by H. H.
IROBINSON 2 2%)52 2S 22 eee ee 109
XII.—Heat of Combustion of Silicon and Silicon Carbide ;
by We G. Mixtmn =<. o2 2222 oo 5h ae Se ee ee

XIU.—Vanadium Sulphide, Patronite, and its Mineral Asso-
ciates from Minasragra, Peru; by W. F. HitLteBranp__ 141

XIV.—Mineralogical Notes; by W. T. ScuaLuEr -_-_--- __-_- 152.

XV.—Thermoelectromotive Forces of Potassium and Sodium
with Platinum and Mercury; by H. C. Barker... -__- 159

X VI.—Reaction between Potassium Aluminium Sulphate and
a Bromide-Bromate Mixture; by F. A. Goocu and
IR? We OSBORNE 2 ee Re a eT

XVIt.—Preparation of Formamide from Ethyl-Formate and
Ammonium Hydroxide; by I. K. Psenps and C. D.

AP) ENGIN Gi es RE ACE ISN «PN it 173
XVIIIL—Lower Middle Cambrian Transition Fauna from
Braintree; Mass:; bya El. W. SHIMER . 27) ee 176

SCIENTIFIC INTELLIGENCE.

Geology and Natural History—Bermuda Islands. Part IV. Geology and
Paleontology, A. E. VERRILL, 179.—Bermuda Islands. Part V, Section I,
A. E. Verritiu: Maryland Geological Survey, Calvert County, G. B.
SHattuck, B. L. Mier and others, 180.—Maryland Geological Survey,
St. Mary’s County, G. B. SHarruck, B. L. Mi~uer and others : Geological
Survey of India: Brief Descriptions of some recently described New Min-
erals, 181.—Chiastolite from Bimbowrie, South Australia, 188.

Obituary—Professor A. HErLprin, 184.

CONTENTS. v

Number 141.
Page

Art. XIX.—Plains in Cape Colony; by E. H. L. Scuwarz._ 185
XX.—Use of Zine Chloride in the Esterification of Succinic

NCIC Dye ley Kevand Me CA Pyaines seu apes) eee te aes 194
XXI.—Volumetric Estimation of Lanthanum as the Oxalate;

lie \AVR aw eNe SI Disa ons) sci 6) a aa oes aise sey 4 le Oe Aree eee LUTE
XXII.—Studies on the Mode of Growth of Material Aggre-

ACCS se NOV Gr ANS eles ORGAN aE IAY rane neta Ale end Sy orts atas 2 nat 199
XXIII.—Catalan Volcanoes and their Rocks; by H. S.

WNEASIEDEN GOINGS Se, Seep ata ooh Te MON me nae uree aly
XXIV.—Anemonella thalictroides (L.) Spach; an anatomical

Ships RO wae ee MLO TMi es tne wet) os Te A eee A eee AS
XXV.—Mineralogical Notes; by C. PatacnE ....---...-. 249

XXVI.—Mercury Minerals from Terlingua, Texas; Kleinite,
Terlinguaite, Eglestonite, Montroydite, Calomel, Mer-
cury ; by W. H. Hiritepranp and W. T. Scuauier ._. 259

XXVII.—Note on the Forms of Arkansas Diamonds ; by
Gr he Mun Z. andl SW ASHINGTON 2. ato es 275

XXVIII.—On a Method for the Observation of Coronas ;
Losyan OAM Sie papemieebantunees cS ils Sune ey Se cS 277

SCIENTIFIC INTELLIGENCE.

Miscellaneous Scientific Intelligence—Las Formaciones Volcdnicas de la Pro-
vineia de Gerona, S. CaLperRon, M. Cazurro, and L. FERNANDEZ-
Navarro: I Vulcani Attivi della Terra, G. Mercauui: Die Mineralien
der Siidnorwegischen Granitpegmatitgiinge, W. C. Broaarr, 282.—Das
Problem der Schwingungserzeugung, H. BarKHAUSEN: Lehigh Univer-
sity, Astronomical Papers, J. H. OGBurn, 283.

Obituary—ANGELO HertLprin: Professor J. M. Sarrorp, 284.

al CONTENTS.

Number 142.
Page

Art. XXIX.—Corpuscular Rays produced in different
Metals by Réntgen Rays ; by C. D. Cooxksry-___.__-_- 285

XX X.—New Species or Sub-species of Hercules Beetles from
Dominica Island, B. W. L., with notes on the habits and
larvee of the common species and other beetles ; by A.

HD VERRILG oo. oa ee ea

XX XI.—Successive Cycles of Coronas ; by C. Barus__--.. 309

XX XII.—Behavior of Molybdic Acid in the Zinc Reductor ;
by D: L. SRANDALI. ooo: 2282 ee ee 313

XX XITI.—Measurement of the Optic Axial Angle of Min-
erals in the Thin Section; by F. E. Wrieur. (With
Plates: Tandell); oi ooo A Soe ee ak 317

XXXIV.—Note on a New Radio-Active Element ; by B. B.
BOLT WOOD 2.22 seas ie) a oe ee Ieee Sie Ge ee 370

SCIENTIFIC INTELLIGENCE.

Chemistry and Physics—Action of Ozone upon Metallic Silver and Mercury,
Mancuor and KampscHULTE: Separation of Fellurium from the Heavy
Metals, BRaUNER and Kuzma, 873.—Rays from Thorium Products, Haun:
Decomposition of Gaseous Hydrocarbons by Ignition with Powdered
‘Aluminium, Kusnerzow, 374.—Studies on the Mode of Growth of Mate-
rial Aggregates; IL (Addendum) Distribution of Variations, A. J. LorKa,
375.—Method for the Observation of Coronas, C. Barus, 376.

Geology—United States Geological Survey, 376.—Carnivora from the Ter-
tiary Formations of the John Day Region, J. C. MERR1IAm, 377.—Lower
Miocene Fauna from South Dakota, W. D. Matrnnw, 379.—Points of the
Skeleton of the Arab Horse, H. F. Osporn, 380.—Hiszeit und. Urgeschichte
der Menschen, H. Ponnia: Physikalische Kristallographie vom Stand-
punkt der Strukturtheorie, E. SOMERFELDT, 581.

Miscellaneous Scientific Intelligence—Carnegie Institution of Washington :
Laboratory Manual of Invertebrate Zodlogy, G. A. DREw, 382.

Obituary—Dr. W. O. ATwaTER, 382.

CONTENTS. Vil

Number 143.

Page
Arr, XXXV.—Electric Arc between Metallic Electrodes ;
DyaaVVer GC Api: ands Hey DY ARNOUD) sere anos see ae 388
XXXVI.—Gibbs’ Geometrical Presentation of the Phenom-
ena of Reflection of Light ; by A. W. Ewen ._----_-- 412
XXXVII.—Decay of Ionized Nuclei in the Fog Chamber,
ime the Wapserot. dime: by. Cy DARUS Ys eae se sae e see 419
XXX VIII.—Crystallographic Combinations of Calcite from
Wrest aterson, Ni Jii;) by EL Es WrirnocKm 222352225. 426
XXXIX.— Preparation of Acetamide by the Action of
Ammonium Hydroxide and Ethyl Acetate; by I. K.
ame Ars PRE EP Sateen ba Ae et cre to RRM es CE 429
XL.—Volumetric Estimation of Potassium as the Cobalti-
MEARE LON AAA als s Dirlurcisa mn yeas ace Meites rei eS ed 433

SCIENTIFIC INTELLIGENCE.

Chemistry and Physics—Atomic Weights of Silver, Nitrogen and Sulphur,
RICHARDS and Forspes: Atomic Weight of Radium, Mdme. Curig, 459.—
Melting-Point of Pure Tungsten, WARTENBERG : Solubilities of Inorganic
and Organic Substances, A. SEIDELL: Practical Chemistry for Army and
Matriculation Candidates, and for Use in Schools, G. Martin: Elements
of Physical Chemistry, H. C. Jones, 440.—Canal Rays, F. PascHEn, etc. :
Propagation of Plane Electromagnetic Waves over Plane Surfaces and
their relation to Wireless Telegraphy, J. ZENNECK, 441.—Infiuence of Mag-
netic Fields on the Resistances of Electrolytes, G. BeErnptT: Change of
Resistance in Metal Wires with Occlusion of Oxygen, G. Szivessy: Atlas
of Absorption Spectra, H. S. Unter and R. W. Woop: Bulletin of the
Bureau of Standards, 8. W. Stratton, 442.

Geology and Natural History—Devonic Fishes of the New York Formations,
C. R. Eastman, 443.—Paleontology of the Niagaran Limestone in the
Chicago Area. The Trilobita, S. WeLuer: Revision der Ostbaltischen
Silurischen Trilobiten, F. Scumipt, 445.—Stratigraphy of the Western
American Trias, J. P. SmrrH: Remarks on and Descriptions of new Fossil
Unionide from the Laramie Clays of Montana, R. P. WuHirTrieLp, 446.—
Palaeontologia Universalis: Die Fossilen Insekten, A. HanpuirscH : Illi-
nois State Geological Survey, H. F. Bain: Connecticut Geological and Nat-
ural History Survey, Bull. No. 8. Bibliography of the Geology of Connec-
ticut, H. E. Grecory, 447.—Tables of Minerals including the Uses of
Minerals and Statistics of the Domestic Production, 8. L. PenrirLp : New
California Minerals, G. D. LouprRBAcK: Elements of Biology: A Practi-
cal Text-Book Correlating Botany, Zoology, and Human Physiology. G. W.
Hunter: Elements of Physiology, T. HoueH and W. T. Srpewick, 448.—
Young of the Crayfishes Astacus and Cambarus, KE. A. ANDREWS: Hyvolu-
tion and Animal Life, D. S. Jorpan and V. L. Ke~tioce : Report on the
Crustacea (Brachyura and Anomura) collected by the North Pacific Explor-
ing Fixpedition, 1853-1856, W. Srrmpson, 449.—Reports on the Scientific
Results of the Expedition to the Eastern Tropical Pacific, in charge of
Alexander Agassiz, by the U. S. Fish Com, Steamer Albatross, from Oct.,
1904 to March, 1905, L. M. Garrett, 450.

Obituary—M. Maurice Lorwy, 450.

Vili CONTENTS.

Number 144.

Page

Art. XLI.—Internal Temperature Gradient of Metals ; by
S. Bo SERVISS = - 02 S00. Ua 451

XLII.—Agegraded Terraces of the Rio Grande; by C. R.
my ns 0 20S) Ses ee a
XLIII.—Waterglass ; Part YI; by J. M. Ornpway-_..---- 473

XLIV.—Action of Dry Ammonia upon Ethyl Oxalate ; by
I. K. Peertes, L. EH: Weep and @€) Ronlousus. eee 480
XLV.—Note on Volcanic Activity; by C. Barus._-...-.. 483

XLVI.—Artificial Hematite Crystals ; by C. EK. Munron_. 485
XLVII.—Anhydrite Twin from Aussee ; by F. Bascom and

V. GOLDSCHMIDT: 272) Gahsoane uy ous ila oe Sle 487
XLVIII.—Occurrence of Olivine in the Serpentine of Chester
and Middlefield, Mass.; by C. PatacuE...---.--.---- 491

SCIENTIFIC INTELLIGENCE.

Chemistry and Physics—Speculations in Regard to Atomic Weight Numbers,
H. Couuins, 496.—Vapor-tension of Sulphur at Low Temperatures, H.
GrevuNER: Helium in Natural Gas, Copy and McFaruanp, 497.—Electro-
Analysis, KE. F. Smita: Fizeau’s Research on the Change of the Azimuth of
Polarization due to Movement of the Earth, 498.—Secondary Cathode Rays
Emitted by Substances Exposed to y-Rays, R. D. KurEMan: Secondary
Roéntgen Radiators from Gases and Vapors, BARKLA, CRowTHER: Abrupt
Limit of Distance in the Power of Positive Kays to Produce Phosphor-
escence, J. Kunz, 499.—Vacuum Bolometer, EK. WarBuRG, G. LEITHAUSER
and EK. JOHANSEN: Ratio of Electrical Units, E. B. Rosa and N. E. Dorsey,
500.

Geology amd Mineralogy—Geology of North Central Wisconsin, S WEID-
MAN, 900.—Research in China; Petrography and Zoology, E. BuackK-
WELDER, 501.—Miscellanea Paleontologica, A. FritscH : Evidences of a
Coblenzian invasion in the Devonic of Eastern America, J. M. CLARKE,
502.—The Geology of Islay, S. B. Witkrnson: Geology and. Water
Resources of the Bighorn Basin, Wyoming, A. FisHER: Geology of
the Guaynopita District, Chihuahua, a contribution to the knowledge of
the structure of the Western Sierra Madre of Mexico, HE. O. Hovey, 508.—
Tertiary Mammal Horizons of North America, F. OsBorn, 504.—Gold
Nuggets from New Guinea, A. LiversIDGE; 505.

Miscellaneous Scientific Intelligence—Annual Report of the Board of Regents
of the Smithsonian Institution, showing the Operations, Expenditures,
and Conditions of the Institution for the year ending June 30, 1906, 506 ;
National Academy of Sciences: American Association for the Advance-
ment of Science, 507: Annual Report of the Board of Scientific Advice
for India for the year 1905-1906: Mendelism, R. C. Punnerr: Les Prix
Nobel in 1904: Memorials of Linneus, 508.—Astronomical Observatory of
Harvard College: New York State Museum: Dew-ponds, Martin, 509.

InpEx to Volume XXIV, 510.

Ld

TUE

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES,

Art. I.— Current Theories of Slaty Cleavage; by Grorce
F. Brcxer.

Tue theory that slaty cleavage is due to pressure normal to
the cleavage is old and very generally esteemed satisfactory.
Sedgwick and others held cleavage to be mainly a phenomenon
due to the crystallization or recrystallization of minerals in an
appropriate orientation, and this idea with modifications has
been advocated of late years by Messrs. Van Hise and Leith.
Mr. Leith’s recent paper* is the most authoritative exposition
of it.

This geologist acknowledges that my theory of slaty cleavaget+
applies in certain cases to which he gives the name of fracture
cleavage. According to him, such cleavage is mainly character-
ized by the presence of actual partings within the mass, but
sometimes shows flow structure as well. He distinguishes flow
cleavage from fracture cleavage, however, attributing to the
former a greatly preponderating importance in nature and
ascribing it to causes distinct from those which produce fracture
cleavage. To this latter he attributes the fissility of those rocks
in which a parallel arrangement of mineral constituents is ab-

* Rock Cleavage, U. S. Geol. Surv., Bull. No. 239, 1905.

Like Tyndall and Daubrée, I consider a parallel arrangement of flattened
grains unessential to cleavage. Rupture takes place (as Messrs. Van Hise
and Leith concede) on planes of maximum slide or maximum tangential
strain. Rupture is a gradual process and cohesion is impaired through flow
before it is destroyed. Impaired cohesion in my theory is cleavage. Cleavage
developes most perfectly when the stress tending to produce it is persistent

in direction, because viscous resistance is then small. In a rotational strain
there are two sets of mathematical planes on which maximum slide takes

Am. Jour. Sci.—Fourts Srrigs, VoL. XXIV, No. 139.—Juny, 1907.
1

2 G. F. Becker—Current Theories of Slaty Cleavage.

sent throughout,* as well as cleavage which is not parallel to
the mica plates of phyllites. Fr acttire cleavage, he says, is
present abundantly in the rocks of the lithosphere. + It is per-
‘haps desirable for me to comment on his paper as succinctly
as possible.

Mr. Leith does not make the theory he supports entirely clear
to me. He states, and repeats in his summary,t that the
parallel arrangement of component minerals in slates showing
flow cleavage is developed by recrystallization “mainly in planes
normal to the greatest pressure.” By greatest pressure, I under-
stand him to mean resultant stress. Other passages, however,
seem to contradict these. In one of them§ he says that the
final position of cleavage ‘may or may not be inclined to the
greater stress depending upon the nature of the strain.” Again
on another pagel he admits that pure or irrotational strains are
of rare occurrence in rock masses and concludes from the nature
of rotational strain that “ the final position of cleavage is usually
inclined to the axis of greatest stress.4| He maintains, however,
that even in such eases cleavage is always tending to develop
normal to the greatest principal stress.

In dealing with the strains accompanying the development of
cleavage, Mr. Leith is more definite and states§ that ‘‘ wherever
the directions of shor tening of a rock mass can be determined
with certainty, any flow cleavage which may be present is normal
to the greatest total shortenmg which the rock has undergone.”

place and both sets are parallel to the axis of rotation. They make with the
greatest axis of the strain ellipsoid angles given by

A being the greatest axis, B the least and C the axis of rotation. The planes
of maximum slide contain the circular sections of the ellipsoid only in a
limiting case. During the progress of strain these mathematical planes sweep
through wedges of the mass, but the two sets of planes sweep at dif-
ferent rates, one set having a relative angular velocity from, say, 20
to an infinite number of times as great as the other. On the pianes which
sweep rapidly, viscosity reinforces rigidity, there is no time for considerable
flow to take place, and, unless actual rupture occurs, so that joints form, the
effect will be small. On the other set of planes viscosity is small, the mass
has time to yield by flow, cohesion is weakened, and cleavage results. Ina
word, the theory is that slaty cleavage is due to solid flow attendant upon
rotational strains. So much of the energy of the system as is not poten-
tialized is dissipated on the plans of maximum slide, and this may or may
not lead to the alteration of mineral constituents, e. g., the transformation
of feldspar into biotite. / It is discussed in Finite Homogeneous Strain, Flow
and Rupture of Rocks, Bull. Geol. Soc. Amer., vol. iv, 18935, p. 18, and in
Experiments on Schistosity and Slaty Cleavage, U. S. Geol. Surv., Bull.
241, 1904.

*Tdem, p. 127. + Idem, p. 184.
t Op. cit., p. 118. SIdem, p. 138.

|| Idem, p. 113. §| Idem, p. 106.

G. F. Becker

Current Theories of Slaty Cleavage. 3

This explanation differs from that proposed in the last para-
graph more than might be supposed. It is analogous to Sharpe’s
theory, but is more “general ; for Sharpe supposed cleavage de-
veloped by an external pressure perpendicular to the induced
cleavage, or to a pure strain, usually accompanied by lateral
constraint; whereas Mr. Leith’s second hypothesis is that,
whether the strain is pure or not, the cleavage is normal to the
least axis of the strain ellipsoid.

I entirely share Mr. Leith’s opinion that pure strains are rare
in nature. Some idea of their rarity may be gained by a little
reflection. The direction of a force with reference to a resist-
ing plane may be regarded as fortuitous. If so, the chance
that the direction will be exactly 90° is infinitesimal, but if a
variation of plus or minus half a degree is tolerated, the chance
will rise to one in 20,626, which is the number of square de-
grees on a hemisphere of unit radius, or 360 radians. On the
other hand, a zone one degree in width on a sphere at a polar
distance of 45° has an area of 255 square degrees, so that the
chance of a force having an inclination of 45° + 30’ to a fixed
plane is 255 times as great as that it should be normal to the
plane. The average “value of all possible inclinations is an
angle of one radian (57° 18’) to the normal. Thus pressures
at less than 45° to the plane are more probable than those at
higher angles and normal pressure is least. probable of all.
Hence a pure strain is a highly improbable limiting case of
rotational strain. Unmodified scission is also a limiting case,
but is 860 times as probable as a pure strain.

It has been assumed in the preceding paragraph that rock
masses undergoing deformation may be regarded as resting
against a fixed support, and this is only partially true. Any
supporting masses must yield by rotation to some extent, though
the amount of such yielding must usually be exceedingly small
as compared with the amount of deformation. When a dis-
location occurs between the Andes and the basin of the Pacific,
the trend of the range is not sensibly changed to accommodate
the rocks adjacent to the fault system. However, so far as the
supporting resistance does rotate, the probability of a pure
strain is increased by the diminution of the rotational strain
component. I shall assume that the probability rises to one in
10,000, though in my opinion this is a gross exaggeration.

Sharpe? s theory i is that cleavage is due to pure strain. The
many geologists who are content “with this theory ought to tell
the rest of us what happens in 9999 cases in which the strain
is not pure.

In the vast exposures of the Archean and early Paleozoic,
millions upon millions of cases of dynamo- metamorphism are
exposed to examination and when such numbers of instances

4 G. F. Becker— Current Theories of Slaty Cleavage.

are to be dealt with the laws of probability become exact.
Again, wherever there are evidences of dynamo-metamorphism,
cleavage appears in the rocks, not always good cleavage, but
still a fissile structure which should be accounted for. With
every possible allowance for yielding of supports, it appears to
me conclusively shown above that the average direction of de-
forming force to the resisting plane cannot have exceeded some-
thing lke 45°, and hence also that schistosity is brought about
as a concomitant of strains in which the rotational element is
large. It follows to my mind that Sharpe’s theory is inade-
quate, for if it were a sufficient explanation, not more than a
ten thousandth part of the strained rocks ought to show cleavage
or foliation.

The hypothesis that crystallization takes place in surfaces
perpendicular to the resultant stress is attractive, but it must
be tested first of all by determining for the simplest cases what
the direction of resultant stress really is. Mr. Leith evidently

GEG

supposes that in the case of pure or irrotational strain in a
homogeneous mass, the resultant stress coincides in direction
with the least axis of the strain ellipsoid. Such a coincidence
will truly exist between the external stress or surface traction
and the axis in question when the strain is pure, but there is
no such agreement between the resultant internal stress at an
arbitrarily selected point and the local orientation of the strain
ellipsoid. In pure strains the resultant stress acting on any
material particle is in the direction of the motion due to this
action. The paths traced out by the particles are called the
lines of dispiacement or the “lines of flow” and the surfaces
which are perpendicular to these lines are the elastic or plastic

G. FF. Becker—Current Theories of Slaty Cleavage. 5

equipotential surfaces. Hence if crystallization takes place on
surfaces perpendicular to the resultant stress, these are the
equipotential surfaces.

The simplest conceivable deformation is (irrotational) shear.
In a shear the lines of flow are rectangular hyperbolas and the
equipotential surfaces are rectangular hyperbolic cylinders.
Little more complex is the case of two shears at right angles to
one another. This corresponds to the axial homogeneous com-
pression of a cube, slab or cylinder of constant volume. In
discussing the mechanics of slate formation the cubical com-
pressibility of the mass is of small importance because after the
limit of elasticity is reached and flow begins, there is no further
change of volume. A cut (fig. 1) showing the lines of flow and
the plastic equipotentials is borrowed from W. J. Ibbetson’s
well-known work on elasticity.* The circular cylinder A, B,
C, D is supposed compressed by uniformly distributed pressure
to the shorter cylinder A’ B’ C’ D’, and during the process A
moves to A’ along the curve connecting the two, B moves to
B’, ete. The equipotential surfaces are hyperboloids of revolu-
tion represented by the equation given by Ibbetson,

2¥Y —e#—vtec’—0

where y lies in the vertical and ¢ is a constant. They are
represented by full lines in the figure.

Add to the cylinder shown in this figure a second inverted
cylinder at the bottom of the first, and suppose the two to
represent only the central portion of a slab. Then the entire
diagram would show the equipotential surfaces on which mica
scales would form if they grew at right angles to the pressure
in a mass subjected to pure strain.

In mere translation, or in rotation, no work is done against
purely elastic or plastic resistances. Hence in a rotational strain
at any given instant the elastic potential is the same as it would
be for a pure strain of equal amplitude. There is an important
difference between the two cases, however, for in pure strain
the system of lines of flow and of equipotentials remains fixed
relatively to the mass, so that the motion of the particles, how-
ever great, is confined to the lines of flow which pass through
them at any instant. On the other hand, in a rotational strain
the lines of flow and equipotentials are not fixed relatively to
the mass, but only relatively to the axes of the strain ellipsoid,
and, like these axes, shift continually with reference to the
material particles of the body undergoing strain. At any in-
stant, however, the equipotentials or surfaces normal to resultant
stress can be definitely assigned.

* Mathematical Theory of Perfectly Elastic Solids, ete. London: Macmillan
and Company, 1887, p. 172.

6 G, EF. Becker—Current Theories of Slaty Cleavage.
Inifigure 2, | have drawn out the equipotentials for a rota-
tional strain which is identical with that illustrated in figures
3 and +. The hyperbolas are the traces on the plane of
the diagram of the hyperboloids of revolution whose equation —
is stated above, when the axes of co-ordinates are the prin-
cipal axes of the strain ellipsoid. The two asymptotes are
the traces of a two-sheeted cone of revolution, so that in
the whole system of surfaces there is not a plane area. It is
upon these surfaces of double curvature that mica should be
deposited at the final moment of strain were it true that this
mineral crystallizes perpendicularly to resultant stress. If such

2

crystallization occurred during the whole progress of stress, the
mica would be found not only on one system of hyperboloids,
but upon innumerable intersecting systems of hyperboloids.

1 am not aware of any lithological phenomena of a character
corresponding to such equipotentials. Nothing more unlike
the structure of a slate belt can be imagined, and I conclude
that the hypothesis under discussion is wholly without founda:
tion.* The fallacy, of course, consists in confusing the forces
acting on the exterior of the es with the resultant of these
and the internal forces. It is this resultant which is actually
exerted on any small group of molecules within the body.

The lines of flow may be considered as representing the ab-
solute motion of particles of the mass. There is also a relative
motion of the elements of mass which is rectilinear and takes
place along the planes of maximum tangential strain. It is to
impairment of cohesion caused by this relative motion that I
suppose cleavage and jointing due.

Although the equipotentials are as far as possible from being
plane surfaces, yet the mica scales in phyllites are arranged in

*Mr. F. E. Wright has published a preliminary note on some experiments
which would seem to offer some support to the supposition that crystalliza-
tion in glass is determined by external forces. In these experiments, how-
ever, insufficient care was taken to ensure uniformity of temperature, and
when they were repeated with more precaution the results were practically «
negative. This Journal, vol. xxii, p. 224, 1906.

G. F. Becker—Current Theories of Slaty Cleavage. 7

planes and must crystallize there in obedience to some law.
Sharpe believed the micas secondary and that they erystallized
most rapidly in the direction of least resistance.* I entirely
agree with him.

“It would appear from Mr. Leith’s discussion that as a matter
of fact the relation sought to be proved is not that mica scales
form normally to local stress, but that they are arranged per-
pendicularly to the shortest axis of the strain ellipsoid. He
offers a variety of evidence that this relation exists in those
rocks which have undergone what he designates as flow cleav-
age, and this evidence is discussed in the fifth chapter of part
one. He takes up first the distortion of pebbles in conglom-
erates, which he alleges are elongated in a direction parallel to
the schistosity of the matrix. A schistose conglomerate is not
a particularly favorable rock for a discussion of this description,
because the schistose lamellae wind in and out between the
pebbles, and it is impossible to assign to them an average direc-
tion with any degree of accuracy.+ My experience, however,
does not coincide with his, so far as observation is concerned.
Where conglomerates have been rendered schistose, and the
pebbles are not so abundant as to touch one another, it is in some
instances possible to break them out with adherent portions of
the matrix. Ihave thus extracted many scores of pebbles from
schistose conglomerates where the conditions appeared favor-
able, and I have found that each pebble came out with an
appendage of schist,a sort of beard, which in almost all cases
stood at a sensible angle to the major axis of the pebble. Never-
theless, I do not foramoment undertake to deny that there are
conglomer ates where there is sensible coincidence between these
directions. The question is what such a concidence would in-
dicate. I do not think that Mr. Leith has put the correct
inter pr etation upon it. In conglomerates, as everyone knows,
there is a strong tendency for the pebbles to arrange themselves
with their shortest axes perpendicular to the plane of bedding,
though there is usually some imbrication or shingling. If the
plane of bedding were parallel to the plane of fixed resistance,
and if also a force were to act on the conglomerate at 90° to
the plane of bedding, then the elongation of such pebbles as
lay quite flat would coincide with the normal to the least axis
of the strain ellipsoid. But each of these conditions must be
of very rare occurrence, and that both should be fulfilled at
once is in the highest degree improbable.

* Geol. Soc. Journ., vol. v, p. 129, 1849.

+On p. 116, Mr. Leith asserts that in a rock undergoing flow ‘‘ the general
effect of rigid particles is to transmit stresses locally i in directions normal to
themselves.” I know of no such theorem in mechanics, and believe the
statement incorrect.

8 G. F. Becker—Current Theories of Slaty Cleavage.

If, on the other hand, either condition fails to be fulfilled,
the position of the strain ellipsoid will not be indicated by the
elongation of the pebbles and the divergence may amount to
any angle. In a general way, this conclusion may be arrived
at In a moment, for the ellipticity of the strain ellipsoid will be
superposed on that of the pebble and the resultant figure will
coincide with neither, either in amount or in attitude.

Specitic examples are perhaps more convincing than general
principles, and I have therefore computed some cases ~ which
are illustrated in figure 3.* In diagram @ are shown a circle
and three ellipses in a square which are to represent the sections
of a sphere and three plastic ellipsoids within a cube. In 6 the
mass 1s supposed to be strained by a force acting at 30° to
the resistance, and this is assumed to be horizontal. “The strain
is the same as that represented in two other diagrams in this

3

SS rr

paper and some further details concerning it will be given
presently. The sphere is of course distorted into the strain
ellipsoid and the major axis of this figure will stand after
strain at an angle of 22° 37’ to the horizontal. The greatest
axes of two of the ellipsoids originally coincided with the direc-
tion of the resistance, but after strain they make with the
horizontal angles 8° 45/ and 6° 5’, showing how the original
eae affects the final orientation. One of these elongated
pebbles makes an angle of about 14° with the strain ellipsoid
and the other about. 163°. No ellipsoid which is originally
parallel to the resistance can have a negative inclination after
strain, but the third ellipsoid illustrated “dips at minus 11° 42’
in the unstrained state and after strain at minus 6° 5’, thus
standing at an angle of nearly 29° to the strain ellipsoid. Had
either its ellipticity or its attitude in the unstrained mass been
different, its final inclination would be more or less than 6° 5’.

This last ellipsoid was computed with a secondary purpose,
for, in its final position, it coincides exactly with the direction
of the cleavage which, according to my theory, would be de-
veloped in the mass by strain. Either of the other pebbles if

*The angles given in the diagram are v = 22° 37’, 0, = 8° 40’, 6. =6°9',
63 = —6°5', 6= — 11° 42’.

G. EF. Becker—Current Theories of Slaty Cleavage. 9

broken out of the metamorphosed mass would bring away a
“beard” of schist standing at an angie to its axis.

These specific instances “merely illustrate the general conclu-
sion that the elongation of pebbles gives no information of any
value as to the position of the strain ellipsoid. The same con-
clusion is immediately applicable to the flattening of angular
mineral fragments.

The evidence afforded by fossils is as a rule no better than
that derived from pebbles. It is easy to imagine an ellipsoid
circumscribed about a trilobite for example, and then infer the
apparent distortion in a given strain. The discordant results
which have been deduced from the measurements of fossils by
various observers are thus easily accounted for.* Nevertheless,
with a sufficient amount of work I believe better results could
be obtained. In a strained fossil there must be two central
sections which are undistorted and possibly these might be
found by accurate measurements in some favorable cases. If
found, they would determine the position and the ellipticity
of the section of the ellipsoid through the greatest and least
axes. To determine with accuracy the undistorted sections of
a fossil would be a delicate job and has not been attempted so
far as I know.

The evidence from volcanic textures, such as the blow holes
in pre-Cambrian lavas, is as poor as that from pebbles, since
blow holes are substantially always ellipsoidal in unmetamor-
phosed lavas.

Mr. Leith also attempts to use the distortion of beds and the
attitude of folds to prove the position of the minor axes of the
strain ellipsoid. In certain cases beds of a composition some-
what different from that of the slate are crenulated in a direc-
tion normal to the cleavage, and Mr. Leith considers this a
proof that the “ greatest shortening ” of the mass is also normal
to the cleavage. In figure 4, which illustrates my own theory
of cleavage, I have drawn a bed before and after distortion,
assuming “that this bed retains its original length unchanged
because of lack of plasticity. The axis of crenulation is exactly

* A part of the difficulty seems to be due to the fact that even flat organ-
isms are by no means always deposited in strict conformity to the stratifica-
tion. This is apparent in recent muds and in unaltered sedimentary rocks.
Among schistose rocks a good example is afforded by the Ordovician slates
of Arvonia, Virginia. They contain Cyclocystoides which were originally
circular flattened discs and have been converted into very regular ellipses
with axes in the average proportion of 3 to 2 or thereabout. I have examined
some specimens of this slate a few square inches in area on which the orienta-
tion of the little ellipses varied by at least 45° and in no regular manner,
doubtless because of original variation in position. Again, where the plane
of cleavage approaches the plane of sedimentation, but does not coincide
with it, fossils may undergo a deceptive distortion. Thus in the Arvonia
slates there are cases where the ellipses representing Cyclostoides are very

regularly oriented, but with their major axes at an angle of some 20° to the
grain of the slate.

10 G. F. Becker—Current Theories of Slaty Cleavage.

normal to the plane of cleavage assigned by my theory and
marked by broken lines, but it is at an angle of 29° to the minor
axis of the strain ellipsoid. The crenulation of a bed merely
shows that the axis of folding lies between the minor axis of
the strain ellipsoid and the direction of unaltered length ; it
does not even tend to prove the actual position of the strain
ellipsoid. On any theory of slaty cleavage it is easy to con-
struct crenulated beds at any of a wide range of inclinations to

4

the cleavage, and therefore no one angle has any valuable
significance.

Mr. Leith, furthermore, advances the view that intrusions of
great masses of igneous rocks are known to compress adjacent
rocks in directions normal to the periphery of the intrusive
mass, and that cleavage is developed in the surrounding rocks
parallel to the periphery y of the intrusive masses. Now, when
batholiths invade a region, they unquestionably produce an
outward pressure which is commonly manifested by fractures
and apophyses in the surrounding rock. The strains set up
must be of an enormously complicated kind and the outlines
of the batholith itself usually show great irregularity. That
anyone should be able adequately to analy ze these strains so as
to determine the pr incipal axes, or even to establish with any
fair degree of approximation the parallelism of the schistosity
to the outlines of the batholith, is to me quite inconceivable.

Mr. Leith’s last argument on this subject is that crystals and
pebbles included in schistose rocks are frequently fractured or
sliced, and that this slicing does not take place paralle! to the
schistosity, but at a considerable angle to it. From this dis-
crepancy, ‘he argues that the cleavage has a different origin
from the slicing, and that while the slicing occurs at an angle
to the direction of greatest negative normal stress, the schistos-
ity is per pendicular to it. Now, if an included pebble or crystal
had precisely the same properties as the surrounding mass, it
would of course yield like the surrounding mass, and would show
the same schistosity and nothing more. On the other hand, if

G. F. Becker— Current Theories of Slaty Cleavage. 11

the included mass were relatively very brittle, it would be cracked
at the inception of strain, and would therefore exhibit a be-
havior of its own. On my theory of cleavage, this behavior
ean be fairly well followed up. Fig. 4 is a diagram supposed
to represent a quadrangular surface of a plastic mass including
a cube of a different character. I suppose this cube to be made
of some substantially isomorphous material, such as glass or
quartzite or some very fine-grained eruptive rock, and that this
cube is also brittle. ‘Then at the inception of str ain, it will be
cracked at angles of 45° to the axes of the initial strain- ellip-
soid. It may crack in two directions, or only in one, and [|
shall suppose that the direction in which it yields is that ‘which,
according to my theory, is characteristic of the master joints
in slates. Now, these lines of fracture will during continued
strain change their inclination, precisely as if they were mere
geometrical lines in the plastic mass. The several slices will
slip over one another and be rearranged. Doubtless at the
edges of these slabs there will be a certain amount of disturb-
ance of the surrounding material, but there appears to be no
reason to suppose that these disturbances will not so balance
one another that the centers of inertia of the several slabs will
behave with simple regularity. If so, these centers of inertia
will also remain on a material line which will be deflected pre-
cisely as if the cube were absent altogether. It is then possible
to compute for certain displacements the position which these
centers of inertia will take, and therefore to exhibit the relative
position of the slices after deformation is complete, and this is
done in the second figure of the diagram.

In constructing this diagram, advantage has been taken of a
little problem solved in my former paper on this subject. I
have there shown that, provided Hooke’s Law holds and Pois-
son’s ratio is assumed at one- -fourth, a force inclined to the
surface of the mass at an angle of 30° ‘will bring about just this
distortion. Now, there are substances for which —Poisson’s
ratio is equal to one-fourth, especially the glasses. Hooke’s
Law is applicable to small strains with perfect accuracy ; for
large strains it affords only a first approximation. The diagram
may therefore be erroneous to some extent, but the only error
which it can contain is in the direction of the applied force, and
this error probably does not exceed one or two degrees at most.
It is impossible to draw such a diagram without assuming some
law between stress and strain.*

It thus appears that my theory of cleavage completely explains
the slicing of a pebble and the inclination in the position of the

*Tf x, y is the position of a point in the unstrained mass, and x’, y’ the
point to which it is brought by strain, then «’=1:0577 x+y; y'=0°7691 y.
It follows that v,=56° 58’, v=22° 37’.

12. G. F. Becker—Current Theories of Slaty Cleavage.

fractures to the resultant cleavage shown in the cut by broken
lines. In fact, if my theory is correct, some further information
might pos ssibly be obtained from such pebbles as to the oper-
ations to which the mass has been subjected. It may be observed
in the diagram that the line of the centers of inertia of the
several fr agments nearly coincides with the direction of cleavage,
but does not do so absolutely. The difference is so slight that
it might be attributed to bad drawing, but this is not the case.
The difference is only two-thirds of one degree, and the sig-
nificance of this difference is this: The lines of the center of
inertia coincide with the direction of the first fibers to undergo
maximum tangential strain, whereas the cleavage represents
the final direction of maximum slide. Now, as has been men-
tioned above, there is another set of planes of maximum
tangential strain which, in this particular case, has wandered
through a wedge of the mass bounded by planes at 28° from
one another ; - thus the particles lying between the direction of
the lines of the center of inertia and the cleavage have been
subjected to maximum tangential strain more than. forty times
as long as the particles in the other dir ection, and it is to this
difference that I attribute the development of the cleavage.
If instead of being absolutely brittle the enclosed cube yielded
plastically to a minute extent before rupture, the fractured mass
would show a trace of cleavability in the direction of the centers
of inertia of the slices. If rupture were to take place simul-
taneously on both sets of planes of maximum slide, double
displacements resembling those shown in my paper on Simul-
taneous Joints,* figure 9, would probably occur. Fig. 5 is
borrowed from Mr.

a Diller’s paper on the

Taylorsville region of
California,t figure 4,
and shows a sliced
quartzite pebble from
aschist. <A sliced rock
pebble is more in-
structive than a
cleaved feldspar be-
cause the cleavage of
the mineral will mod-
ify the angle of sheing.
Mr. Leith presents no further argument from observation
for his contention that “‘ wherever the directions of shortening
and elongation of a rock mass can be determined with certainty,
any flow cleavage which may be present is normal to the great-

* Washington Academy Sci., vol. vii, p. 274, 1905.
+ U.S. Geol. Sury. Bull. on Taylorsville Region, not yet in print.

G. F-. Becker—Current Theories of Slaty Cleavage. 18

est shortening which the rock has undergone.” Certainty
seems to me a strange word to apply to conclusions from such
evidence. Not one of the methods can be depended upon to
determine the position of the strain ellipsoid within many
degrees, excepting the untried expedient of seeking the undis-
torted sections of a fossil.

Mr. Leith discusses at some length the superposition of
fracture cleavage on flow cleavage. He interprets many
observed cases as indicating that after flow cleavage has been
developed, a fracture cleavage has been superimposed upon it,
of course under radically different conditions of stress. He
borrows from a report by Mr. Dale* an illustration of such a
case taken from a thin section, and I shall follow his example

in figure 6. When a double structure is developed in rocks
the hypothesis that each is due to aseparate cause always seems
to me very dubious. If parting, or weakness (cleavage), is
developed in one direction, a subsequent fortuitously oriented
force intense enough to produce rupture at allis almost certain
to lead to movements on the old surfaces, and, if this will not
suffice to relieve strain, something like gr anulation usually
ensues. Hence I am led to believe that the divisions shown in
Mr. Dale’s plate have a common origin and were simultaneously
developed. Mr. Leith concludes that the diagonal fractures in
the cut are developed on my theory. But if if apply my theory
in extenso to this case, the exact orientation of the strain ellip-
soid is at once determined. The major axis should bisect the
acute angle between cleavage and the cracks, and therefore
stand at an angle of 19° to each. The position of the minor
axis follows. Supposing the volume of the rock constant, the
minor axis will be proportional to the tangent of 19° or 0°344
=B. Hence at once the linear compression in the direction of

*U.S. Geol. Surv., 19th Am. Rep., pt. iii, pl. 28, p. 208. From a slide

of a Cambrian roofing slate near West Paulet, Vermont. The black spots in
the lower part of the cut are pyrite.

14.0 G. EF. Becker—Current Theories of Slaty Cleavage.

B has reduced the unit leneth to 0°344 unless the strain was
attended by diminution of volume, in which case the shr inkage
in this direction was somewhat greater. The product of the
other two axes is AC=1/B for the case of incompressibility.
Some further relation must be exactly known to determine the
ratio of A to C. This the cut does not afford, but an approxi-
mation can be obtained without difficulty. The fixed plane in
the case of pure strain is in the direction of the major AXIS,
while in the case of scission on my theory, it coincides in
direction with the cleavage. When pure and rotational strains
are combined the resistance lies between the two extremes
noted, which in this case differ by 19°. Now, as I shall point
out presently, the strain exhibited in the cut bears evidences
of a large amount of rotation and the fixed plane was probably
nearer the cleavage than the major axis. The assumption that
it made an angle of 6 degrees with the cleavage cannot be more
than a degree or two out of the way, and if this were the cor-
rect value, it easily follows that A = 2°36 and C = 1:23.

I have drawn in to the cut, fig 6, the ellipse with these axes
and also shown the two lines representing the traces of the
planes of maximum tangential strain. One of them coincides
accurately with the cleavage while the other has exactly the
same direction as the joints. If this cut is compared with fig.
10 of my former paper, which was issued six years before Mr.
Dale’s plate, it will be seen that the similarity is very great.
According to my theory, cleavage should develop i in the acute
angle between the direction of the applied force and that of the
fixed resistance on one set of planes of maximum slide, while the
other set of such planes will be marked, if at all, ‘by joints.
That appears to be exactly what has happened in the case under
discussion, and it is the great difference between the two struc-
tures which lead me to infer that the element of scission in the
strain is large.

Having an approximation to the values of the axes, it is a
mere matter of detail to compute the component pure strains
and scission. The former are due to the vertical force com-.
ponent, the latter to the horizontal component. To combine
them to a resultant applied force it would be needful to know
or to assume the relations between strain and stress. Hence I
merely indicate a pressure which must lie between the minor
axis and the cleavage, but an uncertain angle.

It would appear that my theory throws far more light on
this occurrence than does Mr. Leith’s, and affords a very satis-
factory explanation of an interesting occurrence which on a
natural scale covers only one square ‘millimeter.

While my theory of cleavage does not necessarily involve
the formation of new minerals, s such as mica, it affords a means

G. F. Becker—Current Theories of Slaty Cleavage. 15

of accounting for the frequent occurrence of mica on at least
one set of plane surfaces wherever the rocks affected by cleavage
have presented great resistance to deformation. During the
progress of a strain in a body which is not ideally brittle,

there is at first a cubical compression affecting the whole mass
alike and becoming constant at the elastic limit. Thereafter
an expenditure of energy takes place and this is confined to the
planes of maximum slide, where it is dissipated through the
viscosity of the mater ial, It is not needful to prove this
statement, which is involved in the very definition of viscosity.
The temperature along these surfaces must rise, and, by increas-
ing the mobility of the molecules , the heat liberated will promote
chemical recombination in so far as compounds of lower poten-
tial are stable at the prevailing temperature and pressure. In
rotational strains the energy expended per unit mass along
those planes of maximum slide which are nearly fixed rela-
tively to the material particles will be many times as great as
along the set of planes which wanders rapidly through the mass,

and the chemical effects which may take place will be in some.
direct proportion to this expenditure. Thus in slates due to
rotational strains it is comprehensible that one and only one
set of planes should be marked by abundant mica scales. In
irrotational strains there may be two or more mica-coated sur-
faces. On the other hand, there is no liberation of heat dur-

ing distortion which characterizes planes perpendicular to the
least axis of the strain-ellipsoid and none distributed along the
equipotential surfaces.

A few pages of the memoir under review are devoted to a
discussion of strain and stress, and these call for some com-
ment lest they should mislead. The author states that when
the form of a body alone changes, the strain is called distor-
tion, which is in entire accord with usage. A. few lines fur-
ther on he adopts from Van Hise a term “ pure shortening”
for “any irrotational strain in which all three principal axes
are changed in length in such a ratio that the volume remains
constant.” As is well known, a “ pure” distortion is synony-
mous with an irrotational distor tion, so that shortening as here
employed is absolutely synonymous ’ with distortion. “Why so
well-chosen a term as distortion should be replaced by one
which does not suggest its definition is not clear. Shortening
is indeed absolutely misleading, for its opposite would be elon-
gation, a term frequently used in the discussion of strains, but
in a very special sense. It means extension in a single direc-
tion without change of dimensions at right angles to this direc-
tion ; so that if the ratio of elongation is 2, the volume of the
mass is doubled. Elongation in “this technical sense is merely
an analytical device. Only such cellular structures as pith,

16 G. EF. Becker —Current Theories of Slaty Cleavage.

cork and sponge have properties which enable them to be

used in approximate illustration of (negative ) elongation. Mr.
Teh not only defines pure shortening, but illustrates it. The
illustration, however, shows a strain in which only two of the
axes have undergone change, and which is neither more nor
less than a (pure) shear. Perhaps, then, shortening is only a
new term for shear.

Mr. Leith says that Thomson and Tait call an irrotational or
pure shear a “simple” shear. This is incorrect. They desig-
nate by “simple” shear the strain I ventured to rename scis-
sion, while they call a pure shear simply a “shear.”* It was the
confusing similarity between the terms for essentially different
strains which led ‘to my innovation. In my nomenclature,
pure is a qualification of shear which is superfluous except.
ing when the reader might possibly forget the definition of
scission.

Farther on, he states that “scission is equivalent to a pure
shortening combined with a rotation of the body as a whole.’
If the verbal definition is accepted, this is wholly untrue.
If the diagram is correct so that shortening = shear, it is
true in a sense, because the two processes may each vive an
equal ellipsoid i in the same orientation ; but the boundary con-
dition of the two ellipsoids would not be the same, so that if
each were elastic their behavior on being set free suddenly
would differ fundamentally. The ellipsoid due to scission would
not only vibrate about the mean spherical figure, but (because
of inertia) spin on its mean axis; while that due to pure defor-
mation followed by rotation would not spin. If the ellipsoid
were plastic, that produced by scission would show a fiber or
cleavage under proper conditions of rupture or etching. Rolled
steel is a case In point; and when blocks of steel are used to
test explosives, the substitution of rolled metal (which has
undergone scission) for forged metal (in which, so far as possible,
scission is avoided) can be detected at a glance.

Mr. Leith states that “stress is the action and reaction
between two adjacent parts of a body.” This definition defines
nothing at all. If stress is action, it cannot be reaction. In
fact, the phrase sounds like a confused echo of the theorem as
to the equilibrium of different stresses embodied in Newton’s
third law. Stress is simply the total force measured per unit
area which is exerted between contiguous bodies or contiguous
parts of abody. It is distinguished from “ bodily” forces such
as gravity, which are measured per unit volume.

The author further states that “ any possible stress may be
regarded as equivalent to three normal stresses whose directions

*See Net. Phil., paragraphs 172 and 682. Love employs the same terms,
which are in accord with the commonly accepted definitions.

G. F. Becker— Current Theories of Slaty Cleavage. 17

are mutually perpendicular.” This is true only for a case of
static equilibrium. In any authoritative text-book on dynamics
will be found Poinsot’s theorem, that the forces acting on any
point are reducible to a resultant and a couple.* If Mr. Leith’s
statement were correct, it would be possible to reduce a couple
toasingle force. Itis, of course, the couple which when resisted
gives rise to the rotational strain, just as it is an unresisted or
partially resisted couple which gives the rotation of a planet
or a rifle ball.

All the arguments which Mr. Leith offers in favor of his
views of slaty cleavage and against my theory have now been
passed in review ; while to his denial of my accuracy in report-
ing the orientation of bubbles in cakes of ceresin and in stating
the directions of cleavage,} I have no reply to make. Messrs.
Van Hise and Leith do not seem to me to have improved upon
Sharpe’s theory, and, so far as I can see, geologists must choose
between the pure strain theory of the able Englishman and my
rotational strain theory. Sharpe’s theory is consistent but
leaves much to be explained, both from a molecular and a
molar standpoint. When the prevalence of rotational strains is
admitted, so that a slate belt becomes tectonically equivalent to
a distributed fault, this theory does not apply and cannot be
adapted.

It appears to me that Mr. Leith is in duty bound to make
public exact reasons for his assertions, to give precise methods
for determining the position of the strain ellipsoid or the equi-
potentials in a slate, to show why there is no cleavage on planes
of maximum slide, and to explain thermodynamically how it
happens that the planes on which the entire energy of defor-
mation is expended are not those on which feldspar is converted
into mica. We are past the stage in which mere opinions or
general impressions should be allowed decisive weight.

The distinction which Mr. Leith draws between flow cleavage
and fracture cleavage is substantially a real one, which [ pointed
out 14 years ago. ‘Flow will tend to take place,” I stated,
“along one set of planes of maximum slide ‘“ because of the
inferior viscous resistance.” The other set of planes of maximum
slide, it was inferred, would be marked by ruptures such as
joints. I did not perhaps sufficiently consider that the scale of
the whole structure might be microscopic, yet the magnified
reproduction of Mr. Dale’s slide precisely resembles a reduced
photograph of the structure which I had in mind.

Washington, D. C., April, 1907.

*Cf. e. g., Thomson and Tait, Nat. Phil., section 559 g.
{Opmcitsnp mle:

Am. Jour. Sci.—FourtTH SERIES, Vou. XXIV, No. 139.—Juty, 1907.

4

18 Definition of the Geologic Term Laramie.

Arr. Il—On the Origin and Definition of the Geologie
Term Laramie* ; by A. C. Vearcu. (An Abstract.)

Iyvestications of the United States Geological Survey
during the summer of 1906, covering the larger part of the
Laramie exposures on the Laramie Plains, examined by the
King and Hayden surveys, have revealed many new and impor-
tant facts bearmg on the Laramie problem.

By detailed areal surveys it was found: (1) That the ligni-
tiferous series which in the Laramie Plains lie between the
Montana below and the Fort Union above, and has a maximum
thickness of about 12,500 feet, is divided about the middle by
an unconformity; (2) that this unconformity is in the same
stratigraphic plane and continuous with the unconformity
which in the vicinity of Carbon and to the southeast separates
all the Laramie beds, studied by the Hayden and King parties,
from the underlying Cretaceous; (8) that the beds above the
unconformity rest, ‘often with great divergence of dip, on all
the underlying nadie down to and including the Dakota; (4)
that the basal conglomerate, locally well developed at “the
horizon of the unconformity, while composed largely of mate-
rial derived from the underlying Cretaceous rocks, notably the
Benton, contains pebbles and bowlders from the pre-Cambrian
crystallines now exposed in the hearts of the surrounding
ranges. This unconformity thus involves the total thickness of
the Cretaceous portion of the beds below the unconformity,
and probably the whole sedimentary series of this region, or
over 20,000 feet of strata.

The Laramie Plains section in brief is as follows:

Generalized section of the rocks of the western part of the
Laramie Plains in Carbon County, Wyoming.

Feet
NorthePark Dertiany oe 5.2 ==. 4,500 4+
Unconformity eee eee etme
Hontilinion’es 23222 Sone See eae 800-2,000
“Upper Laramie? 22 '.- 22 2. 22 een O Od
Unconformity eee eee eee
“Lower Laramie” 4.4002 220. 6 ON OG
( Lewis ee oxtblill seerok ‘the ene § 3,000
Montana -~ Mesa Verde j surveys l 3,200
Pierre shale? pie 1c essere ei a alee ese 3,500 ©

* Published by permission of the Director of the United States Geological
Survey. This subject is discussed in fullin the Journal of Geology, No. 5
(July—August).

+ It is the belief of Dr. T. W. Stanton that the Mesa Verde and part of the
Lewis also belong to the Pierre, as that formation is developed east of the
Rocky Mountains. A local name will therefore be applied to this lowest
division of the Montana in this section.

Veatch— Definition of the Geologie Term Laramie. 19

Feet

INrolbraravees = Slo as amie or 5 me 800

Colorado Bentony eee tech far. Piven yy bt. a rae 1,500

J PD Sei FCO EY 2s en et Sb SA 150

JA [Over RS Na cs Bes ae as SU Deas ge Peay 200

WWISVPOVE SONG) Bee ee a eee 8 15

1 EC0 Ll BACXO Shades Se a ge SI Oy 1,650
Carboniferous sandstone and limestone with

basal conglomeratic quartzite__-.-.--- 1,800

Pre-Cambrian crystallines.

Since returning to Washington, the author has critically

reviewed. the writings of the Hayden and King surveys and
prepared a discussion on the subject, which is now in course of
publication in the Journal of Geology. These investigations
have resulted in the following conclusions:
_ 1. The name Laramie is derived from Laramie Plains in
eastern Wyoming. As commonly used in the early seventies,
this included the plains region extending from the Front Range
to and slightly beyond the North Platte River.

2. The most important locality on the Laramie Plains at
this time was Carbon. It was not only a noted paleontological
locality, but the most important coal mining town on the Union
Pacific Railroad at that time. It was the only locality on the
Laramie Plains where the King Survey critically examined
and distinctly delimited the Laramie beds. The Hayden Sur-
vey recognized Laramie strata at another point on the Laramie
Plains, Rock Creek, but regarded the Carbon locality, including
its southern extension containing the plants labeled from
““ Medicine Bow stage station,” as affording better and more
complete exposures.

3. It was the practice of the Hayden and King surveys to
name formations and groups from localities where the beds
were regarded as typically exposed. While King and Hayden
did not always definitely state that the name was derived from
a certain locality, the source of the name can in all cases be
completely inferred from the context. Thus King used Green
River, Bridger, Uinta, Truckee, and other names “without say-
ing the name was derived from such and such a locality, while
he distinctly states the source of Vermilion Creek, Weber, and
other names. King’s strong feeling in this matter of a ‘type
locality is shown by the fact that he refused to use the prior
name Wasatch and adopted the new name Vermilion Creek
simply because at what he considered Hayden’s type locality
the beds were not completely and typically exposed. The state
of feeling at this time is further shown by the fact that the
name Laramie was proposed and adopted as an exact synonym
of Hayden’s Lignitie as defined by him in Wyoming and

20 Veatch—Definition of the Geologic Term Laramie.

Colorado. If merely a general term without a type locality
had been desired, the term Lignitic, exactly defined for the area
involved, would have served all purposes. The change was
clearly based on recognition of the necessity of having a geo-
graphic type locality. From the above facts it follows irresist-
ibly that the type locality of the Laramie is Carbon, on the
Laramie Plains.

4. A critical consideration of investigations of Hayden and
King parties in this region shows that the actual Laramie expo-
sures studied by them are separated from the Cretaceous by an
unconformity of great magnitude. At Carbon, Hague partic-
ularly and minutely included only the beds above the break.
Both Hayden and Hague regarded these beds as entirely con-
formable with those beneath ; hence the statement by King that
the Laramie beds are those which contormably overlie the Fox
Hills, while correct according to the then existing knowledge,
is not correct at the type locality and therefore without deter-
minative value in this connection. It but illustrates anew the
absolute necessity of a type locality to afford means of finally
and conclusively correcting any inaccurate statement or conclu-
sions of the author or authors of a geologic name. Strictly
considered, the term Laramie, thetefor e, can appropriately be
applied only to the beds above the great unconformity and—
fixing an upper limit in part from our pr esent knowledge—
below the Fort Union.*

5. The attempt to redefine the term Laramie from the
exposures in the Denver region, some 200 miles from the type
locality, is therefore not defensible. It results in the scientitic
anomaly of applying the term Laramie to a series of beds
entirely distinct from those at the type locality on which the
name was based. It completely robs the name of all geograph-
ical significance and gives to it even less meaning or appropri-
ateness than a mere lithologic term such as Lignitic.

6. While strictly speaking the name Laramie can be applied
appropriately only to the upper beds (“ Upper Laramie”) and it
cannot with any propriety be restricted to the lower beds
(“‘ Lower Laramie”), the consideration that it was proposed for

* At Evanston there are several reasons for believing that the base of the
Wasatch of Hayden contains representatives of the Fort Union, Puerco, and
Terrejon. Between the Laramie and the Coryphodon-bearing Wasatch are
some 4000 feet of strata separated from the Coryphodon-bearing beds by an
unconformity. At Black Buttes beds now known to be Fort Union (Knowl-
ton, Bull. Geol. Soc. Am., vol. viii, 1896, p. 145) were referred by King to the
Vermilion Creek. It therefore seems not only logical but in accord with
the original usage to define the upper limit of the Laramie as the Fort Union.
The Washakie beds which Hayden regarded as, in this region, limiting the
Lignitic above (3d Ann. Rept. U.S. Geol. Survey Col. and New Mex., 1869,
p. 90) and which were included by King in his Vermilion Creek, are the

beds from which Knowlton reports distinctive Fort Union plants at Black
Buttes.

Veatch— Definition of the Geologic Term Laramie. 21

the beds between the Wasatch and the marine Montana Creta-
ceous and has been most commonly and extensively used in
this broad sense, has led to the suggestion that the retention of
the name in this original sense will cause the least confusion
and that it therefore might be expedient to define the Laramie
as that series of beds occurring between the marine Montana
Cretaceous and the Fort Union.

In connection with this suggestion of expediency, it should
be pointed out that the continued use of this term in the
“catch all” sense is wholly at variance with the abundance of
strong and wholly logical reasons for the restriction of the
term Laramie to the “Upper Laramie” shown by a careful
consideration of the historical data. If the point of confusion
is regarded as one of great importance, it might be worth while
considering the entire abandonment of the term Laramie.

In either case a new name is required for the beds here re-
ferred to as “ Lower Laramie.” Many considerations suggest
that this name should come from the region of the Laramie
Plains. This would be historically appropriate in many ways,
and would result in placing the type localities of the upper and
lower portions of the beds which have been called Laramie in the
broad sense in the same section. There are reasons for believ-
ing that the enormous development of Lower Laramie beds in
the western , part of the Laramie Plains near the mouth of the
Medicine Bow River, or, as it is more commonly called by the
local people, ‘ The Bow,” where there is relatively very little
evidence of a break between the upper and lower beds, more
completely represents the Laramie deposition than at any other
point. These considerations make the Bow formation or group
avery appropriate designation for these lower beds. On the
other hand, the fact that the region of Golden has been made
classic, in ‘connection with the “Laramie problem,” by the
studies of Cross, Eldridge, Knowlton, and others, ‘raises the
question whether the name Golden formation or group might
not be a more appropriate name.

The discovery of this great unconformity at all points that
have been critically eset over an area 1,000 miles north
and south and 250 miles east and west, the fact that it occurs
on both the east and west sides of the Front Range of the
Rocky Mountains, and its great magnitude, all make it one of
the important mile- posts in the geologic history of western
North America. All these considerations suggest anew the
first conclusion of Cross in the Denver region that this uncen-
formity marks the dividing line between the Cretaceous and
Eocene in this region. On this basis the arrangement of
groups ee ‘above and below the great br eak would be
as follows

22 Veatch—Definition of the Geologic Term Laramie.

Green River

Montana

or Gulf [ Colorado

series

Lower | Knight of Coryphodon beds * ) a7
Tertiary < - Wasatch
Fort Union, Puerco and Terrejont 5
or Eocene
(gall aramie
Unconformity
Upper [ “Lower Laramie”
CHOIABSONS |
|

* The name Knight has been proposed for the upper part of the Wasatch
containing Coryphodon remains. It is taken from Knight Station, a point
near the locality where fossils belonging to this genus were first found in
North America. Further, the typical Upper Wasatch is extremely well
developed around Knight Station. Certain considerations suggest that the
Knight formation may be an exact synonym of King’s Vermilion Creek
formation, but as the writer has not had the opportunity to examine King’s
type section he has proposed the provisional name Knight. pending a study
of the Vermilion Creek section. See Prof. Paper 56, 1907, pp. 87-89, 92-96.

+In the Evanston section, between the beds belonging to the Laramie
(Carbon) group and the Coryphodon-bearing Wasatch, are 4,000 feet of
strata which have the stratigraphic position of the Fort Union and Puerco.
These are here separated from the Coryphodon-bearing portion of the
Wasatch by an unconformity of much less magnitude and importance than
that of the base of the Laramie (Carbon) beds.

Kip—Determination of the Hardness of Minerals. 23

Ill.—A New Method for the Determination of the
Hardness of Minerals; by H. Z. Kre.

1. Tue hardness of minerals is perhaps their most obvious
physical characteristic. It was natural, therefore, that no sooner
had minerals become an object of scientific i inquiry than efforts
were made to determine the relative, and later the actual hard-
ness of various species. Simple as the problem would appear
at first glance, the results for the mineralogist of more than a
century of intermittent investigation, carried on by more than
a score > of difterent investigators, are sur prisingly meager. Every
text-book of mineralogy gives, it is true, a rough statement for
the mean relative hardness of each species deser ibed, as referred
to an empirical scale of ten grades (Mohs’s scale), and for
ordinary determinative purposes this is valuable, though less so
than it might become. We know also ina general way that many
minerals show greater hardness on one crystal face than on
others, though specific information is lacking for the most part
inthe manuals. Differences have also been shown to exist on
one and the same face according to the direction of the test.
The curves obtained by plotting ‘these different values reveal a
direct relation to the symmetry of the crystal and form the
most important result yet attained in the study of hardness,
though comparatively few minerals have been investigated and
the curves established by different investigators are far from
uniform.

The main problem, however, that has engaged attention, the
determination of numerical values for ev ery ‘deer ee of har dness,
has as yet found no satisfactory answer. ue al experimenters
have, it is true, arrived at values for 8, 9 or even all 10 of the
members of Mobis’s scale, but these. results vary so greatly
among themselves that without some method of control or
verification it is impossible to place confidence in any of them.
Thus Iddings’s Rock Minerals, to quote one of the most recent
publications, s, allows us to believe with one authority that the
hardness of gypsum as compared with corundum (assumed to
be 1000) is -04, or with another that it is 1°25, more than 30
times as great, or with a third that it is 12°03, 300 times as
@reatey i

2. Corresponding to the lack of uniformity of result, and
indeed largely responsible for it, we find varying conceptions

as to what hardness really is and of the factors upon which it
depends, no agreement as to what method, theoretically or
practically, would give the most reliable returns, the assumption
of various unpr oved conditions, the confusion of physical with

24. Kip—Determination of the Hardness of Minerals.

chemical terms, and finally, no concurrence as to what force or
forces should be the measure of hardness or how these forces
should be combined.* The problem is indeed so complicated,
primarily by inherent conditions (chiefly the difficulty of
distinguishing hardness from allied physical properties) and
secondarily by the varying methods offered for its solution,
that here and there voices are heard pronouncing the problem
insolvable. Thus Daniell in his Principles of Physics states
that “hardness is a property that cannot be measured” and
Miers in his Mineralogy speaks of hardness as a character not
capable of absolute measurement (ed. 1902, p. 110). While
admitting, and indeed emphasizing the fact that the results
hitherto obtained are anything but convincing, I do not at all
agree with the views just quoted. The investigations that
have been made are not so much failures in themselves (quite
the contrary can be maintained of most of them) as they are
failures when regarded as solutions of one and the same problem.
Viewing the history of these efforts in a general way, one is
struck by the fact that too often the investigators appear as
devotees of a certain method rather than as seekers for a certain
end, employing a given method and moulding it to their pur-
pose. And their effort appears too often as an attempt to refine
a method rather than to establish its serviceability. As in the
history of many mechanical inventions, one observes an advance
from simplicity to complexity but as yet no advance from com-
plexity to refined suey. Thus the metal needles of
Frankenheim, guided by the hand, make way for the weighted
point of Seebeck, under which the mineral is drawn by the hand.
Grailich and Pekarek replace the human hand with a pulley and
weight. Pfaff substitutes seven diamond points for one and
then passes over to the boring method for rapid determination
of a mean value. Jaggar carries the boring method to the
highest degree of delicacy, and one must add of complexity, yet
attained by adding clock-work and the microscope. Static
res re tests have likewise advanced from the simple plan of
* A good bibliography of the somewhat voluminous literature on the sub-
ject of mineral hardness will be found in Professor T. A. Jaggar’s article—
A Microsclerometer, for determining the Hardness of Minerals, this Jour-
nal, Dec. 1897. The article by Franz there referred to (cited incorrectly by
Grailich and Pekarek) is to be found in Pogg. Annalen, Bd. Ixxx, 37-55, 1850.
To Professor Jaggar’s list should now be added: Rosiwal, Quarz als Stand-
ard—Material fiir die Abnitzbarkeit ;, Vienna, Verhandl. Geol. Reichsanst.,
1902 (234-246). J. L. C. Schroeder van der Kolk, Over Hardheid in verband
met Splijtbaarheid, voornamelijk bij Mineralen ; Verhandel. der K. Akad,
Wet., 2 Sect., viii, No. 2, Amsterdam, 1902. Egon Miiller, Uber Hirtebe-
stimmung; Inaug. Diss., Jena 1906. A fairly complete survey of the whole
field may be obtained by consulting the article by Grailich and Pekarek

(Sitzungsber. k. k. Akad., Vienna 1854, xiii) for the earlier period and the
dissertation of Egon Miller for the more important recent contributions.

Kip—Determination of the Hardness of Minerals. — 25

noting the weight required to force a point a given depth into
a surface to the laborious method of Auerbach, who regards
hardness as ‘ the limiting elastic resistance (tenacity) of a body,
in case of contact of one of its plane surfaces with the spherical
surface of another body,” and who would obtain a value for
hardness by multiplying the least value of the (central) pressure
per unit of area necessary to produce permanent set or rupture
at the center of the impressed surface by the cube root of the
radius of the sphere.

3. In taking up the problem once more I have a three-
fold objectin view. 1. Toinvite general acceptance of a single
definition of hardness, based upon the actual constitution “of
minerals rather than upon abstract physical conceptions, which
will serve as a working hypothesis in determining its value.
2. To establish theoretically i in conformity with the detinition
the best method of investigation. 3. To put this method in
practice by means of suitable apparatus and adequate math-
ematical calculation.

It is self-evident that we cannot expect uniformity of result
until we secure uniformity of aim,i.e. an agreement as to
what hardness is. By this is not meant an explanation of the
factors that combine to produce the quality (that is a problem
for pure physics), but merely an agreement as to what force or
forces must be used to overcome hardness. This, it will be
observed, is a simpler task than the measurement of the forces
upon which hardness depends. With these we cannot as yet
deal directly, or more properly singly. Fortunately there is
already a high degree of unanimity along mineralogists as to
what hardness means for them. This conception has been
erystallized in the brief but admirable definition given by Dana:
a Hardness is the resistance offered by a smooth surface to
abrasion,’ and with a slight improvement by the Century
Dictionary, as “the quality of bodies which enables them to
resist abrasion of their surfaces.” To this conception of the
quality, which is not an off-hand generalization but a well-con-
sidered and well-tested definition I hold, not loosely and vaguely
but with all strictness. Other definitions are conceivable and
others have been given, but none defines so accurately what
the mineralogist understands by hardness and none adapts itself
better to the physical constitution of minerals as we conceive
it to be. To a person not familiar with the history of scler-
ometry this insistence upon a single clear-cut definition, in view
of the fact that others are possible, may seem superfluous if not
pedantic. Yet the lack of one or the failure to hold firmly to
one accepted in theory is accountable for much of the confusion
of conception and diversity of aim on the part of those who
have believed themselves to be working on the same problem.

26 = Kip—Determination of the Hardness of Minerals.

Thus Jaggar, while accepting Dana’s definition in theory and
repudiating static pressure tests, reverses this in practice and
actually employs a method which, as Auerbach has pointed
out, is fundamentally only a modification of the method he
condemns. Pfaff likewise passes from abrasion tests, or what
are certainly intended as such, to the boring method , appar ently
without realizing that he is implying thereby a very considerable
eaten of his definition of hardness.

Abrasion being a mechanical process, the question at
once arises what force or forces produce it and how are these
forces to be measured and combined. In fig. 1 let CD rep-
resent a tool producing abrasion upon the surface AB, and let
it represent by its length the least force adequate for this

purpose. The value of CD is evidently ./p@4.@pR. Or,

expressed in words, the force that produces abrasion is resolv-
able into two forces, one perpendicular to and the other in the
1 plane of the surface. Calling the
eee ee ee D former the pressure and the Jatter
the pull, the force producing abra-
sion is equal to the square root of
the square of the pressure plus the
B square of the pull. In practice the
two components are generally ap-
plied as separate forces. It should be noted further that the
lateral component may also be a complex force, though not
necessarily so. Thus in a mineral with striations running at an
angle between 0° and 90° to the direction of CF, this force will
again be resolvable into two forces acting at right angles in the
plane of the surface, and its value will have to be determined
before it can be combined with EC.

Investigators have hitherto assumed that a sufficient measure
for abrasion could be found in pressure alone or in pull alone.
This assumption would be true only when the one force was a
direct function of the other, a condition which may obtain
among amorphous substances but which nowhere else can be
assumed as true, or even probable, without proof. Imagine a
mineral with the molecular structure suggested in fig. Dy oN
tool passing in the direction AB will produce abrasion, let us
say, with a pressure of « grams and a pull of 2% grams. — Pro-
ceeding in the direction BA
it is quite conceivable that a

_R pull of « grams would be

AN \ \ \ \ \ \ \ . \ Ve AMneent but that a pressure
of 2% grams would be re-

quired. In both directions the resistance to abrasion would
be the same, though the factors that combine to produce this
resistance might differ according to the direction of the test,

A

Cham ie

2

Kip—Determination of the Hardness of Minerals. 27

the ieee pull required in the direction BA being offset by the
greater pressure required to overcome the tendency of the
molecules to “shed” the pressure, and conversely in the direc-
tion AB. It is evident at once what untrustworthy values will
be obtained if either component alone be taken as the measure
of hardness. It also becomes clear how futile it is to compare
results obtained from abrasion tests with those obtained from
static pressure tests, since the two forces required to produce
abrasion are in no sense a tunetion of the force required to
produce penetration. It should be remarked in passing that
the force EC in fiy. 1 is not necessarily the same in value as
the corresponding force in a static pressure test on the same
specimen. EO in itself may produce no appreciable effect
upon the surface, molecular dislocation taking place only when
it acts in union with the force CF.

5. A fallacy into which some investigators seem to have
fallen is the substitution of rate for pull, in cases where the
abrading instrument passes over the same portion of the sur-
face repeatedly, i in a single direction or with a rotary movement.
Thus Jaggar speaks of four variables: rate, weight, depth and
dur ation, and states that any one of them may be made a meas-
ure of hardness provided the other three be kept constant.
The value of the lateral component in the abrading force is not
taken into account, although care is taken to specify that it
must operate at a known rate. Thus, it is argued, if with a
given weight and a given rate 50 revolutions of the diamond
point be required to attain a given depth in the case of calcite
and 143 revolutions be required to reach the same depth in
fluorite, then 50 and 148 represent the respective hardness of
these ininerals in relation to the abrading agent. This state-
ment, however, would be true only on the theory that the work
done by the 50 revolutions was exactly 50/148 of the work
done by the 148 revolutions, or in other words, that the resist-
ance overcome by the average revolution on the specimen of
calcite was exactly the same as that accomplished by the average
revolution in the case of fluorite. The mere fact that the rate
was maintained constant in both cases is, of course, no proof
that this assumption is correct. The discrepancy involved
becomes still more apparent when we consider the results
obtained by this method for No. 2 and No. 9 of the Mohs’s
seale, the hardest and least hard of the minerals tested. Doubt-
less scores of the 188,808 revolutions required to reach a depth
of 10m in the case of corundum were entirely or practically
ineffective in producing abrasion, and probably no one of them
represents as great an expenditure of abrasive force as is
represented by the average of the 8.3 revolutions that sufticed

28 Kip—Determination of the Hardness of Minerals.

to reach the same depth in gypsum.* The number of abrading
movements is no true test of hardness unless the effective force
of ee average movement is the same in all cases.

While bringing in one factor which has hitherto been
oe disregarded, I attach, on the other band, much less
importance to another variable which Ptaff, Javgar and others
have been careful to maintain constant or to submit to careful
measurement, namely, depth. It is not too much to say that
instead of keeping depth constant it ought to be left to shift
for itself. Imagine two minerals the dimensions of whose
molecular unit spaces stand in the ratio of 1:3 but which are alike
in respect to intermolecular attraction and molecular form and
arrangement. Evidently the resistance to abrasion will be the
same in both. If, however, we should make depth a constant,
the hardness of the mineral with the lesser molecular yolume
would appear three times as great as that of the other. Pfaff’s
substitution of equal volumes of abraded substance for equal
depth of abrasion is, of course, merely a device for measuring
depth, the length and breadth of the abraded surface being
kept constant. In determining hardness we are concerned with
a molecular phenomenon. Therefore depth and volume, as
used by the auelars quoted (as well as by Rosiwal, Miiller,
Bottone and others), in volving as they do arbitrary spacial units
of measurement, have here as little significance as the size of
crystal faces in crystallographic determinations. A further
objection to the method in which weight and rate are kept
constant and hardness determined by the number of abrading
movements necessary to reach a given depth is the fact that a
point weighted sufficiently to procure molecular dislocation
when drawn over or revolved upon a given mineral will almost
certainly produce mass dislocation when applied to any mineral
softer than the first. Hence by this method the actual differ-
ences between the harder and the softer minerals will be inva-
riably exaggerated. The weight, therefore, in all cases should
be great enough when put in motion to produce molecular
displacement, but no greater.

7. It follows from the above that hardness must be deter-
mined either :

(1.) By observing the least foree (whose components for a
given mineral may be designated as the critical pressure and
the critical pull for that mineral) sufficient to produce LDH
or

(2.) By noting the total force required to wredacs a given
amount of molecular dislocation.

*Tn justice to Professor Jaggar it should be stated that he attaches himself
no great value to the numerical results obtained in the series of experiments

described in his article ; a fact which those who quote him would do well
to mention.

—

Kip-—Determination of the Hardness of Minerals. 29

The two methods are, of course, the same in principle and
one may be used to control and verify the other.

Given any two minerals, to determine the relative amounts
by weight that must be abraded from each to produce equal
molecular dislocation, the following method is suggested :

Multiply the specific gravity of the first over the specific grav-
ity of the second by the density of the second over the density
of the first. The result will be the weight of 7 molecules of
the first over the weight of 2 molecules of the second. Thus
if the specific gravity of fluorite be 3°183 and its density -128,
and the specific gravity of quartz be 2°65 and its density -132,*
then

37188 1382 = -420156
Xx
DO 25 32500
that is, a mass of fluorite that contains as many molecules as a
given mass of quartz will weigh 1°3 of the weight of the
quartz.

Omitting the conception of mass or volume, we can arrive
at the same result more simply by making the weight of the
material abraded from each mineral proportional to the molec-
ular weight of each. It is evident that if the molecular weight
of fluorite be 78 and of quartz 60 the weight of x molecules of
each will stand in the same ratio. If, now, we determine the
total amount of force required to abrade 1 mg. of quartz and
then the total force required for 1°3 mg. of fluorite, we shall
have the respective hardness of the two minerals. These
values should be exactly proportionate to those obtained
by Method 1, provided both are carried out with sufficient
accuracy.

This plan, however, is open to some objection in practice.
It would necessitate constant weighing of the fluorite lest the
amount abraded should exceed 1°3 of the weight of the quartz.
Furthermore in view of the unequal density of the two miner-
als we evidently could not multiply the force employed by the
distance traversed in determining the total force, nor yet by
the time during which the force operated, except on the assump-
tion that the time of passage of the point over the surface was
determined by molecular resistance alone. This assumption is
generally made, but it evidently only approximates the truth.

8. A more feasible plan for amorphous and isometric
minerals, and one which could be adapted to minerals of lesser
symmetry as our knowledge of molecular structure increases,
is the following: It is clear that if we would dislocate the
same number of molecules in two minerals, A and B, of which
B is the denser, we must cause the abrading agent to traverse a
greater distance on A than on B. In general the distance to

* These values according to Schroeder van der Kolk, loc. cit.

380 Aip—Determination of the Hardness of Minerals.

be traversed on the denser mineral, B, is to the distance to be
traversed on the less dense mineral, A, as 4/ den. A is to
4/ den. B.

In practice, then, we should proceed as follows :

Determine the critical pressure, 7, and the critical pull, y,
for a given mineral by direct observation and measurement.
Hardness; then, = Vary. This is Method 1. Establish in

this way the hardness of a series of minerals, a, b, c,d. To
verify the results. Pass the point over the surface of mineral
a, with the critical pressure previously determined, at a con-
stant rate* until a distance p has been traversed. Determine
weight of material abraded. Pass the point over the surface
of mineral 6, with the critical pressure already established for
b, at the same rate as on a, a distance

Weigh the abraded material. This weight should bear the
same relation to the weight of substance abraded from @ as the
molecular weight of 6 bears to the molecular weight of a. Thus
if the molecular weight of a be 75°6 and of 6 50-4 the amount
of matter by weight removed from « by 40 movements or trips
of the abrader should be 1°5 times the weight of that removed
from } in 20 trips, assuming the trips to be all of equal length
and assuming the density of a to be 8 and of } to be 64. If
the substance to be abraded from a should weigh, let us say,
1:6 times that removed from 4, it is evident that the critical
pressure on @ as determined by ] Method 1 has been too great,
or that on 6 too small. These tests should be carried on until
the critical pressure for all the minerals under consideration
has been established. The critical pull can then be easily
determined by direct measurement, and hardness then caleu-
lated as in Method 1.

The fact should not be concealed that in view of our ina-
bility to determine in most cases the weight of the physical
molecule as opposed to the chemical molecule, our chief reliance
at oe will have to be placed on Method 1.

These theoretical considerations make it necessary that
an ee designed to measure hardness should meet the
following requirements : 1. It must produce abrasion, not
merely penetration. 2. It must provide a means of measure-
ment for pull as well as for pressure. 3. It should allow for
regulation of rate. 4. The forces nrodueine abrasion should
“be used solely for that purpose, or if employed otherwise the
amount so used should be easily ascertainable.

*Tt is desirable that a low rate be adopted, yet not so low as to cause
molecular gliding instead of dislocation.

Kip—Determination of the Hardness of Minerals. 31

Statice pressure tests, producing as they do penetration or
fracture but not abrasion, are excluded from the outset.

The rotary method under which are included the Rotations-
sklerometer of Miller and the rotating discs of Jannettaz and
Goldberg as well as the Mesosklerometer of Pfaff and J agear’s
Microskler ometer, is likewise inadmissible, certainly in the form
in which it has been employed hitherto. For when pressure
alone is measured it becomes at once merely a modified form
of static pressure test. And the practical difficulties that
would confront us if we were to attempt to measure the force
of rotation, involving friction at so many points, to calculate
the increment of resistance as depth increased and to maintain
molecular dislocation as opposed to mass dislocation, on the one
hand and polishing of the surface on the other, would be exceed-
ingly great if not insurmountable. Grinding a surface with a
standard sand is open to the objection that no means is offered of
guarding against mass dislocation, the sand becomes at once
adulterated with particules of the abraded substance, there is
no certainly that the sand itself has the same torce of attack in
any two tests and there is no means of determining definitely
when the sand is “dead.”

We are led by a process of elimination to the abrading
method par excellence, at once the simplest and most delicate,
which for want of a better name is known as the scratch
method. This fulfills, or can be made to fulfill, all the require-
ments enumerated at the head of this paragraph, and permits,
furthermore, of distinction between molecular and mass dis-
placement, in so far as we can deal at all with submicroscopic
divisions of matter. In reviewing the devices hitherto em-
ployed we find that none satisfies entirely the requirements that
may justly be made of such an instrument. Several fail to
provide any means of measuring pull as well as pressure, none
of them has been actually used so as to measure both forces
at the same time, and those that might have been so used afford
no means of distineuishing between the forces actually produc-
ing abrasion and those expended in other ways or for other
purposes.

10. To meet the demands imposed by our definition of
hardness and by our conception of the physical structure of
minerals the apparatus described below has been designed.

The diamond point resting on the surface of the mineral is
balanced by a weight hanging directly beneath it and suspended
from four arms running out from the short brass cylinder into
which the diamond point and its holder are screwed. Pres-
sure on the point is regulated by the amount of the weight w
(fig. 3). The mineral m is drawn in the direction ab by means
of a thread passing over a pulley p and ending in a weight z.
A thread attached to one of the above mentioned arms at just

32 Kip—Determination of the Hardness of Minerals.

the level of the diamond point runs in the direction da and joins
a spring balance s, which is suspended between two uprights.
As the mineral is drawn in the direction a) the diamond point
rides with it until the tension in s is so great as to cause the
point to become stationary. A scratch will now be produced
on m provided the weight w (including the weight of the

\D)
o

diamond point and arms) be not less than the critical pressure
for the mineral. The critical pull will be shown at once by
the spring balance s.

It will be observed that the two forces producing abrasion
are used for that purpose alone, with the exception of what
force is absorbed by friction at the point 2. The friction at
this point can be calculated in advance for all weights or ten-
sions hae to occur by the method suggested in fig. 4. Two

weights of # mg. each are attached to the

wa soos ends of a thread 1 running over two pulleys,

one of which is the same as used at the

point 2 in fig. 3 and the other identical in

\ construction. The weight which it is nec-

a fi-g essary to add to either a or 6 in order to

destroy their balance is evidently the meas-

ure of friction in the two pulleys, and one half of this will be
the amount of friction at the point n for a tension of « mg.

All minor details of the apparatus are omitted in order that
its main principles may come the more ciearly into view. The
carriage upon which the mineral rests, the graduated dise by
means of which it can be turned at any angle, the tracks upon
which the carriage runs, screws for levelling the surface of
the mineral, a device for regulating rate and other details can
be easily supplied by the imagination.

Results obtained exper imentally by means of this apparatus
will be published later. Meanwhile the writer would be glad
to have anyone interested in the problem of mineral hardness
avail himself of the contents of the present paper.

Vanderbilt University, Nashville, Tenn.

ewer coe

Crandall—Santa Clara Valley Region in California. 33

Art. 1V.—The Cretaceous Stratigraphy of the Santa Clara
Valley Region in California ; by Roperic CRANDALL.

Introduction.
Loealities and faunas.

Table showing the geographical distribution of the fauna.
The Horsetown horizon.

Distribution in central California.

Absence south of Arroyo del Valle.

Cause of absence in southern California.

Movements during the Cretaceous period.

INTRODUCTION.

Ty this paper are recorded the various collections of fossils,
from the Cretaceous, that have been found in the vicinity of
Santa Clara Valley, and at Mt. Diablo, which is east of this
immediate region. The distribution of the three horizons of
the Cretaceous in this region are discussed, with reference to
their. relations elsewhere.

Localities and Faunas.

In the vicinity of the Santa Clara Valley, there are many
localities where Cretaceous fossils have been found. These
places are given below in geographical order.

1. North Berkeley.

2. Mt. Diablo.

3. Haywards.

4, East of Decoto.

5. Pleasanton region.

6. Jordan’s Ranch, Arroyo del Valle.
7. Crossby Ranch, Arroyo del Valle.
8. Milpitas.

9. Beryessa Canyon.

10, Alum Rock Canyon.

11. Evergreen.

12. Dry Creek, five miles southeast of Evergreen.
13. Whitney Ranch near Gilroy.

14. New Almaden.

15. Pigeon Point.

16. Stevens Creek.

17. Stanford University.

18. Belmont.

The accompanying outline map shows the position of these
localities in the central portion of California.

North Berkeley.—One mile north of Berkeley, in a locality
which comprises less than a square mile, the following Cretace-
ous fossils have been collected.

Ax, Jour. Scl.—FourtTH SERIES, Vou. XXIV, No. 139
°
Oo

JULY, 1907.

34 Crandall—Santa Clara Valley Region in California.

Knoxville horizon :
Aucella Piochi Gabb
Belemnites, sp. indet.
Modiola major Gabb
Lucina colusaensis Stanton
Pecten complexicosta Gabb
Cardinia ?
Myoconcha?
Turbo
Atresius liratus Gabb

Horsetown horizon :
Phylloceras onoénse Stanton

Chico horizon:

Hoplites, sp. indet.
Inoceramus, sp. indet.

Most of the fossils were found in bowlders of limestone lying
upon shale beds, but some of the Awcellae were found in a fine
conglomerate. With the forms listed above are thick, Venus-
like shells which cannot be separated from the matrix for
identification, but are considered by Dr. Merriam to resemble
Paskenta species. It should be noted here that Phylloceras
onoénse is a Horsetown form, and does not belong with the
fossils from the Knoxville.

Fragments of Hoplites and an Jnoceramus, which are the
basis of the identification of the Chico horizon, have been found
in sandstone beds in the hills directly east of the buildings of
the University of California. Little is known of the beds in
between the Chico and the Avwcella-bearing horizons, but it
seems probable that there could be only a very small thickness
of rocks intervening. The whole series of Cretaceous here is
overlain by Tertiary, and in the hills east of Berkeley disap-
pears under the later formations. It reappears in the vicinity
of Haywards, about eighteen miles southeast of Berkeley.

Mt. Diablo.—The Knoxville at Mt. Diablo has furnished

the following characteristic forms :

Aucella Piochi* Gabb
Belemnites
Inoceramus
Gastropodas

The slightly altered Knoxville beds, with a high and variable
angle of dips, rest directly upon the older Franciscan rocks,
and are intruded by dikes of peridotite. The unconformity
between the Cretaceous and the Franciscan is plainly marked.

The Knoxville series at this place is composed of dark shales,
with occasional sandy layers, and small lenticular masses of

* Given as Aucella mosquensis, H. W. Turner, Bull. Geol. Soc. Am., ii, 399.

Crandall—Santa Clara Valley Region in California. 35

limestone. . Shark’s teeth, spines and small silicitied foramin-
ifera are present as well as the fossils given above.

The localities where the Avwcellae were found are Bagley’s
canyon, two miles north of the main peak, and four-fifths of a

ws
RWEZ EAA GikKEmen MAP
of the
SANTA CLARA VALLEY)
oWNt. Diablo RE GION
2 ay
Roderic Gromdoalh
ard Q liveraare : Seat e oO
ts ; OP te OnF Q | OID d ae @ Fossil Localities
Wiles We G cy.
am Aes
San, 05

--“QLos&atos }

Nery Ali

WSALINAS

Del Monte
ou. Marea
MOnterey

mile northeast of EKagle Point in the neighborhood of a large
peridotite dike. ;

The Belemnites were found near the northern end of the
peridotite dike in limestone and also in a ‘coarse sandstone
near by.

36 Crandall—Santa Clara Valley Region in California.

The Gastropods, Awcella and Jnoceramus came from a eal-
careous nodule, one-third of a mile north of the locality where
the Belemnites were found.

There was found, just north of the serpentine, a fragment
of wood of the genus Cupressinoxylon, the ancestor of the
sequoias. Near this locality an Awcella was collected.*

From the Cretaceous above the Awcella Piochi beds of Mt.
Diablo there have been collected the following forms, at a
locality southwest of the mountain :

List or Horsetrown Fossits.

Chione varians Gabb
Cucullaea truncata Gabb
Trigonia aequicostata Gabb
Lytoceras Batesi Trask

List oF Cuico Fosstts.

Anecanthoceras Turneri White
Anchura californica Gab
Baculites chicoensis Trask
Dentalium Cooperi Gabb
Dentalium stramineum Gabb
kriphyla wnbonata Gabb
Mactra tenuissima Gabb
Meretrix nitida Gabb
Nautilus sp.

Pachydiscus sucitéensis Meek
Pecten operculiformis Gabb
Pinna Brewert Gabb
Schluteria diabléensis Anderson
Scobinella Dillert White
Trigonia evansana Meek
Cardium annulatum Gabb

The range of the following species is through Horsetown and
Chico epochs:
Chione varians
Cucullaea truncata
Lriphyla umbonata
Pecten operculiformis
Trigonia evansana
These five forms are known to have been found in the Horse-
town at other places. One of the species, Lytoceras Batesi,
is a characteristic Horsetown form. This is hardly enough
faunal evidence to prove the presence of Horsetown beds at
Mt. Diablo, but argument will be offered later to show the
probability of the presence of this horizon at this place.

*H. W. Turner, Geology of Mt. Diablo, Bull. Geol. Soc. Am., ii, 394.

Crandall—Santa Clara Valley Region mm California. 37

The following are characteristic Chico forms :

Anchura californica
Baculites chicoensis —
Nautilus sp.

Pinna Breweri

The Chico series is composed of dark shales, like those of
the Knoxville, with sandy and calcareous layers. These cannlo
be distinguished from the lower series except by the fossist.
The fossils occur sparingly through the beds, northeast, east,
southeast, and south of Diablo ; but the only place where they
are abundant is at Curry’s Creek. Besides the shales, at Curry’s
there are conglomerates, the pebbles of which are fragments of
metamorphic rocks and quartz porphyry.

In a geologic section, Mr. Turner shows the Chico resting
uncontormably upon the Fr anciscan, but at a very high angle,
and overlain at this place by Eocene beds. In another section
the Chico rests upon the Knoxville, with dips apparently con-
formable; but he says that there is no doubt that a considerable
time elapsed between the close of the Knoxville epoch and the
opening of the Chico epoch. The probable thickness of the
Chico, Mr. Turner gives as about six thousand feet.*

Haywards.—A specimen of Crioceras percostatum Gabb is
given by Anderson} as having been found near Haywards, but
from a locality that is unknown. It is considered by him
as probably representing the Knoxville at this place. In the
first part of his paper he has given this form as a typical
Horsetown form.{ Gabb has classed it as belonging to division
* A,” which does not place it definitely. In connection with
other forms at nearby localities, it will be considered here as a
Horsetown form.

East of Decoto.—In the collection of the University of Cali-
fornia there are two specimens of A wcella Piochi Gabb, marked:
“east of Decoto.” The exact locality is not known, but it prob-
ably is in the southeast corner of the Concord sheet, northwest
of and adjoining the Pleasanton region.

Pleasanton region.—The Or etaceous of the Pleasanton region
covers a large area, which includes about one-sixth of the topo-
graphic sheet of this name. Sunol, Pleasanton, and Walpert
ridges are for the most part composed of Cretaceous beds. From
these beds there has been collected a specimen of Venus varians
Gabb, of the Horsetown, and several Knoxville forms; A wcella
crussicallis Keyserling, ‘Aucella Piochi Gabb, and Ammonites
3 sp. indet. The Tertiary overlies the Cretaceous on the west

*H. W. Turner, Geology of Mt. Diablo, Bull. Geol. Soe. Am., ii, 395.

+ Cretaceous Deposits, Proc. Cal. Acad. Sci., 3d series, ii, No. 1, 45.
{ Loe. cit. p. 42. $ Pal. Cal., i, 77.

38 Crandall—Santa Clara Valley Region in California.

and northwest edge of the Pleasanton Quadrangle. South of
Niles canyon the Cretaceous appears where the Tertiary has
been eroded from the hill tops, east of Niles, and it is also ex-
posed on the west side of Sunol canyon, extending south west-
ward to form the underlying part of the hills rising between
the Calaveras and Santa Clara valleys.

Cretaceous shales are exposed in Niles canyon, showing
numerous folds. The Cretaceous of this region consists of three
series of beds. North of Niles canyon and south of Haywards
pass there are large areas that are covered with massive con-
glomerates, the main constituents of which are bowlders of
quartz porphyry and biotite granite. The conglomerates are
probably the same that Mr. Turner described from Mt. Diablo.

South of these conglomerates is a thick series of hard, thinly
bedded black shales with occasional sandy layers. From these
shales on the north bank of Stony Brook Creek, about three
miles north of Farwell station in Niles canyon, was collected
an indeterminable Ammonite. Another Ammonite came from
these same black shales on the west bank of Sunol canyon, at
the point where the Mission Peak road enters the canyon from
the west.

The third locality where fossils were collected is on the south
bank of Niles canyon about one and a quarter miles northeast
of Niles. and about due south of the station of Meriendo. A
small fossiliferous concretion of hard flinty limestone was found
here, but it’ was possible to get only one good specimen. This
specimen was identified as Venws varians. Several hundred
yards east, up the canyon, Awcella Piochi was found in black
shale.

Aucella crassicollis was found on the north side of Mission
Creek, along the Mission Peak road, about three miles slightly
northeast of the town of Ir vington. With it were fragments
of Venus-like shells. At this place the Tertiary rests uncon-
formably upon the Cretaceous.

The beds in Niles canyon, in which Venus varzans was found,
are several hundred feet higher in the section than those from
which A weella crassicollis were Spread: In the general level-
ling of the Cretaceous sediments, before the deposition of
Tertiary, uneven erosion must have left caps of Horsetown
upon the Knoxville. The beds from which Venus vurians
was taken are conformable with the Knoxville beds in which
Aucella Piochi are found. No specimens of A. crassicollis
were found in the Niles canyon section, but there are several
hundred feet of sandstone between the two horizons determined
by the fossils found. These intervening beds may represent
the horizon of Awcella crassicollis.

Jordaws Ranch, Arroyo del Valle. — From Jordan’s ranch
in Arroyo del Valle, eight miles southeast of Livermore, the

Crandall—Santa Clara Valley Region in California. 39

Cretaceous fauna given here has been collected by Dr. J. P.
Smith and Dr. L. G. Yates.

Horsretown Fossits.

Lytoceras alamedense Smith
Phylloceras onéense Stanton
Lytoceras Batesi Trask

Hoplites Remondi Gabb
Lytoceras ef. timotheanum Mayor
Belemnites sp.

Cuico Fossits

Desmoceras Hoffmanni. Gabb
Hlaploceras Breweri Gabb

Baculites chicoensis Trask

Baculites occidentalis Meek
Placenticeras californicum Anderson
Placenticeras pacificum Smith
Desmoceras cf. selwynianum W hiteaves
Flolcostephanus sucidensis Meek
Lytoceras cf. cala Stoliczka

Phylloceras ramosum Meek

Cinulia obliqua Gabb

Inoceramus ct. vancouverensis Shumard
Trigonia evansana Meek

Pectunculus Veatchi Gabb
Pachydiscus sucidensis Meek
Pachydiscus newberryanus Meek

Of this fauna the following forms are characteristic of the
Chico:
Desmoceras Hoffmanni Gabb
Baculites chicoensis Trask
Placenticeras californicum Anderson
Placenticeras pacificum Smith
Cinulia obliqua Gabb
Pachydiscus newberryanus Meek
Trigonia evansana Meek

The Horsetown is represented by the characteristic forms
below :

Hoplites Remondi Gabb
Phylloceras onéense Stanton
Lytoceras timotheanum Mayor
Lytoceras Batesi Trask

In the general list there is a strong mingling of Chico and
Hor setownl forms, but the total fauna dows more affinity with
the Horsetown ‘Avert with the Chico.

40 Crandall—Santa Clara Valley Region in California.

Crossby Teanch, Arroyo del Valte.—Interest in the Arroyo
del Valle beds led to a measurement of the Cretaceous section
at that place. The Jordan ranch beds were taken as a standard
horizon, and the thickness of the series determined on both
Be of these beds. Southwest of the Jordan ranch beds, there

e black shales and hard sandstones. On the ridge due west
of. Jordan ranch, there is a massive conglomerate, the pebbles
of which consist of granite, quartz porphyry, and similar rocks.
This conglomerate in places is very hard and quartzitic. The
shales, sandstones, and conglomerates are very similar, lithologi-
cally, to the Knoxville beds in the adjoining ‘Pleasanton region.
Search for fossils brought no results, but at a ranch in the
canyon the writer was ‘shown a collection of shells from the
surrounding hills. One of these rock specimens contained gas-
tropods and Venus-like shells which were in rock that resembled
the yellow sandstone at the Jordan ranch. Another piece of
hard black shale contained Awcella Piochi Gabb. The exact
location is unknown, but it came from beds already classified,
in field work with the Knoxville, because of its distinctive
tilhologic character and its dissimilari ity to the overlying Upper
Cretaceous. In no place was an actual contact found between
the Upper Cretaceous and the Knoxville, but wherever the
line was crossed the dips and structure indicated conformable
deposition. The Knoxville beds rest directly upon the older
Franciscan rocks, containing schists, jaspers, serpentines and
igneous intrusives. In no place was any actual contact observed
between the Knoxville and the Franciscan. Here, as in other
places, the change from massive Cretaceous beds with good dips
everywhere, to crushed sandstones, schists, ser pentines, and in-
truded igneous rocks, is marked.

A number of sections were run across the Knoxville and
Upper Cretaceous beds. The conditions for determining sec-
tions are not good, because the Knoxville is covered in most of
the area by Upper Cretaceous and the Upper Cretaceous is in
turn overlain by Pliocene or Pleistocene gravels. The thick-
ness of the Knoxville was found to be about four thousand feet,
the line between this horizon and the Upper Cretaceous being
determined arbitrarily by lithological differences in sandstones
and shales.

With the Jordan ranch beds as a fixed horizon, the thickness
of Upper Cretaceous was found to be about four thousand feet.
The bed from which the fauna was collected by Dr. L. G.
Yates and Dr. J. P- Smith is in the middle of this section.
The fauna given here shows more attinity with the Horsetown
than with the Chico, and still there are two thousand feet of
sandstone below these beds and above the Knoxville.

As far as evidence of structure goes, the beds assigned to the

Crandall—Santa Clara Valley Region in California. 41

Horsetown are entirely conformable in dip with those of the
Knoxville, indicating continuous deposition as proven by Diller
and Stanton through the Shasta-Chico series in northern Cali-
fornia. The upper part of this four thousand feet may be con-
sidered Chico, on the evidence of the Chico fauna. Northwest
of Jordan’s ranch, down Arroyo del Valle, the Chico is over-
lain by massive, light yellow, cavernous weathering sandstones,
conformable in dip. No fossils were found, but the beds re-
semble the Eocene as found in other places in middle California,
and are similar to the series described by Mr. Turner as over-
lying the Chico at Mt. Diablo. At the latter place they have
been identified by the presence of Twurritella uvasana and other
characteristic Eocene fossils.

Milpitas.—The Milpitas locality is southwest of and is prac-
tically a continuation of the Pleasanton region. The fauna is
typically Knoxville in character.

Aucella Piochi Gabb
Hoplites (fragment)
Belemnites (fragment)
Pecten complexicosta Gabb

The exact locality from which these fossils come is on the west
side of the range separating Santa Clara and Calaveras valleys,
and is about 4 miles northeast of Milpitas. Aucellae were
found in several other localities southeast of this place along
the foot of the Monument Peak ridge. The locality, at which
fossils are most plentiful, is on the road from Milpitas to Cala-
veras Valley, about a quarter of a mile east up the grade from
the valley floor. All the fossils in this locality come from a
hard compact sandstone that weathers with a brown concentric
stain, which makes a lithologic character that is very constant.
Above the Cretaceous there is a blue shale, probably middle or
lower Miocene, which is apparently conformable in dip, but
not so in reality. The unconformity is shown by the lack of
Eocene sandstone which is present in Arroyo del Valle, Mt.
Diablo section, several miles east of this place, and by a varia-
tion in the level to which the Cretaceous floor was reduced by
erosion, previous to the depositions of Miocene. In the Milpitas
locality, the Miocene overlies beds containing Awcella Piochi
of the lower Knoxville, and in the Pleasanton region it overlies
beds containing Awcella crassicollis of the Upper Knoxville in
one place, and Venus varians of the Horsetown in another.

Beryessa Canyon.—\n Beryessa canyon the fossils given here
have been found in a coarse conglomerate that is composed of
hard flinty shale nodules.

Aucella Piochi Gabb
Phylloceras knoxvillense Stanton

42 Crandall—Santa Clara Valley Region in California.

The Awcellae here are found both in the pebbles and in the
matrix of this conglomerate. This indicates an elevation of
the land mass in this vicinity during Knoxville times, with rapid
erosion following the uplift. This may mean no more than a
local unconformity, in this place, and is representative of near
shore conditions. There is a considerable thickness of barren
shales, sandstones, and conglomerates interbedded, underlying
this fossiliferous conglomerate, which must belong to the Knox.
ville. Serpentine is intruded into the Knoxville sandstones at
this place. As was found further north near Milpitas, the
Tertiary overlies the Cretaceous unconformably. In Alum
Rock canyon, several miles south of this place, there is a similar
conglomerate, apparently continuous with the Beryessa con-
elomerate, but in the pebbles of which no Awcellae have as
yet been found.

Alum Rock Canyon.—The Cretaceous conglomerate of Alum
Rock, exposed near the mouth of the canyon, has furnished no
fossils, but a fine-grained conglomerate further up the canyon
has yielded the following Knoxville forms:

Belemnites, sp.
Aucella Piochi Gabb *

There are also beds of a heavy massive sandstone which have
yielded no fossils.

The sedimentary beds continue southeast along the base of
the Monument Peak range for several miles and then dis-
appear.

Lvergreen.—The main mass of the Cretaceous sediments in
the ridge between Hall’s valley and the Santa Clara Valley are
massive conglomerates with large bowlders of quartz porpyhry,
and other siliceous igneous rocks. Above this conglomerate
there are a few hundred feet of Cretaceous sandstone upon
which the Tertiary lies unconformably. Below the conglom-
erate is a hard black shale from which Aucella Péochi Gabb
was obtained. The exact locality is a small hill that juts out
from the Monument Peak ridge, about one and a half miles
southeast of the town of Evergreen.

Dry Creek.—The Dry Creek locality is five miles southeast
of Evergreen P.O. The Cretaceous at this place is represented
by black sandy shales which are badly crushed near the ser-
pentine of the Silver hills. These fossils were found near the
Dry Creek road that goes from Evergreen to San Felipe Valley.

Aucella Piochi Gabb
Aucella crassicollis Keyserling

* This was termed Aucella mosquensis by Dr. J. P. Smith in his paper on
the Age of the Auriferous Slates of the Sierra Nevada, Bull. Geol. Soc. Am.,
v, 256. Aucella trigonoides was also mentioned from Stevens creek canyon,

west of San José. Both of these species of Aucella, Dr, Smith has sinee rec-
ognized as being the common Aucella Piochi of the Pacific Coast.

Crandall—Santa Clara Valley Region in California. 43

Near Silver Creek, in the southwest corner of the Mt. Hamilton
sheet, the Franciscan rocks are present, but whether the Cre-
taceous lies upon them or is faulted against them is not known,
as intruded serpentine has obscured the relations.
Gilroy.—Gilroy is the next known locality of Cretaceous in

the Santa Clara Valley region and is the furthest south of those
under consideration. In a road cut on the Whitney ranch,
about four miles west of Gilroy, Dr. J. P. Smith found the
following fossils :

Aucella crassicollis Keyserling

Olcostephanus cf. mutabilis Stanton

These were in a hard black shale similar to the Knoxville shale
elsewhere. Near where the fossils were found, the serpentine
has been intruded into the shales but there has been no ap-
preciable metamorphism.

New Almaden.—Gabb mentions Aucella Piochi from near
the New Almaden mine southwest of San José, but does not
give the exact locality.* Dr. Becker + states that the greater
part of the surface at this place is occupied by the metamorphic
rocks, pseudodiorites, pseudodiabases, phthanites, limestones and
serpentines. The age of these he gives as Cretaceous because
of the Awcella found by Mr. Gabb. It is known now that
some of the rocks classified by Becker as Cretaceous are pre-
Cretaceous. We may assume, then, that there is an area of
Cretaceous or Knoxville age near New Almaden, resting upon
the metamorphic rocks of the Franciscan series, which have not
been properly differentiated from them.

Pigeon Point.—Cretaceous beds exposed along the coast,
in a section from 4 to 24 miles wide, and 12 miles long,
starting from near the meer of Pescadero Creek and extend-
ing southward to Aflo Nuevo Bay, have furnished the following
Chico fossils:

One and a half miles southeast of Pigeon Point on coast.

Turritella chicoénsis Gabb
Trigonia evansana Meek
Panopuea concentrica Gabb
Arca breweriana Gabb
Tellina, sp. indet.

Nucula truncata Gabb
Cuculluea bowersiana Cooper
Ostrea, sp.

One mile north of Pigeon Point on coast.

Trigonia leana Gabb
Mactra, sp. indet.
Glycymeris Veatchi Gabb

* Pal. Cal., ii, 247. + Mon. xiii, U. S. Geological Survey, p. 310.

44. Crandall—Santa Clara Valley Region in California.

Pholadomya subelongata Meek
Pinna Brewert Gabb
Inoceramus subundatus (?)
Cinulia obliqua Gabb
Perissolax brevirostris Gabb
Lunatia, sp.
Two and three-quarter miles north of Pigeon Point on coast,
in conglomerate.
LTurritella chicoénsis Gabb

Half a mile south of mouth of Pescadero Creek.
Mactra, n. sp. (?)

Fragment of coral, unidentifiable.

Bolsa Point, one mile north of Pigeon Point.

Glycymeris Veatchi Gabb
Mactra-like shell

The beds are of hard shales, sandstone and massive conglom-
erates. The conglomerates forming the top of the series are
much disturbed ; the shales are apparently the bottom.

Sand dunes and gravels overlie these Upper Cretaceous beds
along the coast, exposing the Chico only im the cliffs in places.
The relations of Cretaceous to Ter tiary are not pa shown
here, but the two series of beds are supposed to be uncon-
formable.

Stevens Creek.—The Cretaceous beds at Stevens Creek con-
sist of interbedded conglomerates and hard black shales, which
have furnished these Knoxville fossils.

Aucella Piochi Gabb
Belemnites

The fossils were found both in the shale and conglomerate.
The shale has broken off in large blocks from the steep walls
of the canyon and in these the Awcellae and Belemnites were
obtained as well as from rock in place near by. The shale
fragments are large and angular, so they must be from the
walls of the canyon. Besides the fossils given above an un-
identifiable bivalve was found.

Stanford University.—Two specimens of Baculites chicoén-
sis Trask have been found near Stanford University. One
came from a hard yellow sandstone, apparently in place, at the
north end of a quarry about a quarter of a mile south of the
University buildings. The other specimen was found inside
of a block of sandstone that was brought up by an intrusion of
basalt. The basalt cuts rocks of Miocene age, and this specimen
of Baculites may be considered as proof of the presence of
Chico underlying the Tertiary deposits in this neighborhood.

Crandall—Santa Clara Valley Region in California. 45

Belmont Hill—From Belmont Hill, west of the town of
that name, these four Knoxville forms have been collected.

Aucella Piochi Gabb
Gastropods, sp. indet.

Aucella crassicollis Keyserling
Hoplhites fragment

~ There are two different places where the fossils were found

in this general locality. A little west of Belmont Hill, the
Hoplites, gastropods and a fragment of an imprint that might
be either an Awcella or an Inoceramus were found | by Dr.
J. C. Branneyr.

About one-half mile southwest from this place, in the main
ereek bed, Dr. Branner and Mr. R. Anderson found a bowlder
that contained Awcella Piochit. The material of the bowlder
seemed the same as the rock exposed in the creek bed, although
no fossils have yet been found in the rock in place.

The bowlder containing the Awcellae is a fine-grained con-
glomerate, made up of small pebbles of jasper.

On Belmont Hill these fine ‘grained conglomerates rest upon
the jaspers of the Franciscan series. These pebbles of jasper in
the conglomerate here are proof of the unconfor mity between
the Franciscan rocks and the Knoxville.

A table is given below which shows the distribution of spe-
cies from the various Cretaceous horizons of this region.

The Horsetown Horizon.

In the first work done upon the Cretaceous of California,
two divisions of the Cretaceous were recognized—the Shasta and
the Chico groups. The Shasta group was subdivided into two
horizons by Dr. White and the upper part was called Horse-
town, and was considered to have a distinctive fauna.*

The tendency of other geologists has been to class this. hori-
zon with either the Upper or Lower Cretaceous rather + than
consider it as independent.

Mr. Anderson is of the opinion that the Horsetown is a
separate horizon. ¢

Diller and Stanten have shown that there was continuous
sedimentation throughout the Cretaceous in rorthern Cali-
fornia; that the fauna of Knoxville, Horsetown, and Chico
intergrade, but still the three horizons are considered to have
sufficiently distinctive faunas to be separated.§

*Correlation paper, Bull. 82, U. S. Geol. Survey, p. 184.

+G. F. Becker, Early Cretaceous of Calif., Bull. Geol. Soc. Am., ii, a
J. 5. Diller, Geology of Calif. and Oregon, Bull. Geol. Soe. Am., iv ie

?

{ Cretaceous of Pacific Coast, Proc. Cal. Acad. Sci., 3d series, ii, 1, p. 47.
$ The Shasta-Chico series, Bull. Geol. Soc. Am., v, 464.

x

x

x X

x Soy ypeSeo etooobooesccS Qqex) sunipna aU0LY)
Ne SSS |e ichacets aor sities UOJURIG asuaguo sy.ia00p hy

VNOVY NMOLASHOF,

se “== === YOJURIG asuappracouy spLa00y hy
ye Sie Ns Shee al coca dacs Saomeace spo g esas osc saqudoyy
[eG DUG SENOUIONNION |
eras - Sulploshoy s27700¢ssnuo DION TF
tonne eee ee e- spodo.ajsp 9
SUE SESS Sis eee ee ae eee S70 LLU LOOOLU IE

XxX

xX X

gn (eee ee Ogun,
52 loconer 2 = (1) Hy pUon Oli:

2 [bce Seob dogs cone de soceses (2) nenping
Tor TTT TT QQRX) DIs0IIWaIdLUOD UAI0%T

MK [eore error UOJUVIG sesuansnjoo DUuLdWT

An ee -- qqexy soln njorpoyr
Moree cre crete cree ete SAPUMaOT

x
x

lax
x
x

xx
x

|
|

|
|

Lt |

GL

Korey &T

(mWeeTSIBAW “S
‘ut @) YoorQ AIq

quowpegd gt | Xx

uepeupy MeN FL! X

Yoo SUPAG}IS OT |
yulog Woasig cL.

Crandall—Santa Clara Valley Region in California.

“ATU PLOFUBIS

]

|

|
|
|
|
|
|

Sy apres s See Fes Vqey 2Y901g Dany

G

VNOVY WITIAXONY,

weer310 Aq IL|x
uodury Yoo unTy OL x xX
essektog 6
seyidtiq § XX
youry Aqsso1g p,
aT[e@A oAOILW
youvy uvpioe 9
(0Jooe(qf FO Sey F
spriemisvy &
oTqeid “FN
Aapeyteg T

ae A OAOITW
UWOISaT WOJUBSBITG CG

46

“VNOVY AHL FO NOILNATALSIG: TVOIHdVAD0KL) AHL -ONIMOHS alavV YT,

ta. AG

Californ

won im

Crandall—Santa Clara Valley Reg

8I

XXX

xxx x

VSG PR IRS ROR OK ON PR ROR ORK OK OK

Ria ie eee YO VUDSUDAD DIUOHLT
UN A OL OCHO)
Sie arias UOS.LOPUY sesuao7gnip DIayINJYyIY
eee nen ghia Vx) LANnaLg DUUIT
Dot qe srmuofynoiado uaa
Re eee Yoo seswanzons snosiphyong
Bee ee ee Sr SE OS Oi
Rae ee ee qqen nprpew wr0jpa0ayy
She ae eee qqey vulassinuay DLO IT
Be ee qqes vypu0qun nihydruy
eer iee re qQqey winawun.s 5

gee eee qqexy 2adoo) wnaynquag
TT tras "= GQVKH DIDIUNL DEDIINING
Rise ip Ea ee QQVyy supa awory,d
"oro ">= SgBIT, sesua0n2y— sajunoODg
sitet Coneeriaes qqeg vorwiofiyns DEnyoup
5) SNRUDLODOUT
eres ate Sine eee tay quouseay saqadozy

VNAV OOIHO

he gant Sen yr ee cn eat ‘ds a2uwajag
Ser ee LoALT, Wenuvayj0wue spLa00IhT
Neen eee Re qqey wpuowary sazydozyT
~o ooo > YYUIG aswapaunpyy spsa00jh'T
Hoa QR) UlNID98OI.1AG $N.L00090L)
Smee es “=> qqexy vqngsooinban nrU0brt 7
SS SSS A), CEINGL SIOLOONOTE
zat rrtt = Gage nypu0gun mphiydiuug
Save Baise ce al QQvyy DIDIUNL) DADINING

61

a

IT OF

oO

(qU00) VNOV JT NMOLASUOP

|
|
|

a .

Torna.

f

Cal

gion im

e

XX KKKKKKK KKK KK

y R

é

ita Clara Vall

ay,
a7

S

Crandall
SOCK XE KE XI

48

Be aes Gee OSS S03) LM DOUNTT
Rec ee Penn OBL) S2U9SO.M20ALQ LDIOSSLMAT
Geant Ee as (2) srgppungns snUinla00Uuy
soo oop pypbuojagns DA WOpyp OY
ae eae qqey eyonag satawliohi 4
Rees eis es ata oe ‘yoput “ds nun yr
Dies ee OOK 9) OE) (OOWIOINE MN,
ee ee ae REE OS EAS)
Be aa ae LIMOOD, DUNISMANOG DEDINING
pS = GOR) DIDIURUT DINONINE
eee eee ey ro NS OULU ome,
re ees Sa “== QV DUDLLANAG DIL
Sees Seer kas qqey voiyuaou0s podoung
Sete cee QQBx) S28229001Y9 DIAIMLN TL,
x ">TO AA CHOWN SOMUOUIULT
------ > yoo, sruplusaqnan snos~phyon gd

scrips Sees qqey 2yaqvad snpnounpoogy
~ PAVTUNYS S2S71AMIANOIUDA “JO SNUDMIIOUT
Sips nee re TIPE qqeg vnbiygo pynuy
poe eer “- yooyy wnsownes spsav0phy.
Sr oie ator Yoo sesugnions snunydajsooojy
Seren oe YU wnayiond sy.aoVquaovy
Ae ee LYZOIOIG BjV9 “Jo sy.lavoghT
- SOABOILY AA WeeeupLUhayas “Jo sP.dao0wsaqe
~*~ UOSMOPUY WMIVUMOLI]DI SDAIDYUIOD

See Geta ress SUDIUIPLOIO SOPYNID
SS oer ey cater qqry eamatg sp.a00jdnyy
Tagsg eae ies qqeg wuupuffoyy spdav0wsaq
w= a == Qe DyngsooInbay DIUOHILT
= GRY WINQDINUUD UNEP)

x xX

6 G I ((u00) VWNOAV OOIHD

Crandall—Santa Clara Valley Region in California. 49

The type section of Horsetown, near the place of that name,
is considered by Mr. Diller to represent only the upper part of
the horizon.*

Dr. L. F. Ward has lately identified plant remains from
Horsetown, which he has pronounced Lower Cretaceous in age.t

It is evident that the Horsetown is an independent horizon.

Distribution of the Horsetown Horizon.—The Horsetown
has long been known in northern California, but has not been
recognized in central or southern California. A few species
representative of this horizon which have been found in the
vicinity of Mt. Diablo, have been referred to the Chico. { Dr.
Becker has expressed the opinion that there was an unconformity
between Chico and Knoxville at Mt. Diablo,§ and Dr. Fair-
banks | has described these formations as unconformable in
southern California.

Mr. Turner has found a conformable series of 10,000 feet
of Cretaceous at Mt. Diablo with Aucella Piochi at the bottom
and lower Chico forms near the top.4 There are five thousaud
feet of unfossiliferous beds in between these two places, which
Mr. Stanton thinks must represent Upper Knoxville and the
Horsetown. Sedimentation was slower at Mt. Diablo than
in northern California, or else Mt. Diablo was out of water
during the Horsetown period.**

All of these statements have tended to the general opinion
that the Horsetown was absent in the Mt. Diablo region and
that Knoxville and Chico are unconformable.

From the Santa Clara Valley region there are forms which
show the probable presence of Horsetown at certain localities.

The fauna, given in the table, shows the possible presence
of Horsetown at Pleasanton, Arroyo del Valle, Haywards,
Berkeley, Benicia and Mt. Diablo. Further proof of the pres-
ence of this horizon at the last-named place is, that there are
five thousand feet of beds between lower Knoxville and
lower Chico, and field work has shown the whole Cretaceous
series to be conformable here, as is true at Arroyo del Valle
and Pleasanton. The thickness of Chico Turner gives as six

* Geol. Cal. and Oregon, Bull. Geol. Soc. Am., iv, 213.

+ Mesozoic Flora of the U. S. Monog. xlviii, U. S. Geol. Sur., Pt. I, p. 227.
ote Westenion, Fauna of the Knoxville Beds, Bull. 133, U. 5. Geol. Sur-

Uateatiaraphy ofCal bull Ol sOe Ss. Ge Soaps LO:

| H. W. Fairbanks, Pre-Cretaceous Age of Metamorphic Rocks of the Coast
Ranges, Am. Geol., ix, 165; also Stratigraphy of Calif. Coast Ranges, Jour.
Geol., ili, 426.

*| Geol. of Mt. Diablo, Bull. Geol. Soc. Am., ii, 401.

oe W. Stanton, Fauna of the Knoxville Beds, Bull. No. 188, U. S. G. S.,
p. 21.

Am. Jour. Sci1.—FourRTH SERIES, VoL. XXIV, No. 139.—Juxy, 1907.
4 ;

50 Crandall—Santa Clara Valley Region in California.

thousand feet; but in the thickest known section of Creta-
ceous on Elder "Creek the Chico is only four thousand feet ; and
south of Mt. Diablo at Arroyo del Valle four thousand feet i is
the thickness of combined Horsetown and Chico, with the same
amount for Knoxville. Thus, at Mt. Diablo, the six thousand
feet must represent more than the Chico and some of it may
be considered Horsetown. The series seems to be conformable
and sedimentation was probably slower at this place than in
northern California.

South of Arroyo del Valle, the Horsetown has not been rec-
ognized* and the Chico is found resting unconformably upon
Knoxville shales and sandstones, into which ser pentine has been
intruded in pre-Chico times.

We (ji &

. Section at Elder Creek.

Legend,
TERTIARY,
CHICO.

.- HORSETOWN.
KNOXVILLE.

FRANCISCAN,

V.Section across the Santa Lucia Range.
Scale unknown,

Comparative sections showing the relations of the Cretaceous in California.
Cause of Absence of Horsetown in Southern California.—
The lack of Horsetown in Southern California shows that at

* A specimen of Hoplites Remondi has been picked up in the hills near
Del Monte, but nothing definite is known about its occurrence. :

C a Valley Region in California. 51

the end of Knoxville time a movement took place which lifted
southern California out of water, and submerged a greater area
of northern California, allowing the transgression of Horsetown
beyond the eastern limits of the Knoxville sea. The sections
given here (fig. 2) show the relations of the three Cretaceous
horizons in northern, central and southern California.

Section I is plotted with uniform dips from Dr. Diller sec-
tion on Elder Creek.*

Section II is from Mr. Turner’s work at Mt. Diablo, and
shows a conformable series.+

Sections III and IV are from field work at Arroyo del Valle,
and show the conformable relations of the Cretaceous series,
and the position of the Jordan ranch beds, from which the
mingled Horsetown and Chico fauna was taken.

Section V is from work done by Dr. Fairbanks in southern
California.

An outline map is given below to show, diagrammatically, the
relations between the various Cretaceous horizons in California.

The Chico shore line is taken from a map of Diller and
Stanton.§ The Horsetown shore line is partly from this same
map, and partly from the known distribution of that horizon
in this immediate region.

The Knoxville shore line is drawn in from the localities at
which this horizon is recognized in the region under discussion,
and the occurrence of the same horizon further south at Mt.
Pinos, and further north of this immediate region or Napa
County.

The overlap of Horsetown upon Knoxville, and Chico upon
Horsetown in northern California, as demonstrated by Diller
and Stanton, is shown here. The lack of Horsetown, and un-
conformity of Chico upon Knoxville in southern California, is
also shown. The point where Horsetown deposits stop, or the
approximate location of the axis of differential movement of
Horsetown times, which raised southern California and sub-
merged a greater area of northern California, is indicated.

Movements during the Cretaceous —At the beginning of
Knoxville there was an epeirogenic movement that extended
from Alaska to southern California, which was a non-uniform

es and Stanton, The Shasta-Chico Series, Bull. Geol. Soc. Am., v,

+ Geol. of Mt. Diablo, Bull. Geol. Soc. Am., ii, 400.

tH. W. fairbanks, Geology of Northern Ventura Co., Santa Barbara Co.
ete., Reprint 12th Ann. Rep. State Mineralogist, 1894, p. 2.

§ The Shasta-Chico Series, Bull. Geol. Soc. Am., v, 454.

| H. W. Fairbanks, Geol. of Northern Ventura Co. , ete., 12th Ann. Rep.
State Mineralogist, 1894, Reprint, p. 20.
re Diller and ‘Stanton, The Shasta-Chico Series, Bull. Geol. Soc. Am., v,
50.

52. Crandall—Santa Clara Valley Region in California.

sinking of the western part of the continent. Deposits of Knox-
ville age are known to occur from Alaska to Mexico.*

At the end of the Knoxville, or the beginning of the Horse-
town period, there must have been another large movement.
The extent of this cannot be determined now, but appears to

OUTLINE MAP
showing.

‘SHORE LINES in CALIFORNIA
during the

“CRETACEOUS,

‘Scale,

010 25° 50

have affected mainly the southern portion of the Pacifie Coast
by uplifting it above the Knoxville sea.

* Diller and Stanton, Shasta-Chico Series, Bull. Geol. Soc. Am., v, 456.

Crandall—Santa Clara Valley Region in California. 53

Tn northern California, the Horsetown overlaps the Knox-
ville and the Chico overlaps the Horsetown, butin the vicinity
of Mt. Diablo the Horsetown cannot be considered to overlap
the Knoxville. At the time of the movement or some time
during the Horsetown epoch, the large intrusions of serpen-
tine through the Coast Ranges must have taken place, as they
are found in numerous localities intruded into the Knoxville,
but not into the Chico beds.*

At the end of the Horsetown period there was a gradual sub-
mergence of the whole coast, which allowed the transgression
of the Chico to the foot of the Sierras, innorthern California,
and permitted the Chico sea to cover the present Sacramento Val-
ley. In southern California, the Chico shore line, as given by
the authorities previously quoted, is close to the present shore
line, indicating a fairly uniform subsidence of the coast, but
with the southern end of California relatively higher than. the
northern, as were the conditions during Horsetown time. In
northern California, the Chico has transgressed nearly to the
Nevada State line, but in southern California the shore line,
from the present evidence, could only have been about as far
east as the base of the present San Jacinto range.

Résumé and conclusions.—In the Santa Clara Valley region
there are scattered localities where the Cretaceous is present.

The three horizons of the Cretaceous are represented among
these various localities.

The Knoxville is present at fourteen, the Horsetown at five
and Chico at five. At three of the five localities where Horse-
town is present Chico is present also.

The Knoxville and Chico are both on the east and west
sides of the Santa Clara Valley. The Horsetown is found only
upon the east side in the vicinity of Mt. Diablo.

The Knoxville and Chico extend farther south than the most
southerly of the localities considered in this paper. The Horse-
town is not known in any place farther south than the Arroyo
del Valle. .

The Cretaceous is represented at Mt. Diablo and Arroyo del
Valle by about one-half the thickness of the maximum section on
Elder Creek.

There is no unconformity between Horsetown and Knoxville
or Horsetown and Chico in this immediate region.

There is nothing to show interrupted sedimentation at the
last mentioned localities.

The continuity deposition of the Cretaceous series in northern
California, as advocated by Diller and Stanton, is accepted.
Continuity of deposition is considered true of the series at Mt.
Diablo and Arroyo del Valle.

*F. M. Anderson, Cretaceous Deposits of the Pacific Coast, Proc. Cal. Acad.
Sci., 3d series, ii, 1, p. 54.

54 Crandall—Santa Clara Valley Region in California.

The Horsetown sea transgressed over northern California be-
yond the Knoxville shore line; at Mt. Diabloit almost coin-
cided with the Knoxville shore line, and in southern California
appears to be absent.

The shore line of the Chico sea shows that when subsidence
took place that the southern end of California was relatively
higher than northern California, as was the condition of affairs
during the Horsetown period.

Stanford University, California, Dec. 16, 1906.

A. H. Verrill—Solenodon of San Domingo. 55

{
Arr. V.—Wotes on the Habits and Kxternal Characters of
the Solenodon of San Domingo (Solenodon paradoxus) ;
by AA. Hyarr VRRILL.

AxtrroueH Solenodon paradoxus of San Domingo and Haiti
was discovered and imperfectly described as early as 1839,
several years before the Cuban species (Solenodon cubanus)
was known to science, it is still practically unknown to recent
zoologists. The published descriptions of this rare and inter-
esting mammal are vague and unsatisfactory. For many years
it has been commonly considered extinct, and when, in Decem-
ber, 1906, I undertook a collecting trip to San Domingo with

San Domingo Solenodon (Solenodon paradoxus).

the avowed intention of obtaining the Solenodon, prominent
zoologists stated that the quest was hopeless, one of them say-
ing that I would be as likely to secure specimens of ghosts as
of Solenodon paradocus.

During the five months spent on the island, I devoted a
great deal of time to hunting for the Solenodon, and in inter-
ens natives from the remote and little-known parts of the
island.

56 A. H. Verrill—Solenodon of San Domingo.

[ soon found that the animal was well known to the 1atives
in certain isolated localities, but that over the greater portion
of the Republic it was absolutely unknown.

This is readily accounted for by the presence of the mon-
goose in most parts of the country, and it is only a question of
time when this pest will overrun the entire island end the

Solenodon will become actually exterminated.

The natives have several names for the Solenodon, calling it
“Orso”, Milqui’, ‘* Homigero”, and “ Juron”, while thé Eng-
lish-speaking negroes from the British West Indies Know it as
“Ground Hog’. The name “Juron” (ferret) is also applied
to the mongoose, and for some time I was misled by this con-
fusion of the two animals. In its habits the Solenodon resem-
bles a hog, rooting in the earth and cultivated grounds, tearing
rotten logs and trees to pieces with its powerful front claws,
and feeding on ants, grubs, insects, vegetables, reptiles, and
fruit, and at times proving destructive to poultry. On several
occasions it has been known to enter the houses in search of
roaches and other vermin, and has been captured in rat-traps.

It is strictly nocturnal, and spends the day in caves, holes in
the coral limestone rocks and in hollow trees and logs. It is a
slow, stupid creature. It is unable to run rapidly, but shambles
along with the zigzag, sidewise motions of a plantigrade. It is
doubtless owing to ‘this that it obtained the native name of
“Orso” (bear).

Its long snout and stout front feet, with their curved claws,
and its thick, short neck, prove impediments to forward pro-
gress. According to the natives it is incapable of running
straight. They also claim that when pursued it frequently trips
itself and tumbles heels over head. When hunted with dogs
it thrusts its head into the nearest hole or shelter and allows
itself to be captured without resistance.

The only specimen that I obtained was a female which was
captured alive and uninjured. A few days after its capture it
gave birth to three naked young. These the mother promptly
devoured, and she died three days later.

This specimen (see figure), as preserved in formol, is 14 inches
in length, exclusive of ‘the tail, which measures about 13 inches
in length.

The body and head are covered with sparse, coarse hair,
which is reddish ferruginous from the eyes to the shoulders,
and dusky brown on the rest of the body.

The hair becomes very thin and scattered on the hind quar-
ters, which for some distance on the back and sides are naked,
roughly corrugated, and warty, with a sparse, short, wooly
growth between the excrescences.

The legs, snout, and eyelids are naked and with the bare

A. H. Verrill—Solenodon of San Domingo. 57

skin of the rump are pinkish white. The ears are short, thin,
rounded, and are bluish gray with light edges. The heavy,
rat-like tail is dark brown and naked. The claws are horn-
color. The front feet and claws are large, heavy, and mole-
like and well adapted to digging and tearing asunder rotten
wood, ete. They are much smaller in proportion than in the
Cuban species, however. The snout is also more flexible than
in S. cubanus, from which it also differs in the naked skin of
the rump, the color, size, and other characters.

58 C. H. Gordon—Mississippian Formations

Art. VI.—WMississippian (Lower Carboniferous) Formations
in the Rio Grande Valley, New Mexico;* by C. H.

GORDON.
Introduction.

Tue observations upon which the following notes are based
were made during the summer of 1905 in connection with an
investigation by the U. 8. Geological Survey of the mines and
mining districts of New Mexico under the direction of Mr.
Waldemar Lindgren, the results of which are to appear in a
forthcoming report by the Sur vey. For the identification of
the fossils collected, the writer is indebted to Mr. George H.
Girty.t

Exposures of rocks belonging to the Mississippian series
occur at a number of places in New Mexico. They have long
been known to occur at Lake Valley, from which circumstance
they early received the name of the Lake Valley Limestone.
The observations of the writer show that exposures of the Lake
Valley limestone occur in many places in the region about the
southern extension of the Black and Mimbres ranges, and rocks
apparently identical were observed in ‘the Caballos Mountains
but fossil evidence of the age of these beds is not at hand. In
Socorro County there are but two small areas in which outerops
of Lower Carboniferous rocks are known to occur, one in the
Magdalena Mountains, where they constitute the principal ore-
bearing formation of the lead and zinc mines at Kelly, and
another on the Arroyo Salado at the base of the Sierra La-
drones, discovered in 1905 by W. T. Lee,§ of the U. 8. Geo-
logical Survey.

The limestones at Kelly are seemingly unfossiliferous, though
Lower Carboniferous crinoids are reported | to have been found
in them. The evidence on which this announcement is based,
however, is lacking, and in the absence of satisfactory data these
beds can not well be correlated with the Lake Valley lime-
stone. Herrick*] gave to these beds the name Graphic-Kelly
limestone. A hyphenated name of this kind is objectionable
and they will be here referred to as the Kelly limestone, from
the town in the vicinity of which they occur.

* Published by permission of the Director of the U. S. Geological Survey.

+ A bulletin of the U.S. Geological Survey treating of the fauna of the
Lake Valley formations is now in preparation by Dr. Girty.

+t Cope, E. D., Eng. and Mining Jour., vol. xxxiv, p. 214, 1882.

§ Personal communication.

| Herrick, C. L., Am. Geol., vol. xxviii, p. 310, 1904; Jour. Geol., vol.
xii, p. 188, 1904. Keyes, C. R., Proc. Iowa Acad. Sci., vol. xii, p. 169, 1904.

§] Loe. cit.

in the Rio Grande Valley, New Mexico. 59

North of Socorro County the Lower Carboniferous forma-
tions disappear by overlap, the rocks of the Pennsylvanian
series resting directly upon those of pre-Cambrian age.

Lake Valley Limestone.—Above the Percha shale (Devo-
nian) in western Sierra County is a series of limestones with
some shales, the upper beds of which at Lake Valley and Hills-
boro are filled with fossils. The Carboniferous age of these
beds was first recognized by E. D. Cope,* who, on the authority
of Dr. C. A. White, referred them to the Middle Carboniferous
(August, 1881).

S. A. Miller, to whom Cope submitted the fossils collected
by him at Lake Valley, publishedt+ a description of them, in
which he announces their Lower Carboniferous age and de-
scribes several species new to science.

A little later (January, 1882), B.S. Silliman,{ on the author-
ity of Mr. Arnold Hague and Mr. C. D. Walcott, likewise as-
signed these beds to the Lower Carboniferous. This conclusion
was strengthened by the investigations of I. M. Endlich § in
1883, whose excellent section of the formations at Lake Valley
is reproduced on page 61. Ina paper published in 1884, F. M.
Springer | described the Lake Valley beds, and gives a list of
the fossils obtained from them.

Ten years later Ellis Clark, at that time manager of the Lake
Valley mines, published a paper, illustrated by map and sec-
tions, in which the stratigraphy and ore deposits are described
with considerable detail.

Following is a section of these beds obtained by the writer
at this locality :

9. Capping of andesite.
Lake Valley Limestone. ( Mississippian.)

8. Coarse, subcrystalline, yellowish-white limestone in moder-
ately thick beds, more shaly below. Abounds in crinoids
and other fossil forms. Some beds cherty. Full thick-
ITE SSeMIO LL SILO wens aN EN ure Sw ly age) 29 SIAN Vs ey INS DN el) 60/

7. Blue shale including thin beds of bluish limestone contain-
ing the same fossils as No. 9, but crinoids not so
| OUT ANS EU aN ras eFat G RI e aesd eclae a ee 75!

6. Grayish blue hard, compact limestone, more or less siliceous
at top. This is called the Blue limestone, and locally is
known as the ‘‘ Footwall Lime” from the fact that it

* Cope, E. D., Am. Naturalist, vol. xv, pp. 831-832, 1881.
+ Miller, S. A., Journ. Am. Soc. Nat. Hist., vol. iv, pp. 306-315, 1881.
¢ Silliman, B. S., Trans. Am. Inst. Mng. Engrs., vol. x, pp. 424-444, 1882.
§$ Endlich, F. M., Am. Nat., vol. xvii, pp. 149-157, 1883.
| Springer, F. M., this Journal (3), vol. xxvi, pp. 97-103, 1884.
oe Clark, Ellis, Amer. Inst. Mng. Engrs., Trans., vol. xxiv, pp. 188-169,
1884.

60 C. H. Gordon—Mississippian Formations

underlies the ore deposits. The upper portion in places
consists of flint breccia. The flint fragments, sometimes
carrying silver, are gray, brown, chocolate, pearl and
green, the ereen 9 yielding ‘a higher grade of ore than the
other colors. At the base isa bed 5 feet thick of coar sely

erystallized yellowish-white limestone ---.---.-------- 25,
5. Compact grayish limestone filled with nodular chert.
Shale partings usually rather thick .........._.-...... 50
Percha Shale. (Devonian.)
4. Grayish yellow and blue shales _- ot ee ee 60’
3.) Black ‘fissile shale... 9. lp) 2 J ie 100’

Mimbres Limestone. (Siluro-Ordovicjan.)

2. Pink limestone upper beds siliceous with drusy cavities;
contains Silurian’ fossilss 222-22. ae
1. Quartzite and limestone below.

By Clark, beds 7 and 8 were called “ crinoidal limestones,”
No. 6 “blue limestones,” and No. 5 “nodular limestones.”

The recognition of the Lower Carboniferous (Mississippian)
facies of the fossils obtained from beds 7 and 8 is confirmed by
Dr. Girty. In the collection from the blue shaly beds (No. 7),
Dr. Girty identified the following species :—

Fossils from Lake Valley.

Laphrentis, sp.
Fawvosites, sp.
Platyceras pileiforme
Megistocrinus evansii ?
Physetocrinus planus
Trematopora vesiculosa
Fenestella, sp.
Pinnatopora, sp.
Crania, sp.

Leptaena rhomboidalis
Rhipidomella ? sp.
Productus semireticulatus

Productus burlingtonensis ?

Productus aff. scabriculus
Productus aff. arcuatus
Productus, sp. a
Productus, sp. 6b

The following
(GA Koyo)
Laphrentis, sp.
Crinoid stems
Megistocrinus evansii ?
Platyceras, sp.

Spirifer imbrex ?

Spirifer aff. grimest.
Spirifer att. peculiaris
Delthyris nova-mexicana
Syringothyris, sp.
Reticularia temeraria
Athyris lemellosa

Athyris aff. incrassata
OCleiothyris roissyt

Retzia, sp.

Camarotoechia occidentalis
Rhynchopora aft. pustulosa.
Platyceras, sp.

Orthoceras, sp.

Goniatites ? sp.

Phillipsia peroccidens ?

were obtained from the overlying beds

Trematopora americana
Rhombopora, sp.
Leptaena rhomboidalis
Schizophoria swallow ?

in the Rio Grande Valley, New Mexico. 61

Platyceras peculiare Productus aff. arcuatus
Platycrinus parvinodus Spirifer afk. grimesi
Dorycrinus unicornis Spirifer inbrex ? ia
Stegonocrinus sculptus Reticularia cooperensis
Physetocrinus lobatus Athyris lamelliosa
Periechocrinus whitet Athyris ? sp.

Fenestella sp.

Crinoids are more abundant in this bed both in species and
individuals than in the shaly beds below. Concerning the re-
lations of this fauna, Dr. Girty says*:—“ The crinoid-bearing
beds at Lake Valley have long been known to paleontologists,
and they are generally regarded as of lower Burlington age.”
Beds 5 and 6 contain corals, crinoid plates and stems, but good
specimens are difficult to obtain.

w
w
z
=
>
wy
=i
=
=
>
[ry
we
=
=i

MONUMENT PK.

Fic. 1. Sketch profile section at Lake Valley, New Mexico, showing the
stratigraphic relations of the formations and the position of ore deposits
(represented by the heavy black line) within the Lake Valley limestone.
After F. M. Endlich.

Two miles east of Hillsboro a good exposure of the Lake
Valley formation shows the lower beds 5 and 6 to be wanting,
and the upper crinoidal beds 7 and 8 are seen resting upon the
eroded surface of bluish gray calcareous shales carrying an
abundant Devonian fauna. Following is a list of fossils ob-
tained from the crinoidal beds at this locality:—

Fossils from Hillsboro.

Michelinia? sp. Rhipidomella dalyana
Zaphrentis sp. Productus semireticulatus
Amplexus sp. Productus aft. seabriculus
~Periechocrinus whitei Spirifer inibrex ?
Rhodocrinus wortheni var. Spirifer aft. peculiaris.
urceolatus Delthyris nova-mexicana

* Personal letter.

62 C. H. Gordon— Mississippian Formations

Cacocrinus multibrachiatus Spirifer aft. S. forbesi
Cactocrinus proboscidialis Spirifer aft. S. grimesi
Steganocrinus pentagonus Spiriferina sp.
Platycrinus sp. Athyris lamellosa
Platycrinus subspinosa Cleiothyris sp.
Physetocrinus lobatus Platyceras 3 sp.
Physetocrinus copet Phillipsia aft. per occidens
Trematopora vesiculosa Phillipsia aff. loganensis
Leptaena rhomboidalis

At Kingston nine miles west of Hillsboro, the Devonian
shales (Percha formation) are overlaid by ive bedded blue
limestone, nodular cherty beds and shaly thin-bedded lime-
stone, fhe total thickness of the formation being about 100 to
125 eon From these beds were obtained the following rep-
resentatives of the Lake Valley fauna :—

Fossils from Hingston.

Leptaena rhomboidalis
Schizophoria swallowi ?
Productus aft. scabriculus
Spirifer imbrex ?
Athyris aff. tncrassata
Cleiothyris roissyi
Orthoceras sp.

Laphrentis sp.
Crinoid indet.
Fenestella sy.
Rhombopora ? sp.
Spirifer aff. peculiaris
Syringothyris sp.
Athyris lamellosa

Near Cooks on the north side of Cook’s Peak, at the south
end of the Mimbres Mountains, the Lake Valley beds have a
thickness of 275 to 300 feet and show the same lithological
characteristics as at Lake Valley. The following fossils were
obtained from them at this locality :—

Fossils from Cooks Range.

Zaphrentis sp.
Crinoid stems
Schizophoris swallowi ?

Productus sp.
Spirifer centronatus
Athyris lamellosa

Productus semireticulatus Cleiothyris sp.

Kelly Limestone.—Resting upon the eroded surface of gran-
ites and schists at Kelly in the Magdalena district at the north
end of the Magdalena range, occur 120 to 125 feet of massively-
bedded, coarsely er ystalline limestone, which, on the evidence
above referred to, is usually regarded as Lower Carboniferous
in age. This reference receives support also in the general
lithologic and stratigraphic relations of the beds. For the
most part the bedding is massive and devoid of shale partings.
Near the middle of the formation is a dark bluish, weathering
to yellowish drab non-crystalline or compact stratum five feet
thick: which by the miners is known as the “Silver Pipe Lime-

in the Rio Grande Valley, New Mexico. 63

stone.” Just below this bed in the Graphic and Kelly mines
are located the most important ore bodies, in following which
the stratum furnishes a convenient guide. The beds dip from
30° to 40° south 65° west, and have been affected to a notable
degree by faulting. The areal extent of the Kelly limestone
is limited, being confined so far as demonstrated to the Magda-
lena region.

Ore deposits in the Lake Valley Limestone.—The -discovery
of ore at Lake Valley was made by a cowboy in 1878 While
tightening the girth of his saddle his attention was attracted
by a peculiar piece of stone whose weight surprised him. On
having it assayed it proved rich in silver. The fame of the
discovery soon spread and the district quickly became noted as
a producer of silver. At present no ore is being mined, opera-
tions haying practically ceased in 1894.

The best description published of the nature and occurrence
of the silver ores of Lake Valley is that of Clark in the paper
above cited. He classes the ores as (1) siliceous, (2) neutral, and
(8) more or less basic. They occur chiefly along the bedding
plane between the blue limestones and the overlying shaly beds
at the base of the crinoidal limestone formation.

The ores consist principally of gray, brown, chocolate, pearl,
and green flints, the last named usually ‘yielding a orade of ore
higher in silver ‘than the other colors. The richest ore bodies
occur in cavities of varying size containing galena with massive
crystalline structure, and sometimes in the form of a pulver-
ulent mass known locally as “orey metal.” The largest of
these ore cavities, called the Bridal Chamber, pr oduced about
2,500,000 ounces of silver. Manganese and iron oxide have a
variable but usually prominent development as accessory min-
erals, while the silica content ranges from 5 to 60 per cent.

The form and relations of the ‘deposits and their occurrence
in pockets and as a siliceous replacement of the limestone along
the contact with overlying shales strongly suggests their origin
from ascending hot solutions. From the nature of the region
the presence at no great distance of an intrusive igneous body
from which the solutions come may be safely predicated.

The region is one of extensive volcanic flows, to the erosion
of which is due the limited areas of sedimentary rocks exposed
in the vicinity of Lake Valley and northward along the slopes
of the range. The earliest of these eruptions consisted of
andesite, which occurs resting upon the eroded surface of the
sedimentary formations.

Ores in the Kelly Limestone—At Kelly the ores occur
at several horizons along the bedding planes of the Carbonif-
erous formations, the principal deposits, however, being
found within the Kelly beds just below the “ Silver

64 C. H. Gordon—Mississippian Formations, ete.

Pipe” stratum above mentioned. The chief ore values in this
district consist of lead and zinc, with little or no silver.
With the exhaustion of the oxidized ores near the surface,
large bodies of sulphide are coming to light below, with
which is associated a considerable amount of copper. The asso-
ciation with the ores of certain minerals, such as pyroxene,
magnetite, and specularite, intergrown with quartz and calcite,
suggests the derivation of these deposits likewise from ascend-
ing hot solutions. The region is one of marked igneous activ-
ity, as shown by the presence in the immediate vicinity of
several bodies of intrusive rocks, chiefly monzonite-porphyries,
while extensive flows of andesite and rhyolite cover the range
as a whole.

April 20, 1907.

Gooch and Heath—Ilodometric Determination of Copper. 65

Arr. VII.—The Lodometric Determination of Copper; by
Rae Goocn and feeb. HmarE.

[Contributions from the Kent Chemical Laboratory of Yale University—clix. |

WueEn potassium iodide is added to a suitable solution of a
cupric salt, cuprous iodide is precipitated while iodine equiva-
lent to the amount of iodine fixed in the cuprous iodide is
liberated. This reaction has been made the basis of an iodo-
metric method for the determination of copper. The first
suggestion of such a method appears to have been made by
De Haen in 1854. In this process cupric sulphate.was treated
in solution with potassium iodide and the free iodine deter-
mined by sulphurous acid according to Bunsen. From the
amount of iodine thus found the copper was calculated, accord-
ing to the equation

2CuSO, + 4KI—~>2K,SO, + Cu, I, +1.,,.

This method was mentioned in the following year by Mohr,*
with the modification suggested by Schwarz that the free iodine
be determined by sodium thiosulphate instead of by sulphurous
acid. E. O. Brown,+ apparently without knowledge of De
Haen’s previous work, proposed, in 1857, similar procedure,
and in 1868 the method with slight modification was presented
again by Riimpler.t Concerning the utility of the method
opinions have varied. Mohr never favored it. So late as
1877 Mohr,§ after quoting Meidinger to the effect that cuprous
iodide freshly precipitated and washed is capable of taking up
iodine, and Carl Mohr’s criticism that potassium iodide acts
upon cuprous iodide according to the concentration, states that
the method is not exact and has nowhere found practical appli-
cation. On the other hand, Fresenius| recommended the
method for the determination of small amounts of copper,
noting that ferric salts and other substances capable of setting
free iodine from an acidified solution of potassium iodide must
not be present, and indicated the most favorable procedure.
The copper salt treated, he says, should be the sulphate, pre-
ferably in neutral solution, though a moderate amount of sul-
phuric acid is not objectionable. Much free sulphuric acid
and any free nitric acid should be neutralized by sodium carbon-
ate, and the precipitate dissolved in acetic acid, an excess of
which does no harm in the iodometrie process.

Of recent writers some have favored the method while
others have commented upon it unfavorably. Low] has been

* Titrirmethode, p. 387. + Journ. Chem. Soe., x, 65.

¢ Journ. prakt. Chem., cv, 193. § Titrirmethode, 5 Aufl., 288.

| Quant. Anal., 6te Aufi., 335, 1875.

*| Journ. Amer. Chem. Soc., 18, 468 ; 24, 1083.

Am, Jour. Sct.—FourtH Series, Vout. XXIV, No. 139.—Juty, 1907.
5

66 Gooch and Heath—lodometric Determination of Copper.

outspoken in praise, to the extent of declaring a preference
for this method in the most aceurate technical work over all
other methods, even the electrolytic method.

According to Low’s earlier modification, metallic copper is
dissolved in nitric acid, the solution is freed from nitrogen
oxides by boiling, a considerable amount of zine acetate is
added, and in the solution having a volume of 50° an excess
of solid potassium iodide is dissolved. Zine acetate is pre-
ferred to sodium acetate to take up the free nitric acid. It
is said that an excess of potassium iodide is necessary to insure
rapidity of: action and is harmless. According to the later
modification of this method Low preee the cupric salt by
dissolving the metal in nitric acid (sp. g. about 1°20), boils the
solution, adds bromine water to destroy the nitrogen oxides,
boils to expel the bromine, treats with ammonium hydroxide
in excess, adds acetic acid and boils again if necessary to get a
clear solution. The advantage of using an excess of potassium
iodide is emphasized, and the statement is made that unless an
excess of this reagent is present the reaction does not proceed
to completion until the titration of the free iodine takes place.
Low recommends the use of 1 gm. of potassium iodide, an
excess of 0°6 grm., for every 0:075 grm. of copper.

Various criticisms haye also been made of the reaction when
employed in gravimetric estimations of the cuprous iodide pre-
cipitated. Pisani* notes that potassium iodide can be used to
effect the precipitation of cuprous iodide and that satisfactory
separations may thus be brought about.

Flajolot+ states that potassium iodide cannot be used as the
precipitant since it dissolves cuprous iodide, and recommends
the precipitation of cuprous lodide from the solution of copper
sulphate slightly acidified with sulphuric acid, by treatment
with sulphurous acid and hydriodic acid: Kohner + atlirms
that cuprous iodide is solul ole both in hydriodic¢ acid and in
potassium iodide.

Brownings has shown that cuprous iodide may be satisfac-
torily pr ecipitated and separated from a cadmium salt by add-
ing to a solution of cupric sulphate a moderate excess of
potassium iodide (1 grm. to 4 grm. in all), expelling iodine and
hydriodic acid by evaporating the solution to dryness, and
treating the residue with water, filtering off the precipitate ong
weighing upon asbestos in the perforated crucible.

As a result of elaborate study Moser | has reached the con-
clusion that the reaction by which cuprous iodide is formed

* Compt. rend., xlvii, 294.
+ Journ. prakt. Chem., xi, 105.
t Ztschr. anal. Chem., xxvii, 215.

S$ This Journal [4], xlvi, 280, 1893.
| Zeitschr. anal. Chem., xliii, 597, 1904.

Bee.

Gooch and Heath—lodometric Determination of Copper. 67

from potassium iodide and cupric sulphate in neutral sorution
is complete at very high concentration of the solution; that
the completeness of the reaction is greatly affected by the
volume of liquid; that the amount of potassium iodide em-
ployed is almost without influence either in neutral solution or
in acid solution; and that the presence of free sulphuric acid
even in large amounts or of hydrochloric acid present in
amount equivalent to the cupric sulphate is advantageous.
Moser recommends, therefore, the addition of sulphuric acid
for the purpose of bringing the reaction to completion. To
the cupric sulphate (about 0-6 gm.) dissolved in 50° of water

10N
contained in a 300°™ stoppered flask are added 5°” of ona Ok.

and 2 gm. of solid potassium iodide, the mixture is shaken
frequently for two minutes, and the free iodine is titrated by
sodium thiosulphate, with stirring, to the end-reaction of the
starch indicator.

According to Fernekes and Koch,* an excess of acetic acid
does not influence titrations, while a certain amount of potas-
sium iodide—1°5 grm. to 2 grin. for 0:0038 grm. of copper, and
2°5 grm. for 0:0939 grm. of copper—must “be added to bring
about complete action in a volume of 100°,

Quite recently Cantoni and Rosensteint have tested the
reaction between potassium iodide and a cupric salt under
various conditions; but these investigators do not give the
absolute values of the amounts of copper taken and found,
merely recording the relative effects of varying conditions.
From the record of their results it would appear that a five-
fold increase of the minimum amount of potassium iodide
added to portions of 100™* of solution containing the same
amount of copper salt is without influence upon “the result ;
that increase of volume from 100° to 350°, other conditions
being the same, may affect the results by as much as 5 per cent
of their value. The authors conelude that the method gives
good results under properly controlled conditions.

So evidence and opinions as to the effects of various con-
ditions in the process are contradictory.

The chief matters of difference concern the influence of an
excess of potassium iodide used as the precipitant, the dilution
at which the precipitation should take place, and the effects of
acids upon the formation of the cuprous iodide. We have
thought it desirable, therefore, to again study these points
experimentally.

In the experiments detailed in Table I, small amounts of a
solution of pure copper sulphate, standar dized by the electrolytie

* Jour. Amer. Chem. Soce., xxvii, 1229.
+ Bull. Soe. Chim. [3], xxxv, 1067-73 (1906).

68 Gooch and Heath—lodometric Determination of Copper.

method and containing 0:0020 grm. to 1°™*, were drawn from
a burette and treated with potassium iodide in solution. In
some of the experiments the iodine set free was titrated with-
out previous dilution, while in others the mixture was diluted
previous to the titration. The volumes at precipitation and at
the end of the titration are noted. In Series A is shown the
effect of twice the amount of potassium iodide theoretically
required, at volumes varying from 30°™ to 80°™ at precipitation
and from 36° to 86°? at the end of the titration.

In the experiments of Series B the effect of increasing the
amount of potassium iodide under conditions otherwise similar
is studied. In Series C is shown the effect of large dilution of
the solution containing the amount of potassium iodide used in
the experiments of Series A.

TABLE I.
Liffects of Volume of Solution and Concentration of Potassium
LTodide.
Volume
= SN Copper
Copper At At end equivalent Error in
taken precip- of titra- KI to I found terms
as CuSO, itation tion used by NazS.03 of copper
grm. em? em® erm. grm. erm.
A
0°0400 30 36 4 0°0591 —0:0009
0:0400 40 46 “4 0°0387 —0°0013
0°0400 50 56 °4 0°0388 —0°'0012
0:0400 60 66 4 0°0391 —0°0009
0:0400 80 86 4 00391 —0:0009*
B
0:0400 40 46 0°8 0°0400 0:0000
0°0400 30 36 8°0 0:0399 —0°0001t
0'0600 45 54 13°0 0°0599 —0-0001F+
C
0°0400 30 200 4 0°0038 —0°0367{
0:0400 30 300 4 0:0004 —0:0396f
0°0400 30 500 *4 0:0004 —0°0395f
0:0400 30 1000 “4 0°0005 —0°0395f

From the results recorded in A it appears that, though the
excess of potassium iodide is about 0-2 grm., the amount used
being approximately twice that required by the theory, the
reaction resulting in the formation of cuprous iodide and libera-
tion of iodine is not quite complete. On the other hand, the
results recorded in B show plainly that at similar dilution the

*End-point slow in coming.

+ The Cu.l, was completely dissolved in KI before titrating.
{ Visible precipitation of Cus, took place on titrating the free iodine.

Gooch and Heath—lodometric Determination of Copper: 69

reaction yields excellent indications of the amount of copper
handled when the amount of potassium iodide is considerably
more than the theoretical amount, varying from four to sixty
times the amount required by theory, the absolute excess varying
from 0°6 grm. to 12°7 grm.; and this is true even though the
amount of potassium iodide is sufticient to dissolve completely
the cuprous iodide formed.

So itis plain that the amount of potassium iodide used has
within limits an influence upon the result. In a volume of
about 50° an excess of 0-2 germ. of potassium iodide is not
enough, while an excess of 0°6 grm. appears to be sufficient.
Beyond this limit the addition of potassium iodide has no
appreciable effect. It is natural to suppose that at high dilu-
tions a larger excess of potassium iodide would be needed
to complete the reaction than is required at lower dilutions.
Table II contains the results of experiments made to test the
efficiency of potassium iodide in precipating 0:0010 grm. of
copper, taken as sulphate, in a volume of 100°.

TABLE II.

Effect of Potassium Lodide in Neutral Solutions at a Fixed
Volume of 100%™.

Copper
takenas KI Acid Copper
CuSO, used added Volume found Error
grm. erm. em? em?® erm, erm.
0°0010 1 none 100 0:0008 —0:0007
ee 2 sf is 0°0006 —0°0004
$s 3 x wy 0°0009 — (0001
se 4 ee eet: 0.009 —0°0001
as 3 rg os 0°0013 +0°0003

From the results of these experiments it appears that while
the action upon a milligram of copper, taken as the sulphate in
100° * of solution containing no free acid, is not completed by
1 grm. or 2 grm. of potassium iodide, it is practically complete
when an excess of 3 to 5 grm. of potassium iodide is present.
The fact is again emphasized that up to a certain proportion,
increasing with the dilution, the amount of potassium iodide
influences the completeness of the reaction in neutral solution.
An excess of potassium iodide amounting to 0°6 grm. to 1
erm. in a volume of 50°, and to from 3 gr. to 5 grm. in a
volume of 100°, will precipitate completely 0°0020 erm. of
copper. In the practical application of these facts it must be
borne in mind that it is the excess of potassium iodide and not
the full amount added which is important.

So we may very properly fix upon 2 grm. as the uniform
amount of potassium iodide suitable for the precipitation of

70 Gooch and Heath—lodometric Determination of Copper.

cuprous iodide in a volume of 50°™* of a neutral solution con-
taining 0°2 grm. of copper; and upon 5 grm. as the amount of
potassium iodide suitable in neutral solutions having a volume
of 100".

We have now to study the effect of free acid upon potassium
iodide.

TABLE IIT.
Effect of Acids upon Potassium Iodide.
Copper
equivalent
KI Acid Volume to I set free
grm. em? em? grm.
H.SO,
cone.
2 1 50 0°0002
2 2 50 0°0005
2 3 50 0°0007
2 5 50 0:0019
5 ul 100 0°0002
5 2 100 0°0002
5 3 100 0°0002
5 5 100 0°0014
HCl
cone,
2 ul 50 0:0002
2 2 50 0°0008
2 3 50 0:0006
2 5 50 0:0016
5 1 100 0°0002
5 2 100 0°0002
5 3 100 0:0002
5 5 100 0°0008
HNO;
cone.
purified
2 1 50 0°0025
2 2 50 0:0094
2 3 50 0°0230
5 1 100 0°0002
5 2 100 0:0002
5 3 100 0°0002
5 5 100 0°0294
HC.H;02
50%
2 25 50 0°0002
5 2 100 0°0002

5
5 50 100 0°0003

Gooch and Heath—Iodometric Determination of Copper. 1

So it appears that a trifling amount of iodine is in every case
set free, due no doubt to presence of traces of iodate. It
appears also that no more than 2™* of concentrated sulphuric
acid or hydrochloric acid may safelv be present with 2 grms.
of potassium iodide in 50°™* of solution, and the presence of
1°™ of pure nitric acid makes error. The tendency to liberate
iodine is manifestly less at the higher dilution, and it appears
that in a volume of 100° of solution containing 5 grms. of
potassium iodide 3°° of concentrated sulphuric acid, hydro-
chloric acid or nitric acid free from nitrogen oxides may safely
be present. Acetic acid of 50 per cent strength may apparently
make up half the solution at either dilution. When either
sulphuric acid, hydrochloric acid or nitric acid is present,
obviously the higher dilution is preferable.

Following are the results of experiments in which various
amounts of copper were determined by titration of the iodine
set free in a volume of 100% * in presence of 5 grms. of potas-
sium iodide and free acid.

TABLE IV.

Effects of Acids upon the Determination of Small Amounts
of Copper.

Copper
taken Total Copper
as Cu(NOs)2 KI Acid volume found Error
grm. grm, cm?, cm, grm. grm.
H;SO, cone.
0:0010 5 ] 100 0°0016 + 0:°0006
0°0010 5 2 100 0:0014 +0°0004
0°0010 5 3 100 0:0019 + 0°0009
HCl cone.
0°0010 5 1 100 “ 0.0014 +0:'0004
0°0010 5 2 100 0:0014 +0°0004
0:0010 5 3 100 0°0015 +0:0005
HNO; cone.
purified
0°0010 5 1 100 0:0014 +0:0004 .
00010 5 2 100 0:0015 +0:0005
0°:0010 5 3 100 00015 +0°0005
HC.H;3Oc.
50 per cent
0°0010 5 10 100 0.0012 +0:'0002
00-0010 5 20 100 0°:0012 +0°0002
60010 5 30 100 0:0010 +0:°0000
0:0010 5 40 100 0.0010 +0:°0000
0°0010 5 50 100 0:0010 + 0:0000

It appears that so much as 50°* of 50 per cent acetic acid may
be present with 5 grms. of potassium iodide in 100° of solution

72 Gooch and Heath—Ilodometric Determination of Copper.

without interfering appreciably with the estimation of 0-0010
grm. of copper and that the error introduced by the presence
of 1°7*, 2°" and 38° of sulphuric acid, hydrochloric acid and
nitric acid (free from nitrogen oxides) in 100°" of solution is
scarcely appreciable.

In Table V are given the results of similar procedure applied
to larger amounts of copper.

In the experiments of Series B and C the material for each
test was metallic copper standardized electrolytically. Portions
of this material were weighed and converted to the nitrate by
acting with nitric acid. The solution of the nitrate was evapo-
rated nearly to dryness and the residue dissolved and titrated
in the manner indicated.

TABLE V.

Liffects of Acids upon the Determination of Larger Amounts
of Copper.

KI Volume

Copper Approx- at begin- at end
taken as Pres- imate ning of of titra- Copper ,
Cu(NO;). ent excess Acid titration tion found Error
7 grm. orm.) rm: ems cm, cm?. grm. ~ grm.
A
Final volume between 110°™° and 120%.

H.SO, conc.
0-1200 5:0 45 25 100 + 119 071200 +0-0000
90-0900 5°0 4°5 3°0 100 114 0°09038 +0:°0008
0-9900 5°0 4°5 3°5 100 114 0°0905 +0°0005

HCl cone.
0:0900 5°0 4°5 2°0 100 ey, 6:0897 —0:0003
071200 5°0 Ad 2°0 100 119 0°1195 —0-°0005
0:0900 5°0 4°5 3°0 100 114 0°0901 +0°0001
071200 5°0 4°5 3°0 100 119 0°1200 +0°0000
0°1200 5°0 4°5 3°5 100 . 119 071197 —0:0008
0°0900 5°0 4°5 4°0 100 114 0:09038 +0:0003

HNO; conc.
0:0900 5:00) 405 On ealO0 114 = 0:0900 +=+0:0000
0°1050 5°0 4°5 1195) 100 117 0°1051 +0°0001
0°0900 5'0 4°5 225 100 114 0:0901 +0:0001

50 per cent

2 3

071200 5°0 4°5 3°0 100 119 0°1195 —0-0005
0:0900 5°0 4°5 5:0 106 114 0°0898 —0:0002
0°1050 5°0 4°5 10°0 100 117 071048 —0'0002

Gooch and Heath—lodometric Determination of Copper. 73

TABLE V (continued).

KI Volume
Copper Approx- at begin- at end
taken as Pres- imate ning of of titra- Copper
Cu(NO;)2 ent excess Acid titration tion found Error
grm. grm. grm. cm’. cm’. cm’. grm. grm.
B

Final volume between 140°? and 155°"° without increase of KI.

03336 5:0 3:5 ee 100) 1530 03315) 0-001
02818 50 4:0 tae 1OOW) adi 4 0-2797% 6-000
03320 5:0 35 ce 100 152 0:3290 -—0:0030
02541 5:0 3:5 iene 100 140 0:2523 —0-0018

C

Final volume increased to 182°"? and 150°™3,
with corresponding increase of KI.

H.S0;
cone.
0°2218 70 6°0 2 100 135 * 0°2214 --0°0004
0°5231 8:0 6°4 3 100 150 0°3226 —0:0005
HCl e
cone.
0°2023 7°0 6:0 2 100 132 072016 —0:0007
0°2581 7°8 6°7 3 100 141 0°2574 —0-°0007
HNO; conc.
purified
0°2023 8°0 7°0 1 100 132 0°2017 —0-:0006
0°2520 10°0 8°5 3 100 148 0°2512 —0'0008
F HC2H;02
50 per cent
0°2125 75 ae 5 100 133 0°2119 —0°0006
0°2064 8:0 2h 8 100 132 0°2058 —0-:0009

The N/10 sodium thiosulphate used in estimating the iodine
liberated added appreciably to the initial volume, 100°, of
the solution. In series A the increase of volume, less than
20°", did not affect appreciably the accuracy of the determi-
nations. In series B the increase of volume to 140°™*, without
corresponding increase in the amount of potassium iodide
present, did affect the indications unfavorably.

In series OC, however, the unfavorable effect of similar dilution
was overcome by the addition of more potassium iodide.

It is apparent that at any volume a very considerable excess
of potassium iodide above the theoretical equivalent involved
in the reaction is necessary,-and that the necessary excess

74 Gooch and Heath—lodometric Determination of Copper.

increases very materially with the dilution of the solution. It
appears also that the noted small amounts of sulphuric acid,
hydrochloric acid, and nitric acid (free from nitrogen oxides)
exert no appreciable influence upon the indications of the pro-

cess carried out at a volume approximately 100°; and that
acetic acid may be present in amount equivalent to at least 25
per cent of the absolute acid.

We find no ground for the inference of Moser* that the
presence of acid, best sulphuric acid, is necessary to the attain-
ment of good results at all volumes excepting the most con-
centrated: and there appears to be no reason why the addition
of small amounts of acid should increase the amount of
iodine liberated if the potassium iodide is free from iodate
or other oxidizer. We are wholly unable to offer any ex-
planation for Moser’s extraordinary observation, ae con-

trary to our own, that variation in the amounts of a SOE
from 1™* to 100° (0-49 grm. to 49 grm.) for 50°" of a solution
of copper sulphate, is practically without effect in the treat-
ment by potassium iodide.

The best general procedure in determining by the iodometric
method amounts of copper not exceeding about 0-3 grm. seems
to us to be covered by the following ‘directions :—The solu-
tion of the cupric salt, containing no more than 3° of
concentrated sulphuric acid, hydrochloric acid or nitric acid
(free from nitrogen oxides), or 25°™* of 50 per cent acetic acid,
is to be made up to a volume of 100°", 5 orm. of iodate- free
potassium iodide are to be added, and the titration of the free
iodine is to be made by sodium thiosulphate in the usual man-
ner with the use of the starch indicator at the end. In case the
end reaction has not appeared when 25° of the thiosulphate
have been added, 2 grm. to 3 grm. more of potassium iodide
are to be added before continuing the titration.

The error of the process, properly conducted, should not
exceed a few tenths of a milligram in terms of copper.

* Zeitschr. anal. Chem., xliii, 597, 1904.

Benton—Strength and Elasticity of Spider Thread. 75

Arr. VIII.—TZhe Strength and Elasticity of Spider Thread ;
by J. R. Benton, Pu.D.

Tue writer of this note happened to notice a spider thread
of such extraordimary thickness and length (0°01 in diameter ;
2-5" long} as to suggest the idea of measuring its physical
properties, since this could easily be done with ‘the thread in
question, though it would be scesedinel difficult with spider
threads of the usual size.

The results may be of interest for the following reasons :
(1) as furnishing a test of the popular idea that spider threads
are composed of a substance of extraordinarily great strength ;
(2) on account of the occasional technical application of spider
threads in the reticules of scientific instruments; and (3)
because the material of the thread, in its chemical nature, is
placed among the proteids, which usually have peculiar
mechanical properties. On account of the great complexity
of chemical structure of the proteids, their physical properties
may be especially useful in determining their relationships
with one another. In the present instance, the mechanical
properties of spider thread may furnish eround for deciding
whether or not its material is identical with silk, as is some-
times asserted.

A spider thread, as is well known, is not usually composed
of a single fiber, but of a number of fibers adhering together
more or less closely. In the thread used for the present study,
the number of component fibers was very large. It was not
feasible to count them directly; but single fibers, when sepa-
rated from the main thread, could be seen to have a diameter
less than one-twentieth of that of the main thread. This
would indicate several hundred fibers in the main thread, if it
can be assumed that all of the component fibers have the same
size. The component fibers appeared to adhere together only
very loosely ; at some places bundles of them were distinctly
separate from the rest of the thread. Under such circum-
stances the apparent cross-section of the thread varied greatly
from point to point. The true cross-section must be known in
order to determine the mechanical pr operties of the material ;
and as it was not feasible under these circumstances to ascer-
tain the true cross-section, the thread was twisted, so as to
bring all of the component fibers into one compact mass.
The twist applied amounted to three revolutions for each cen-
timeter of length, and it resulted in bringing the fibers together
into a cross-section of fairly perfect circularity. The diame-
ter, as measured by a micrometer microscope, varied from
point to point between the limits 0076 and :0103°",

76 Benton—Strength and Elasticity of Spider Thread.

Tensile Strength.—TYhe following results were obtained for
the tensile strength, or stress to produce rupture, the thread
having been tw isted as described :

Stress at rupture
Number of Load at rupture Cross-section at Dynes per cm? Pounds per

trial in grams weight point of rupture sq. in.
1 40 5-28 < Om em 74x 10° 11000
2 82 5°28 15°2 22000
3 85 4°55 18°35 27000
4 85 4°55 18°3 27000
5) 88 4°55 18°9 28000
6 98 4°95 19°4 28000

Final value: (Mean of trials 2, 3,4, 5, and 6) 18:0 x 10° 26200

It would seem justifiable to assume that the first trial, giv-
ing a value less than half the mean of the other, represents
some anomalous condition (such as a flaw in the material). It
is accordingly rejected.

It appears then that the material of this thread possesses
quite a high tensile strength, about double that of most kinds
of wood (the value for pine ‘being about 10,000 and for other
woods ranging between 6,000 and 23,000 pounds per square
inch).

Variations in Length.—The length of the thread was found
to vary irregularly from day to day, the stretching force being
eonstant. This was probably due to absorption of moisture,
but was not definitely investigated.

Elastic After-effect.—lf the stretching force was changed,
the thread stretched (or contracted); but the strain so pro-
duced did not at once assume a constant value, but gradually
varied. Thus in one set of experiments, the stress was
increased by 5X10’ dynes per em* (730 pounds per square inch)
or 1/36 of the mean breaking strength, and readings taken at
intervals of 30 seconds after. applying the load gave the fol-
lowing values for the strain (mean of three trials) :

00200
"00208
00207
°00210
"00218

Upon removing the load as soon as these readings were taken,
the following values of the strain were observed at inter vals
of 30 seconds:

"00048
“00040
"00037
"00034
“00029

Benton—Strength and Elasticity of Spider Thread. 7

Similar, but much less distinct, effects were observed when the
change in stress was only half as great.

On account of the hygroscopic | properties of the substance
it was not feasible to follow these changes over any great
length of time.

Youngs Modulus.—In view of the variations of strain
just described, it is clear that any value obtained for Young’s
modulus, or the ratio of stress to the corresponding strain,
must involve some arbitrariness depending upon the time
when the strain is observed. <A similar situation exists for
most other substances, however, though it is not usually
emphasized in publications as much as, in the writer’s opinion,
it ought to be. Most (f not all) substances exhibit some
elastic after-effect after any change of stress, even when the
change of stress is very small. It is customary, in computing
the modulus of elasticity, to use the value of the strain observed
immediately after changing the stress. The result obtained
in this way will be approximately the same as if any later
value of the strain were used, the approximation being more
or less good according as the elastic after-effect is more or less
small.

In the present experiments a load was applied long enough
(about three minutes) to obtain five readings for the increase
in length ; then the load was removed, and five readings were
immediately taken. The change in length, as determined from
the mean of the five readings with the load, and the mean of
the five readings after removing the load, was made the basis
of computation for Young’s modulus,

The first experiments were made with a thread 6-2 long
and 6°19<10~° em* in mean cross-section. The mean of ten
experiments gave for Young’s modulus 3:27x10" dynes per
cm’ (480,000 pounds per square inch), the different determina-
tions varying from 2°98 to 3°62 10" dynes per em’, The
change in stress employed was 2°5 10" dynes per em*. With
larger change in stress the after-effects were so great as to pre-
vent proper determination of the modulus.

Ten experiments were also made with another piece of the
thread, 16°5™ long, and 5°30 107° em’ in mean cross-section,
twisted with a specific twist of three revolutions per centi-
meter of length, the same range of stress being used. They
gave a mean “value for Young’s modulus of 2°70x10" dynes
per em’ (890,000 pounds per square inch), the extremes being

2°10 and 3°29 10" dynes per cm’.

Adopting as the final value for Young’s modulus 3:0 10"
dynes per cm’, it appears that the material of the spider thread
is, like most or ganic materials, much more stretchy than the

78 Benton—Strength and Hlasticity of Spider Thread.

metals (modulus about 10”), but does not at all approach such
a substance as rubber (modulus about 10° dynes per em’).

Elongation at rupture amounted to about 20 per cent of
the original length of the thread.

The thread showed a distinct elasticity of torsion; but in
view of its complex structure, data on torsion can not have
any very definite significance, and accordingly are not pre-
sented.

Specific Gravity—A 10™ length of the thread, of mean
cross-section 5X 107° em*, weighed 0°33™8, which gives for the
specific gravity, 0°66.

Comparison with Silk Fibers.—The mechanical properties
of silk fibers have been studied by Beaulard,* who found
for the tensile strength 2°85 10° dynes per cm’, and for
Young’s modulus 6°50 10" dynes per em*. As these values
differ from those for the spider thread by amounts greater than
would be accounted for by experimental errors, it seems prob-
able that the material of the spider thread is different from
silk.

University of the State of Florida, Gainesville.

*F. Beaulard: Comptes rendus, cxxxv, pp. 623-626, 1902; Journal de
physique, ii, 785-795, 1908 ; quoted in Science Abstracts, 1908, p. 129, and
1904, p. 187.

Chemistry and Physics. 79

SCIENTIFIC INTELLIGENCE.

I. CHEMISTRY AND PHYSICS.

1. A New Intermediate Product in Thorium.—tit is well
known that the radio-activity of ordinary thorium has been attrib-
uted by Haun to radio-thorium, an element so closely related
chemically to inactive thorium that the separation of the two
presents the greatest difficulties. Boltwood and Dadourian of
Yale University both found that the activity of thorium minerals
depends directly upon the amount of thorium in them, from which
it follows that radio-thorium must be a product derived from ordi-
nary thorium. The same investigators found, however, that
commercial thorium nitrate possessed an activity less than one-
half as great as it should have, according to the activity of the
minerals. This striking result led at first to the belief that radio-
thorium must have been separated to a large extent in the process
of manufacturing the salt, in spite of the fact of the well-known
difficulty of this separation. Hahn has now explained the cause
of the low radio-activity of the artificial salt in a very ingenious
and satisfactory manner. Upon examining samples of thorium
nitrate, which had been kept for various lengths of time, he found
that products recently made from monazite gave the activity to
be expected according to the amounts of thorium in them, while
older preparations gave a much diminished activity. The activ-
ity appears to diminish for about three years,’then to remain
nearly constant for a considerable perigd, and finally to increase
gradually. Hahn explains this behavior by supposing that a prod-
uct intermediate between thorium and radio-thorium, which emits
no a-particles, is separated in the technical production of thorium
nitrate ; the radio-thorium, which appears to have a period of decay
of about two years, gradually disappears until a sufficient amount
of the more slowly formed and more slowly decomposing inter-
mediate substance has been formed to cause the radio-thorium to
increase towards equilibrium. At present the data are uncertain,
but Hahn estimates that the period of existence of the interme-
diate product is about 7 years, and that equilibrium would be
reached in about forty years. Hahn believes that he has positive
proot of the existence of the intermediate product in some prep-
arations which, while containing no appreciable quantity of tho-
rium, are increasing in activity and are producing radio-thorium.
Mesothorium is the name suggested for the intermediate body.—
Berichte, x), 1462. H. L. W.

2. Lhe Interference of Fluorides with the Precipitation of
Alumina.—Yor the determination of the bases in silicates, the
-method sometimes used is the treatment of the substance with
hydrofluoric and sulphuric acids, followed by evaporation and
heating until fumes of sulphuric acid come off, in order to remove

80 Scientific Intelligence.

the silica and get a solution of the bases. Following this method,
Hrnricusen observed that the results obtained in determining
alumina were sometimes very much too low, even as much as 30
per cent of the whole amount. He traced the difficulty to the
circumstance that the fluorine is often removed very incompletely
by evaporation with sulphuric acid in such cases, and to the very
surprising fact, apparently hitherto unknown, that ammonium
fluoride interferes with the precipitation of aluminium hydroxide
by means of ammonia. In fact, when ammonia was added to a
solution of aluminium sulphate, which had been acidified with
hydrofluoric acid, no precipitate came down while the liquid was
hot, but upon cooling a crystalline deposit of ammonium alumi-
nium fluoride, (NH), AIF, ‘analogous to cryolite, Na,AlF,, was
obtained. Hinrichsen has made. a quantitative study of this
solvent effect of ammonium fluoride and found that 0.6 g. of
NH,F was sufficient to keep in solution 0.14 g. of Al,O,. He rec-
ommends avoiding the decomposition of silicates for analysis by
means of sulphuric and hydrofluoric acids except in cases where
the presence of alkalies makes it “ unavoidable.” (Like most of
the German chemists, he is evidently unfamiliar with the excellent
method of J. Lawrence Smith for the determination of the alka-
lies.) When the method is used he advises evaporating off the
sulphuric acid completely and gently igniting the residue, in
order that the fluorine may be completely removed.

The interesting and important fact disclosed by Hinrichsen’s
work led the writer of this notice to confirm it by qualitative
experiments. It was found that while ammonium fluoride readily
prevents the precipitation of aluminium hydroxide by ammonia,
this is not the case with ferric hydroxide, and further that the
addition of a soluble phosphate to such a solution causes the
precipitation of aluminium phosphate. It appears possible that a
method for the separation of iron and aluminium might be based
upon this behavior, the aluminium to be precipitated as phosphate
and weighed as such.— Berichte, xl, 1497. H. 1. W.

3. Magnetic Compounds of Manganese with Boron, Anti-
mony and Phosphorus.—Until a few years ago the only magnetic
metallic substances known were iron, nickel and cobalt. Then
several alloys of non-magnetic constituents were found to be
strongly magnetic, all of which contained manganese, viz., man-
ganese-tin, manganese-aluminium and manganese-aluminium-
copper. It appears, however, that Wohler had noticed nearly
50 years ago the first case of a magnetic compound of non-mag-
netic elements in the oxide of chromium Cr,O,. WaADEKIND has
recently found that the boride of manganese, MnB, has remark-
ably strong magnetic properties, probably one- quarter to one-
half as oreat as soft iron, while the compound MnB, does not
appear to be magnetic at ‘all. The antimonide, MnSb, seems to
be about twice as strongly magnetic as the botide, while the
phosphide Mn,P, shows this property to a much shighter extent.—
Berichte, x1, 1259. H. L. W.

Chemistry and Physics. 81

4, A New Variety of Chromiwn.—lIt has been observed by
J ASSONNEIX that metallic chromium, or also this metal containing
boron or carbon, when heated in the electric furnace in contact
with much copper, gives a solution containing about 1°6 per cent
of chromium. Upon cooling, the chromium separates in a mossy,
crystalline condition, and this can be separated from the matrix
of copper by dissolving the latter in nitric acid. The product is
almost pure chromium in a very fine state of division. It is stable
in the air at ordinary temperatures, but it burns like tinder in
contact with a flame, and ignites in oxygen at about 300°, giving
a very brilliant incandescence.— Comptes Rendus, exliv, 915.

: H. L. W.

5. Memoir and Scientific Correspondence of the late Sir
George Gabriel Stokes. Selected and arranged by Josepu Lar-
mor. In two volumes, Vol. I, pp. xi + 475; Vol. II, pp. vi +
507. Cambridge, 1907 (The University Press).—From 1854 to
1885 Sir George Stokes was Secretary of the Royal Society and
had much to do with the scrutiny to which that society subjects
papers which are offered for publication in its transactions. He
performed these duties so carefully, with so much kindness and
critical acumen, and offered so many helpful suggestions to the
authors of papers that he gradually came to occupy the place of
adviser and friendly critic to most of the British physicists of his
own and of the younger generation. His correspondence was
thus very large and he saved all letters which came to him, so
that at his death more than ten thousand letters were found.
Out of this great number Professor Larmor has made a careful
selection and, in many cases, has been able to obtain Stokes’s
replies, which are inserted in their proper places. We have the
result in the two volumes under review, which no student of
physics can read without keen delight and very great profit.
Stokes’s life-time covered that great era of physical discovery
which (if second to any other) is second only to the time of New-
ton; he, and some of his correspondents, contributed more, prob-
ably, to these great results than any other group of men in any
country. And in these volumes we find the intimate, contempo-
rary history of this active time; we read the first informal
accounts of important discoveries, or of tentative and hesitating
speculations which have since become the classical theories of
physical science. One gets thus a feeling for historical perspec-
tive and an appreciation of how discoveries are made which could
hardly be obtained in any other way. Another thing which
comes out quite clearly, is the effect which Stokes’s activity had
upon the progress of his science indirectly, through his advice
and assistance to others, and quite independently of his published
contributions. Many investigators have gratefully acknowledged
their indebtedness to him, and these letters show very clearly
that the gratitude was well-deserved. Lord Kelvin, in particu-

Am. Jour. Sci.—Fourtu SErigs, VoL. XXIV, No, 139.—Juty, 1907.
6

82 Scientific Intelligence.

lar, has more than once insisted upon his own obligations to his
friend for advice and suggestions ; and though the correspond-
ence between these two great men is not included in the present
volumes, one is glad to learn from the preface that the letters on
both sides have been preserved and will be published separately
ays) CS memorial of the life-long friendship and collaboration of
the authors.”

Besides the scientific correspondence there are letters of bio-
graphical interest, and the first volume begins with an admirable
chapter of personal “Notes and Recollections” of Sir George
Stokes by his daughter, Mrs. Humphry. A useful index is
appended to each volume. 16 Ny 1

6. Elektrische Fernphotographie und Ahnliches ; von ARTHUR
Korn. Pp. 87, with 21 figures and 1 table. Zweite Auflage.
Leipzig, 1907 (S. Hirzel).—This second edition follows the first
issue after an interval of some three years. The subject with
which it deals is one in which so much active work has been and
is still being done that in many respects, especially from the
practical side, the situation is materially changed. he author,
who has already presented the prominent advances in wireless
telegraphy in recent numbers of the Physikalische Zeitschrift,
now gives them to a wider public in this volume. Although
brief in extent, the subject is clearly and adequately presented.

II. Gronoey.

1. United States Geological Survey.—Recent publications of
the U.S. Geological Survey are included in the following list
(continued from vol. xxiii, p. 226) :

Topocrapuic Atias. One hundred and forty folio sheets.

Mtverat Resources of the United States: Calendar Year
1905. Davin T. Day, Chief of Division of Mining and Mineral
Resources. Pp. 1403, with one figure.—This is the twenty-second
volume of this valuable series ; many of the chapters have already
been placed in the hands of the public as separate pamphlets.

ProresstonaL Pargrrs, No. 57. Geology of Marysville Mining
District, Montana: A Study of Igneous Intrusion and Contact
Metamorphism ; by JosepH Barrett. Pp. x, 178, with 16 plates
(two in cover) and 9 figures.—See p. 85.

Buiietins, No. 279. Economic Geology of the Kittanning
and Rural Valley Quadrangles, Pennsylvania; by Cuartus
Burns: Pps 19s: with 11 plates and 14 figures.

No. 286. Economic Geology of the Beaver Quadrangle, Penn-
sylvania (Southern Beaver and Northwestern Allegheny Coun-
ties); by Lester H. Wootsry. Pp. 132, with 8 plates and 35
figures.

No. 287. The Juneau Gold Belt, Alaska; by Artruur C.
Spencer: and A Reconnaissance of Admiralty Island, Alaska, by
Cuarves Witt Wretent. Pp. xii, 161, with 37 plates and 41 figures.

No. 294. Zine and Lead Deposits of the Upper Mississippi
Valley; by H. Foster Bain. Pp. xi, 155, with 16 plates and 45
figures.

Be

Geology. 83

No. 296. Economic Geology of the Independence Quadrangle,
Kansas ; by Frank C. Suraper and Erasmus Hawortru. Pp.
74, with 6 plates and 3 figures.

No. 297. The Yampa Coal Field, Routt County, Colorado ;
by N. M. Fenneman and Hoyr 8S. Gate: with a chapter on the
Character and Use of the Yampa Coals by Marius R. Camp-
BELL. Pp. 97, with 9 plates and 2 figures.

No. 303. Preliminary Account of Goldfield, Bullfrog, and
other Mining Districts in Southern Nevada; by FREpERIcK
Lestie Ransome; with Notes on the Manhattan District by G. H.
Garrey and W.H. Emmons. Pp. 98, with 5 plates and 15 figures.

No. 305. The Analysis of Silicate and Carbonate Rocks; by
W. F. Hittesranp. Pp. 200, with 24 figures.

No. 314. Report on Progress of Investigations of Mineral
Resources of Alaska in 1906; by Atrrep H. Brooks and others.
Pp. 235, with 4 plates and 9 figures.

No. 315. Contributions to Economic Geology 1906. Part I—
Metals and Non-metals, except Fuels. §. F. Emmons, E. C. Ecker,
Geologists in Charge. Pp. 505, with 4 plates and 20 figures.

Water Suppty and Irrigation Papers, No. 182. Flowing
Wells and Municipal Water Supplies in the Middle and Northern
Portions of the Southern Peninsula of Michigan; by Frank
LEvVERETT and others. Pp. xl, 292, with 5 plates and 69 figures.

No. 187. Determination of Stream Flow during the Frozen
Season; by H. K. Barrows and Roserr E. Horton. Pp. 99, ©
with 1 plate and 14 figures.

No. 188. Water Resources of the Rio Grande Valley in New
Mexico and their Development ; by Wituis T. Lee. Pp. 59,
with 10 plates and 2 figures.

No. 191. The Geology and Water Resources of the Western
Portion of the Panhandle of Texas; by CuarLes N. Goutp. Pp.
70, with 7 plates and 3 figures.

No. 192. The Potomac River Basin ; by Horatio N. ParKer,
Baitey Wits, R. H. Botstrer, W.W. Asue, and M. C. Marsu.
Pp. vi, 364, with 10 plates and 2 figures.

No. 193. The Quality of Surface Waters in Minnesota; by
R. B. Dotz and F, F. Wessroox. Pp. 171, with 7 plates and
4 figures.

No. 194. Pollution of Illinois and Mississippi Rivers by Chi-
cago Sewage; by Marsuatt O. Leicuron. Pp. 369, with 2
plates and 5 figures.

No. 196. Water Supply of Nome Region, Seward Peninsula,
Alaska, 1906; by Jonn C. Hoyr and Frep F. Hensuaw. Pp.
52, with 6 plates and 1 figure.

No. 200. Weir Experiments, Coefficients, and Formulas (Revi-
sion of Paper No. 150); by Roserr E. Horron. Pp. 195, with
38 plates and 17 figures.

2. Wisconsin Geological and Natural History Survey. Bul-
letin No. xv. Economic Series No. 10. The Clays of Wisconsin
and their Uses, by Heinricu Ries; with a Report on Molding

84 Scientific Intelligence.

Sands, by H. Ries and F. L. Gariup. Pp. xii, 259, with 30 plates
and 7 figures. Madison, Wis., 1906.—The geological and chemical
relations of the clays of Wisconsin were described in a Bulletin
of the State Survey by Dr. E. R. Buckley, issued in 1900. A
second part dealing with physical and other tests of the bricks
made from the clays was projected but not completed. The pres-
ent Bulletin by Dr. Ries takes up this latter aspect of the matter,
as well as dealing independentiy with the whole subject of the
clay industry. The observations upon which the report is based
have been accumulated by the writer with the assistance of Mr.
F. L. Gallup. It may be noted that the value of the products of
the clay-working industries of the State in 1905 was $1,382,000,
of which more than 90 per cent was in common brick. The
author discusses the probable future of the industry and concludes
that there is likely to be a steady increase in the annual produc-
tion of clay products and raw clays in the State. The second
part of the volume is devoted to a report on the molding sands
by the two gentlemen already named.

3. Geological Survey of Alabama. The Underground Water
Resources of Alabama; by Euernr A. Smiru. Prepared in
cooperation with the United States Geological Survey. Pp. xvi,
388, with 30 plates and 23 figures. Montgomery, 1907 (The
Brown Printing Co).—The material of the present volume was
largely developed by work carried on jointly by the State Survey
and that of the general Government. Water Supply Papers Nos.
102 and 114 (U. 8. G.S.) contain most of the well records here
given, as also a sketch map of the Artesian Water System. The
subject has been further investigated, however, and the results
are now published under the auspices of the Alabama Survey.
A somewhat full account of the geology of the State bearing
upon this subject is given, and the subject of Underground Waters
and Artesian Wells in general is also discussed. Then follows an
account by counties of the wells of the State, followed by data
in regard to their chemical character. It is interesting to note that
of the 1414 wells described, 86 per cent are in the Cretaceous,
about 10 per cent in the Tertiar y, and the remainder, 4 per cent,
in the older, forms of the Appalachian division. The larger part
of the wells are in the outcrop of the Selma chalk and obtain
their supply mainly from the Eutaw sands, although some pene-
trate as low as the Tuscaloosa, strata. A series of maps shows
the geological and geographical relations of these underground
waters. j

4. Geological Survey of Western Australia, Mt. Margaret
Goldfield ; by Cuas. G. Gipson, Assistant Geologist. Bulletin
No. 24. 77 pp., 7 figs., 26 pls., with 13 geological and mining
maps. Perth, 1906 (Fred. Wm. Simpson, Government Printer).
—This report is the most recently issued of the series of special
bulletins dealing with the mineral fields of Western Australia.
The bed-rock throughout the region is granite, which occurs in
dikes and bosses, and greenstone,—both massive and foliated

Geology. 85

varieties. The weathering of the granite rocks to a great depth
is one of the marked features of the area. In the Laverton belt
the ore occurs as hematite-quartz lodes and as reefs; the Burt-
ville mines are in small quartz reefs of exceptional richness, and
the Erlistoun belt contains ore in veins and mingled with hema-
tite-bearing quartz lodes.

5. Geology of the Marysville Mining District, Montana ; by
JOSEPH BaRRELL. Prof. Paper No. 57, U.S. Geol. Surv., pp. x,
178, with 16 plates and 9 figures, including maps in cover. As
stated in the sub-title, this is chiefly “a study of igneous intru-
sion and contact metamorphism.” The Marysville region lies 17
miles northwest of Helena, and its importance as a mining dis-
trict is due to the deposition of ores in the contact-zone of a
small bathylith or stock of quartz diorite about 24 miles long by
14 broad at its widest part. The map which accompanies the
work, on a scale of 4 mile to the inch, presents the detailed geol-
ogy of an area about 53 by 8 miles in length and breadth. The
sedimentary rocks are upturned limestones and shales of Algon-
kian age, through which the igneous mass has been intruded
with peripheral sheets and dikes of aplite, microdiorite and dio-
rite-porphyry, with some small intrusions of gabbro in another
place. All of these varied features receive adequate treatment,
but the special character of the work lies in the detailed investi-
gation of the bathylithic body, of the method of its intrusion,
of its form and of its relations to the surrounding rock masses
both past and present. Added to this is a study of the contact
phenomena produced, from various points of view, all of which
are treated in detail. The presentation of the geological facts
shown by the very careful and complete study of the district is
accompanied by a very full discussion of them on the theoretical
side, so far as they have a bearing on the main problems of the
area.

In regard to the manner of intrusion, Professor Barrell believes
this was accomplished, not by the filling of a predetermined
chamber, nor by marginal assimilation and melting of the walls,
nor by laccolithic invasion, but by magmatic stoping, that is by
a process of passive invasion of the sediments, by the breaking
off of blocks from the roof and their subsidence into the body
of the magma, the latter thus eating its way upward. He is led
to this view by the insufficiency of other hypotheses and by cer-
tain positive facts which are presented. This view, as a method
of explanation for the “mise en place” of many bathylithic
bodies of granite, has been advanced by other geologists, but
this is the most complete and adequate presentation of it,
founded upon details of actual occurrence carefully worked out,
that has yet appeared. Thus the work, though devoted to the
local geology of a small area, is of wide interest since the
problems treated are of great general importance.

Tes Vignes

86 Scientific Intelligence.

6. Invertebrate Paleontology of the Upper Permian Red Beds
of Oklahoma and the Panhandle of Texas ; by J. W. Brxpe.
Kansas Univ. Science Bull., vol. iv, 1907, pp. 115-171, pls. v—vill.
—Here is presented a long and detailed historical review of the
Permian in the area named in the title. The author concludes
that “the Wichita (including the Albany) and Clear Fork beds
of Texas are unmistakably Permian. * * * The light-colored
Permian rocks of Kansas and Texas have their equivalents in the
red strata of Oklahoma.”

The fauna described is from the Red-Bed series at Whitehorse
spring, west of Alva, Oklahoma, and near Dozier, Texas. “Some
of the species seem directly derived from the Kansas Permian or
Pennsylvanian, while others * * * are derived from the
European Permian, especially that of Russia. There seems to be
comparatively little resemblance to the Indian or Chinese forms.”
These latter faunas are of the normal type, while those of Russia
and this American region are of a very salty sea yielding Red
Beds, gypsum, and salt. It is therefore probable that the phys-
ical condition rather than faunal migration has in the main caused
the resemblance pointed out. Further, as ‘“‘ the preservation of
the specimens is such as to obliterate many of the critical char-
acters,” 1t seems, at least forthe present, safer to assume that the
invertebrates of the Red Beds are a direct outgrowth of the pre-
vious Pennsylvanian species of the same region.

The named species number 28, of which 23 are new. It is
almost entirely a molluscan assemblage, there being 13 Pelecy-
poda, 14 Gastropoda and 1 brachiopod. The difference be-
tween this assemblage and that described many years ago by
White, from the Wichita, is very striking, as the latter is a biota
of decidedly Pennsylvanian character plus 4 ammonites with
Mediterranean affinities. These differences are hardly satisfacto-
rily explained by differences in facies alone, as ‘‘the Wichita
material is a limestone fauna, while the fauna [here described ]
* * * is a sandstone fauna.” All in all, paleontologists are
thankful for this contribution, but it is evident that much yet
remains to be done in the field before a complete unraveling of
the physical and faunal conditions during the Pennsylvanian and
Permian of the Southwest, now bound up in the Red-Beds for-
mations, is at hand. Oy

7. The Stromatoporoids of the Guelph Formation in Ontario ;
by W. A. Parks. Univ. Toronto Studies, Geol. Ser., No. 4, 1907,
pp- 137-172, pls. 1-6.—This is the first of a much-desired pro-
posed work on the American Stromatoporoids, and treats of all
the recognizable species in the higher Silurian, the Guelph of
Ontario. The illustrations are heliotype reproductions of good
drawings by the author, showing the microscopic structure.
There are, of Actinostroma, 1 new species; of Clathrodictyum, 3
species ; Labechia, 2 new ; Rosenella, 1 new; Stromatopora, 1 ;
Stromatoporella, 2 uew ; Hermatostroma, 1 new species. . S.

Miscellaneous Intelligence. 87

III. MisceLLANEOUS SCIENTIFIC INTELLIGENCE.

1. Bulletin of the Bureau of Standards. S. W. Srrarron,
Director. Vol. 11, No. 3, pp. 319-483 ; vol. i111, No.1, pp. 1-161.—
The two numbers above noted complete the second volume of the
Bulletin of the Bureau of Standards and begin the third volume.
This Bureau, since it was established by Act of Congress in 1901,
has not only accomplished its organization on a practical basis,
carried on the regular routine work of the office, and issued a
series of twelve Bureau circulars, but it has also completed a
large number of investigations in various related lines in physics
and chemistry. The papers which have thus far appeared in the
Bulletin number fifty-four. A considerable number of these
deal with electrical problems, including theoretical discussions, as
for example that by Dr. E. B. Rosa (vol. ii, No.3) on the formu-
le for the mutual inductance of coaxial coils; also measurements
of inductance, capacity, electromotive force; descriptions of instru-
ments, as the electrodynamometer and potentiometer, etc. The
different aspects of photometry are repeatedly presented, with
also the problems of radiation, absorption and spectroscopy, the
phenomena of incandescent lamps, etc. Still other papers deal
with subjects under heat, the testing of thermometers, and so on.

A large number of workers have contributed to the list of
papers referred to ; among these the following names are often
repeated: E. B. Rosa, E. P. Hyde, P. G. Nutting, K. E. Guthe,
G. K. Burgess, C. W. Waidner. W. W. Coblentz has an article
in vol. ii, No. 3, on a vacuum radiomicrometer, and another on
the investigation of infra-red absorption and reflection spectra.
This last-is a somewhat novel method of investigating the chem-
ical constitution of various compounds, chiefly minerals. The
fact is brought out that minerals containing water of crystalli-
zation show quite uniformly water bands, while those containing
hydroxyl groups generally have a marked band at 3u; further,
sulphates have a strong band at 4:55, one less constantly shown
at 9:1u, while silicates lack such definite bands, thus suggesting a
lack of uniformity in the structure of the silicate radical. The
fact that the 3p band is absent with talc, while if is given by ser-
pentine, confirms the general view that hydroxy] is present in the
latter but not in the former species.

2. Carnegie Institution of Washington.—Recent publications
from the Carnegie Institution are included in the following list:

Variation and Differentiation in Ceratophyllam ; by Raymonp
Peart with the assistance of OL1ve M. Pepper and Frorence J.
Hacie. Pp. 136, with 26 figures and 67 tables.—The immediate
purpose of this investigation is stated to be an “ attempt to work
out as exactly and completely as possible for a particular organ-
ism the laws according to which post-embryonic differentiation
and growth occur.”

88 Scientific Intelligence.

The Collected Mathematical Works of Grorere WILLIAM
Hitt. Volume Four. Pp. 460, 4to.—This volume contains me-
moirs No. 51 to No. 84.

Research in China. In three volumes and Atlas. Volume
I in two parts. Part I: Descriptive Topography and Geol-
ogy; by Bartey Wits, Exior Birackwetprer, and R. H.
SarGEenT. Pp. xiv, 353, xvi, with 51 plates and 65 figures.
Accompanied by a folio Atlas embracing 41 pages of Geograph-
ical and Geological Maps, largely colored.—A notice of this
highly important work will appear later.

Selection and Cross-breeding in Relation to the Inheritance of
Coat-pigments and Coat-patterns in Rats and Guinea-pigs ; by
Hansrorp MacCurpy and W. E. Castir. Pp. 50, with 2 plates
and 31 tables.

Further Researches Concerning Atomic Weights of Potassium,
Silver, Chlorine, Bromine, Nitrogen, and Sulphur; by T. W.
Ricuarps in collaboration with ARTHUR STAEHLER, GEORGE 8.
Forses, Epwarp MuELLER, and GRINNELL JONES. Pp. 88.

The Compressibilities of the Elements and Their Periodic
Relations ; by THroporre W. Ricuarps, in collaboration with
W.N. Sruri, F. N. Bring, and F. Bonnet, Jr. Pp. 67, with
7 figures.

3. Field Museum of Natural History.—The official title of
the Field Columbian Museum, as it was formerly known, has
recently been changed to the form here given. The Annual
Report of the Director, Dr. Frederick J. V. Skiff, to the Board of
Trustees for 1906, contains besides general administrative matter
of particular importance to those engaged in museum work, a
portrait of the late Mr. Marshall Field, the founder of the Museum,
and some remarks about him and his work. Other publications
of the Museum recently issued include the following :—

Botanical Series. Vol. II, No. 4, Publication 117. Studies
in the Genus Citharexylum ; by Jesse More GreEenMAN. Pp.
185-190. Vol. IT, No. 5. (No. 118.)

Flora of the Sand Keys of Florida ; by Coartes FREDERICK
Mitispaueu. Pp. 191-243.

Geological Series, Vol. III, No. 5. (No. 120.) Analyses of
Iron Meteorites ; compiled and classified by Otiver CumMINGS
Farrineton. Pp.59-110.—The author has here brought together
all reliable analyses of iron meteorites classified according to the
physical structure of the specimens, the classification being that
ordinarily in use as developed particularly by Brezina and Cohen.
This compilation is not only of very great interest and value in
itself, but it serves to bring out clearly the relation between
chemical composition and structure. In regard to this the author
states :—

“The most striking feature brought out by the analyses is the
relation shown between chemical composition and structure. This
seems to be definite and general. All the meteorites of a hexa-
hedral structure have a nearly uniform composition, while among

Miscellaneous Intelligence. 89

the octahedral meteorites, fineness of structure increases with
increase of nickel. This conclusion can best be shown by obtain-
ing the averages from the analyses of the different groups, omit-
ting all obviously faulty analyses. The results thus obtained are
as follows :—

Width of
Class. No. of lamellze Per cent
analyses. in millimeters. Fe.
nlexahednitesmys: ssa 29 Tran ee 94°12
Coarsest oxtrahedrites _-_- 2 ee Deb 93°18
Coarse cc fk 22 2°0-1'5 92:28
Medium sé Manes 88 1:0-0°5 90°64
Fine re PEG 41 0°4-0'2 90°18
Finest & ee 13 0:2- — 88°51

“Tt is worthy of note that these averages are not means between
wide limits, but are derived from nearly uniform values. Prac-
tically all of the members of the classes conform in composition
to the average. Were all the groups equally well known, it is
probable, too, that the gradation of percentage of Fe would be
even more uniform than here shown. The medium octahedrites,
for example, while numerous, have been as a whole imperfectly
analyzed. Moreover, some of the meteorites classed as medium
octahedrites, which are characterized by low percentage of iron,
such as Algoma and Glorieta Mountain, have width of lamelle
such as to place them near if not in the fine octahedrites.

“The apparent conclusion from the above results is, that the
content of nickel influences the structure. It may also account
for the change from a hexahedral to an octahedral structure,
since the irons with a hexahedral structure have the lowest per
cent of nickel. So constant and definite does this relation hold,
that given a certain structure the per cent of nickel can probably
be stated more accurately by this principle than it has been de-
termined in some analyses.”

4, Twenty-fourth Annual Report of the Bureau of American
Ethnology to the Secretary of the Smithsonian Institution, 1902-
1903 ; by W. H. Hortmss, Chief. Pp. xl, 846, with 21 plates and
1112 figures. Washington, 1907.—The opening pages (i—x]) of this
volume are given to details of the administration and work of the
Bureau. Following this report is an exhaustive and most inter-
esting paper (pp. 1-846) by Mr. Stewart Culin upon the games of
the American Indian. The amount of material collected by the
author (in his early work, associated with the late Mr. F. H Cushing)
is surprisingly large and varied, and it is presented with admirable
fullness of description and illustration; his studies thus throw new
light upon the mental characteristics and life of the Indian. The
games are either games of chance or of dexterity, those involving
pure skill and calculation being absent. Further, the games are
found to be essentially identical among all the tribes. It is shown
also that they are either connected with definite rites or have
come from ceremonial religious observances.

90 Scientific Intelligence.

5. Bulletin of the Imperial Earthquake Investigation Com-
mittee.—Vol. I, No. 2. March, 1907, Tokyo, Japan.—This see-
ond number of the Bulletin, the inauguration of which has already
been announced (xxiii, 322), is devoted to a series of earthquake
papers by Professor F. Omori. One of these discusses the Val-
paraiso and Aleutian earthquakes of August 17th, 1906, remark-
able as occurring almost simultaneously, although their centers
were separated by a little more than two-thirds of the earth’s

seismic circumference. These disturbances were doubtless not

independent, but are to be regarded as manifestations of seismic
force at the extremities of the earthquake zone which extends
along the entire Pacific coast of North and South America.
Another article of more than local interest is devoted to the dis-
tribution of recent Japan earthquakes; among other points it is
shown that the most active seismic zone of Japan forms a con-
necting link between the Mediterranean-Himalaya zone and the
great American zone before alluded to.

6. Studies in Plant Chemistry and Literary Papers; by
Heten Assotr Micuast, with a Biographical Sketch. Pp. 416.
Cambridge, Mass., 1907 (The Riverside Press).—Dr. Michael
(Helen C. DeS. Abbott) has left a noteworthy record as the
result of a brief scientific career, aside from her endeavors in
other fields of activity. As a chemist she was, as Dr. Wiley has
pointed out in an appreciative introduction, among the very first
investigators of this country who began in a systematic way to
study the relations of chemical composition to species of plants
and plant growth. The relations which she endeavored to estab-
lish between morphological and chemical features in plants are
pointed out in some of the papers reprinted in this volume. There
are also included certain more popular addresses which indicate
the advances in the biochemical study of plants and their con-
stituents, especially the influence of environment upon plant com-
position.

The biographical sketch of 100 pages, by Nathan Haskell
Dole, deserves particular notice. In it are reproduced Miss
Abbott’s detailed impressions of the great teachers whom she
met during her travels abroad. One finds most delightful
sketches of many of the distinguished chemists and botanists of
twenty years ago. The reminiscences of Hoffmann, Liebermann,
Ladenburg, Pringsheim, Wislicenus, Pfeffer, E. von Meyer,
Bunsen, Kekulé, Crookes and many others make one wish that
more of these entertaining impressions were available. They also
serve to indicate the difficulties which woman has encountered in
gaining scientific recognition, and recall Dr. Michael’s pioneer
efforts in overcoming current prejudices.

The volume is primarily attractive as the record of the scien-
tific and literary endeavors of a woman “emancipated from the
shackles of conservatism.” A series of literary papers are
appended to the scientific reprints. L. B. M.

Miscellaneous Intelligence. 91

7 Beitrdge zur chemischen Physiologie, herausgegeben von
F. Hormerister. IX. Band. 1906-1907, Braunschweig (Fr.
Vieweg und Sohn).—Among the papers which deserve special
notice in this volume are the studies of Wiechowski and Wiener
on the fermentative destruction of uric acid and the products
arising therefrom. New methods of preparing organs for the
investigation of the purine enzymes are described and allantoin
was isolated as an enzymatic product resulting from the metabo-
lism of uric acid. Papers from Prof. Hofmeister’s laboratory
(by Sasaki and Savaré) deal with the non-diffusible substances in
the urine and indicate that the compounds included in this group
may have significance through the quantitative alterations found
in pathological conditions. An elaborate investigation of the
debated specific-dynamic action of proteins in nutrition is pre-
sented by Falta, Grote and Staehlin. Almost every important
department of physiological chemistry is represented in the pres-
ent volume; e. g. the chemistry of the liver (Bang, Tiirkel,
Goodman), muscle (Saxl, Urano, Comessatti), digestion (Slowt-
zoff, Schréder, von Fiirth), the blood (Loeb, Lefmann), ete.

i. BoM.

8. The Common Bacterial Infections of the Digestive Tract
and the Intoxications arising from them; by C. A. HeErRtTER,
M.D. Pp. vi, 360, New York, 1907 (The Macmillan Company).—
This book of 357 pages is of special interest to the student and
practitioner of medicine. It gives a clear and concise résumé of
recent investigations on the pathological conditions of the gastro-
intestinal tract. The bacteria of the digestive tract in health
and disease are widely discussed, but most of the emphasis is
placed on the relation of the bacterial products to various kinds
of intestinal diseases. Excessive putrefactive and fermentative
processes, and methods of studying and controlling them form a
large part of the discussion. The book embodies the views
recently presented in a lecture before the Harvey Society for the
Diffusion of Medical Knowledge. L. Fe R:

9. Handbook of American Indians North of Mexico. Edited
by Freprrick Wess Hoper. Part I. Pp. ix, 972, with one
colored map. Washington, 1907. Bulletin 30 of the Bureau of
American Ethnology, Smithsonian Institution.—This is the first
part of an encyclopedic work on the American Indian. It is
arranged alphabetically and the topics give descriptions not only
of the various stocks and tribes and other divisions of the aborig-
ines, with the origin and derivation of their names, but it also
includes a host of other terms, both general and special, covering
all the subjects connected with Indian life and history. <A large
number of illustrations are introduced, so that the compendium
is one of more than usual completeness. The colored chart, due
to the late J. W. Powell, shows the distribution of the linguistic
families of American Indians.

10. A University Text-Book of Botany ; by Dovatas Hoven-
TON CampBeLL. Second Edition. Pp. xv, 579, with 15 plates

92 Scientific Intelligence.

and 493 figures. New York, 1907 (The Macmillan Company).
—This is a new etlition of a work which has been before the
public for some time ; the changes introduced are in general un-
essential in character.

11. The Birds of the Chicago Area; by Frank M. Woop-
RuFF. The Chicago Academy of Sciences. Bulletin No. VI of
the Natural History Survey. Pp. 221, with 10 plates. Issued
April 15, 1907.—The area embraced in this catalogue is about
fifty miles square and includes all of Cook and Du Page counties
with the nine north townships of Will County and a : portion of
Lake County, Indiana. It is of varied character and is particu-
larly notable for its birds, since it lies on the border between the
eastern ranges of many species as well as being in the path of the
Mississippi Valley migration. It is not surprising, therefore, that
the list of birds here included is unusually large for a single dis-
trict.

12. The Seventh International Zoological Congress. —The
International Zoological Congress will meet at Boston on Aug.
19-23. The meetings of sections will be held at the Harvard Medi-
cal School on the successive days named. Saturday, August 24,
will be devoted to an excursion to Harvard University, and on
the succeeding days arrangements have been made for visits to
Woods Hole, Columbia University, and the American Museum,
to Yale and Princeton Universities, with excursions also to Phil-
adelphia, Washington and Mt. Vernon, and other points ; these
will occupy the time from Aug. 25 to Sept. 6. In addition there
will be a trip to Niagara Falls, across Lake Ontario to Toronto,
Sept. 7-9. Finally, an excursion to Bermuda is planned, in case
fifty members enroll themselves for it. This last extends the
meeting of the Congress until September 22d. Information may
be obtained from Charles H. Townsend, Director of the New
York Aquarium, Battery Park, N. Y.

13. Centenary of the Geological Society of London.—The
Geological Society of London, the oldest geological society in
existence, was founded in 1807. It is now to celebrate its Centen-
ary, on the 26—28th of September next, and the occasion promises
to be one of very great interest. The Society had its beginning
when geology was in its infancy, and during the past one
hundred years it has been always in close connection with the
development of the science and by its active work has con-
tributed very largely to its advancement.

14. Neues Jahrbuch fiir Mineralogie, Geologie, und Paliion-
tologie.—The honored Jahrbuch, which for a century has been
one of the chief supports of working mineralogists and geol-
ogists, is pow in its one hundredth year. This notable event is
to be celebrated by the publication of a Festband, which will
soon be presented to all its subscribers.

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES. ]

Art IX.—On the Radio-Activity of Thorium Salis; by

Berrram B. Boutwoop.

MeasvreMents of the a-ray activities of a number of min-
erals containing thorium have been described in an earlier
paper.» A measure of the ionization produced by known
weights of the finely-powdered minerals in the form of thin
films was obtained by introducing the films into an electro-
scope and determining the rate of leak of the charge in terms
of the fall of the gold-leaf in scale divisions per minute. On
dividing the rate of leak by the weight in grams of the
mineral in the film a number was obtained expressing the
specific activity (activity per gram) of the given mineral.
The minerals examined contained uranium as well as thorium.
The activity of one gram of uranium with its equilibrium
amounts of disintegration products (actinium, radium, etc.) has
been found to be a constantt which will be called the normal
specific activity of the uranium-radium series. The value of
this constant for the particular electroscope used was deter-
mined by measurements of the activity of certain minerals
containing uranium and no thorium. Knowing the content
of uranium and the constant for the normal specific activity
of the uranium-radium series, it was possible to calculate for
each, of the thorium minerals that portion of the specific
activity of the mineral which was due to the thorium and
thorium products which it contained. Dividing this by the
weight (in grams) of thorium contained in one gram of mineral,
a number representing the activity per gram of thorium was

* This Journal, xxi, 415, 1906.
+ McCoy, Phil. Mag., ix, 176, 1906; Boltwood, loc. cit.
M. JOUR. Sci.—Fourta SERIES, VoL. XXIV, No. 140.—Aveaust, 1907.
‘

94 Boltwood—Radio-A ctivity of Thoriwm Salts.

obtained. It was found that, within the limits of experi-
mental error, the activity corresponding to one gram of thorium
in a mineral was the same for the different minerals examined,
which indicated that the activity of one gram of thorium with
its equilibrium amounts of disintegration products—the normal
specitic activity of the thorium series, as it can be called—was
a constant.

Measurements were also made of ‘dhe activities of a number
of specimens of thorium oxide which had been separated by
chemical methods from the minerals and from certain thorium
salts prepared by the Welsbach Company. The relative
ionization produced by a known weight of each of these oxides
in the form of a thin film was determined in the electr oscope.
The specitic activity of the oxides prepared directly from the
minerals was found to correspond to the normal specific activ-
ity of the thorium series found in the minerals. The activity
of the oxides prepared from the Welsbach salt was, however,
found to be much lower than the normal. These results led
to the conclusiou that the chemical process employed by the
Welsbach Company was in some way peculiar since it appar-
ently resulted in the separation of over one-half of the radio-
thorium corresponding to the thorium present.

The question of the radio-activity of thorium in minerals
and salts has been examined also by Dr. Dadourian,* who
employed a method based upon thé measurement of the
activity of the deposit obtained by exposing a negatively
charged plate to the emanation evolved by a solution of the
thorium salt or mineral. The results obtained by Dadourian
and the writer were in close agreement and led to similar con-

clusions. Results of a similar character were also obtained by
McCoy and Rosst and by Eve.t

Among the salts examined by. Dadourian were two specimens
of thorium nitrate, the one prepared from North Carolina
monazite and the other from Brazilian monazite, which had
been supplied by the Welsbach Company to the writer nearly
two years before. Mr. H. 8. Miner, the chemist of the Wels-
bach Company, stated that the former salt was about two
years old and the latter at least one year and a half old at the
time they were sent to me. A third salt examined by
Dadourian was a specimen of thorium nitrate which had been
purchased from Eimer & Amend about three years before
the time at which he tested it. It had, however, been used in
the meantime by the writer for the preparation of thorium-X,
that is, the original salt had been dissolved in water, the
thorium had been precipitated as hydroxide with ammonia,

* This Journal, xxi, 427, 1906. + Ibid., xxi, 433, 1906.
t Ibid., xxii, 477, 1906.

Boltwood—Radio-Activity of Thorium Salts. 95

and the washed hydroxide had been reconverted into nitrate.
In order to again obtain the solid salt the solution of the
nitrate had been evaporated to dryness under conditions iden-
tical with those under which a considerable number of other
salts have been prepared and which give a salt containing
about forty-eight per cent of thorium oxide. The number
given by Dadourian as expressing the specific activity of the
thoria in this salt is therefore undoubtedly too high and the
correct value would be about 9°0. The reason for this low
value will appear later in this paper.

The specimen of Welsbach salt examined by the writer con-
sisted of a kilogram of thorium nitrate made from North
Carolina monazite and had been received about fifteen months
before the time of my experiments.

The three Welsbach salts examined by Dadourian and the
writer were therefore at least four years old, three and one-
half years old and one and one-third years old, respectively.
The oldest salt contained about forty per cent of the radio-
thorium in equilibrium with the thorium present and the
youngest salt must have contained at least thirty per cent of
its equilibrium amount of radiothorium. The difference in
ages of these two salts was about three years and they had
both been prepared from the same mineral by the same pro-
cess. If radiothorium was a product formed directly from
thorium it was obvious that its period of decay (recovery)
could not be less than half-value in about six years and might
be somewhat longer.

In April, 1906, the belief that the above data might have
an important bearing in indicating the rate of disintegration
of radiothorium was privately communicated to Dr. Otto
Hahn, the discoverer of radiothoriam. Dr. Hahn replied that
the data supplied were difficult to reconcile with the results of
his own experiments, made directly with a preparation of radio-
thorium, which seemed to show a half-value period of about
two years for this substance. He made, however, at the same
time the interesting suggestion that the lack of agreement
could be explained if a ‘vayless product having a slow rate of
change intervened between thorium and radiothorium.

The rate of disintegration of radiothorium has since been
determined by Blane, * who finds that the half-value period is
737 days, and the existence of a product intermediate between
thorium and radiothorium has recently been demonstrated by
Hahnt in a very convincing manner.

Hahn made a careful examination of a series of samples of
the intermediate chemical products obtained in the technical

* Rend. della R. Accad. d. Lincei, xvi, 291, 1907.
+ Berichte d. chem. Ges., xl, 1462, 1907.

96 Boltwood—Radio-Activity of Thorium Salts.

preparation of pure thorium nitrate from monazite by the firm
of Dr. O. Knéfler and Co., in Berlin. He found only i insig-
nificant differences in the specific activity of the thorium series
in these samples and no differences indicating the separation
of any considerable proportion of the radiothorium.*

On testing samples of the pure thorium nitrate which had
been prepared by the same firm some years before, he found
that these showed a conspicuously lower activity than the
freshly prepared salts, and this decrease in activity appeared to
continue for about three years, after which the activity
remained fairly constant for some time. He also states that he
has been able to obtain preparations which are free from thor-
ium, but which show with the passage of time a marked
increase in their activity and in their power to produce thorium
emanation. These preparations must contain the intermediate
product between thorium and radiothorinm: He therefore
reaches the conclusion that it is this inter mediate product, for
which he suggests the name “ mesothorium,” and not the radio-
thorium, which is separated from the thorium in the technical
process of preparing pure thorium nitrate.

The films of thorium oxide which were originally used for
the determination of the activities given in my earlier paper,
and also a number of others which had been prepared and
measured at the same time or shortly afterward, have been
carefully preserved. The activity of these films has been
recently re-measured in the larger of the two electroscopes
described in the earlier paper and it has been possible to com-
pare their present activities with their activities at the time of
the first measurement. It was found that the specific activities
of all the oxides has decreased by an amount equal to from 15
to 30 per cent. of their former values.

The results obtained are shown in the following table. In
column I the decrease in the activity of each oxide is given in
terms of its activity when first measured. In column II is
given the approximate time which elapsed between the two
measurements. In column III is given the decrease to be
expected in the activity of each oxide if the activity had been
falling at a rate corresponding to half-value in 737 aay) (the
rate of decay of radiothorium). Oxides numbered 1 to 8 are
the oxides given corresponding numbers in my earlier paper.
Oxide No. 9 was prepared from thorianite by a chemical pro-
cess differing only slightly from that used in preparing oxide
No. 5. Oxides No. 10 and 11 were separated from certain

* During the past year I have made an examination of similar chemical
products kindly supplied by the Welsbach Company. The results obtained

were similar to those obtained by Hahn, and no marked differences were
noted in the thorium specific activity of the freshly prepared materials.

Boltwood— Radio- Activity of Thoriwm Salts. 97

technical products supplied by the Welsbach Company and
obtained by them in the process of preparing pure thorium
nitrate from monazite.

TABLE.
I II III
Number Source of oxide Decrease Time Decrease

% days. calc,

lea Viantlexdiste: 22h 2S eS 500 38%
Ie NVelsbachi nitrate. 2 2) Uae oe 32 489 38
5 es os ec ie Uhog ly A a a) 347 28
4 : fs FT ca ees! ual atl Me 22 365 30
EMM OTATIUGC hice cance ye emer 30 489 38
6 GAZ Ur Uy i ca GG GINS ED Re LEAP 26 438 34
ide NUOMAZULC aii, 2. oR oe che ee 26 498 34
Sie VME Te Sy ONC Gaui eee rst sey (0) 365 30
Oe NOrIaAnite cesun eee tine eu ore 30 408 33
10: Welsbach residue No. 3 _..-_--- 23 33 Dil
it VWelsbach residuenNow4) | 2s 19 331 27

It will be noted that none of the oxides has lost its activity
at a rate greater than that.corresponding to a fall to half-value
in 737 days, while most of the oxides have lost their activity
at a lower rate. The behavior of the oxides indicates that at
the time they were first measured they each contained much
less mesothorium than the quantity in equilibrium with the
thorium present. Oxides numbered 2, 5, 8, 9 and 10 in par-
ticular must have contained but a small proportion of their
equilibrium amounts of mesothorium since the rate of fall of
their activity has so closely approached that of radiothorium
itself.

The results which have been obtained clearly indicate that
the low specific activity of the thorium series in the salts pre-
pared by the Welsbach Company can not be attributed to any
special peculiarity in the chemical methods employed in the
Welsbach works. It also appears that the chemical separation
of thorium from mesothorium can be effected without difficulty
by a variety of reactions.

A portion of the Welsbach nitrate from which the three
oxides Nos. 2, 3 and 4 had been indirectly prepared has been
preserved in its original crystalline form. A small amount of
this nitrate was converted directly into oxide by ignition, and
the specific activity of this oxide was determined a few days
later. This was done shortly after the second series of meas-
urements of the older oxides had been carried out. It was
found that the specific activity of the thorium series in the
original nitrate is now practically the same as it was at the time
when the first series of measurements was made. The obser-
vation of this fact at once suggested the possibility that even

98 Boltwood— Radio-Activity of Thorium Salts.

the precipitation of thorium as hydroxide from a solution of
the nitrate is effective in separating thorium from mesothorium,
since this was the only treatment to which the thorium in oxide
No. 2 had been subjected.

Over two years ago I had made some experiments with the
object of obtaining a more definite knowledge of the chemical
behavior of thorium-X. At that time a considerable number
of thorium-X residues had been prepared by the well-known
method of precipitating the thorium as hydroxide from a solu-
tion of the nitrate, filtering off and washing the hydroxide,
evaporating the filtrate to dryness, and gently igniting the resi-
due to remove the ammonium salts present. The amount of
thorium nitrate used in some of these experiments was as much
as a hundred grams and the volume of the filtrate was often
more than two liters. In these experiments it was always
found that the residue, after the removal of ammonium salts,
contained very appreciable amounts of thorium. This thorium
was finally removed from the residue by a second treatment
with ammonia in a solution of small volume. A number of
such thorium-free residues had been prepared and measure-
ments of their activity had shown a steady fall for a period of
about 30 days. After 30 days the residues still retained a defi-
nitely measurable activity, which was comparatively small,how-
ever, and amounted to a leak of only a few divisions per
minute in the electroscope. This residual activity, which was
observed further in some instances for a period of about one
week, appeared to be fairly constant during that period, after
which its progress was not further followed. It was attributed
at the time to the presence of a little radium in the original
nitrate.

Several of these old thorium-X residues have been preserved,
and as soon as the possibility that they might have originaily
contained mesothoriim suggested itself they were introduced
into the electroscope and their activities noted. The activity
of each of them was found to have risen enormously, until it
has now reached a value whieh is many times greater than the
minimum value to which it had originally fallen. These thor-
ium-free residues, weighing together not more than a few milli-
grams, have now an activity equal to that of several grams of
thorium oxide of normal activity and spontaneously evolve an
emanation which is identical with that of thorium, falling to
half-value in 54 seconds and producing the characteristic thor-
ium-active deposit. These residues therefore contain radio-
thorium (and its products) which has been formed by the
disintegration of the mesothorium originally present. If the
half-value period of mesothorium is about 7 years, as suggested
by Hahn, then these residues must still contain mesothorium
and their activity will continue to increase for some time longer.

Boltwood—Radio-Activity of Thorium Salts. 3/8)

It is obvious, therefore, that the chemical process first
described by Rutherford and Soddy for preparing thorium-X
from thorium, consisting in the precipitation of the thorium as
hydroxide from a solution of the nitrate, can be employed also
for the separation of mesothorium from thorium. It has the
advantage that the mesothorium is obtained in a relatively con-
centrated form, but it can be applied with advantage only
where thorium nitrate of some age is available, since fresh
thorium nitrate will contain little if any mesothorium.

The fact that the two oxides numbered 8 and 4, which were
obtained through the intermediate formation of the normal
and ‘ basic” sulphates, respectively, also show a decrease in
their activities is not in itself very suggestive, since in the pre-
paration of the sulphates the nitrate was first converted into
hydroxide. Also in the cases of the other oxides, No. 5 to
No. 11 inelusive, the chemical operations through which the
thorium passed from the first decomposition of the mineral to
the final separation of the pure thorium salt were too compli-
cated to make it possible to determine at what particular step
the mesothorium was removed. From various data it appears
probable, however, that the precipitation of thorium by sodium
thiosulphate is effective in separating thorium from mesotho-
rium.

A further pomt which appears to be worthy of notice in
passing is the similarity in chemical behavior shown by thorium
and radiothorinm on the one hand, and by thorium-X and
mesothorinm on the other. We have no good evidence as yet
of the chemical separation of thorium and_ radiothorium.
These two elements appear to remain together most persist-
ently through elaborate chemical operations which result in
the separation of the thorium from every other known element.
In contrast to this is the facility with which thorium-X and
mesothorium can be separated from thorium and radiothorium.
The chemical similarity of mesothorium and thorium-X is fur-
ther indicated by what follows. Nearly two years ago an
attempt was made to separate radiothorium from thorium by
precipitating barium sulphate in a dilute solution of a thorium
salt. This experiment was performed because it was thought
that the entrainment of radiothorium by barium sulphate
might explain the presence of radiothorium in the radium-
barium sulphate residue for thorianite where Hahn had first
obtained it. The precipitated barium sulphate was highly
active when first prepared, but its activity fell regularly ata
rate corresponding to half-value in 4 days, until at the end of
about 40 days it had reached a negligible value. At the start
it therefore contained thoriam-X but no appreciable amount
of radiothorium. This precipitate of barium sulphate has

100 Boltwood—fRadio- Activity of Thorium Salts.

recently been examined and is now quite active. It gives
off thorium emanation but contains no thorium. Its present
activity is therefore due to radiothorium formed from the
mesothorium which was precipitated with it. From this it
seems probable that the radiothorium which Hahn separated -
from the residue of insoluble sulphates, obtamed some time
before in working up a considerable quantity of thorianite,
had not been present in the residues when first prepared, but
had been formed in them later through the disintegration of
mesothorium. This supposition is further strengthened by
Hahn’s statement* that his radiothorium probably contained
some mesothorium.

It appears quite likely, therefore, that the entraining action
ot barium sulphate on mesothorium was directly responsible
for the presence of radiothorium in the thorianite residues.

Summary.

Measurements or the changes which have taken place in the
activity of certain thorium preparations have given results
which are strongly in support of Hahn’s assertion ‘that an inter-
mediate product having a slow rate of change and not emit-
ting a-rays exists in the thorium series between thorium and
radiothorium.

Certain methods for the separation of this intermediate
product from thorinm have been described.

Sloane Laboratory, Yale University,

New Haven, Conn., June 16, 1907.

* Loe. cit.

E. EF. Lawton—Bands in the Spectrum of Nitrogen. 101

Arr. X.— Wavelengths and Structural Relation of Cer-
tain Bands in the Spectrum of Nitrogen; by Ki. KE. Lawton,
PhD.

Waite engaged upon a study of the spectrum of nitrogen,
it became necessary to photograph in Deas aaaie the solar
spectrum and the spectrum given by the capillary portion— of
a nitrogen Pliicker tube, which had been exhausted to about
ee CO ipom investigating, it was found that so little has been
done on this spectrum in ‘recent years that most of the measure-
ments of the wave-lengths of the lines in the spectrum are far
inferior in precision to the measurements which can be made
with the instruments now available.

The measurements of Angstrém and Thalén,* Boisbaudran,t
and Hasselberg,} expressed in Angstrom units, give only one
decimal place; in general, the error varies from 0-2 to 0-5 of
that unit. Deslandres§ has published a table of wave-lengths
of nearly the whole negative Spec um but these results give
only one decimal place. and Hermesdorf§ have also
studied the same spectrum. Ames, however, measured only
the second line in each band, while Hermesdorf obtained the
wave-lengths of six bands, particularly those of the band with
its head at X 3577. The other bands measured have their heads
at X18805, 8755, 8710, 8536, and 3371. These measurements
have been. obtained with a strong dispersion, and the accuracy
of measurement is given as 0-01 Angstrom unit. Hermesdorf’s
accurate investigation of the wavelengths of a part of the
spectrum is the only one that the writer has been able to find
in the literature that brings our knowledge of the wave-lengths
apace with the present instruments. Hence, the object of the
present work has been to extend, if possible, our knowledge in
this direction.

My own results have been obtained from measurements
made on photographs taken with a large concave Rowland
grating, which is installed in Sloane Phy sical Laboratory. The
grating has 20,000 lines per inch, and a radius of 21°5 ft. By
means of an adjustable shutter. the spectrum given by the
capillary part of a Pliicker tube, such as is ordinar ily used in
spectroscopic analysis, was placed side by side with the stand-

* Acta Soc. Upsala, iii, 1875.

+ ‘‘ Spectres Lumineux.”

{ British Association Heist, p. 188, 1886.

$ Comptes Rendus, ci, p. 1256, 1885; ibid., ciii, p. 375, 1886; Annales de
Chien: et de Phys., xv, p. 5, 1888.

| Phil. Mag., p. 58, 1890.”

“| Annalen der Physik, xi, p. 161, 1903.

102 E. FE. Lawton— Bands in the Spectrum of Nitrogen.

ard spectrum. The solar spectrum, with Rowland’s values, was
used as the standard, and the unknown wave- -lengths were
obtained by interpolation. The nitrogen tube had been ex-
hausted to about 1™™. Many photographs were obtained with
exposures varying from forty-five to ninety minutes, the longer
time always giving the best results. These photographs* show
for this spectr um a complicated structure. The heads of bands
lie toward the red, with the tails extending up the spectrum,
and the lines of the bands are degraded on the side toward the
head. Each band offers at the head an intense triplet. In the
bands measured, the second and third lines of this triplet are
clearly doubles, and it is quite probable that the head of the
band which forms the first line of the triplet is a double also.

Considering a single band—near the head there is a confu-
sion of lines showing no regularity of intensity; but as the
lines become more and more remote from the head, the lines
are seen to be grouped in series of threes, or triplets, and while
the distance between the lines of a single triplet diminishes
the distance between triplets increases. The lines of. each trip-
let show the same intensity, which gradually diminishes with
an increase of distance from the head until they finally dis-
appear, or, as is sometimes the case, become hidden in the next
band. The irregularity of intensity shown by the lines near
the head has been pointed out by Deslandres,t+ using the band
X 3577, as due to several secondary series, if | may be per-
mitted to call them such, which are quite distinct from those
series of lines which form the triplets of the tail. To these
secondary series belongs the strong triplet seen in the head of
each band. In the same paper Deslandres considers that the
third line of the triplets in the tail is a double, but I have seen
no indications that this is so, and neither do Hermesdorf’st
measurements indicate this to be the case.

The second group of this spectrum extends from 25000 to
38000 and is made up of five series of bands.§ All except
the lower part of this group has been photographed during
this investigation. The first attempt to secure a photograph of
the lower part showed some faint lines at about 14600, but a
later attempt proved unsuccessful, owing to the weakening of
the tube. Eye observations, however, showed some strong
lines in this lower region. Using the photographic method,
some good plates of “the region 24200 to %3000 have been

*Tt may be worthy of remark here, as a matter of interest, that there are
no telluric lines in the solar spectrum corresponding to the lines of this
nitrogen spectrum. Nothing more than what might be termed accidental
coincidences were observed.

+ Comptes Rendus, exxxviii, p. 317, 1904.

t Loc. cit.
S$ Comptes Rendus, cili, p. 575, 1886.

FE. EF. Lawton— Bands in the Spectrum of Nitrogen. 108

obtained. At 4059 and 23998 there are two bands which
present beautifully the phenomena of triplets. Just above
there is a group of six bands which overlap each other in such
a way as to entirely obliterate the triplets.

The wave-lengths of the bands beginning at 74059 and
23998 have been measured and the results are given below.
The stronger dispersion has made it possible to measure nearly
a hundred more lines in these two bands than did the early
observers. The measurements have been made with the aid
of a micrometer-microscope especially arranged for the pur-
pose, and in each case measurements have been made by set-
ting on the center of intensity of the lines. Each result given
in the table is the mean of at least eight measurements.
Taking 10°-7™™" as the unit, the precision of measurement
should be at least equal to 0- O1 of that unit, except in the case
of the extreme triplets, which being rather faint, the setting
is more uncertain. The measurements of the wave-lengths
follow.

Wave-lengths of the bands beginning at 4059:458 and 3998 419.

4059°458 4053°790 4046:227
207 "347 4045-781
015 UO “446
4058°895 4052°898 350
“694 *730 ‘016
“511 ‘370+ 4044-727
403 “271 "336
"O84 4051°987 4043°878
4057°883 657 °185
7 291 4042°997
“351 "059 “744
119 4050°863 "399
4056979 "482+ 208
"628 ‘268+ 4041°853
“504 7179 “571
°266 4049°781 ‘200f
030 16M 4040°815
4055°831 4048 °943 “509
524 449 4039-902
"359 4047°959 “189
077 597 4038°731
4054°639 "164 4037'780
“408* . 4046°964 “A447
"245 679 40367981
4053°896 “494 “281
* Very faint.

+ A heavy double.

{ Faint and difficult to measure.

104 EF. EF. Lawton —Bands in the Spectrum of Nitrogen.

4035°655 4025°869 4015°792
4034:°938 “361 4013°991
‘374 4024:004 556
"065 4023°552 “199
4032° ses ‘072 4011°333
326* 4021°613 4010°879
40381:°771 ‘175 534
4030°779 4020°704 4008589
236 4019°160 altel!
4029°672 4018°730 4007°830
4026°580 “317 4005°774
‘099 4016°641t “394
4027°589 °195 4004°984
4026°357
3998'°419 3989°126 3978°229
"289 3988°978 3977°941
8997°946 681 "557
696 587 “294
59] 331 3976°774
*360 3987°839 "351
‘138 °281 ‘072
3996°959 3986°993 3975°521
"829 "498 OTL
"505 "043 3974°918
"048 3985'687 "563
3995°889 [282 314
608 OB7/ 3973°680
"494 3984°802 3972°961
3994°617 "284 378
*382 3983°884 3971°781
“moi "543 3970°988
3993°639 “418 °393
372 Nei 3969°821
"258 3982°831f 3968°864
3992°912 "405 Boo
ay) LY “001 3967°801
"180 3981°569 3966°693
3991:997 ‘274 “130
"D40 3980°883§ 3965°689
*335 "492 3964°474
“Oyivis "285 3963°990
3990°907 3979°934 487
‘496 643 3962°157
“B37 SD) 3961°681
"059 3978°904 OAK,
3989°802 *565 3959°792
"426 397 347,

* This line hazy and setting uncertain.
+ All the triplets above this were very faint.
t Very broad line and difficult to measure. § Very faint.

kh. E. Lawton— Bands in the Spectrum of Nitrogen. 105
3958 °870 3954°017 3949°259
3957°392 38952°254 3948°826
3956°932 3951°866 3946°972
482 “457 562
3954°892 3949°694 173
419

To Deslandres* we owe most of our knowledge of the laws
governing the bands in the nitrogen spectrum, and the lines
in a band. These laws, three in number, are briefly as follows:

1. Ina given band the interval from one line to the follow-
ing in any series, calculated in vibration numbers, are in arith-
metical progression, 1. e., the lines are connected by a relation
of the form,

]

ae =N=a+0n’

where @ and 6 are constants, and m is a series of positive
integers.

2. When two or more series arise from the edge of a band,
they are similar in all respects, and all bands belonging to the
same substance have the same number of series.

3. Ina series of bands the vibration numbers of the edges
form a series similar to that of the line in a single band.

With these laws as a basis, Deslandrest has, quite recently,
investigated the band at X 3577 in the second group of the
negative spectrum. He finds that there are seven series of
lines and all can be expressed by the formula

N=A(m+") +e

where A, c, p, g, are constants, and m a series of integers.
The application of the formula to the wave numbers of a band is
not given by Deslandres, hence the above measurements furnish
an opportunity for the application of the formula. Taking
the band beginning A 3998, the formula has been applied to the
triplets of the tail. The constants have been determined by
the method of trial and error giving the following formule :

Ist line of triplet.
N=2501°457 +. 0:0251257(m + 0°85)”
2d line of triplet.
N=2502°145 + 0°025296(m + 0°5)*
3d line of triplet.
N= 2502-786 + 0:02490(m +0°5)?
*Comptes Rendus, ciii, p. 375, 1886. Annales de Chemie et de Physique,

xv, p. 5, 1888.
+ Comptes Rendus, cxxxvili, p. 317, 1904.

106 £. E£. Lawton— Bands in the Spectrum of Nitrogen.

The results as calculated by these formule are given in the
following tables :

1st line of the triplet.

Calculated Observed Difference
1 1
Nea Nay

1 SNe) ABO ETASY 2507°623 +0°146
16 2508°591 2508°467 + °124
17 + 2509°462 2509°387 + °075
18 2510°384 2510°328 + °056
LO Zoe sion 2511°300 + '057
20 2512°380 2512°383 — ‘003
DSO Mae yD 2513°469 — ‘017
DOM ASMA owile 2514°601 — ‘026
23 2515°749 2515°776 — 027
94 2516°972 2517°000 — -028
Diy, Dy iieouz KG} 2518°297 — ‘051
96 2519°571 2519611 — ‘040
27 2520-945 2520°992 — °047
28 2522°370 2522 403 — ‘033
JO DOA 384:5 2523'°878 — 033
30 2525°460 2525°385 + :075
31 2526°945 2526°916 a IOs)
Bie Da evo 2528°514 ALY COSY
B33) Dingo a7 2530°201 + 046
34° 9531:973 2531°842 Seis:
85 2533°749 2533°588 ee ONG}

2d line of the triplet.

M = lo 25083222 2508'155 +0:067
16 2509°032 “978 SE Oey!
4 "892 2509°S61 a OO Sil
18 2510°702 2510°776 — -074*
WO eas Whee? 2511°758 + -005
DOs 5M elie 2512-788 OID
PL DIESE} 2513°863 — °025
Y2 2514°958 2515042 — ‘084+
23s oles 2516°157 — °042
D4 Dalye3i29 2817°384 — ‘055
95) Yolsio94 25187642 — ‘048
26 2519°909 2519°947 — ‘088
Hike DEO 2521°350 — ‘075

2 Oy BZOO ROO! 9522°710 — ‘018
29 2524°159 2524°181 — °022
BO) Vay as7¢ 7 2525°669 + -008
ie Po OE24G DvoX (OMI se 0)335)
32 2528 864 2528°817 ae Olly)
BS. VISORS 2530°450 ae O SY
SAO 2eLAe 2532-120 se oy Qat
35 2034°024 2533°851 + 173

* Broad line with the center of intensity uncertain.
+ A second measurement verified this result.

E. EF. Lawton— Bands in the Spectrum of Nitrogen.

3d line of the triplet.

Calculated Observed Difference
2 m
m=_14 2508:021 2507°969 — 0°052
15 ‘778 2508°754 + “024
16 2509°565 2509°535 + °*030
Lefeon Zon saa? 2510°407 + °005
18 2511°308 2511°300 + °008
19) 2512-294 ZOD? oy, + 042
DO Assert 2513°255 — ‘011
91 -9514°296 9514°317 — ‘021
DIZOL OM 2515°394 — ‘003
DMO Movooil 2516°560 — *023
AO ilieaita 2517°763 — °031
25 25183977 2519°006 — ‘036
NG) DOD 2520°287 — ‘015
Nie ie Gale 2521°631 — °014
28 2o2 30 '9523:°030 — ‘019
29 9524°455 2524°480 — °*025
30 2525°949 9525°973 — ‘024
BL Dayo) 7 e203} 2527°498 — °005
32), 2529-083 2529°073 + °010
SO JOGO 30 2530°713 + °O17
384 2532°4293 2532°398 + +025
35 06.28.4160 2534°101 + *059

107

It is seen from these results that the formula applies fairly
well to the third line of the triplet, but the agreement between
the calculated and observed values is not so good in the case of
the other two lines. Another thing to be noticed is that values
of N for small values of m do not exist. The same result was
obtained by the writer when Deslandres’* formula was applied
to the measurements of Hermesdorf. If the residuals are
plotted, they show in each case that the curve approximates a
parabolic form. Considering the complex structure of the
nitrogen spectrum, the agreement is as close as could be antici-
pated, and it is evident that it will require a more complicated
formula, than has as yet been proposed, in order that the resid-
uals may be brought within the limits of observation.

The points of chief interest which have been brought out by
the foregoing are :

IL; The wave-lengths of the lines of the bands beginning
24059 and 13998 have been measured, nearly half of them for
the first time; and all the lines have been measured with a
degree of precision hitherto not attained for this portion of the
spectrum.

2. Application has been made of Deslandres’ formula to the
band 73998, and the constants determined giving as close an
agreement between the calculated and the observed values as can
be expected, considering the complex nature of the spectrum.

* Loe. cit.

108) EE. Lawton—Bands in the Spectrum of Nitrogen.

Accompanying this article is an enlarged drawing of the
band » 38998. The drawing is made to scale ; ; and the intensity
of the different lines indicated in the drawing. (I) is the
band itself (last triplet estimated, too faint for measurement).
(11), (111) and (LV) show the triplets resolved into series.

La I] =| aon
< < = = —]

O28
000+

96%
066

QS6¢
gece

VAS

“Me

In conclusion, it is a pleasure to acknowledge my indebted-
ness to Professor Arthur W. Wright, who obtained the excel-
lent photographs used in this investigation, and whose kind
criticisms and advice were continually of great assistance.

Robinson—Tertiary Peneplain of the Plateau District. 109

Arr. XI.— The Tertiary Peneplain of the Plateau Dis-
trict, and Adjacent Country, in Arizona and New Mexico;

by H. H. Rosryson.

Introduction.—The existence of a base-level of erosion, or
as it is now interpreted, a peneplain, in the Grand Canyon
District of Arizona at the close of the period of the great
denudation and just previous to the inauguration of the can-
yon cycle, was strongly insisted upon by Dutton in his “ Terti-
ary History of the Grand Canyon District.” His point of
view may be illustrated by the following quotation :

“Thus there is a general accord of “testimony that at the
period of the older basaltic er uptions very large bodies of Per-
mian strata lay upon the Carboniferous platfor melas eruth.
it seems as if the summit of the Permian then constituted the
surface of the country, just as the summit of the Carboniferous
does now. The fact that the older basalts, wherever found,
rest upon the same geological horizon, suggests to us the fur-
ther inference that the region near the (Colorado) river was
then flat and destitute of deep canyons and valleys, such as
now exist there, and, therefore, destitute of great hills, buttes,
or mesas, The meaning of this is a base-level of erosion .
This, it is true, looks at first like drawing a very broad and
rather remote inference from a very slender basis, and would
not be justified at all if it were not in general harmony with a
wide range of facts. Many facts take form and coherence
around it which would otherwise seem mysterious.’’*

It is desired to present in this article evidence from the
vicinity of the San Francisco Mountains, which are located on
the San Francisco Plateau+ south of the Grand Canyon, in sup-
port of Dutton’s conclusion, and to show that the peneplain he
described existed not only in the Grand Canyon District, but
extended over the greater part of the southern Colorado Pla-
teau and into the region south of the present plateau, now
known as the Mountain District of Arizona. <And it is further
desired to point out the influence that this peneplain, in con-
nection with later faulting, may have had upon the drainage
system of the Grand Canyon District.

Tertiary History of Southern Plateau District.—Before
describing the dceetal localities where this ancient peneplain is
now exposed, a brief summary of the Tertiary history of the

*Mon. II, p. 224, U. S. G. S., 1882.

+ Name appled to that portion of the Colorado Plateaus, exclusive of

Coconino Plateau, bounded on the north by the Colorado River, on the east
by the Little Colorado and on the south and west by the Aubrey Cliff.

Am. Jour. Scl.—FourtH Series, Vou. XXIV, No. 140.—Aveusrt, 1907.
8

110 Robinson —Tertiary Peneplain of the Plateau District.

Southern Plateau District will be presented in order that the
problem under discussion may be given its proper setting. The
data for this have been taken from the reports of Gilbert,*
Dutton,+ Davis,f and especially from the more detailed work
of Huntington and Goldthwait§ in the Toquerville District of
Utah. Although the sequence of events is based principally
on observations in the southwestern part of the plateau, it is
believed to be equally applicable to the entire southern portion
of the platean country, since the facts where best known, as at
the Mount Taylor Mesa and adjacent region in New Mexico,
fit into it without apparent discord.

During early Eocene time, certainly, the southern plateau
country was a land area on which sediments were deposited

partly in lakes, partly probably as fluviatile deposits. It
is possible that the region was uplifted and eroded in the
latter part of the Eocene, since only strata of lower Eocene
age are found in the southern part of the High Plateaus of
Utah. The extensive volcanic activity present in the Basin
Range country of Nevada and Arizona, and the High Plateaus
of Utah, was absent, as there are no traces of volcanic plugs
or dikes, which can be referred to this period, to be seen any-
where in the southern portion of the plateau. At this time
the plateau had been differentiated but shehtly, if at all, from
the Basin Range country, as is rather clearly indicated by the
character and distribution of the Eocene deposits, especially in
the High Plateaus of Utah, although it may be noted that
Lindgren| places the initial separation of the Sierra Nevadas
-and Wasatch Mountains from the Basin Range country in
middle Cretaceous time.

In the latter part of the Eocene, or possibly in early
Miocene time, uplift occurred by monoclinal folding, which,
on the whole, appears to have raised the western part of the
plateau country above the eastern. In the southeastern section
this folding gave rise to an elevated tract of country, the axis
of which had a trend somewhat west of north, while a second
elevated region extended westward for an indefinite distance
from the Echo Cliff monocline situated on the east side of the
Colorado River. This tract appears to have had its principal
extension northward, and did not affect the extreme south-

* Survey West of the 100th Meridian, vol. iii.

+ (a) Tertiary History of Grand Canyon District, U. S. G.S. Mon. II, 1882.
(6) Mt. Taylor and the Zuni Plateau. U.S. G.S. 6th Ann. Rept. 1885.

t+ (a) An Excursion to the Grand Canyon of the Colorado, Bull. Mus.
Comp. Zool, Harvard Coll., vol. xxxviii, Geol. Ser. vol. vi No. 4, 1901.
(6) An Excursion to tho Plateau Province of Utah and Arizona, Bull, Mus.
Comp. Zool. Harvard Coll., vol. xlii, Geol. Ser., vol. vi, No. 5, 1904.
~§ The Hurricane Fault in the Toquerville District, Utah. Bull. Mus. Comp.
Zool. Harvard Coll. vol. xlii, Geol. Ser., vol. vi, No. 1, 1903.

| The Age of the Auriferous Gravels of the Sierra Nevada. Jour. Geol.,
vol. iv., pp. 894-896, 1896.

Robinson— Tertiary Peneplain of the Plateau District. 111

western part of the plateau country. The folding presumably
gave rise to closed basins with the resultant formation of lakes,
and deposition of sediments in them, and also to open basins,
in which fluviatile deposits were laid down. The cycle of erc-
sion begun at this time has continued unbroken to the present.

With the next period of uplift, known as the “ First Period
of Faulting,” it would appear that the plateau district was
rather clearly marked out, since the faulting on both the east
and west borders was such as to lower the country on either
side of the piateau, or to raise the plateau itself. The date of
this first period of faulting cannot be definitely fixed, but from
what is known of the history of the Basin Ranges, it may be
provisionally placed at the close of the Miocene.

By far the greater part of the erosion of the southern portion
of the plateau country had been accomplished by the time the
“Second Period of Faulting” occurred near the close of the
Pliocene. At the close of this eycle of erosion, known as
“The Period of the Great Denudation,” the relief produced
by the monoclinal folding and the faulting of the first period
had been obliterated to a large extent, and the surface of the
entire southern plateau country, and even a greater area, had
been reduced to a peneplain.

Immediately atter the development of the peneplain, and
before the second period of faulting, widespread eruptions of
basalt occurred over much of the southern portion of the
platean, and it is to this capping of lava that the peneplain
owes its preservation. In the San Francisco Mountains, as
determined by the writer,* a second period of volcanic activity
ensued in very late Pliocene, or early Quaternary time, during
which San Francisco Mountain and the neighboring large
cones were built up by flows of andesite, trachy-andesite, and
rhyolite. To this same general period it seems pr obable that
Mount Floyd, the Sierra Blanca and Mount Taylor should be
assigned.

The second period of faulting near the close of the Phocene
gave rise to the Plateau district essentially as it stands to-day,
and introduced the “‘ Canyon Cycle of Erosion,” during which,
in Pleistocene time, the Grand Canyon of the Colorado has
been cut, a considerable amount of ‘the soft strata overlying
the present surface rock of the plateau—the resistant Upper
Aubrey cherty limestone—has been stripped off and cliff pro-
files refreshed. The amount of erosion accomplished during
this cycle is, however, insignificant when compared with that
of the period of the great denudation.

During the canyon cycle of erosion a third period of volcanic
activity occurred in the San Francisco Mountains characterized
by eruptions of basalt from many small cones. These erup-

* Unpublished report ; to be published by the U.S. G.S

112 Robinson—Terliary Peneplain of the Plateau District.

tions took place principally before or during the period of
glaciation that is marked by the morainal material and waste
fans of San Francisco Mountain, while a few of the cones and
flows are of very recent geological age, and may possibly date
from historic time. To this same period of eruption are
assigned the recent basalts, not only on the plateau, but in the
surrounding Basin Range country. They may be easily dis-
tinguished in the great majority of cases from the older basalt
capping the peneplain by their freshness, uneroded condition,
and the fact that they have not suffered displacement by the
faulting of the second period.

The ‘history of the Plateau as outlined differs from that of
Dutton* for the Grand Canyon District in carrying the period
of the great denudation through the Pliocene and of restrict-
ing the erosion of the present Grand Canyon to the Quater-
nary. The fact that the relief produced by the faults of the
first period has been effaced, while that resulting from faults
of the second period has not, appears to make this change
necessary, if but one period of peneplanation exists. The date
of the second period of faulting, which inaugurated the canyon
cycle of erosion at the close of the Pliocene, is the same as
that assigned by Dutton for the cutting of the inner gorge of
the Grand Canyon, and is coincident with the very “general
uplift that affected the western part of the United’ States at
that time. All recent work in the plateau region has shown
that numerous complications exist that were not originally sus-
peeted and that have not as yet received sufficient attention to
permit their being given proper weight. With the history of
the region thus incompletely, if not incorrectly, outlined, the
description of the peneplain foes at the close of the
period of the great denudation may be taken up.

Fie. 1. Black Point Monocline.
1 Upper Aubrey, 2 Moencopie, 3 Shinarump, 4 Basalt.

San Francisco Plateau.—On the San Francisco Plateau
the peneplain is graphically displayed in natural cross-section
at Black Pointt in the Little Colorado Valley by the erosion
of the strata along the Black Point monocline.

The strata visibly involved in the monocline extend from
the Upper Aubrey through the Moencopie and certainly well

* Op. cit. (a), chap. xii.

+A high point eight miles north of Black Falls on the Little Colorado
River; it is uot shown on the maps.

Robinson—Tertiary Peneplain of the Plateau District. 113

into the Shinarump formation.* In the more distant views of
Black Point from the south the slope shows no outcrops east
of a heavy bed of sandstone, corresponding to the conglomer-
atic or coarse sandstone member situated near the base of the
Shinarump, which is located about one-third the distance
between the stripped surface of the Upper Aubrey cherty
limestone and the end of the point overlooking the valley,
measured from the cherty limestone. This lack of outcrops
is due to the fact that that portion of the Shinarump consists of
very soft and easily weathered marls. Between the sandstone
and the cherty limestone exposures are not infrequent and are
oceasioned by the sandstone beds of the Moencopie, a forma-
tion composed principally of soft argillaceous and calcareous
shales.

The feature that attracts attention, as soon as the inclined
position of the beds is recognized, is that an exceedingly
smooth surface, outlined by the black basalt, has been cut
across the basset edges of the strata. It is in fact a very per-
fect plane of erosion and it would be hardly possible for so
smooth a surface to be developed across strata varying in hard-
ness from a compact sandstone to a barely consolidated marl

except ata base-level of erosion. To the west of the steeply

flexed portion of the monocline the peneplain is located on a
small thickness of the nearly horizontal shales of the Moen-
copie formation, but whether the beveling finally reaches the
Upper Aubrey ‘limestone cannot be said. It does not do so
within a distance of 10 miles, beyond which the peneplain is
hidden from view by a cover of recent basalt.

The exposure at Black Point is of interest because it per-
mits the peneplain to be traced directly from a region of dis-
turbed into one of undisturbed strata where its surface is
very closely parallel to the plane of stratification. Evidence
is thus obtained which permits the peneplain to be extended
so as to include the large areas of horizontally bedded Moen-
copie shales protected by basalt that occur especially in the
Little Colorado Valley, and which substantiates the general
conclusion of Dutton.

Another locality where the planation of the strata may be
seen is at the north end of Anderson Mesa some eight miles
southeast of Flagstaff.

On the west side of the mesa the basalt caps cherty lime-
stone; on the east side 400 feet of red shale and conglomerate

23 rie stratigraphic sequence in this region is as follows:

Shinarump formation = Triassic
Moencopie vi = Permian (?)
U. Aubrey cherty limestone

** eross-bedded sandstone - Upper Carboniferous.
L ‘red sandstone. \

114. Robinson—Tertiary Peneplain of the Plateau District.

lie between the lava and the limestone. The contact between
the limestone and shale may be traced around the north end
of the mesa for two miles, in which distance it gradually rises
270 feet, thus making it certain that the mass of shale is wedge-
shaped. Here then is a smooth surface developed across such
widely different rocks as the very resistant Upper Aubrey
cherty limestone and the soft shales of the Moencopie forma-
tion. The section thus shows the cherty limestone to be

4

is a ar —
eT ea ne paren

Fic. 2. Anderson Mesa, north end.

1 Upper Aubrey cross-bedded sandstone, 2 Upper Aubrey limestone,
3 Moencopie, 4 Basalt.

involved in the planation and furnishes the clue to the correct
interpretation of the surface underlying the basalts of the
Black and Mogollon mesas farther south.

The presence of the Upper Aubrey formations at Slate
Mountain, 14 miles northwest of San Francisco Mountain
and six miles southwest of Cedar Ranch, makes it evident
that the red shales and marls at Cedar Ranch lie either as a
beveled wedge between the cherty limestone and capping
basalt, as at DAmndercon Mesa, or as a degraded down-faulted
block ; a point that could probably be made clear by tracing
a bed of conglomerate by its numerous pebbles w estward
along the face of the mesa. Whichever be the case, it
is certain that the level contact surface between the basalt and
the underlying sedimentaries is here, as elsewhere, a peneplain.

At the three localities that have been described the surface
ot the peneplain, as judged by the line of contact between the
sedimentary rocks and the overlying basalt, is remarkably
smooth. Of equal smoothness is the contact between the
Moencopie shales and basalt at all points in the vicinity of
the San Francisco Mountains. Judged by the character of
this contact, of which many miles are exposed in the Little
Colorado V alley, it may be said with a high degree of certainty
that the surface of the peneplain was one of practically no
relief. The contact between the cherty limestone and basalt
was not so extensively seen as that between the Moencopie
shales and basalt. Where displayed on the west side of Ander-
son Mesa, on both sides of Oak Canyon and in the Aubrey
Cliff west of Williams, the contact shows the same degree of
smoothness that characterizes the shale contact and creates the

Robinson—Tertiary Peneplain of the Plateau District. 115

impression that the peneplain as developed on the resistant
Upper Aubrey cherty limestone was of extremely low relief.
The fact, too, that so far as known no monadnocks rise above
the surface of the basalt covering the peneplain points to a
very small amount of relief, as the maximum thickness of the
lava, where shown in the walls of canyons and about the edge
of the field, is not over 100 feet.

The peneplain may be actually traced over very considerable
areas by means of the basalt which caps its surface. A general
acquaintance with the region leads to the belief that the basalt
may be safely used to estimate the present and former extent
of the planation, since wherever the peneplain is unmistakably
present the basalt rests upon its surface.

The peneplain may now be traced on the San Francisco
Plateau over the western side of the Little Colorado Valley
between the Atchison, Topeka and Santa Fé Ry. and Coconino
Point. The eastern boundary is everywhere one of erosion,
while westward the recent basalts cover its surface, the bound-
ary being an irregular line ranging from 8 to 15 miles distant
from the river. The exposures at ‘Cedar Ranch and Anderson
Mesa are at the summit of easterly-facing mesa cliffs now 400
to 700 feet above the surface of the plateau. This means, of
course, that the peneplain once had a greater extension eastward,
though it cannot be stated positively that the surface was
formerly continuous between these localities and the Little Col-
orado Valley. From the section at Anderson Mesa it is possible
to trace the peneplain westward to the Aubrey Cliff, especially
as there is evidence of its existence under the old lava on the
down-faulted block of Moencopie strata at Sycamore Canyon.
The peneplain is definitely known to cover between 1,000 to
2,000 square miles and originally had a much greater exten-
sion in all directions, except the northeastern. In that direc-
tion, on the east side of the Little Colorado, higher land is
encountered within 10 to 15 miles, which must have marked
the limit of the peneplain. This is on the assumption that
there has been no marked later faulting or warping which
would raise the region east of the Little Colorado above that
on the west. This remains to be proved, but it is believed,
as the result of a study of the region west of the river, that
no large faults are present:and that the tilting has tended to
lower the region east of the river rather than raise it.

Mohave Penepiain. — Huntington and Goldthwait,* in
connection with their study of the Hurricane and Grand Wash
faults, have described an area of some 3,000 square miles, to
which they have given the name Mohave peneplain. The
evidence for planation, as in the San Francisco Mountain

* Op. cit., pp. 226-245.

116 Robinson—Tertiary Peneplain of the Plateau District.

region, consists largely of areas of smoothly beveled strata of
widely different degrees of hardness, whieh have been pro-
tected by basalt. A typical case is that of Bellevue Ridge, 4
miles south of Dry Canyon (figure 3), concerning which the
authors say:

“Here the level surface on which lies the uplifted lava is
composed of Moencopie shale, Shinarump conglomerate,
Painted Desert shale and sandstone, and Kanab sandstone, all
dipping strongly to the east. Where the strata are exposed to
erosion the hard Shinarump and Kanab form strong cuesta-
like ridges separated by valleys excavated in the soft Moencopie

Fic. 3. Bellevue Ridge.

1 Upper Aubrey limestone, 2 Moencopie shales, 3 Shinarump congl.,
4 Painted Desert shale and sandst., 5 Kanab sandst., 6 Basalt.

and Painted Desert. Only under conditions approaching
closely to base-levelling would it be possible for a level surface
such as that beneath the lava to truncate smoothly strata of
such varying hardness.”

The surface of the peneplain has been displaced, as shown
in the figure, from 1,000 to 1,500 feet by the recent Hurricane
fault, which belongs to the second period of faulting.

The Mohave peneplain, as described, did not extend beyond
the present western boundary of the plateau, as the Basin
Range country consisted of “ancient mountains well dissected
and mature, and presenting nearly the same appearance as to-
day.’ To the east. in the vicinity of Toquerville, the High
Plateaus marked the boundary, but farther south the peneplain
presumably extended eastward to the Kaibab plateau.

Black Mesa.—Black Mesa is the name applied to the elevated
belt of country on the San Francisco Plateau, which extends
from the San Francisco Mountains as far as Clear Creek Canyon,
a distance of 50 miles. The eastern boundary of the mesa is an
erosion cliff that has an average height of 500 feet ; the western,
except for the part along Oak Creek Canyon, is more strikingly
marked by the Aubrey Cliff—the limit of the plateau—w hich
rises fully 1,000 feet above the Verde Valley. The surface of
the mesa is entirely covered by basaltic lavas, both of the older
and more recent periods of eruption, and diversified by many

Robinson—Tertiary Peneplain of the Plateaw District. 117

small cones. It is necessary, therefore, to form an idea of the
surface underlying the lavas from the outcropping strata at the
edges of the mesa.

On the western side these strata belong predominantly to the
Upper Aubrey limestone and sandstone formations. On the
eastern side, however, the overlying red beds of the Moencopie
and Shinarump formations outcrop beneath the lava cap. The
strata possess a slight dip to the northeast, and it is evident
that the lava rests upon.the eroded edges of the beds. In
view of the peneplained condition of the contact between the
lava and underlying sedimentaries at the north end of the mesa,
Anderson Mesa section, figure 2, and in the region farther
north, it is regarded as practically certain that the peneplain
exists under the older lava throughout the entire extent of the
mesa, although the surface may be slightly more irregular in
the vicinity of Beaver Creek.

The presence of isolated lava-capped buttes beyond the
eastern edge of the mesa makes it certain that the peneplain
once had a greater extension in that direction, and it is con-
sidered as highly probable that it formerly extended, as did
the peneplain in the San Francisco Mountain region, somewhat
to the north of the present position of the Little Colorado
River. That it originally had a greater extension westward
into the present mountain district of Arizona, a point that
will be referred to later, would appear equally certain, since
the western boundary is now marked by a cliff due to erosion.

Mogolion Mesa.—This mesa is a continuation of the Black
Mesa and extends in a direction a little south of east for a dis-
tance of 160 miles from Clear Creek Canyon to the White
Mountains. On the north the boundary of the mesa is not
always distinctly marked. The southern edge, on the contrary,
is sharply limited by the Aubrey Cliff, which, where it over-
looks the Tonto Basin, closely duplicates the upper 2,000 feet
of the walls of the Grand Canyon of the Colorado.

Marvine traversed the mesa in journeying from the Little
Colorado River to Fort Apache, and his report, with cross-
section, Is given in Volume III of the Survey West of the
100th Meridian.* He states that the lava capping the mesa is
“identical with that upon the Black Mesa.” His section shows
that the lava on the north side of the mesa is underlaid either
by the Upper Aubrey sandstone or by the uppermost beds of
the Lower Aubrey sandstone. On the southern side the strata
ave ‘the Lower Aubrey sandstonest of a horizon below the
beds that occur on the north side of the mesa.

* Pages 215-218 and pl. iv. Note that the south end of cross-section is

incorrect. See foot-note on page 217.
+ Reagan, Geology of the Fort Apache Region, Am. Geol., Nov., 1903.

118 Robinson—Tertiary Peneplain of the Plateau District.

Marvine’s description shows that there is a beveling of the
strata in the Mogollon Mesa similar in character to that found
in the Black Mesa and suggests that the surface underlying the
older lava of the mesa may very likely be peneplained. It is
to be noted, however, that several outerops of Cretaceous
strata have been found among the lavas in the vicinity of Min-
eral (St. Johns, Ariz., topographic sheet) and north of Fort
Apache, from which it is to be inferred that the surface was
not so flat as farther west.

It seems probable that the peneplain originally stretched
northward beyond the Little Colorado, as was the case to the
west, and that it may also have extended somewhat farther
south, Eastward the distance to the next locality where traces
of the peneplain are found is 125 miles, and as the intervening
region has never been geologically studied or topographically
mapped, it is not possible to say whether the peneplain for-
merly existed there. It is safe to say, however, that if it was
not developed, the region must have possessed an extremely
mature topography.

Lhe Mount Taylor region of New Mexico—The peneplain
thus far described occurs in the southwestern and southern
portions of the plateau. There is evidence, also, that it exists
in the eastern part—in the Mount Taylor plateau and sur-
rounding country. Dutton in his report on the Zuni Moun-
tains* gives the following description:

“Just north and northeast of Grants the map shows two
detached lava-capped masses, separated from the great volcanic
mesa on which Mount Taylor stands. These two masses are
separated from each other by a high, narrow saddle. The two
are really one long tongue of sedimentary rock. Directly in
the saddle the junction of the Cretaceous and Jura-Trias
appears, but it is exposed in a well-developed monocline dip-
ping eastward. The angle of dip is about 16° to 18°. The
monocline, before the eruptions of lava,t was beveled off
smoothly by erosion, so as to form a nearly horizontal surface
across the basset edges of the upturned strata. The course of
the monocline is a little east of south, and it just touches the
sharp southwestern angle of the Mount Taylor mesa, where
the edges of the Cretaceous beds beneath the lava cap are
flexed up by it.”

Gilbert states,{t in regard to this same fold, that it may be
seen in the Acoma mesa, south of Mount Taylor, which is
capped by lava. Dutton also describes the San Mateo mono-
cline, north of Mount Taylor, which carries Cretaceous strata,
dipping 15° east, under the Mount Taylor plateau. On account

* Op. cit. (b), p. 151 and pl. xiv.

+ The italies are Dutton’s.
t Op. cit., p. 559.

Robinson—Tertiary Peneplain of the Plateau District. 119

of the pronounced dip of the strata at the localities described,
the beveling of their edges was very apparent. It is to be
noted, however, that the strata under the Mount Taylor pla-
teau lava have a gentle dip north and west, thus making the
peneplain coextensive with the plateau. But the peneplain
now stands through erosion on an average of 700 feet above
the surrounding country, which indicates that it must origi-
nally have had a much greater extension than at present.

Evidence of the existence of the peneplain in the region
south of Mount Taylor is found at Tres Hermanos buttes and
Mesa Lueera, in addition to the Acoma mesa. It is derived,
as elsewhere, from the relation that exists between the eroded
strata of folds and faults and the overlying basalt.

Gilbert has described* this relation at Tres Hermanos
buttes, located 50 miles south of Mount Taylor, as follows:

“From the buttes a fold runs 10 miles northwest to the
Acoma plateau, beneath which it disappears. . . The throw is
between 1,500 and 2,000 feet, and is to the southwest. The

60°E —
Fic. 4. Monocline at Tres Hermanos Buttes.
C Cretaceous, T Triassic, B Basalt

fold is older than the basalt of the vicinity. The eroded
edges of the strata upturned by it support the lava caps of the
Hermanos buttes and of the Acoma plateau. The level line
of the Acoma lava shows that the folding has not continued
since the eruption, and the antiquity of the eruption is meas-
ured by a general denudation of the country of more than 500
chee tea

Gilbert’s cross-section of the monocline, in which propor-
tions were estimated, is given in figure 4.

It shows that the strata involved in the fold had suffered
great erosion before the lava covered them. The situation
appears to duplicate that at Black Point (figure 1), except that
the lava cap has been largely removed, and as the result the
underlying strata have been considerably dissected. As the
capping lava may be traced to the Acoma mesa and hence to
the Mount Taylor plateau, it is believed that Dutton’s descrip-
tion of the peneplain at the latter locality may be applied to
this locality.

At Mesa Lucera, 35 miles southeasterly of Mount Taylor,
the relation of the basalt to the displacements is also described
by Gilbert :*

*Op. cit., p. 557.

120 Robinson—Tertiury Peneplain of the Plateau District.

“Upon the San José, a few miles west of its junction with
the Puerco of the East, there is exposed a complex disturbance,
consisting of a number of faults and fol ds,. that were not suf-
ficiently studied to warrant an attempt at their representation.
The residual throw is to the east and the trend of the dipping
strata north and south. The basalt of the Lucera plateau,
and Mesa Redonda, rests on the disturbed strata, and is not
tilted with them.”

Nothing definite can be said in regard to the original extent
of the peneplain in this part of the plateau, since topographic
maps are lacking and much of it has never been geologically
studied. The opinion only may be stated that it must have
covered a very large area, at a minimum estimate possibly 5,000
square miles.

The Mountain District of Arizona.—The peneplain in the
Black and Mogollon mesas is terminated on the west and
south by an erosion cliff overlooking the Mountain district,
which makes it highly probable, as noted, that it formerly
extended somewhat into that district. The facts on which
this conclusion is based are meager, and what is here presented
may be considered largely suggestive and as indicating a point
for future study.

In the Black Hills and Bradshaw Mountains, west of the
Black Mesa, it seems possible to obtain a rather definite idea as
to the presence of the peneplain. The topography of the
Black Hills and of considerable areas about the Bradshaw
Mountains* is of the mesa type and has been produced by
the erosion of horizontal flows of basaltic lava. The eroded
eastern edges of the flows, capping the Black Hills from
Jerome for 40 or more miles southward, overlook the
Verde valley and face the eroded and lava-capped Aubrey
Cliff, from 10 to 20 miles distant. To explain the present
position and condition of the lava flows it is necessary to
assume that the surface on which the lava rests originally con-
tinued across the Verde valley, and has since been obliterated,
in connection with faulting, by the erosion of the valley.

The lava in the Black Hil Ist rests either upon a small
thickness of limestone or upon granite, while in the Bradshaws
it lies on granitic or metamorphic rocks. The formations are,
therefore, of a very much lower horizon than those upon
which the lava rests in the Aubrey Cliff. This difference
may be explained by the erosion, essentially to grade, of the
relief produced either by a monocline or fault. The first case
would duplicate the condition found at the Black Point mono-
cline (figure 1); the second, that near the point where the

* Jerome and Bradshaw Mountain topographic sheets. U. 8. G. S.

+ Reid, Sketch of the Geology and Ore Deposits of the Cherry Creek
District, Ariz. Econ. Geol., vol. i, No. 5, 1906.

Robinson— Tertiary Peneplain of the Plateau District. 121

Virgin River cuts the Hurricane fault in Utah,* except that
the throw of the fault would be reversed. It seems probable
that the displacement has been caused by faulting rather than
by folding, and that there have been two distinct movements.
The first would account for the difference in the stratigraphic
horizons on which the lavas rest, respectively, in the Black
Hills and Black Mesa, while the second faulting displaced the
lavas and initiated conditions favorable to the development of
the present topography.

The surface underlying the basalt in the Black Hills has not
been definitely described, but as the character of this surface
is known both to the east and west, it may be supposed that it
is about as thoroughly peneplained as is the corresponding sur-
face in the Black Mesa. In the Bradshaw Mountainst is
found the southwestern limit of the peneplain. The floor on
which the lavas rest, in the region east of the mountains
proper, is thus described :

“ The thickness of the wide eastern basalt flows is very vari-
able; the upper surface is relatively horizontal, but the bottom
fits the hollows in the underlying granite topography. Thus,
near Richinbar the granite reaches the level of the surface of
the mesa at several + points, while near Bumblebee, 2 miles to
the west, the contact of the lava and granite lies 800 feet
lower.”

Elsewhere, if one may judge from the structure sections, the
planation was more thorough. This is shown on those sections
whieh cut Yavapai schist areas. On section C-—C, for instance,
the present surface of the schist departs from a straight line
joining the Bigbug Mesa lava and that east of Oedar ‘Canyon
on an average of about 300 feet, the maximum being 800 feet
at Cedar Canyon. These figures indicate an amount of erosion
that compares favorably with that in the neighboring plateau
district since the eruption of the older basalts and confirms the
impression that considerable areas about the Bradshaw Moun-
tains may formerly have been peneplained.

The entire belt of country immediately south of the Mogol-
lon Mesa is, from the point of view of this article, an unknown
region. The only description suggesting planation is that of
Marvine yt of the region west of Fort Apache. It is as follows:

“West of the post isa more open country, the floor of which
tends to be of the Red Wall limestone, though the preserva-
tive effect of the hard basalt has been to keep it covered with
mesa-like buttes and ridges of the banded red and yellow
Lower Aubrey beds...

* Huntington and Goldthwait, op. cit., p. 225 and fig. 5.

+ Jagger and Palache, Bradshaw Mountains folio; Aziz:, No. 126, U. 8!

G.S., 1905.
tOp. cit., p. 219 and pl. iv.

122 Robinson—Tertiary Peneplain of the Plateau District.

Where crossed, the Salt River flowed in a canyon in the
Red Wall limestone, the northern side of which, being
descended after night-fall, was not examined. It was much
lower than the southern side, however, which reached a height
of nearly 1,800 feet above the stream. Of this about ten or
twelve hundred feet were of Red Wall, capped by some sand-
stone, the remainder being basaltic lava. The limestone
must all be affected by a strong, constant northeast dip, for,
passing over the high basalt top southward 2 or 3 miles, the
Lower Tonto (vitreous) sandstone is encountered at a height
equal to that of the limestone near the river, the whole thick-
ness of the latter, together with that of the Tonto shales, being
covered with lava.

The Bradshaw Mountains, and the locality just described,
lie about 30 miles beyond the border of the plateau and appear
to mark the limit of the peneplain toward the south. That it
did not exist in the vicinity of Clifton, Arizona, nor apparently
at Globe, is indicated by the following quotation :*

nauluive ‘country W was now (in the Tertiary) a land area with
rough topographic features, to which the faulting contributed
important elements. An active erosion following the epoch of
faulting has not been able to entirely efface its influence, and,
though obscured by the Tertiary lava flows, these fault carps
are still dominent features visible in the granite bluffs of the
Coronado and Copper mountains, and the down thrown valley
between them.”

It would seem, however, from the thoroughness of the plana-
tion of the plateau, as though in veneral erosion must have
reduced the relief of the Basin Range country to a mature
type, and it would follow that many intermontane valleys were
thoroughly graded contemporaneously with the development
of the ‘peneplain i in the plateau country.

Conclusion. —The area covered by the remnants of the
peneplain in the southwestern part of the plateau is about 5,000
square miles. Such isolated masses of lava-capped red shale
and sandstone as Red butte and others, at present unmapped,
clearly indicate that the peneplain originally extended over a
much larger region. The boundaries of this area may roughly
be considered on the west as the boundary of the plateau, on
the north and northeast the Kaibab Plateau and the high eliffs
that limit the Grand Canyon District and San Francisco Plateau
in those directions,.on the south and southwest a line about
30 miles beyond the border of the plateau and parallel to it,
comprising in all some 20,000 square miles of country. In the

* Lindgren, Copper Deposits of the Clifton-Morenci District, Ariz., P. P.,.
No. 43, U. 8. G.S&., p. 95, 1905.

Robinson—Tertiary Peneplain of the Plateau District. 123

southeastern part of the plateau about Mount Taylor the area
is estimated at not less than 5,000 square miles. The region
over 100 miles wide inter vening between these two areas may
have been a pene pale or lowland. The encanyoned course of
the Zuni River,* and also of the Puerco of the West, across
the Defiance monocline and against the dip of the str ata, sug-
gesting a superposed origin for these streams, points to this
conclusion. In fig. 5 the location of the various localities that

ARIZONA

Fic. 5. Southern portion of the Colorado Plateaus in Arizona and
New Mexico.

have been described is shown and the approximate position of
the boundaries of the peneplain indicated by the dotted line.
While it is possible to show the existence of the peneplain
at many localities, the question whether the several parts
represent a single peneplain or not is more difficult to answer.
Incomplete knowledge concerning the extent of the later fault-
ing and warping by which the peneplain was raised from a low
altitude to its present height and different portions of it rela-
tively displaced, introduces an important element of uncer-
tainty. The Mohave peneplain, as represented by the surface
underlying the older basalts of the Sheavwits and Uinkaret
plateaus, is certainly a unit, as is also the area in the Little
Colorado Valley. The portion in the Black and Mogollon
mesas and Black Hills is a unit, and the same appears to
be true of the Mount Taylor region. It is believed that they

* Dutton, op. cit. (b), pp. 145, 146.

124 Robinson—Tertiary Peneplain of the Plateau District.

all are the eroded remnants of a single pe pele, with the
possible exception of the Little Colorado Valley area. There are
a few isolated facts regarding the actual elevation of the pene-
plain above the present plateau surface and the relation of the
different parts to one another, that point toward the existence
of two separate peneplains, ‘the older of wide extent, the
younger local in character. The conclusions that would fol-
low were such the case, involving the extension of the older
peneplain over an area much larger than that described and
introducing greater detail into the geologic history of the
region, make it desirable not to go into this matter until
the facts that appear to support it have been verified and much
more field work has been done.,

Influence of the peneplain and later displacements upon the
drainage system of the Grand Canyon District.—That the
Grand Canyon District was formerly thoroughly peneplained
appears certain, and the surface was developed for the most
part upon the soft strata- marls, shales and sandstones—oft the
Moencopie and Shinarump ‘formations. ne slope, as deter-
mined by Huntington and Goldthwait for the Mohave pene-
plain, was toward the south, and it would seem natural that
the principal streams, fol lowing the maximum slope, likewise
should have had courses to the south or southwest. After the
faulting, which raised the peneplain above base-level, the well-
recognized tendency of streams to readjust themselves under
such conditions would become operative, and an entirely new
drainage system would develop.

The effect of the faulting was to raise the eastern part of
the district above the western and apparently to produce a
slight, though not constant, tilting of the surface to the north.
The streams that would develop on this new surface might at
first be controlled, for instance, by the lava-capped areas, but
eventually they would cut down to the Upper Aubrey cherty
limestone, or other resistant rock. Upon this they would either
be superposed or become adjusted to the structure produced
by the folding and faulting. The large streams would most
likely be partly superposed, partly subsequent, while the small
streams would be subsequent, provided the rate of uplift were
not too rapid.

The Colorado River may be considered as a superposed
stream in the vicinity of the Kaibab Plateau. The peneplain
died out against ths flanks of the Kaibab at what is now an
elevation of about 7 ,000 feet, though originally, of course, the
elevation was iat less. When the faulting and tilting occur-
red a depression was formed between the peneplain and upland
that marked out the course of a master stream. As a result
of this uplift there would be a slight ponding on the east side

Robinson— Tertiary Peneplain of the Plateau District. 125

of the Kaibab until the waters rose to the highest point on the
boundary at the southern end of the plateau, “when overtlowi ing
this point they would continue northwestward. What may
have determined the southwestern course from Kanab Creek
to the Uinkaret Plateau is uncertain, altliough it may have been
influenced by the lava flows now ‘vepresented by the eroded
remnants in Mounts Emma, Logan and Trumbull. At the west
side of the Uinkaret Plateau the river turns southwest and
then south to Diamond Creek, where it changes its course
abruptly to the northwest, which direction it holds until it
reaches the Grand Wash and passes beyond the border of the
plateau. The course of the river was observed by Gilbert*
in 1871 from Diamond Creek, and his description may be here
quoted :

“ At the mouth of Diamond Creek center four valleys of
denudation, whose relation to the structure of the Plateau is
very interesting. Two of these valleys are occupied by the Colo-
rado and are portions of the Grand Canyon; that is to say, the
river turns here abruptly at a right angle, and its two courses
appear to have been determined by distinct causes. The third
valley is the canyon cut by Diamond Creek, and is a prolonga-
tion eastward of the lower course of the Colorado. The fourth is
that of Peach Spring Wash, and prolongs southward the upper
course of the Colorado .... The character and remarkable
straightness (of Peach Spring Wash)—I refer to its general
course, and not to its Tee due to the fact that it was
primarily determined by a fault; . . . its throw is to the west,
and the amount of its dislocation near the river is about 600
feet. Its strike is north 25° east: For a few miles at least it
is included in the upper course of the Grand Canyon, and, as
I looked down Peach Spring Wash, and commanded with my
eye a long vista of the canyon beyond, I was strongly impressed
with the idea that the dislocation that had determined the one,
had also marked out the other for a long distance. Later geo-
graphical determinations show that from the neighborhood of
the Uinkaret Mountains to the mouth of Diamond Creek, a
distance of thirty miles, the general direction of the canyon is
straight and coincident with the observed trend of the fault.
If what now appears probable shall hereafter be demonstrated
—tlhat the canyon for a long distance follows closely the line
of faulting—the necessary deduction that the fault antedates
the beginning of the canyon will be an interesting addition to
the cronology of the river.

“The identity in direction of Diamond Creek with the lower
course of the Grand Canyon is not a mere coincidence, but

* Op. cit., p. 79. See also Mt. Trumball and Diamond Creek topographic
sheets, U. g, Gass

Am. Jour. Sci.—Fourtu Series, Vout. XXIV, No. 140.—Avaust, 1907.
9

126 Robinson— Tertiary Peneplain of the Plateau District.

depends on their common relation to the Plateau structure.
The Aubrey Cliff, which crosses Arizona in a northwest diree-
tion, here intersects the Colorado. Since the general dip of
the strata is to the north, and the escarpment is Gue merely to
their unequal denudation, there lie, at the foot of the escarp-
ment, a series of cnoraarelinal valleys, of which the Tonto Basin,
the Upper Verde Valley, and Aubrey Valley are examples.
Diamond Creek runs, in like manner, parallel to the cliff, and
differs from the others only in having excavated a deep gorge,
which its low level of discharge enabled it to do. The same
Aubrey Cliff that follows its northern and northeastern mar gin
reappears beyond the Colorado, and, for forty miles, bears the
same relation to the lower course of the Grand Canyon, leading
to the belief that the stream was here guided, at ‘the first, by
the monoclinal valley, and that the Aubrey Cliff, as a topogr aphic
feature, is more ancient than the Grand Canyon. The cliff
now rises from three to five miles back from the brink of the
canyon, and may be supposed to have retired to that position
by slow waste during the excavation of canyon.’

Gilbert’s observation in regard to the course of the Colorado
between Diamond Creek and the Uinkaret Plateau—that it
coincides with a fault—has been verified by the later work of
Dutton* on the Hurricane fault. While the river from Diamond
Creek to the Grand Wash, in the ight of present knowledge,
should be regarded as adjusted to the structure underlying the
peneplain rather than guided by a preéxisting valley. The
description is very clearly that of a subsequent stream, and as
it applies to one-third of the course of the Colorado in the
Grand Canyon District, its bearing on the origin of the river is
evident.

If the boundary between the peneplain and lowlands in the
Little Colorado Valley lay not far to the east of the river, as
the observations at Black Point seem to show, then, when the
uplift and tilting occurred, a northwestw ard sloping depres-
sion would be formed along the boundary. The waters of the
streams on both slopes of this depression would be gathered
along its line of lowest level into a trunk stream, which would
flow northwestward until the unfavorable southward slope at
the northern boundary of the peneplain was met. At that
point it would be turned sharply south, joining the westerly
flowing consequent that would there be developed. Or, if the
peneplain extended still farther east than is supposed, the
Little Colorado may be even more simply explained as a con-
sequent that developed on the uplifted peneplain, and is now
strictly governed by the underlying rock structure. As the
dip of the strata is in general northeastward, the river in

* Op. cit. (a), chap. vi.

Robinson—Tertiary Peneplain of the Plateau District. 127

either ease would occupy a monoclinal valley, like the Colorado
from Diamond Creek to the Graud Wash, the eastern wall of
which would be well marked where resistant strata were
encountered, or indistinct where they were soft.

The actual conditions appear to agree closely with the
inferred. The profile of the river is quite unique in that it is
a distinctly reversed curve, the upper and lower ends being
much steeper than the intermediate portion. This may be
illustrated by the following figures: For the first 40 miles,
measured from the mouth of the river, the average grade is
34 feet per mile, for the last 40 miles it is 55 feet, while for
the intervening stretch of 200 miles the grade is but 9 feet per
mile. In more detail the grades of the first 200 miles are:
(a) 25 miles at 35 feet per “mile, (0) 15 miles at 31 feet, (c) 35
miles at 15 feet, (7) 100 miles at 7°5 feet, (¢) 25 miles at 18
feet, with the remaining 80 miles at increasingly steeper grades.
The two points to notice are the steep grade for the first 25
miles (a), and the low grade for the 100- ile stretch (2). The
first shows the effectiveness of the heat y Carboniferous forma-
tions, in which the river has cut a precipitously walled can-
yon, in preventing a normal grade from being developed in the
lower course of the stream ; “the second illustrates the same
point, since it has resulted from the river being obstructed by
the same formations in the vicinity of the “Crossmg” north
of Canyon Diablo. The river is thus closely controlled by the
strata it flows upon, and is far from uniformly graded. The
character of the river, near Winslow, in section (d ) has been
described by Marvine.* He said: “ "At this point, and above,
the river is not at all eral of this region, inasmuch as it
does not flow ina canyon. On the contrary, its flood plain is,
in places, vver a mile or more in width, through which it
winds its muddy current in quite a tortuous course. It prac-
tically occupies a monoclinal valley between the Carbonifer-
ous and Triassic.” The river continues in this monoclinal
valley until the deep canyon at its mouth is reached, and of its *
course in the vicinity of Tanner’s Crossing, Davist has said
that “It may be plausibly regarded as a subsequent stream ;
such was certainly its habit where we crossed it.”

The river in its course northward passes through the Shina-
rump and Moencopie formations, and especially in its lower
course, though also between Grand and Black Falls, through
Carboniferous. The western slope of the valley is clearly

a stripped structural surface, and, except near the river, is
marked by the Upper Aubrey cherty. limestone. The north
side of the valley is limited by a series of prominent cliffs,

* Op. cit., p. 214. + Op. cit. (a), p. 158.

128 Robinson—Tertiary Peneplain of the Plateau District.

mostly of Triassic sandstones, from Moencopie Wash southward
to the vicinity of Winsiow. From there, south and east, the
soft shales and marls overlying the sandstones are encountered
and the cliffs become much diminished in height.

The smaller streams of the Grand Canyon District appear
to be, without exception, consequent in origin. Davis* has
already ascribed such an origin to several on ‘the north side of
the Colorado. Among others are Peach Spring Wash and
Diamond Creek, deseribed by Gilbert in the passage quoted.
All the washes that head in the Black and Mogollon mesas
and run northeasterly into the Little Colorado are clearly con-
sequent. In the vicinity of San Francisco Mountain the
courses of Hull Wash, Oak and Sycamore Creeks are governed
by faults. A similar example is Bright Angel Creek, as shown
on the recently published B right Angel topographic sheet, and
the fault that has marked out its course may be traced south
of the canyon by the line of the Grand Canyon Railroad.

The antecedent origin of the Green and Colorado rivers
advanced by Powell and Dutton, has been generally Aeon
for many years. It is to be noted, however, that Emmons,
as the result of his work on the Tanta Mountains in 1871,
ascribed a superposed origin to the Green River, while Gilbert,
in 1876, said that :

“A large share of the drainage of the Plateaus is not conse-
quent. How much is superposed, and how much antecedent
remains to be determined. With the solution of the problem
are involved the determination of the antiquity and history of
the Green and Colorado rivers, and the physical history of the
great Tertiary lakes.”

Jefferson, in 1897, advanced “an independent and confirm-
atory argument for the consequent origin of the Green-
Colorado... based on the curvature of the river and especially
on its meanders. The meanders are surprising in themselves
from the steepness of the river and their deep incision; yet
more surprising is the location of strong meander reaches just
up stream from structural displacements.” Davis, in the
same year, pointed out that the origin of the Green River was
still an open question, and after visiting the Grand Canyon
District in 19009 wrote:

“That the facts now on record, combined with such knowl-
edge of the region as our party was able to gather, warrant
the consideration of at least one hypothesis alternate to the

* Op. cit. (a), pp. 153-157.

+ Explorations of the 40th Parallel, vol. ii, pp. 194, 205.
{This Journal, 3d ser., vol. xii, 1876, p. 102.

§ Science, vol. vi, 1897, p. 298.

|| Ibid., vol. v, 1897, p. 647.

§] Op. cit. (a), p. 166.

Robinson—Tertiary Peneplain of the Plateau District. 129

theory of antecedence, as an explanation for the origin of the
drainage lines in the Grand Canyon District. I do not, on the
one hand, consider the antecedent origin of the Colorado
disproved, but, on the other hand, such an origin does not seem
compulsory.”

It would seem, however, that certain facts, such as the
thorough peneplation of the Grand Canyon District, with the
uplifting and warping of the peneplain by later faulting, the
dependence of the course of the Colorado upon structure, as
described by Gilbert, and the consequent origin, so far as
known, of all branch str eams, furnished sufficient evidence for
definitely abandoning the idea of the antecedent origin of the
Colorado River and for substituting a consequent origin in
which superposed and subsequent conditions are both present.

While it is believed that the general question of the origin
of the drainage system of the Grand Canyon District rests on
a rather definite basis, it is realized that the number of facts is
at present much too small to allow exact reasons being given
for the location of considerable portions of the courses of the
Colorado and Little Colorado, though not of the smaller
streams. The problem involves the study of a large area, and
it is hoped that whoever may have occasion to visit the region
will succeed in gathering information that will assist in its
correct solution.

Yale University, New Haven, Conn., June, 1907.

130 Mixter—Combustion of Silicon and Silicon Carbide.

Art. XII.—TZhe Heat of Combustion of Silicon and Silicon
Carbide; by W. G. Mrxrer.

[Contributions from the Sheffield Chemical Laboratory of Yale University. ]

Tue direct determination of the heat of combination of
silicon with oxygen presents many dithculties. The reaction
occurs only at high temperatures; the product is not volatile
and hence encloses silicon and prevents complete oxidation.
Troost and Hautfeuille* from the thermal effect of the reaec-
tion of silicon tetrachloride with water obtained 227000° for
the formation of silicon dioxide from erystalline silicon. Ber-
thelot’st result by the same method was 179600 for the forma-
tion of gelatinous silicic acid from crystalline silicon. Ostwaldt
states, “die Bildungswi irme siimmtlicher Siliciumver bindungen
ist noch sehr unsicher.”

Preliminary tests made by burning a mixture of silicon and
carbon in the bomb showed that more than half of the former
could be burned. The method would give good results if the
solid products contained only silica, silicon insoluble in hydro-
fluoric acid and silicon carbide. The last was always present;
hence its heat of formation was determined, as described later,
in order to have the data needed for the work on silicon.

The writer is indebted to the Carborundum Company,
Niagara Falls, for an abundant supply of carborundum powder
and silicon for the investigation. Crystallized silicon was pre-
pared as follows: One part of silicon was dissolved in four
parts of molten aluminum and the mixture allowed to cool
slowly: the metal was dissolved out with hydrochloric acid, and
the silicon, which was in small crystals, was digested atin hot
hydrofluoric acid. The product at this stage contained 2°8 per
cent of aluminum, Treatment with molten potassium pyro-
sulphate reduced the aluminum content one-half. Next the
crystals were pulverized in an agate mortar and the finer por-
tions separated by levigation for one minute in water. Finally
the powder was digested with a mixture of hot sulphuric and
hydrofiuoric acids, washed and then heated to drive off the acid
not washed out. Thus prepared, the silicon contained a little
aluminum, a trace of carbon and 99°95 per cent of silicon. The
last was determined by dissolving the powder in a hot concen-
trated solution of pure sodium hydroxide and making the usual
separation of silica. The carborundum was purified by means
of molten pyrosulphate and, after washing, it was floated in
water and only the finer por tion retained. It was next treated
with nitrohydrofluoric acid to remove silicon and silica, 40

=O. Re lexeoe.
+ Thermochimie ii, 125, Ann. de Chimie et Phys. [5], xv, 213.
t Grundriss der allgemeinen Chemie, dritte Auflage, 268.

Mixter— Combustion of Silicon and Silicon Carbide. 131

grams losing only 0°5 gram. It was finally digested with only
hydrofluoric acid, but there was no further loss in weight. The
purity, of the product was found as follows: 0°5254 gram was
fused in a platinum crucible with 5 grams of potassium nitrate
and 10 of sodium carbonate, and the silica was separated with
the usual precautions. Silicon found 69°9, caleulated 70°3 per
cent for SiC; Si=28-4.
Combustions with Sodium Peroxide.

The bomb used in all of the work was one described in the
paper on carbon.* It was lined with silver cups 1™™ in thick-
ness. Owing to the danger of melting the lower cup the com-
bustible mixture was usually placed in a thick silver dish
resting near but not in contact with the bottom of the bomb.
This dish was usually melted by the combustion. Silver, as is
well known, is oxidized by molten sodium peroxide, but the
amount of oxide formed is small and the error caused is slight.
A nickel dish was tried, but this was attacked much more than
the silver one. After a combustion with sodium peroxide the
fused mass resulting was dissolved in a large quantity of water
and the insoluble tnburned substance was collected, purified
and its weight deducted from the amount of material used for
the experiment. In case of silicon some may have dissolved
in the dilute alkaline solution, but this was not likely, as silicon
is taken up slowly by concentrated alkali. In order that the
iron used for ignition should burn well the bomb was filled
with oxygen at atmospheric pressure. The carbon used was
the acetylene modification described by the writer.+

Experiment 1.

(Geel OO aS es ac A ING cD, debt ca he 1:0530 grams.
Cre MOtsOUEME Ci aat vcs salen aii he tase se —0°0148
Se MD UTM CMe ee tes is Sone loins sae eed as OS Sika neice
mono omition = sss lest ie Nees ent ons 0-072 st
Sodium peroxide wy wee ey Meera ik 19 “
NV NfeaiCre Nei elses poet SR ue rn MRI YL AN 2954 af
Sy equivalent ofrcalorimeter 9222 a2 a Su s
me 5 * sodium peroxide and carbon 4 Me
2 r PALS ISLES) Oy ores eee esate: Me veg n ea riaalta 3269 “
Minutes. Temperature. ’ Temperature interval.
0) 18°452 22°012—18°458 + 0:009=3°563°
1 18°454
2 18°458 Heat observed 3269 x 3°563=11648°
3 22°012 ‘* of oxidation of iron —115°
9 22-012
10 22°009 11533°
11 22-006 :
12 22°003 For 1 gram of carbon 11103°

*This Journal, xix, 435. + Loe. cit.

132. Mixter—Combustion of Silicon and Silicon Carbide.

Experiment 2.

Carbone. jon = ves Rc Sane reas pia Sai era 1°2214 grams.
not burned’: 2.22025 ste eee ee 0500 ane
See NUTR yo lit a He We Saag Seat APN OA
iron for vomitionse sae] 22 oe eae 0:060 aS
Sodium peroxide yew yee eee eae eee 21 rie
Water equivalentyol system) se .2s es) ae ae 3244 ¢
Minutes. Temperature. Temperature interval.
0 18241 22°405 —18°245 +.0:032=—4:199°
1 18°243
2 18°245
15 22-405 Ileat observed 3244 x 4:192=13599°
16 22-404 ** of oxidation of iron — 96°
17 22-399 ——
18 22-397 3503°
19 22°394
20 222391 For 1 gram of carbon 11149°
21 22°388 The motor stopped for a time after the

ignition, hence the interval before the
maximum temperature was observed was
longer than usual.

Experiment 3.

SHleom ys Ses Bee a es eS Ee ee EO 3 OMS

Cowes TNO. UNI e Chine sic s2 756 es anes ye ee ter eh wea —0:0003. “

CANOE ON) 0X21 | area eernn nna enh MIS nae ys ae KN es Dis QiBil Taree
Tronsforomition ies C,2 ae ae oe ee eee ner OOO. sf
Sodium peroxides acer wees ie Uae ee Gra os
Water equivalentiof system)2 424-222 2232). ose 3380°6 a
Minutes. Temperature. Temperature interval.

0) 18°176 21°242 —18'184 + 0.009=3:067°

1 18°180

2 18°184 Heat observed, 3380°6 x 3:067=10368°

8 21-949 “of oxidation of iron —93°

9 21:240 mead SE aEO

10 21°238 10275°
1] 21°236
12 21°234 For 9 grams of silicon _—8342°

13 21°238

Mixter— Combustion of Silicon and Silicon Carbide. 185

Experiment 4.

DS VULT GVO my cc Se aa as ee Ua CU aan 1°4934 grams,
SE TIONG | OBIE) OVO bss att RUSS a cle al ORO BANE 2
Ss) ORT AY 2X Ue iat me is a alee 2 14903
Mr OMBHOT LOMUGTO ME A Sete Gy Sa gene 0°050
Sodinimmeperoxider ease e eee 15 hs
Water equivalent of system... .---.4.._--- 3454 ‘S
Minutes. Temperature. Temperature interval.
0 18-425 22°060—18'431 + 0°008=3°637°
1 18°428
2 18°431 Heat observed, 3544 3°637 12562°
8 22-060 ‘* of oxidation of iron —80°
9 22°059
10 22°058 12482°
11 92°056
12 22°054 For 1 gram of silicon 8379°
13 22°052
Experiment 4,
nC OnmCaT bide aie) wae LS Ou ie 1:105 grams.
Co IIe oad OCOL ENG ONDE Ol O(EU0 Vahaiodlian + Tia ae MeN EE ele OnOMor is:
oe SSibe UO UTI CO: satiate eM SLO TO 1:030 se
Imongtorionition 90) wus Mis, OsO72 hss
DPOdiumperoxide we il een Al ae 11 «6
Wiaterequivalent oftsystem 2640220 22.05 ls. 3565°6 te
Minutes. Temperature. Temperature interval.
0) 18°442 91°120—18°450+ 0:007=2°677°
1 18°446
2 18°450 Heat observed, 3565°6 X2°677 9545°
8 212120 “of oxidation of iron —115°
9 217120 :
HORE 21°118 9430°
11 21°117 :
12 21:116 For gram of silicon carbide 9156°
13 21°415
Experiment 6.
Silicompear birders + Arse e ei s eC Eyes 1°3987 grams.
as Ce NO tIDUEINE Css ince ec eea ua anew O;OOM3 bie
Be SSeS UT IN i peter a xa Se Aaa Be a THORS NG
Tron forte mitiomesroe pee yuk a ae uate ape 0-0534 5 5
Sodium: peroxides spss ye ay ee ane ne 15 as

Water equivalent of system.__-......-...-- 3432°8

134 MWMixter

Combustion of Silicon and Silicon Carbide.

Minutes. Temperature. Temperature interval.

0 18-126 21°836—18°132 +0°013=3°717°
1 18°128
2 Use Heat observed 3432°8 X3:717=12760°
8 21°836 * of oxidation of iron —85°
9 21°832

10 21:829 12675°

aL 21°827

12 21°825 For 1gram of silicon carbide 9110°

13 21:923

The results of the combustions with sodium peroxide are as
follows for 1 gram of each substance :

Acetylene Crystaliine Silicon

carbon silicon carbide

11103° 8342° 9156°

11149° 8379° 9110°

Mean 11126° 83605 9133°
For 12 grams of acetylene carbon --------=-- 133512°
ip 20s4 Wie ea chy Stallime siliconmecmar era 237424°
370936°
For'40°4 grams of silicon carbide.2 222... 522055 22 368973°
IDifferenGess6 2 eis eee a ees Ne aera 1963°

The above results indicate that the formation of silicon
carbide is accompanied with but little thermal change. The
heat of combustion of the diamond according to Favre and
Silbermann is 93240°, 1488° less than that of acetylene carbon in
oxygen; assuming, ‘if the diamond had been burned with
sodium ‘peroxide, that the result would have been 1488° less, we
have 515° for the heat of formation of silicon carbide from
the diamond and crystalline silicon.

Using Thomsen’s* data for the union of sodium oxide and car-
bon dioxide, deForerand’st for the heat of formation of sodium
peroxide from the oxide, and the heat of combustion of acety-
lene carbon, the ther mal effect of carbon and sodium peroxide
may be calculated thus :

Na,O, CO, =e 2759206
C70; = 94728°

170648°
NOOSE
2Na,0,, C = 12s

*Thermochemische Untersuchen iii, 233.
+C. R. exxvii, 514.

Mixter— Combustion of Silicon and Silicon Carbide. 135

The agreement of the calculated with the experimental result
indicates, in the reaction between carbon and an excess of
sodium peroxide, that sodium metacarbonate is formed and not
the orthocarbonate. The latter is not known, but if it may
result by the union of the metacarbonate and sor hania oxide the
thermal effect would be about the same as that of the oxidation
of sodium oxide.

Combustions in Oxygen.

The combustions were made as follows: A mixture of acety-
lene carbon and silicon carbide or silicon was placed in a
shallow silver foil dish supported near the middle of the bomb.
The object was to expose a large surface of the mixture when
burning to oxygen. Naphthalene and anthracene were tried
but did not answer as well as finally divided carbon. That the
temperature was sufficient to volatilize silicon was shown by
the deposition of silica on the sides of the bomb in the form of
a white powder. It settled very slowly in water and passed
through filter paper. Under the microscope it was found
to be spheroidal and non-crystalline. The silica from burn-
ing silicon carbide was all in a finely divided form, while
in that from. silicon elobulles 22) on 322 ian diameter were
found. The residue left by the combustion was treated
with nitric acid to dissolve silver present and, after thorough
washing, was repeatedly digested with hydrofluoric acid
until its weight when dry was constant. Finally the residue
was treated with nitrohydrofluorie acid, the loss being silicon
and the powder remaining silicon carbide. The thick platinum
electrodes, placed 4 or 5°™ above the combustible mixture in
the bomb, were more or less melted by the combustions. In
one instance a globule of platinum silicide formed which was
found to contain 7 per cent of silicon. In no instance was
silicon found in the platinum melted by the burning of silicon
carbide.

Beperiment 7.

SUlTCoOmecat 1d eis he eee eater rene Se aera 0°9820 grams,
ce CC OTT ST CU Cee eae eae ah anceasheN 0°1700 =
Silicon 33 Ose tik one RSE ea a da 0:0068 =“
Silreonscarbide: burned vse eee et eae ae 0°8052 *
fronforicmitrony s aeeire aie ees 0:°0466 *
Oxyoen oie: BSCR basalt etNos eave tee AR. 11 a

Water and water equivalent Of Systema!) 3). 3433 ¢

136 Mixter— Combustion of Silicon and Silicon Carbide.

Minutes. Temperature. Temperature interval.
0 18013 21°520—18°0144+ 0°015=3°521°
I 18°014
2 18:014 Heat observed, 3433 <3°521 = 12088°
7 21°520 —
8 21-517 ‘** of oxidation of iron, —75°
9 ZS Seca |S Ss * carbon —6331°
10 21°509 ————
ll 21°505 Scopes’ cf “* silicon carbide 5682°
For 1 gram of silicon carbide 7057°

Experiment 8.

SMSO AWKCL WL) KO (2 eres Mareen eas A ANC EU he. Ae, 1°5029 grams.

eS 9G May. URASIMONI saa, gw Se es oS DO) cs

Silicon OC SSH = Bec ERNE Nein SeTESER IRD Neen 0°0470 se

@Warbon 2 ie Tan eer 2 OMe iP eee eee Ooh ee) Oil eames
For 1 gram of silicon carbide 6659°

Experiment 9.

Silicon:carbide sae sot eee eee Oa Aamo

ce a TV NESUCI Gy aes ae se 07158 ec
Silicon Ob SST Lier DL cmt ah oan We Ne SAS AM 0:000 OG
Dilicon*canrbidesbmEmed sees ee pale een ate O°5 86sec
@arbon ys eek eeu) a ee ete heen eee eee EOS ec
Irom Forse nit om cpt ss Sastre egies cee en nea aaeoee 8 0049 =“
Oxy Cenees Lay Sea Escher aspen wei eee Tie “
Water and water equivalent of system .------- 3448.6 ee
Minutes. Temperature. Temperature interval.

(0) 18°485 21°818—18°488 + 0:007=3°337°

1 18°488

6 21°818 Heat observed, 3448.64+3°337 = 11508°

iT 21°816 “ of oxidation of iron —78°

8 91°814 SG) 50 OS “ carbon —7349°¢

9 Aes SS

10 21°S811 <a ce “ silicon carbide 4081°
For 1 gram of silicon carbide —6 963°
Experiment 10.

MLITCOMMC AL DICE 08: ee Becerra en pa 11125 grams.

& GG alate ek IG Hh (epee iy eae net he os koe es 0°0436 Ge
Silicon cc COG WIN COs Fae ape ee ropeatee ak ek eae 0°0422 Se
GE Fk G0) 0 Wah eaten As ren eee renee Une Nye bce COs 0°944 UG

For 1 gram of silicon carbide 6882° |

Mixter— Combustion of Silicon and Silicon Carbide. 1387

Experiment 11.

Silom aN Ne este Steeler ome ier il hatha Ney ORY Abeer ie Aue 1:0130 grams.

ce ue TiO PESREVSITCO IUD ayy ce ace ies ILS Na 0°0984 “
Silicon se CO ea ORANG or. I al eM MaapaDnay eaneatel oa 0:0089 a
SiliconecarbotdemouEmeduns 0 quo ose mmc lees smen ne (OO () if “¢
(Berra one ae lane Va EL RUN Reece UN 1°1220 &
IOnMPOR Tomtblone ee UN ee AON REN Ghee s2 0°0530 ec
Org ioce rae S ee eh eet O00 tic A ee 11: 6<
W ater and water equivalent of system asia 3387°5 Gs
Minutes. Temperature. Temperature interval.
0) 18°308 22°573 —18°314+0°012=—4:271°
1 18311
2 [S34 ge eatTobsenvied: stees) ice Milanese sei 15322¢
Li 99-513 Le JOMMOXIGAGLON OF MIRON eae ae — 85°
8 OORT Scns ss OH Wel OXO) Ay ees — 8857°
9 22°567 ; Ce cael
10 92°568 Sera se “ silicon carbide 6380°
1l 22°561

For 1 gram of silicon carbide 7044"

If the weight of carbide equivalent to the free silicon left by
a combustion be deducted from the amount of carbide taken
for an experiment, and the carbon equivalent to the free sili-
con be added to that taken, the results caleulated on this basis
will be essentially the same as those given. The following
table shows the proportions of silicon canbe and free silicon
remaining after the combustions:

Silicon carbide. Silicon.
Experiment 7 7057° 18 per cent. 0-7. per cent.
66 8 6659 95 66 G6 3 ce ce
(<4 9 6963 920 ce 66 0) ce ce
ae 10 6882 Sis screamo SeSugiee nies
(73 li "044 10 (73 6¢ 0°38 ce ce

It was found in the work on silicon that where the products
of combustion contain much free silicon there is a variable
error. For this reason the figures of experiments 8 and 10
should not be included in the final result. The mean of the
other three experiments is 7009 calories for the heat of combus-
tion of 1 gram of silicon carbide in oxygen at 20°. For 40-4
grams it is 282164 calories at constant volume and 283750 cal-
ories at constant pressure.

icon and Silicon Carbide.

d

FSi

zon O

ter—Combust

q

M

138

vu

oh 869

6670-1

0096-0

06 F0-0

OF0<6-0

1606-1

8

ay} JO s[Byop oy} Uosvot sty] AtoZ puw ‘sy~nsea tood poppetd uosAxo Ut

GL06-9

8910-0

6916-0

GEG6L-T

AT

09609

9498-0

8860-0

POST-O

8970.1

OT

06469

0693-0

F89L-0

+660-0

O¢TT-0

8906-0

98999
0089-0
6990-1
8640.0
LEGEé-0
FO6F-T

WAL

00904

LPOF-0

0660-0

0060-0.

166-0

08099

6600-1

080-0

920-0

110-1

oL

:S][NseI oY} JO JUOUOJLIS SULMOT[OF ot}

---UOdI[IS JO WRAD [ IOF YN
Seedo coos ceaoudeds Udyvy UO
Se eee See eee poumnq Uodt{Ig
oe Ole so
ones tee: anpised UL ”
eed wees eens eee Woyv} UOOI[IG

spuauUlnsadany

UL PozJIWUO DIB yy LOM DLIFOUULLO[BO

UODITIS JO SUOTISUG WO0d OLE

Mixter— Combustion of Silicon and Silicon Carbide. 1359

It was observed that the products of some of the combustions
of silicon gave off a gas when digested with cold dilute hydro-
fluorie acid. This gas was presumably hydrogen or a hydro-
earbon resulting from the interaction of a soluble modification
of silicon or a silicon compound and the acid. This is the
most probable explanation of the wide variation in the results.
The substance noted as silicon carbide in the foregoing table
was the residue insoluble in nitrohydrofluoric acid. It was
repeatedly found to yield carbon dioxide when fused with lead
chromate and in one portion 68°7 per cent of silicon was ob-
tained: calculated, 70°3 per cent for SiC. The carbide from
the combustions was in the form of a light grey amorphous
powder.

The heat of oxidation of silicon may be calculated as
follows :

SiC, 20, = 283750°
Sia, = 1963°
285713
CxO. == 94728°¢
Si, O. = 190985°

Berthelot’s* result is 180593° for the formation of gelatinous
silicic acid from 28°4 grams of silicon.

Lithium orthosilicate is known and it is probable that sodium
orthosilicate results when silicon burns in an excess of sodium
peroxide. On this assumption the thermal effect of the combi-
nation of one molecule of silicon dioxide with two molecules of
sodium oxide may be calculated thus :

oNa,O,, Si = 237000

2Na,0, —O, 38000°

275000
Si, O, = 191000
2Na,0, SiO, = 84000°

Experiments were made to determine the heat of combina-
tion as follows: Silicon dioxide in the form of a bulky impal-
pable powder, obtained by igniting silicic acid resulting from the
decomposition of silicon fluoride in water, was used. A mixture
of sodium peroxide, carbon sufficient to form an excess of
sodium oxide, and silicon dioxide was ignited in the bomb. The
temperature of the reaction was sufficient to melt silver. Less
than 50° of oxygen were evolved. No free silica was found

* Loe. cit.

140 Mixter—Combustion of Silicon and Silicon Carbide.

after dissolving the fused mass in a large volume of cold water.

The mean of two results which agreed blocel y was 71000° for one
molecule of silicon dioxide. The difference of 13000° between
the calculated and observed result is due either to an error of 5
per cent in the heat of oxidation of silicon or to the formation
of a mixture of silicates in the experiments with silica. In the
latter instance the conditions are less favorable to the forma-
tion of only the orthosilicate. Hence we may consider that
84000° deduced from the reaction of silicon with sodium per-
oxide is the better result.

Summary of results.

SiC, 20, = 283800°
Si cry stalline, Camorphous = 2000
Si ee 5 Oy. == OOO?
2Na,0, SiO amorphous = 84000°

The figures given for the heat effect of the union of silicon
and carbon are too small to indicate more than this, namely,
that the formation of silicon carbide from its elements is accom-
panied with very small evolution or absorption of heat.

W. FF. Miilebrand — Vanadium Sulphide, Patronite. 141

Arr. XIII.—The Vanadium Sulphide, Patronite, and its
Mineral Associates from Minasragra, Peru; by W. F.

HILLEBRAND.

Iy the Engineering and Mining Journal, Sept. 1, 1906,
p- 3885, and in Informaciones y Memorias of the Society of
Engineers of Lima, Peru, vol. vii, pp. 171-185, 1906, two
accounts are given by Foster Hewett and José J. Bravo,
respectively, of a remarkable vanadium occurrence at Minas-
ragra, about 46 kilometers from Cerro de Pasco in Peru and 48
from the railway. The two accounts agree in all essential
points that are common to both, though that of the Peruvian
writer is much the more detailed as to the geography and geol-
ogy of the region. The data immediately following are drawn
trom the publications referred to, and from additional infor-
mation furnished by Mr. Hewett.

Cretaceous sedimentaries—shales, sandstones and limestones
—dipping at about 45° have been intruded by two (Hewett)
or three (Bravo) series of eruptive dikes, and at the point of
greatest frequency of these intrusives occurs the vanadiferous
deposit, which had as yet been opened up only very superfi-
cially by a few pits along an extension of 400 feet on the
outcrop. The vanadiferous materials occur in vein formation
and are three in kind, aside from alteration products that cover
the surrounding surface. Under the hanging wall is a thick-
ness of about eight feet of an amorphous material of complex
mineral composition which will be designated hereinafter as the
“ore.” Its color is dark, almost black (dark green like oliven-
ite, according to Hewett, bright lead-gray with metallic luster
on fresh surfaces but soon tarnishing, according to Bravo).
Adjoining this material, without distinct line of demarkation
beneath, is a singular hard coke-like carbonaceous matter, from
eight inches to two feet in thickness, which blends on the far-
ther side into a lustrous black substance of from four to six
feet thickness, designated as asphaltite by both the above
named writers, although it is a sulphur compound of carbon
with very little hydrogen.

These three substances will now be considered in detail,
beginning with the last, which from its unique position among
carbonaceous mineral substances seems worthy of a specitic
name for the purpose of more ready separation from those
bitumens and coals which it so strongly resembles in its super-
ficial aspects. The name guisqueite is suggested by Mr.
Hewett, after the settlement nearest to the locality of occur-
rence. The materials described in the following pages were

Am. Jour. Sci.—FourrtH Series, Vout. XXIV, No. 140.—Aveusr, 1907.
10

142 W. EF. Hillebrand — Vanadium Sulphide, Patronite.

furnished by Mr. Hewett, and a portion of the chemical data
published by him were based on incomplete preliminary tests
by myself. The reason for examining them in detail in the
laboratory of the U. 8. Geological Survey is not only that they
represent an occurrence and association so unique, but chiefly
that an as yet undescribed association of vanadium and carbon-
aceous matter in eastern Utah presents some points of similar-
ity with that of Peru, and by reason of the higher concentration
in the latter they seem to offer a convenient stepping-stone to
future research upon the North American occurrence.

As to the source of the materials composing this vein deposit,
nothing can at present be said, nor is a satisfactory explanation
at hand for the comparatively, ‘shar p separation, in what would
seem to be one vein, of such unlike materials, Mr. Hewett

remarks that ‘the appearance of the materials as well as their
occurrence strongly suggests that they were forced into the
shales in a plastic or even liquid condition. Subsequent meta-
morphism has altered the composition of the asphaltite and the
condition of the sulphur in the sulphide material.” It is con-
ceivable that the injected material was originally homogeneous
and that segregation took place after injection. Possible
evidence in favor of this view is afforded by the fact that the
rich vanadiferous ore contains several per cent of carbonaceous
matter similar in composition to the quisqueite, that the carbon-
aceous matter is itself vanadiferous, and that both are extremely
rich in sulphur. In order to account for the coked matter
between the two, it would seemingly be necessary to assume
that the requisite heat came from “the side of the vanadium
sulphide ore. The adjacent dike undoubtedly played the chief
role in this connection.

This is not the only known association of vanadium with
carbonaceous matter, an association which is undoubtedly sig-
nificant though we are as yet unable to explain it. Aside from
the above mentioned occurrence in Utah, vanadium has been
reported in numerous coals, in erahamite that probably came
from West Virginia, and Mr. Hewett tells me that it is a constit-
uent of the grahamite from Oklahoma. The enormously high
proportion of vanadium in the Peruvian occurrence is of special
interest, not only on that account but because it offers an oppor-
tunity to determine its state of combination in at least this
particular case. It is possible that in some instances such
sources may have furnished the vanadium of deposits of sec-
ondary minerals like those of the western United States.

Quisqueite.

This material, the asphaltite of Hewett and Bravo, though
rich in carbon, is worthless as a combustible because of its

W. F. Hillebrand — Vanadium Sulphide, Patronite. 143

content in sulphur, which exceeds that of the carbon in the
specimen examined. Mr. Hewett reports 45 per cent of
sulphur and that, though not holding a flame, the material
burns when heated with the blue flame of sulphur. It is brit-
tle, with density 1°75 and hardness about 2°5. It is infusible
and apparently not affected by the usual solvents for bitumen,
except that carbon disulphide extracts much sulphur. It leaves
a variable amount of highly vanadiferous ash—3-°31 per cent,
containing 0°544 V,O, (Hewett), 0°80 per cent, likewise high
in vanadium (Hillebrand). Much sulphur sublimes when the
mineral is heated in a closed tube and later hydrogen sulphide
is evolved. It is oxidized by long boiling with strong nitric
acid. The following is the composition of the sample analyzed :

Se solulbille mine @ See ss sees chy 15°44
S_combinedi ais 4 2) aL Se 17
(BUREN SA Secale se RAN en OES oye ep el 42°81
1B Aether ae Cele eae 0-91
ING ee patic aiid 8 heres Sin eran thc tanke sen 0°47
OF Oya diiiey oenpeni ol Sites See cae oO
IM Moristaniren cial MOG eae ys aula Sean 3 01
FANS) Oe Sa OS i ae ean eer ni 0°80

100°00

The ash afforded roughly: Si0,, 0°04; A1,0,, 0:08; Fe,O,,
0:10; NiO, 0:06; V,O,, 0°52. These are, however, results of
minor quantitative value because of the minute amount ana-
lyzed. It is practically certain, for reasons to appear later, that
the vanadium, iron and nickel existed in the original material
as sulphides, and, therefore, assuming that the main analysis is
correct, that the oxygen reported is too low.

The composition of this material is, so far as my knowledge
extends, unique, and there are no data known that throw light
on its chemical structure.

The Coke-like Material.

The quisqueite seems to have been converted on its upper
side into a sort of coke, though this differs markedly from arti-
ficial cokes. Its hardness is 4°5. Glancing blows of a hammer
cause much sparking. It is somewhat vesicular, though without
large cavities. The openings are mostly short thin contraction
cracks without orientation, which are best observed under a
lens. They are for the most part lined or filled with a thin
brilliant pitchy black coating, often reticulated, of what appears
to have been originally fused matter and to which the luster of
much of the specimen when freshly broken open is due. There
are also rather numerous minute blowholes lined with or some-

144 W. EF. Hillebrand — Vanadium Sulphide, Patronite.

times loosely filled with the brittle pitchy matter just men-
tioned. Elsewhere the surface is rather dull and of a color
having a suggestion of red when contrasted with the pitch
black of adjacent parts. The texture is extremely fine-grained
like that of steel aud the fracture rather conchoidal. These
peculiarities of texture often give rise to great unevenness on
large fracture surfaces. In the closed tube there appears a
sublimate of sulphur and then the odor of hydrogen sulphide.
The composition of the sample analyzed follows :

® soluble win S. 2s pees een:
SG allo in © Cl sapere tae ons a ace ey nc Le 5°36
(Os Ne ERA, HIST tie lets Ree ye etl Ol 86°63
|G eeu inate ten bss ais We GS dele Gy) 0°25
IN se A oo pk A res ee O51
Oebyodifiiec ss) et Nc en eereGd
IMOTSture tee Stay, oe pan Niet er ipsa een none
1/2) spies he ninaek eran ap EN Uh ae real re 1:97

100°00

The ash is highly vanadiferous.

THE VANADIUM ORE.

The remainder of the deposit, about eight feet in thickness,
is a complex mixture of mineral substances. The color in mass,
as earlier mentioned, is dark, almost black, with perhaps a
suggestion of green as stated by Hewett. It breaks for the
greater part unevenly, but often somewhat smoothly along
planes of weakness, portions of which show permanent brilliant
metallic luster. This is due to a thin coating of presumably
carbonaceous matter like that noticed in the coke-like material.
Bravo reports enclosures of brilliant black “ asphaltite” and
also sulphur visible to the naked eye. No sulphur was to be
seen in the samples at my disposal,* but there were many patches
and points of a dull pyritous mineral, whiter than pyrite and
usually of a faint reddish cast, apparently due to tarnish. These
consist of an iron-nickel sulphide, of which more later, carry-
ing vanadium and siliceous and titaniferous matter. Heated
in the closed tube, the ore gives off sulphur copiously and then
hydrogen sulphide. It does not melt.

The chief component was given in both the earlier publica-
tions as a vanadium sulphide, to which Hewett gave the name
patronite and Bravo that of Rizo-Patronita, in honor of Senor
Antenor Rizo-Patron, the discoverer of the ore and its vana-
dium contents. The paper of Hewett was published Sept. 1,

*Specimens received from Mr. Hewett since this was written do show
free sulphur on some of the parting surfaces in the form of scales.

ae ae ee

W. F. Hillebrand — Vanadium Sulphide, Patronite. 145

1906, and that of Bravo bears the date Aug. 31 of the same
year, hence must have appeared a little later, although in the
August number of the Boletin de la Sociedad de Ingenieros.
Hewett’s name, therefore, has the right of priority and being
shorter is preferable, even though that given by Bravo may
better conform to Spanish usage as to derivation, since it
embodies the whole patronymic.

The following analyses given by those authors show the
composition of the ore incompletely, since appreciable percent-
ages of nickel and titanium were overlooked, and, moreover,
they also show that it varies much iu respect to the relative
amounts of siliceous and sulphide components.

Analysis by Analysis by
J. O. Handy H. Bunting
(Pittsburg) (Lima)
SiONat ak lah ema is 10°88 22°29
AOR Con 3°85 8°32
JENS is oe sare ere Di CN WP UB A 2°45 1°98
Vitel ee cca ei ma ere tae GOS 15°36
WONG 2 ete oe esl st tah Dea 0°50 Shige
S soluble in CS ee ale nveunt GoD Mee
SRcombinedsaaweae tac ea 54°06
(OO ick Ba see Un ORM lela carga 0°33
Moisture: siete menial an trace ya
Windetermineds =e s eee so [5°63 |* 9°98t
[100-00] 100:00

Both I (in Mr. Hewett’s paper) and Bravo reported the
solubility of the vanadium in fixed caustic alkalies, and Bravo
likewise in ammonia. I may say that ammonia does not act so
quickly as the fixed alkalies. Bravo analyzed the alkaline
extract for sulphur and vanadium, after first removing free
sulphur, and reported 32°89 per cent vanadium and 67-20 per
cent sulphur, corresponding approximately to the formula VS,

the ratio as calculated by me from his values is 1:3°28).

The complete analysis of this ore is extremely difficult and
in some respects the results given on p. 147 is lacking in precision.
The following behavior threw light on the nature of the min-
eral components.

Repeated treatment in a Soxhlet extractor with carbon
disulphide removed considerable sulphur. After thorough
drying of the capsule and contents, hot water was passed
through. The aqueous filtrate was blue- -green and held a sul-
phate or sulphates of vanadium. This operation was followed
by extraction with warm caustic soda as long as the filtrate
came through colored. Hereby nothing but vanadium, sulphur
and some of the siliceous components were taken out, and the

* Largely carbonaceous. + Carbon,

146 W. F Hillebrand — Vanadium Sulphide, Patronite.

solution was of a cherry-red color. In it the proportions of
vanadium and sulphur were determined.

It may be said here that the analysis was of necessity made
with interruptions, and it was noticed that the values obtained
for free sulphur increased with lapse of time. The first deter-
mination gave 45 per cent, the second 5:5 per cent, and the
third 7 per cent sulphur, each on a separate portion of the same
sample. This seemed to indicate a progressive oxidation of the
vanadium sulphide with liberation of free sulphur, or else that
the ore contained insoluble as well as soluble sulphur and that
the former on exposure of the powder to air underwent rather
rapid conversion to the soluble modification. Heating the
powder for some hours with a little carbon disulphide in a
sealed tube in a steam bath did not result in an increased yield
of free sulphur, nor did heating the powder to 60° for some
hours in the air do so, hence the latter explanation seems
improbable. To test the former the portion of powder that
had afforded 7 per cent sulphur and had then been extracted
by hot water was allowed to dry thoroughly and left exposed
to air under an inverted beaker for three weeks, after which
time it was again tested for free sulphur and sulphate, both of
which were found, thus apparently confirming the supposition
of oxidation. But the free sulphur found in this test was so
far in excess of that converted to sulphate that it would seem
as if the liberation of sulphur took place much faster than the
oxidation of that which remained in combination with the
vanadium (see page 148).

It was impracticable satisfactorily to examine the material
remaining in the Soxhlet capsule after extraction by carbon
disulphide and water, but separate portions of the ore, treated
directly with alkali, were used in part. From the residue thus
obtained strong nitric acid after a few minutes action took out
all the iron-nickel sulphide together with most of the remain-
ing vanadium, all the molybdenum and some titanic oxide,
silica and alumina, besides acting to a slight degree perhaps on
the carbonaceous component of the ore. The sulphur of the
metallic sulphide was apparently wholly oxidized by the nitric
acid. The solution was analyzed, but for more exact deter-
mination of the composition of the sulphide large amounts of
the ore, only coarsely broken up, were thoroughly extracted
by alkali, the black residue was dried and rubbed down gently
with the finger, and from it the sulphide was extracted by an
electromagnet and purified by repetition of the treatment.

The residue not dissolved by nitric acid is black and con-
tains most of the carbonaceous matter of the ore besides the
titaniferous and siliceous matters not removed by the forego-
ing operations. The carbonaceous material contains much
sulphur and seems in this respect to resemble the quisqueite

W. F. Hillebrand — Vanadium Sulphide, Patronite. 147

already described. On burning it off there is left a straw-
colored residue in which nothing mineralogically definite was
recognized except a few quartz grains. It is not unlikely that
it consists chiefly of a clay-like substance and titanium dioxide
in an amorphous state. Some vanadium in an oxidized form,
perhaps a constituent of the clay, is also present.

Tt thus appears that the sulphur and vanadium exist in four
different forms in the powdered ore. Mr. Hewitt states that
lead, zinc, copper, arsenic, antimony, calcium, magnesium,
silver, gold, and platinum were sought but not found. Uran-
ium and germanium were looked for in vain by myself. Prof.
B. B. Boltwood, who kindly examined for me both the vana-
dium ore and the quisqueite, reports hardly any evidence of
radio-activity.

Bulk Analysis of the Ore.

The results in the followimg analysis are not in all cases
highly accurate, and the summation is affected by indetermin-
able errors, among which is too high water, because some of it is
derived from the hydrogen of the carbonaceous matter. This
error is to a large extent offset by the oxygen of the carbona-
ceous matter. The vanadium sulphide constitutes about two-
thirds of the ore, the metallic sulphide over one-tenth. It is
not only an ore very rich in vanadium and from which the
latter can be very readily extracted, but it also carries a per-
centage of nickel that may prove commercially of much value,
since it will be left wholly in the residues after extraction of
the vanadium.*

Sy (otal woe ak en eles uiL: 58°79 (4:5 free)
AVE ate Pa 1 O88 nS NG AR WET 19°53
ANd Way ee aes ae aed oes My A eg 0°18
Ips tren ct canter mente sin tae ge 2°92
ENGI a soph are thn ene NN ue aM Ny al Sl Oy
(Op Es AE) EO ae ae ee aN OC al Sal t 3°47
SiO WN A 6°88
AIS @) ei PA Se ea Ae 1°53
PAL OTERO) RS Cte SE at SEI iy AT 2°00
UO aes ate Mina ae pelea eA EN 0°20
IY i OTR | Ce Rs EA Ea ck ates emer hele trace
LONE ue Walid ua lscea any Re Reet el trace?
Ae eee Bn sehtin (Wea tek esl ONO
1a Oars aap setae ER nc 1:90
O from V. Sulphatesses sone 0°38
99°75

es Specimens received from Mr. Hewett since this paper was written
showed in some cases no visible metallic sulphide, in others but a few specks.
Tests were not made to ascertain if the amount actually pr ese is less than
in the specimen first received.

148 W. F. Hillebrand— Vanadium Sulphide, Patronite.

Free Sulphur and Water-Soluble Extract.

Several determinations of the sulphur extractable by carbon
disulphide were made at considerable intervals of time. The
first gave 4°5 per cent, the second 5-5, and the third 7 per cent,
thus, as already said, indicating a progressive liberation of free
sulphur. The extractions were made in a Soxhlet apparatus
tillan extract was found free from sulphur.

The first determination of the water-soluble extract gave
0-41 per cent SO, and the solution held vanadium, as shown by
its blue-green color and by qualitative test. Much later, in
another portion, there was found 1°23 per cent SO, and 0:92
per cent V, with ratio V to SO, of 1 to 0-71. The vanadium
was in an intermediate stage of oxidation between V,O, and
V,O,, as ascertained by direct titration and again after reduc-
tion by sulphurous acid.

When the powder that had served for the last test, and
which had afforded 7 per cent free sulphur, was allowed to dry
and left exposed to the air for three weeks and then again
extracted by carbon disulphide and hot water in turn, there
was found additional 0-8L per cent of SO, and 0°40 per cent of
V, but the vanadium of the aqueous extract was all in the V,O,
state, for after titrating directly and then again after reduction
by sulphurous acid, the results were identical. The ratio of V
to SO, was not, however, the same as in the previous test, being
1to1:07. In this last test there was found 3°83 per cent of free
sulphur in addition to the 7 per cent found at first, and since the
sulphur in the SO, is but 0°27 per cent, or but one-fourteenth
of the free sulphur formed during the three weeks, the ground
will be apparent for the earlier remark (p. 146) concerning the
ereater rapidity of splitting off of free sulphur from the origi-
nal sulphide than of oxidation of the lower combination of
vanadium and sulphur that presumably resulted from this
division. But the variation in the two ratios given above
between vanadium and SO, has not been cleared up. These
approximate very roughly to 1 to 1, which indicates a similar
ratio in the lower sulphide, to whose direct oxidation the sul-
phate is supposed to be due, that is, VS. But much more
knowledge is needed, based on numerous and varied tests,
before the reactions involved in the changes above indicated
can be written. :

Patronite.

Several determinations of the sulphur and vanadium in the
alkaline extract, after first removing free sulphur and the
water-soluble constituents, gave the following results :

Vi See ee ee 19:16, 18°89 19:09) 1s46
De FeN aiey Gh Leas 47°74 47°84 45°65 44°74
Ratio Wied Seee een Sc Ole alee Angele RRS The 3 assy

W. F. Hillebrand— Vanadium Sulphide, Patronite. 149

The determinations with the two highest ratios were earlier
analyses in which least free sulphur was found and presumably
least oxidation had taken place. They point to a formula
VS, for the sulphide instead of VS, as reported by Bravo,
whose actual ratio was 3°28. A sulphide of this composition
seems very singular and the possibility suggested itselt that the
compound might be ortho-sulphovanadie acid, but this would
probably be soluble in water and an amount equal to that in
the ore would need, moreover, above 1 per cent of hydrogen,
whereas the total as represented by the water found is less
than 0°20 per cent, and this unquestionably belongs largely
with the siliceous and aluminous matter of the ore.

In the light of the observation relating to the splitting off of
free sulphur from the sulphide, a process that is evidently
connected with exposure to air, the possibility has to be
considered that the free sulphur found in the test showing
least may all have been originally combined with the vanadium
to form a still higher sulphide than VS,. The ratio calculated
on this basis is less than 1 to 4:5, indicating a formula not
above V,S,. Lacking, as we do, any knowledge regarding the
higher sulphides of vanadium, speculation at present would be
premature.

Whatever may be the true formula, the mineral is new and
the name patronite may well be applied to it.

The Iron- Nickel Sulphide, a Highly Nickeliferous Pyrite.

The bulk analysis of the ore indicates a greater percentage
of the iron-nickel sulphide than would be supposed by inspec-
tion, and than was obtained by separation in the manner
already outlined. Much of it is evidently in a state of fine
division and escapes in the washings. Indeed, it was sought
to recover only the grains of some size. It was noticed
after separation that while some of the lar ger grains especially
showed the pyritous color on fracture surfaces, all were more
or less covered with a black coating of what is ‘assumed to be
the carbonaceous material of the ore and that most of them
were black throughout, apparently from the same cause since
on solution in nitric acid a considerable carbonaceous residue
was left. In afew instances apparent crystals, seemingly of
octahedral habit, were observed, but otherwise the fragments
were quite destitute of crystalline structure. The color is
rather whiter than that of pyrite, as said before, becoming
of a reddish cast by tarnish. The mineral is not attracted by
a hand magnet but attaches itself readily to an electrom agnet.
The density of the material analyzed was 4°33, but this value
is of little worth since, despite repeated purification, the frag-

150 W. F. Hillebrand — Vanadium Sulphide, Patronite.

ments contained a good deal of carbonaceous, siliceous and
titaniferous matter in a state quite invisible to the eye. Heated
in a closed tube much sulphur is evolved. The mineral is very
insoluble in hydrochloric acid but is readily attacked by nitric
acid with complete oxidation of the iron, nickel and sulphur.
The other constituents remain more or less unattacked accord-
ing to the duration of contact with the acid.

The results given below are for the four chief constituents
the mean of concordant determinations. It is to be noted that
if the vanadium was counted as the element, every analysis
gave a summation of approximately only 97 per cent. This is
accounted for in part probably by the extreme oxidizability of
the mineral when exposed in powder form to the air, especially
when wet. Hot water-extracts from the powder ferrous and
nickelous sulphates with only a very little vanadium. If the
powder is left moist on the filter, relatively large additional
quantities of iron and nickel can be extracted next day, and
for several successive days, with very little vanadium. But the
discrepancy cannot be apparently in large part thus accounted
for. It seems necessary to assume that the vanadium is in an
oxidized form, perhaps as a constituent of the siliceous-alumin-
ous material. In this case its condition would be that of triva-
lency in all probability, a condition which also affords a more
satisfactory summation than that of quinquivalency. Against
its presence in combination with silica and alumina is its great
excess over that in which it is known to occur in such forms
in nature. On the other hand, the matter insoluble in nitric
acid still contains some vanadium.

The following experiment seems to put it beyond question
that the vanadium is not a constituent of the metallic sulphide.

Mean Atom. Ratio Calc. to 100
Digs e yee 45°06* +32°06=1°455 = 2°02 52°31
ere See 25°38 +55°9 =0°454 ) 29°46
: 3 7 Mle
Ne ees 15°70 +58°7 =0°267 ie DOOM teres
) Ope weane ayes trace 100-00
Ges Ana eat 4:3 16°33 VEO MOL iOOI ROE
Joy ae a 0:09
lis A ane wes 0°47
EO pe oH) 1°38 (partly from H of carbonaceous matter)
THOME EP AN 0°93
SLO ye eee 1:93
PON eee 245) (withsalittlegi.©>)
97°70

* This does not include the sulphur that may have been held by the undis-
solved carbonaceous matter.

W. F. Hillebrand — Vanadium Sulphide, Patronite. 151

When the powder was exposed over night in a covered plati-.
num crucible to the action of strong hydrofluoric acid, nearly
half the vanadium and much of the silica, alumina and titanium
were extracted, but only about one twenty-fifth of the iron and
nickel, these two showing exactly the ratio afforded by the
main analysis below. Boiling with dilute sulphuric acid yielded
results somewhat similar.

If the vanadium is disregarded, the iron-nickel sulphide has
the formula (FeNi)S, with iron to nickel as 1°70 to 1, or nearly
5 to 8. A pyrite with such a high proportion of nickel is
unknown. But a single one of the analyses quoted in Hintze’s
Handbook of Mineralogy shows anywhere near 6 per cent
nickel, this one, however, showing also about 3 per cent of
cobalt. The nearest approach to the present case is seen in the
mineral gunnarite 3FeS,,2NiS, incompletely described by Land-
strém (Geol. Foren., ix, 868, 1887,) with density 4°3 and a tin-
white color with a tinge of yellow, tarnishing yellowish brown.
The present mineral “would seem to be quite distinct from
gunnarite and for the present is to be regarded as a highly
nickeliferous pyrite. Should it seem proper ‘to give it a specific
name later, bravoite is suggested, after Senor José J. Bravo
the Peruvian writer on the vanadium occurrence at Minasragra.

The Carbonaceous Material of the Ore.

A determination of carbon and sulphur in 0:2459 gram of
the residue insoluble in alkali and nitric acid gave 0:0475
carbon and 0:0254 sulphur, of which last 0:0036 was soluble in
carbon disulphide. These figures indicate a similarity to quis-
queite, although the sulphur is lower.

The Oxidation Products of the Ore.

Oxidation products are reported to cover the surface of the
ground in the vicinity of the vein outcrop. The specimen
examined by me resembles porous limonite in appearance.
There was found in it by a rough partial analysis about
45 per cent V,O,, 14-15Fe,O,, 15H 0, 20 or more of siliceous
gangue, nearly 1 of MoO, and a little SO,. Nickel is absent,
or practically so. The material does not represent a single
species, for it contains probably more than one vanadium com-
pound, among them doubtless the minute micaceous scales,
greenish sometimes and sometimes yellowish, that may be seen
with a lens.

Lab. U.S. Geological Survey,
May 22, 1907.

152 W. T. Schaller—Mineralogical Notes.

Art. XIV. — Mineralogical Notes; by Watpremar T.
SCHALLER.*

Purpurite from two new Localities.
Manganotantalite from Mt. Apatite, Maine.
Evansite from two American Localities.
Tourmaline from Elba.

Zinnwaldite from Alaska.

Forms of Pisanite—a correction.

ED EOS 2S I

1. Purpurite from Two New Localities.

South Dakota. — Some specimens of ore found near Hill
City, S. Dak., were seen to contain a considerable amount of
purpurite,t and through the courtesy of Mr. J. N. Smith, who
sent in the first specimens, a larger quantity of material was
obtained. Almost all the fresh cleavage surfaces of the black
iron-manganese phosphate which chiefly forms the ore are
covered with a film of purpurite. On breaking the biack phos-
phate transverse to the cleavage but little purpurite is seen,
and it seems probable that the latter is formed from the black
mineral.

The properties of the mineral as seen under the microscope
agree with those determined on the original mineral (from North
Carolina). One small piece showing erystal outline was noticed,
and is shown in the accompanying ‘sketch (fig. 1). The extine-
tion is parallel and the pleochroism is
as follows: Parallel to the cleavage lines
(vertical direction), rose-red ; normal to -
the cleavage lines, blue purple. The
absorption parallel to the cleavage lines
ss* (rose-red) is less than at right angles to
this direction. On account of the fre-
quent occurrence of small platy masses
showing cleavage lines, the mineral has
probably two cleavages at right angles
to each other, one more perfect than
the other. It is probably the imperfect
one which is normal to the figure and
whose traces show in the vertical lines.

On account of the larger quantity of
material, a determination of the density was made, the value
pr eviously given being unsatisfactory. The material was broken
into small pieces and the fragments of purpurite picked out.
The final sample was not pure, as a small amount of the black
substance could not ave separated. The density determination
was made by weighing the mineral in water in a small test tube

1

Yo rose red,

=
e

Se
>

* Published by permission of the Director of the U. S. Geological Survey.
+ This Journal (4), xx, 146, 1905.

W. T. Schaller—Mineralogical Notes. 153

suspended from the balance arm. The result obtained is 3°40,
whiclt shows that the figure previously given, about 3°15, is
entirely too low.

The sample was air-dried for several weeks and then anal-
yzed, with the results given below:

Ratio.
ONE tenets 715045 306 1:00
Were oe 38°36 240
MOR se ca Me . 12:08 76 1:06
CAO Gere vine al 3 24
FIG O ee Ag) 268 — "88
SOU een recs a 19

MgO,Na,0,Li,O, trace

100°27

The ratios agree fairly closely with those found in the
original paper, giving as the formula of the mineral
(Fe,Mn),O,.P,0,+H,O. In the South Dakota mineral, the
iron gr eatly pr edominates over the manganese,

Connecticut A small specimen from the well known min-
eral locality of Branchville, Conn., was kindly furnished the
writer by the late Prof. Penfield. This specimen is coated
with a purple mineral which in its optical properties as seen
under the microscope agrees well with purpurite.

Only a very small amount of material (0:0197 gram), which
was fairly pure, could be obtained for analysis. A direct de-
termination of the water was made but the phosphoric acid had
to be determined by difference. The results obtained are as
follows :

Ratio

Hes Oe een Sette 27 17 39

Min OEE SS crema antes: 23 15

| BAO epi pate ilar ae gS [44] 31

Oye ee ee ie 33

100

The ratios are nearly 1:1:1, giving the same formula as before,
namely : (Fe,Mn),O,.P,0,-+-H,0.

Considering now the three analyses of purpurite from North
Carolina, South Dakota and Connecticut, it is seen that the
iron and manganese vary reciprocally, so that purpurite is,
without question, an isomorphous mixture of

Fe,0,.P,0,+ H,O.

Mn,O,.P,0,+H,0.
In the South Carolina mineral, the manganese predominates ;
in that from South Dakota, the iron predominates, while in the
Connecticut mineral the iron and manganese are present in

154 W. 7. Schaller— Mineralogical Notes.

nearly equal amounts. These relations can also be shown by -
the following table:

Purpurite from Fe.,03.P20;.H.20. Mn.O;.P20;.H20.
South Carolina Lontains (approx. ) 35 per cent 65 per cent
Connecticut 53 “ Aa «
South Dakota cf 76 ef 24

Should the two end products be found at some future time, it
would not be inappropriate to designate them as ferripurpur ite
and manganipurpurite, retaining the name purpurite for either
the isomorphous mixture of the two end products or else for
the whole group, in a way similar to that in which the terms
garnet or mica are now used.

Manganotantalite from Mt. Apatite, Maine.

The crystals here described were received from Prof. C.
Palache, who suggested they were manganotantalite. They are
associated with the minerals so characteristic of these lithium-
bearing pegmatites, as albite, lepidolite, ete. The crystals are
mostly small, several millimeters in length, and are usually tabular
to the a face (Dana’s 8 orientation). A specific gravity determi-
nation gave a value of 7:14, showing that in these crystals the
tantalic acid lar gely predominates over the columbic acid. A
partial analysis ‘further showed that iron was almost entirely
absent, so that the crystals may well be termed manganotantalite.
The analysis g cave the following results:

Co S585
He @ ware ee oy es 16
ALi Bay @ Jython sey ten pet el Mee eA Ac 14:49 (by diff.)
100°00
The forms present are: 6= {010i}, a= 11005, C= VOM .
ote es 530t d= {730}, k=}108},
wu = §1331, m= 3163}. The angles measured are shown be-
low. Most of the faces were dull and gave poor reflections.
Measured. Calculated.
Oy i OUND) BENING 5Oma eu AO De
b NG = OO | 130 1 247 21 55
DY Xe =| OO. 2 SO 63 10 63 34
Oo OOM TSO HO ess 702 O8 20° 27
Caine ke O01 0s eee) 19 42
Dare — OO, 163 30 approx. 30 50
x TOSS NCR 2 G83 118 approx. 118 20
DX TO = NEB FS 22 approx. 19 54
OR OhON 133 50 45 n0)- O8
Oh yp ss) BBS oe HEB} Sil 15 Ho) Bah
nN Gy = 183 8 1383 30 O04 29 5

W. 7. Schaller—Mineratogical Notes. 155.

Measurements for the new prism 7 = {320} are as follows :
Cryst. Size of

No. _ face. Ref. Meas. Cale.
1 narrow p- 110- +320 Oo 39h lOn 44!
2 broad p- 010 : 320 Glee il Gil. Of
2 i g. 320 : 320 Sie ale 57 48
3 ee te O10) 320 Gill 61 06
4 es g. 010 : 320 Gil 183 61 06

The faces of {133} and {100} are large, those of {163},
{103%, {110}, {320}, small, while the others are usually very
narrow faces.

3. Evansite from Two American Localities.

Idaho.—Some specimens of evansite, reported as being found
in the immediate vicinity of Goldburg, Idaho, were received
through the courtesy of Mr. C. R. Potts of that place. The
massive, anorphous mineral occurs in seams and is very brittle,
with a conchoidal fracture. The hardness is about 3, and the
color, while usually brown, varies considerably, becoming at
times colorless, also yellow or white or a dark red. The brown
specimens resemble common brown opal very much.

An analysis of the brown material gave the following re-
sult :

HEL Oe Se a 36°96 Loss of water
HAO ee ers etre a 19°14 at 107° 20°00
etn eta CORT 540 175 7°36
PAO atom ote.) 34°48 255 3°13
CAO Bie eee en eS 4°32 290 “94
INI ORS 2p a a trace to low redness 3°90
BECOME Ze ie. sane ie ee none blasting 1°61

100°39 36°94

A determination, by means of Thoulet solution, of the rela-
tion of the density to the amount of iron present gave the
following values :

Color of mineral Amount Fe.O; Ay. density ~ Limits
dark red 6°60 per cent 2°00 1:990-2°016
brown 5°49 Bee 1°98 1:972-1:990
yellow 2°15 oe 1:94 1°927-1:947

By plotting these values, it is found that the density of the
mineral free from iron should be between 1°88 and 1°91, though
this value does not necessarily represent that of pure evansite,
as the material on which these determinations were made
contains appreciable calcium. Forbes* gives the following
determinations of the density :

* See literature at end of paper for references.

156 W. T. Schaller—Mineralogical Notes.

Colorless, translucent, _.-.....---- 1822
Colorless, 2) h42 eee ee eee,
Faint, yellow, o-2 322 eer eee 2-099
Nemi-Opaque,. = ss a2s= pe eee 1°965

the average of which is 1°939, though possibly the lowest figure
more nearly represents the density of the pure mineral. Smitht
gives as the density of colorless to milky white, translucent
mater ial, the value 1°842. Kovar gives for white = 1°874
(°87 per cent Fe,O,) and for yellow 1: 937 (1:92 per cent Fe,O,).
These values give an average for all the determinations of
1:93, while an average of the lowest values, representing’ pos-
sibly the purer mineral, gives 1°86.

A determination of the index of refraction of the colorless
evansite from Idaho by Dr. F. E. Wright gave 1°485, the index
for the colored varieties not varying more than ‘01 from this
value.

Alabama.—The second sample was received through Prof.
Clarke from Mr. Charles Catlett, and the locality is given as
“from the coal seam just west of Columbiana, Alabama.” It is
associated with coal and shows a light yellow color with a resin-
ous luster, and is transparent in small pieces. It is very brittle
and easily ‘breaks into small fr agments. An approximate analysis
serves to identify the mineral; the phosphoric acid being
determined by difference.

AO ee eee 38°33

Ca Oe cise Pie igk Pee emo 1°03

IU Dd © pe res ee Pees “75

(iosssonaen seer ee see 38°19

PQ re so. ae emer ae ere 21-70 (by diff.)
100°00

Literature.

1864. Forbes, D. Phil. Mag., xxviii, 341.

1883. Woodward, A.S. Min. Mag., v, 333.

1893.. Smith, H. G. Proc. R. Soc. N. 8. W., xxvii, 382.

1896. Kovar, Fr. Abh. Bohm. Akad., xv, 1.

1896. Petterd,* W. F. Cat. Min. Tasmania.

1902. Petterd,* W. F. Papers and Proc. Roy. Soc. Tasmania,
18-33.

1902. Kretschmer,* F. Jahrb. d. k. k. geol. Reichsanst.
Wien, 52, 353-494.

1903. Wernadsky,* W. and Popoff, 8. Centralbl. Min. Geol.,
331.

* These references only mention the occurrence of evansite.
+ Wrongly given as 1:939 in Dana’s App. I, p. 29.

W. 7. Schaller—M ineralogical Notes. 157

4. Tourmaline from iba.

While analyzing a number of tourmalines from San Diego
County, California, it was suggested by Prof. Clarke that it
would be advisable to analyze the pale pink variety of tourma-
line from the island of Elba. Rammelberg’s analysis of tour-
maline from this locality indicated that these tourmalines were
probably near, in composition, to one of the end products of
the isomorphous mixtures which form the mineral tourmaline.
A number of these crystals were purchased from the Foote
Mineral Company, who wrote in regard to them: “ All of these
were secured in exchange from the late Prot. Bombicci, who
collected them in Elba many years ago.” The crystals were
picked over and after crushing them to haat fragments, each
piece was examined under a hand lens and only the transparent
pure material was selected. These were then washed with
cold dilute HF to remove any albite, mica, etc., that might be
adhering to the tourmaline. The sample finally powdered and
analyzed was pure. The density of the mineral (determined
by Thoulet solution) is 3:04-8-05, as the mineral sank at 3°043
and floated at 3°050. The analysis, made with all care, follows:

Ratio
SKC) reas poe eneaaeeat 37°89 632
BeOS r et eee O28 147
ANIA) Martha eases 43°85 2580 |
ADRS hauls Sena 04 002
BeOrs 8.4.2 ot all "0038
Min Ops oye 32 SS ged 003 | These ratios give the
CAO aes Bee O'T 003 ‘hydrogen atoms equivalent.
NIE Onis See 2 3 2°43 079 | to the metals.
Li,O SOKO See ueGe 1°66 lal
EIE Opes xg ee a8 3°47 386
ES An "10 005 |
Mi ORK Oyees ater: none
100°01
= (0) ea ea ee 04
99°97

Ratio SiO,: B.O, : total H =-4:00 ; :93 : 20:08

This ratio agrees very well with that proposed by Penfield,
namely, 4:1: 20. A discussion of the results will be given
in the paper on the California tourmalines, as at present it is
only desired to place the above analysis on record. The states
of oxidation of the titanium, iron and manganese are arbitra-
rily given.

Am. Jour. Sci.—FourtH Series, VoL. XXIV, No. 140.—Aveust, 1907.

uit

158 W. T. Schaller—Mineralogical Notes.

5. Zinnwaldite Jrom Alaska.

The occurrence of zinnwaldite with cassiterite and topaz in
the York region, Alaska, has already been mentioned* and the
results of a quantitative analysis of this mica are here pre-
sented. The sample analyzed was probably fairly pure.

Analysis Ratio
SiOie: Oe Lene seit e 46°80 no 775
CANOE ee tice n er earele 24°50 "240 943
Hel One eins eae 50 "003 ®
Fe Quad Se Pee hares 6°35 088
Mini @ eae eee One 1°38 "019 111
CaOy ae ee Ns eS ahs ee 24 "004
Nai Oe sistas ee euae 1°73 033 l 131
1 EGO) Recah sean pa a en cede Ses, 4) 9°20 098 |
IV OVe sk Ree Aa aie renee 3°73 "124 124
PTO et ete Ss el 88 098 55
1 OS epee LU UALR tear an 8°63 454 9

103°94
PA Gat) Khe Sai a ee em 3°63

100°31

In the ratios, the soda and potash are taken together and the
small amount of water present is considered as hydroxyl and
added to the fluorine ratio. The empirical formula gives :

Si,,,Al,,,Fe, ,K,,,Li,,.F,,.0

The total oxygen is very slightly greater than three times the
silica, the ratio O: Si being 3°06:1. The mineral is therefore
considered as a metasilicate and the formula may be written

(SiO )) 7. ( ANE rare Pls ey CAO) Allee.
or
(Si0,) ac) eat gla He en (AlO) Ee Ale

The above formula shows that the analysis may well be
interpreted as a metasilicate.

6. Forms of Pisanite—a Correction.

In a paper on “ Minerals from Leona Heights, Alameda Oo.,
California,”+ the writer described a number of new crystal
forms for pisanite (Fe,Cu) SO,.7H,O. One of these, g= {205}
has to be withdrawn as its determination is based on an error.
The face measured and described as {205 is the base {001}.

meas. {205!, 6=90° 03’ p= loen004
cale.{001}, @=90 00 p=15 11

* Bull. 262 U.S. Geol. Survey, 129, 1905.
+ Bull. Dept. Geol. Univ. Cal., iii, 191-217, 1903.

Barker—Thermoelectromotive Forces of Potassium, etc. 159

Art. XV.—The Thermoelectromotive Forces of Potassium
and Sodium with Platinum and Mercury ; by Haron C.
Barker.

Fottowine Seebeck’s discovery in 1821, many investigators
have measured the thermoelectromotive forces produced when
the two junctions of a given metallic conductor with a second
metallic conductor are maintained at different temperatures.
But comparatively few attempts have been made to determine
the values of these thermoelectromotive forces for couples
consisting of sodium or potassium combined with other metals,
for the obvious reason that the ease with which the alkali
metals oxidize renders their manipulation difficult.

Matthiessen* compared the thermoelectromotive forces of
silver-sodium and silver-potassium couples with that of a silver-
copper couple, and expressed his results in terms of the latter.
The two couples were placed in series with a galvanometer,
and the ratio deduced from the observed deflections when the
two couples were alternately arranged, by a commutating
device, to assist or oppose each other. The difference of tem-
perature between junctions did not exceed 26° in any case
given. The sodium and potassium were contained in straight
wide thermometer tubes into which the metal was introduced
melted, in an atmosphere of hydrogen. Into the ends of the
tubes short platinum wires had been fused, and to these the
silver wires were soldered. The results cannot be regarded as
of great value at this time; they merely express the ratio
between the thermoelectromotive forces of, for example, silver-
sodium and silver-copper for a particular small temperature
difference between junctions.

Taitt made some measurements with couples of which one
member was sodium or potassium in order that the lines of
these metals might be included in his “ First Approximation
to a Thermoelectric Diagram.” I can find no record of the
observations on which his results were based, nor any state-
ment as to the means of measurement adopted for the particu-
lar cases. It may be presumed, however, that the method was
similar to that used by him in other cases, and consisted essen-
tially in a measurement of the thermoelectric current-strength
based upon the observed deflections of a sensitive galvanome-
ter. As to the construction of the couples, he states that he
had prepared for him ‘a long quill tube of German glass
with platinum wires inserted near the ends; exhausted it by

* Pogg. Ann., ciii, p. 412, 1858.
t Proc. Edinb. Soe., viii, pp. 350-362, 1873-1874.

160 Barker—Thermoelectromotive Forces of Potassium

means of a Sprengel pump, and drew in melted sodium from
a bath of paraffin.”

Naceari and Bellati* measured the thermoelectromotive
forces of sodium-copper and potassium-copper couples, and, in
order to refer to lead, of a lead-copper couple. The cireuit
was simply completed through a galvanometer of high sensi-
bility, and the deflection noted. “Then, the resistance of the
couple and of the galvanometer being determined, and the
galvanometer calibrated by a Daniell “cell of known electro-
motive force and resistance, the thermoelectromotive force was
calculated. The metals were melted under petroleum of high
boiling point in wide vertical tubes, and narrower tubes filled
by simply i inserting, with the upper end closed with the finger,
and then removing’ the finger and thus allowing the pressure
of the petroleum in the large tube to force the molten metal
up into the narrow tube. When the metal had cooled and
solidified, the tube was withdrawn and the ends of the copper
wires, which were threaded, forced into the solid metal. The
sodium tube was straight; the potassium tube, in order that
measurements might be made with one junction at a tempera-
ture above the melting point, was terminated by a U. The
junctions were surrounded and protected by petroleum, con-
tained in glass globes provided with the necessary tubulures.
Heat was applied by means of a water-bath.

I have not had access to the original communication, and
have had to depend on an abstract in the Journal de Physique,
1877, which, while quite full in many respects, gives but few
numerical results, stating only the values found for the “spe-
cific heat of electricity ” and the temperatures of the neutral
points.

This is the most recent investigation of the thermoelectric
properties of potassium and sodium of which I have found a
record.

The present communication deals with the measurement of
the thermoelectromotive forces of potassium platinum and
sodium-platinum couples for varying temperature-ditferences
between 0° and about 90°, and, for the purpose of comparison,
of a mercury-platinum couple through the same range, by a
potentiometer method. There was available for this use a
Leeds potentiometer, adjusted to read directly in volts when
used in connection with a Weston standard cell of 1:0193 volts.
The potentiometer was carefully tested and some corrections
made. The certified electromotive force of the cell was veri-
fied by a comparison made by the National Bureau of
Standards.

The galvanometer used was a high sensibility D’ Arsonval

*Nuovo Cimento, xvi, pp. 120-180, 1876.

and Sodium with Platinum and Mercury. 161

instrument manufactured by Leeds and Northrup, of most
excellent performance. It was used, of course, with telescope
and scale.

The form of the thermo-couples finally arrived at after
many trials and rejections was as shown in the accompanying
diagram. For clearness, the branch EF is represented in the
plane of the paper; it was actually bent ina plane at right
angles thereto, as toward the reader.

This apparatus was constructed of ordinary glass tubing of
about 0:4° diameter, with platinum wires fused in at A and

1

B. The distance between A and B is about 55™ and the
vertical distance AC or BD about 10™. The branch EF is
about 15° long. The platinum wires were of number 32
B & 8S gauge, and each was about 100° in length. Three
such pieces of apparatus were built, as nearly as possible, alike ;
the six pieces of platinum wire required were cut from a single
piece of about 600 length supphed by Eimer & Amend.

The tubes were filled respectively with potassium, sodium
and mercury, these metals being drawn in at the end F by
suction applied by the mouth to a rubber tube connected at G.
Mercury, of course, was simply drawn in at the ordinary room
temperature. To introduce the other two metals, the tubes
were heated, filled with melted paraffin and again emptied, and
then filled with the metal, previously melted under paraftin in
a test tube. In drawing in the alkali metals, the mouth was
protected against accident by a glass trap interposed in the
rubber tube between the mouth and the opening G. The par-
affin used was ordinary white paraffin of commerce melting at
about 45° to 50°.

It was found that the potassium and sodium in lumps of sev-

162 Barker—Thermoelectromotive Forces of Potassium

eral cubic centimeters could be introduced with perfect safety
into a mass of melted paraffin, and, when melted under it,
showed in each case a brilliant metallic surface, which remained
untarnished indefinitely. Sodium when thus melted could be
simply poured from vessel to vessel without apparent risk;
potassium was transferred when needful by means of a narrow
glass tube closed at one end by the finger, into which the metal
was drawn by suction. Paraffin oil was first tried, but aban-
doned in favor of solid paraffin, as the oil was found not with-
out action on sodium, especially when the latter was melted ;
moreover, the solid paraffin was more convenient and cleanly
to handle.

A tube thus prepared will remain unimpaired indefinitely,
the precaution being taken to protect the free ends of the alkali
metal by a plug of paraffin a fraction of a centimeter long.
The surface of both metals in contact with the walls of the
tube is beautifully clean and lustrous, sodium being distinetly
crystalline in appearance, while potassium is not crystalline, or
obscurely so.

The sodium and potassium were the commercially obtain-
able metals, presumably of fair purity. Cooling curves gave
the melting point of the potassium at 63°5°, and a rough spec-
troscopie examination showed no impurity other than the
expected trace of sodium. The mercury was the pure redis-
tilled metal supplied by a reputable firm.

The purpose of the mercury tube was to afford a compari-
son with a metal whose thermoelectric behavior is less subject
to variation than is the case with commercial platinum, differ-
ent specimens of which have long been known to vary widely
in this respect. Mercury was chosen as readily obtainable in
the pure state and free from crystalline or other structure, fol-
lowing the example of K. Noll* and also as readily introduced
into a tube of the form described.

When placed in position for measurement, the vertical parts
of a tube were inserted in wide test tubes containing mercury
so that the junctions were deeply immersed. These mercury
receptacles were surrounded respectively by water and by
melting ice contained in double walled copper vessels. In
each test tube a thermometer was also inserted, graduated to
0:2°, with the bulb as closely adjacent as possible to the actual
platinum-metal junction. The platinum wires were led to thin
glass mercury cups floating in a large mass of mercury, and
from these cups copper wires were used to connect with the
binding posts of the potentiometer.

The thermometer at the cold junction was found to be indis-
pensable, the temperature being always a little above zero, and

* Ann. der Physik, 1894.

and Sodium with Platinum and Mercury. 163

2

800

T00

a
fo}
°o

MICROVOLTS

>
8

300

DEGREES

subject to slow but considerable variation, in spite of much
precaution.

To obtain a series of observations, the other junction was
surrounded, as stated before, with water, which had been

164. Barker—Thermoelectromotive Forces of Potassium

heated, generally to boiling, or which was colder than the air
of the. room, and corresponding values of temperature and
electromotive force read at intervals as the temperature of this
junction fell or rose, chiefly through radiation. Inasmuch as
the least difference in electromotive force between successive
points at which the potentiometer, as it was constructed, could
be set was 10 microvolts, it was the uniform practice to set the
potentiometer at the point whose corresponding electromotive
foree would next be reached, in following the variation in tem-
perature in either direction; and, depressing the galvanometer
key shortly before, watch the slowly decreasing deflection until
it was reduced to zero, when the temperatures of the hot and
cold junctions were read in close succession. No difficulty was
met with by reason of variation of zero point, and none from
variation of electromotive force of the storage cell used to sup-
ply the potentiometer current, this latter variation being very
slow and of small extent. The deflection of the galvanometer
followed the variations in temperature closely and regularly.
Under the circumstances of actual measurement a variation of
the setting of the potentiometer of one step of 10 microvolts
caused a difference of deflection of the galvanometer of over
20 scale divisions (millimeters), each of which was very easily
divisible by the eye into at least five parts; I conclude, there-
fore, that I would be quite justified in stating the electromo-
tive forces measured to 0-1 mier ovolt, if this seemed desirable.

The results obtained are embodied in the subjoined tables
and in the accompanying curves. In the tables the column
headed E gives the electromotive force in microvolts at which
the potentiometer was set; that headed T gives the tempera-
ture difference between junctions, one of which was invariably
near zero. The thermometer readings have been corrected for
the determined variation in the positions of the fixed points,
but in no other way. The value of T given is the mean ofall
the observed values.

In each of the three cases the electromotive force developed
was such as to cause current to flow from the enclosed metal
to the platinum at the heated junction. The electromotive
forces are therefore all of the same sign and are treated as
positive.

Under N is stated the number of observations made at the
corresponding potentiometer setting. This must not be
regarded as indicating the number of separate series of obser-
vations made with a given couple, or even with a given couple
in a given range. Thus, i in the case of the potassium-platinum
couple, eight series of observations were made at different
times during a period of about two weeks, and four of these
embraced the ge range of temperature in common. In no

and Sodium with Platinum and Mercury. 165

case, however, are there more than three observations at a
given point, as the aim was rather to supply points wanting in
preceding series than to obtain additional values for the same
point.

The column headed D gives the mean of the absolute values
of the deviations of the individual observations from their
mean, expressed in degrees, and also in microvolts. It will be

TABLE I,
Potassium-Platinum

E 4h N D E T N D

—— os.

@) (eo)
840 89°60 1 Papas INS hyn 460 48:71 2, 0°21 2°0
Bee: mabey eet Capa pap eee 450 47°32 1 eek EAs
820 87°52 2 0°04 0°4 440 46°73 2 0°01 O'l
810 86°38 1 he pee 430 45°58 1 ed pes
800 85°28 3 OO7 Opry 420 44°46 1 ees nan
790 84°09 2 O°ll 1:0 410 43°36 1 ae eee
780 83:00 1 eyed kas 400 42°36 1 SATS pai
770 81°91 2, O,O SR ORS 390 41°26 1 my ae
760 80°72 1 Base peel 380 40°24 1 Sete alk
750 79°80 1 Spee eae 370 39°56 1 dee Epes
740 78°76 2 0°06 0°6 360 38°37 2, 0:07 0:7
730 77°54 2 0°08 0°8 350 37°34 1 Weta ne
720 76°54 2 0°14 1°3 340 36°24 2 0:06 0°6
710 75°47 2 0°01 O°l 330 35°14 ] DEN eyo
700 74°40 3 0°08 0°8 320 8©34:°34 1 Sean, Eteus
690 73°24 1 Eas Series SO “83:02 1 ete jae
680 72:02 ] Na ASA is 300 32°40 1 sy Be ee
670 70°86 1 ta Lae eS ee eaate ra Ae eae
660 69°80 2 0°06 0°6 270 28°64 el ave
650 68°64 3 0°16 1°5 260 27°70 pases ares
640 67°55 2 0:07 0-7 anes ppt oii Sse su
630 66°52 1 Lipsy ba Ms 200 21:22 il ig tp See
620 65°44 1 apie So 20d ia a A RIESE Wirt FOO eas
610 64°30 1 ans Bye 170 17°70 1 Sie Sans
600 63°32 2 0°08 0°8 160 16°40 1 a a Nes
590 61°92 i Gath ieee 150 15°34 1 ele epee
580 60°98 1 eee he ee eae: tools iu TS Scare wae
neues EA Pe ae sgr| eee aes 130 13°56 wi aie Neni-s
560 59°20 3 0°00 0°0 120 12°44 1 Sey ee
550 58°14 2 0°20 1°9 eet wale Fea Soa pe
540 57°19 2 0°05 0°5 100 10°18 1 Leica Syens
530 56°01 2 0:09 0°8 90 9°02 1 Bee Sree
520 54°93 2 0:09 § 0°8 80 7°88 1 ze i Pues
510 54°12 1 eis Suite! 70 6°82 1 Bae ways
500 52°96 1 cig pe we
490 51°88 1 Bee. ee
480 50°84 1 ye pee B
470 49°62 1 ns aes,

166 Barker—Thermoelectromotive Forces of Potassium, etc.

TaBLeE II.
Sodium-Platinum Mercury-Platinum
E T N D E Av N D
SSS = FS
° °
170 84°24 4 0°54 1:0 50 84°42 6 2°30 09
160 79°29 84 O;6O) gall 40 58°56 5 124 06
150 73°60 5 O;6255 p72 30 8=639°31 2 0:255 Ot
140 6814 4 DSBS NO) 20) 22:69 2 O Toes
130 63°46 3 0°36 8 =60°7 10 10°64 38 0:29— S055
L200 00729 ao 0°34 06
IIOe, Bee 5 O72 1:4
10 OMe O18 eet ean Oneness
90 41°51 2 0:33). 0°%
80 37°32 2 0°04 0°]
70 32°16 1 aces ee

30 12°64 2 0-40 0°8
noticed that, as expressed in microvolts, the highest values for
each of the three couples are not widely different.

The successive observations of a given series were fre-
quently and even prevailingly much more concordant than
those of different series. A similar observation was recorded
by W. H. Steele in his well known paper,* who says that
“One is led to the suspicion that the thermoelectric constants
are not really constants, but that they vary in a given speci-
men in a manner which, if not arbitrary, yet arises from
changes in condition which are inappreciable.” While not
perceiving in just what sense a natural phenomenon can be
regarded as “arbitrary,” I can otherwise fully appreciate the
comment quoted, since I have myself failed to trace such vari-
ation to its causes with any certainty.

The curves are virtually self-explanatory; in addition to
those obtained by direct experiment, the curves deduced from
them for potassium-merecury and sodium-mercury are indicated.

I hope to find opportunity at some future time to make fur-
ther measurements with a potentiometer system better adapted
to the measurement of such small electromotive forces, and
with metals of whose purity I shall have more assurance. The
foregoing values are therefore presented as preliminary approx-
imations rather than as final results. It is also highly desirable
to extend the temperature range in both directions.

My thanks are due Prof. A. W. Goodspeed, both for sug-
gesting the domain of investigation, and for placing the
resources of his laboratory most freely at my disposal; and I
should be most remiss if I failed to acknowledge my indebted-
ness to Dr. R. H. Hough for his constant advice and aid, par-
tieularly in the construction of the special apparatus required.

University of Pennsylvania, Philadelphia.

* Phil. Mag., Feb. 1894.

Gooch and Osborne—Potassium Aluminium Sulphate. 167

Art. XVI.—The Reaction between Potassium Aluminium
Sulphate and a Bromide-Bromate Mixture; by F. A.
Goocu and R. W. Ossorne.

[Contributions from the Kent Chemical Laboratory of Yale Univ.—clx. |

Wuen the water solution of a neutral aluminium salt is boiled,
hydrolysis may take place until an equilibrium is brought
about between the free acid formed and other active products.
If a product is inert toward the acid set free in the liquid, or
if the acid is continuously removed or destroyed, the process
goes on until no further hydrolytic action is possible under the
conditions.

In the basic acetate process for the separation of alumina,
the insoluble aluminium compound is a basic salt inert toward
the free acetic acid formed under the conditions.

In Chancel’s method of precipitating alumina, by the action
of sodium thiosulphate upon the boiling solution of the soluble
aluminium salt, the insoluble product is likewise a basic salt, in
this case the sulphate, and the sulphuric acid set free is fixed
as sodium sulphate while the thiosulphurie acid formed in meta-
thesis is destroyed at the temperature of the reaction with
formation of the comparatively inactive sulphur dioxide and
sulphur; though as Norton has shown* the basic salt is not
quantitatively insoluble until the mixture is superheated under
pressure.

In Stock’s processt for the determination of alumina, alumin-
ium sulphate in solution is heated at the boiling temperature
with a mixture of potassium iodide and potassium iodate.
The sulphuric acid produced hydrolytically is fixed by the
‘iodide-iodate mixture and comparatively inert iodine is set free
and taken up by sodium thiosulphate to form sodium iodide
and sodium tetrathionate. In this process, as Moody has
shownt, the first insoluble product is a basic salt; but pro-
longed boiling of the mixture of aluminium sulphate, potas-
sium iodide and potassium iodate finally brings about the
complete hydrolysis of the insoluble basic salt, so that the
entire amount of iodine set free in the process, when properly
collected and titrated, may serve as an accurate measure of
the acid produced and therefore of the acidic ion of the alumin-
ium salt.

In an extension of this process to other elements Moody
found that a sufficiently prolonged treatment with the iodide-
iodate mixture completely liberated the acidic ion from the

* Ber. Dtsch. Chem. Ges., xxxiii, 548, 1900.

+ Norton, this Journal, xii, 118, 1901.
¢ Moody, this Journal, xx, 181, 1905.

168 Gooch and Osborne—Potassium Aluminium Sulphate.

sulphates of iron, chromium, cobalt, nickel, ammonium and
from stannic chloride. The hydrolysis of zine sulphate under
similar conditions comes to an end before the acidic ion is
entirely liberated and when the product is one-fifth sulphate
and four-fifths hydroxide.

In the work of which the present paper is an account, the
reaction of a mixture of potassium bromide and potassium
bromate upon aluminium sulphate has been studied. Assuming
that the acidic ion is entirely liberated from the aluminium
salt, the reaction should follow the equation

2K Al(SO,), +5KBr+KBrO, +3H,0
—9 Al(OH), +4K,SO, +3Br,,

For this work a bromide-bromate solution was made up by
dissolving 36 grms. of potassium bromide and 10 grms. of potas-
sium bromate (the proportions indicated in the equation) in
water and making up to 500™*. Pure crystallized potassium
alum was the aluminium salt used and this was weighed out
exactly in every experiment.

The first set of experiments was made to find out whether
precipitation is complete when the bromide-bromate mixture
acts upon the aluminium salt. In every experiment a portion
of the alum was weighed out, dissolved in a small amount of
water ina beaker anda portion of the bromide-bromate solu-
tion was added. In experiment (1) the mixture was boiled
over a flame until at the end of an hour no more bromine
seemed to be liberated and the cooled liquid was colorless.
The precipitate was, so far as possible, transferred to an ash-
less filter paper, washed, ignited and weighed as Al,O,; but to
get the entire amount of precipitate it was necessary to dissolve
from the beaker, by means of hydrochloric acid, a closely
adherent film of precipitate; and from the solution the dis-
solved alumina was recovered by precipitation with ammonia
and determined separately. In experiments (2) and (8) the
precipitation was bronght about by passing a current of steam
into the solution, interrupting the process to concentrate the
liquid by evaporation without boiling and again passing in
steam.

TABLE I.
KA(SO.)2.12H.0 Time of KBr KBrO; Al2.0Os3 Al.Os
taken heating taken taken taken found Error
grms. hrs. grms. grms. grms. grms. grms.
(1) 1:0012 ls 3°6 1: 071078 -1070 0:0008—
(2) 0:5017 2+ 2°5 2s 0°5 0705389 -0529 0.0010—
(3) 0°5029 3+2°5 2° 0°5 0°0541 °0559 0°0018+

In these experiments the precipitation proved to be complete,
or nearly so; though in (3) there is obvious contamination of

Gooch and Osborne— Potassium Aluminium Sulphate. 169

the precipitate due perhaps to the prolonged treatment in glass.
As will appear, however, the process of hydrolysis did not go
to the point of forming aluminium hydroxide.

The degree to which the acidic ion is removed hydrolytically
from the salt under the conditions is shown in the following
series of experiments in which portions of the alum were
treated with the bromide-bromate mixture under conditions
which permitted the collection and titration of the bromine set
free in the reaction.

The apparatus used consisted of a Voit flask, used as a distil-
ling flask, sealed to the inlet tube of a Drechsel wash bottle,
used as a receiver, to the exit tube of which was sealed a Will and
Varrentrapp absorption tube. The Voit flask was charged with
0-5 grm. of alum dissolved in a little water and 25°™ or 80™°
of the bromide-bromate mixture. ‘The solution was boiled for
the time indicated and the bromine liberated passed in a cur-
rent of hydrogen, sent through the whole apparatus, to the
receiver containing in the cylinder and trap a solution of
4 orm. of potassium iodide in 200° of water. The iodine set
free by the action of the bromine liberated in the reaction was

titrated by te

sodium thiosulphate with the aid of a starch

10
indicator.
TABLE II.
A
Theory 6Br to 2K Al(SO,)3.12H.0.
Error in Error in
KAI(SO,)o.12H20 KBr KBrO; Bromine Theory for terms of terms of
Time taken used used liberated bromine bromine Al,O3
hrs. grm. erm. grm. germ. grm. grm. _ grm.
0°5 0°5000 1°8 0°5 0°2052 0°2527 0°0475— 0:°0101I—
11955 0°5000 2°16 0°6 0:2064 0°2527 0°0463— 0:0097—
2°0 0°5000 DOUG 0°6 0:2084 0°2527 0°04483— 0°0094—
2°5 0°5000 2°16 0°6 0°2099 0°2527 0:0408— 0:0091—
BG, 0°5000 2°16 0°6 0°2119 O27 0:0408— 0:0087—
4° 0°5000 Z2Ali6 0°6 0°2120 0:2527 0:0407— 0:0086—
B
Theory 5Br to 2K Al(SO,4)2.12H.0.
0°5 0°5000 1°8 0°5 0°2052 0°2106 0°:0054——s- 000 LI —
<5: 0°5000 2°16 0°6 0°2064 0°2106 0°0042 — 0:0009 —
2°0 0°5000 2°16 0°6 0°2084 0°2106 0:0022— = 0:0005—
2°5 0°5000 AUG 0°6 0°2099 0°2106 0:0007 — 0:0001 —
Bi) 0°5000 2°16 0°6 0°2119 0°2106 O-OO11 + 0:0002 +
4°0 075000 2°16 0°6 0°2120 0°2106 0O:0014 + 0°0003 +

Se ee

170 Goochand Osborne—Potassium Aluminium Sulphate.

From these experiments, in which approximately three times
the amounts of potassium bromide and potassium bromate,
corresponding to the complete hydrolysis of the alum, were
used, it is obvious that the hydrolysis is not complete. At the
end of an hour nearly five-sixths of the acidic ion have been
liberated and this proportion is not much exceeded after four
hours boiling, the liberation of bromine proceeding very slowly
during the greater part of this time.

In the following series of determinations the proportions of
potassium bromide and potassium bromate with reference to
one another and to the aluminium salt were varied in order to
test the effect of such variation on the course of the action.

TABLE III.
A
Theory 6Br to 2K Al(SOx4)..12H,0.
Error in Error in
KAI(SO;)212H,0 KBr KBrO; Theory for Bromine terms of terms of
Time taken taken taken bromine found bromine Al,Os
hrs. grm. grm. grm. grm. grm. grm. grm.
3 0°5000 *0°32 _ *0:09 0°2527 0°1437 0:1090— 0:0232—
3 075000 oe 0°09 0°2527 0°1814 0:0713— 0°0152—
3 0°5000 0°32 0°30 0°2527 0°1853 0°0674— 0°0144—
3 0°5000 By} 0:90 0:2527 0°2182 0°0345 — 0:0073—
3 0°5000 6° De 0:2527 0°2199 0°0328— 0:0069—
4 0:°5000 6 2 0°2527 0°2229 0:0298— 0:0063—
5 0°5000 6 2° 0°2527 0°2320 0°0207— 0:0044—
B
Theory 5Br to 2K Al(SO;)2.12H20.
3 0°5000 *0°32 *0-09 0:2106 0°1487 0:0669 — 0°0143—
3 0°5000 3°20 0°09 0°2106 071814 0°0292— 0°:0062—
3 0°5000 0°32 0°90 0:2106 0°1853 0°0252 + 0°0054—
3 0°5000 3°2 0°90 0°2106 0°2182 0:0076+ 0:0016+
3 0°5000 6° ye 0:2106 0°2199 0:0090 + 0:0019+ .
4 0°5000 6° 2 0°2106 0°2229 0°0123 + 0°0026 +
5 0°5000 6° 2 0°2106 0°2320 0:0214+ 0°0045 +

In the result of experiment (1),in which the amounts of
bromide and bromate were equivalent to one another and to
the acidic ion of the aluminium salt, it will be noted that the
liberation of the bromine falls considerably short of five-sixths

* These amounts are nearly equivalent to the acidic ion of the aluminium
salt. :

Gooch and Osborne— Potassium Aluminium Sulphate. 171

of the equivalent of the acidic ion. In experiment (2) the
bromide was increased tenfold and in experiment (8) the
bromate tenfold with similar though incomplete effects. In
experiment (4), in which both the bromide and the bromate
amount to ten times the theoretical equivalent of the acidic
ion, the results show a liberation of bromine somewhat in
excess of the five-sixths proportion. In experiments (5) (6) (7)
further increase in the amounts of bromide and bromate and
in the time of boiling advanced the reaction so that amounts
of bromine were liberated considerably in excess of the five-
sixths proportion. These results seem to indicate that the
deficient hydrolytic effect is advanced by large increase in the
proportions of either the bromide or the bromate or both, and
would seem to confirm the natural inference that the products
of the action of the bromide and the bromate upon aluminium
sulphate take part essentially in the hydrolytic action as well
as the sulphate itself. These products naturally increase
with the concentration of the reacting substances. With a
very large increase in the amounts of the bromide and bro-
mate and prolonged boiling, it would be natural to expect the
completion of the hydrolysis to the point of liberating bromine
equivalent to the entire amount of the acidic ion.

Experimental difficulties such as the necessity and difficulty
of supplying water frequently during the boiling, the periodical
titration of the iodine liberated in the receiver, made of doubt-
ful value experiments extended over many hours. When,
however, the residue left in the Voit flask in experiment (7)
was treated with a mixture of 1 grm. of potassium iodide and
0°5 grm. of potassium iodate, the iodine set free on boiling for
an hour longer indicated that the acidic ion had been very
nearly liberated.

The iodine-iodate mixture is extremely sensitive to the action
of small amounts of free acid. Experiment showed that the
bromide-bromate mixture is also a very delicate indicator of
free acid; 0°00018 grm. of sulphuric acid proving to be sufti-
cient to liberate bromine from the bromide-bromate mixture
when boiled in the Voit flask under the experimental conditions.

That the progress of the hydrolytic reaction is so much more
rapid in the case of the iodide and iodate is apparently due to
the specific action of these substances upon the basic aluminium
sulphate rather than to a more exact removal of the free acid.

It will be recalled in this connection that, as Moody has
shown,* the iodide-iodate mixture, taken in moderate amount,
does not further hydrolyze the basic salt produced in the action

* Moody, this Journal, xxii, 184, 1906.

172 = Gooch and Oshorne—Potassium Aluminium Sulphate.

upon zine sulphate. A reasonable excess of the bromide-bro-
mate mixture is able in a moderate time to carry the hydrolysis
of aluminium sulphate to a fairly definite point cor responding
nearly to the removal of five-sixths of the acidic ion, while the
iodide-iodate mixture under similar conditions of ‘prolonged
action removes practically all the acidic ion.

Some experiments to test the effect of a mixture of potassium
chloride and potassium chlorate indicated that the hydrolysis
of aluminium sulphate under otherwise similar conditions is
very slight compared with that produced by the bromide-
bromate mixture or by the iodide-iodate mixture.

Phelps and Deming—Preparation of Formamide. 178

Arr. XVH.—The Preparation of Formamide from Ethyl
Formate and Ammonium Hydrowide; by 1. K. Pumres and
C. D. Demine

[Contributions from the Kent Chemical Laboratory of Yale University—clxi. |

Tre action of ammonia on an ester is a typical method for
the formation of an acid amide. Hofmann* states that this
reaction is particularly easy when the ester itself is somewhat
soluble in water; but that in the case of ethyl formate and
aqueous ammonia, when allowed to stand at the ordinary temper-
ature, a considerable part of the acid amide is hydrated to
ammonium formate, and the amount of acid amide formed is
accordingly diminished. He further states that this hydration
goes on to such an extent that a yield much in excess of 70 per
cent is impossible.

The work given here records the results of experiments
which show that such hydration during reaction may be
avoided and the formamide may be obtained in amount equal
to that indicated by the theory for the action from ethyl formate
and ammonia.

Ethyl formate for use in this work was prepared by treating
commercial sodium formate with a mixture of absolute alcohol
and concentrated sulphuric acid. The crude product obtained
as a distillate was fractioned and portions found to boil between
53° and 56° were treated with fused calcium chloride to remove
water and alcohol, and redistilled after filtering from the eal-
cium chloride. This distillate was then treated with dry
ammonia gas obtained by boiling concentrated ammonium
hydroxide in a flask connected with a return condenser and
leading the ammonia gas obtained in this way through a lime
tower. The ammonium formate precipitated was separated by
filtration, and then, by fractional distillation, pure ethyl formate
distilling within two-tenths of a degree was obtained free from
ammonia gas or other impurity.

In all the experiments recorded in the table detinite weights
of the pure ethyl formate, chilled in an ice and salt mixture,
were treated with ammonium hydroxide of known strength
similarly chilled, and the temperature of the mass was not allowed
to rise above 0° until after the reaction between the formate
and the ammonia solution had been largely completed. The
formamide produced in each case was obtained by separating
from impurities by fractioning in vacuo in the usual way and
distilling the residue after the low-boiling impurities, largely
ammonia, ethyl formate, alcohol, and water, had been re emoved.

* Berichte, xv, 977.

Am. Jour. Sci1.—Fourt# Series, Vou. XXIV, No. 140.—Aveust, 1907.
12

174 Phelps and Deming—Preparation of Formamide.

In certain experiments the ethyl formate and ammonium
hydroxide were mixed in a 250°" distilling flask and the mix-
ture was fractioned in vacuo as soon as it had become homo-
geneous. In this process the 250° distilling flask containing
the mixture was connected to a 100°™* distilling flask used as a

..er and the formamide was separated from low-boiling
. purities by fractioning in vacuo in the usual way, the distilling
flask being heated in a “water bath at 60° for fifteen minutes
after the pressure on the manometer registered 15™™. The
formamide was then distilled over, by heating the distilling
flask ina bath of sulphuric acid and potassium : sulphate,* and
caught in the receiver chilled by a current of cold water.

In other experiments the distillation was not made as soon
as the ethyl formate and ammonium hydroxide had become
homogeneous. In each of these experiments a definite weight
of ethyl formate held in a 250° glass-stoppered reagent bottle
was treated with pure, commercial ammonium hydroxide of
known strength, after chilling each in a mixture of ice and
salt, slowly enough so that at no time during mixing did
the temperature of the mixture rise above 0°. The mix-
ture at this time consisting of two layers became homogeneous
in about five minutes, but was allowed to stand in the ice mix-
ture for fifteen minutes; for the heat given out showed that
the reaction progressed after the mass had become homogene-
ous. This solution was kept tightly stoppered for a definite

Ammonium Formamide

Ethyl hydroxide Time of standing oa —
formate Sp. gr. °90 ——_~* an Theory Found

No germ. em’, hr. min. germ. grm.

A
(1) 50 30 4 30 30°41 19°06.
(2) 50 30 240 30°41 23°40
(3) 50 50 30°41 23°57
(4) #50 50 30°41 22°60
(5) 50 50 i 30°41 28°08
(6) 50 50 J 30°41 27°40
(7) 50 50 2 30°41 29°33
(8) 50 50 2 30°41 29°78
(9) 50 0 2 30°41 29°80
(10) 50 50 5 30°41 30°48
(11) 50 50 5 30°41 30°45
(12) 50 50 6 30 30°41 30°50:
B
Treatment with NH,OH and saturation of mixture with NHs.

(13) 59 10 4 30°41 30°45.
(14) 50 10 4 30°41 30°00
(15) 50 20 + 30°41 30°25.

time, after which it was transferred to a 250% distilling flask
connected to a 100° distilling flask used as a receiver with the
*H. Scudder, Jour. Am. Chem. Soc., xxv, 161.

Phelps and Deming—Preparation of Formamide. 175

use of the least amount of alcohol to remove the last traces
from the side-walls of the bottle. The formamide was separated
from low-boiling impurities by fractioning in vacuo in the
manner described, and the formamide distilled and w eighed.

It is obvious from the work recorded in the table that ‘three
factors are of influence in the quantitative formation of il.
mide by the interaction in the cold of ethyl formate a4.d
ammonium hydroxide,—the excess of the ammonium hydroxide
present, the concentration of the ammonia in the solution, and
the time during which the interaction of the two is allowed.
From experiments (1) and (2) of section A of the table it is
clear that with fixed amounts of ethyl formate and ammonium
hydroxide an increase in the length of the time of the reaction
is Naeadeale advantageous. Thee comparison of experiments (1)
and (2) with experiments (5) to (12) shows that increasing the
amount of ammonium hydroxide gives a yield in excess of that
found with the smaller proportion of ammonium hydroxide.
A comparison of experiments (5) to (12) with 6ne another fur-
ther emphasizes the point that length of time allowed for the
completion of the reaction after the mixture has been kept cool
during the first of the reaction long enough to prevent hydra-
tion is an important factor. It is clear that it is easily possible
to obtain the theoretical amount of amide by allowing the
mixture to stand five hours or longer. Two hours standing is
not quite sufficient.

In section B of the table the results shown were obtained by
saturating in the cold, at —10° to 0° for about four hours, the
ethyl formate and ammonium hydroxide chilled before mixing
in a stoppered reagent bottle fitted with a rubber stopper
carrying Im one perforation a thermometer dipping into the
mixture and in the other the delivery tube for the ammonia gas.
This mixture on fractioning in vacuo as described above and
distilling the formamide left in the distilling flask gave the
theoretical amount of formamide.

An experiment showed that a detinite weight of formamide
and water when fractioned in vacuo as in the process outlined
gave an amount of formamide differing from that taken by less
than 0-05 grm.

It is evident that it is possible to obtain from ethyl formate
by the action of ammonium hydroxide the theoretical amount
of formamide. This may be accomplished by treating the chilled
ethyl formate with small amounts of chilled ammonium
hydroxide and saturating for some hours in the cold with dry
ammonia gas; or, more easily, by treating chilled ethyl formate
with larger proportions of chilled ammonium hy droxide and
allowing “the mixture to stand some hours before distilling in
vacuo. The essential thing in the operation is to keep. the
mixture of the ethyl formate and the ammonium hydroxide at
a temperature so low that ammonium formate is not a product
of the action.

\
176 Shimer—Lower-Middle Cambrian Transition Fauna.

Arr. XVIIL.—A Lower-Middle Cambrian Transition Fauna
From Braintree, Mass. ; by H.W. Sumer.

Some time ago, while having a driveway excavated at his
home on Quincy avenue, in East Braintree, Mass., Mr. Thomas
A. Watson found a rather angular slate bowlder, about two feet
in diameter. He kindly turned it over to the Massachusetts
Institute of Technology.

The slate is quite similar in appearance to that of the cele-
brated Paradoxides quarry on Hayward Creek; it is similarly
metamorphosed but is lighter gray in color and lacks the pecu-
har purplish tinge of the Hayward Creek slate.

In it the followi ing fauna was found :

Name Abundance Previous occurrence Age
Acrothele gamagei (Hobbs) ____.__.--.- ie Hayward Creek Middle Cambrian
Hyolithes shaleri Walcott...._.__------ r Hayward Creek Middle se
Paradoxides harlani Green._---.___---- C Hayward Creek Middle BY
Strenuella strenua (Billings) -------___- R North Attleboro, Lower iy

North Weymouth
and Nahant

2Strenuella strenua (Billings)__-----___- R

Olenellus (Holmia) broggeri Walcott ..-. R North Weymouth Lower Cambrian
Ptychoparia rogersi Walcott ---.___-__- r Hayward Creek Middle i
Agraulos quadrangularis (Whitfield)... C Hayward Creek Middle sf

C = very common ; ec = common ; r = rare; R = very rare.

This fauna includes five species of the Middle Cam-
brian, two of which are very abundant, and two of the lower
Cambrian. There is thus a great pr edominance of the Middle
Cambrian element, though it indicates a persistence of the Lower
Cambrian element into Middle Cambrian times. So while we
have here a transition fauna, the rock must be assigned to the
Middle Cambrian period. This very interesting transition
fauna is the first recorded from this region to show in any way
a passage from the Lower to the Middle Cambrian.

The size and angular character of the bowlder would indi-
cate that it had very probably not traveled far from the parent
ledge. Since the glaciers which transported it came from the
northwest, we examined the drift in that direction. Under
the leadership of Mr. Watson we found on the granite hill
south of Quincy Adams very many slate bowlders, some similar
in appearance to the one containing the transition fauna, though
no fossils were found in them. These decreased in number
southeastward ; on the granite hill they composed most of the
drift. Their oreat abundance here would indicate their prob-
able derivation from the drift-covered valley of Quincey Adams,

a suburb of Quincy. Some support likewise is given to this

Shimer— Lower-Middle Cambrian Transition Fauna. 177

theory by the east and west strike and southern dip of the Mid-
dle Cambrian Hayward Creek beds lying a short distance to
the southeast.
Description of species.
Acrothele gamageti (Hobbs).
The two individuals found agree very closely with the type
description.

Hyolithes shalert Walcott.

Parts of several specimens of this species were found. They
represent individuals somewhat below the normal in size.

‘Olenellus (Holmia) bréggeri Walcott.

One individual was found. This includes most of the gla-
bella, and one eye-lobe. The eye-lobe is narrow, elongate, and
crescent shaped; it begins to arch outward from the most
anterior glabellar furrow. The glabella in outline and furrows
is very similar to the type description. Its surface is covered
with “ the inosculating, fine, raised fretwork that as far as known
is confined to the genus Olenellus.”’*

Strenuella strenua (Billings).

Our specimen consists of the characteristic ends of two
pleura. They belong without doubt to this species, as they are
similar in every respect to an individual described from the
Lower Cambrian of North Weymouth.t

? Strenuella sirenua (Billings). Fig. 1.

Referred to this species with extreme doubt is a specimen
that may be a thoracic axis. It
: consists of eight strongly arched
segments, and tapers to an unusual
degree for an axis. It bears a median
row of tubercles, one upon the cen-
tral portion of each fold, the series
becoming more marked toward the
broad end of the axis. The middle
of each fold is marked by the groove-
like line produced by the overlapping
of the anterior portion of the fold
upon the posterior.

The evidence in favor of referring
this specimen to Strenuella strenua
is its marked convexity, the median
row of tubercles, and the presence of
the pleura of that species near it,
though it is very doubtful if they are near enough to have
formed part of the same individual.

* Walcott, 10th Ann. Rept. U. S. G. S., part 1, p. 640.
+ This Journal, xxiii, 199-201.

178 Shimer—Lower-Middle Cambrian Transition Fauna.

Dr Goeike Matthew, who kindly examined the photograph
of the specimen, suggests that it may be a group of pleura
belonging to one of the spinous trilobites. This would explain
its curved character, indicated by the lack of parallelism
between the transverse grooves.

This specimen is in the collection of the Boston Society of
Natural History; catalogue number 13, 341.

Agraulos quadrangularis (Whitfield).

This species is exceedingly abundant but most of the indi-
viduals differ from the normal mature form of the Hayward
Creek Middle Cambrian. In transverse view the glabella is
higher and more arched; the longitudinal contour is of uniform
elevation in the posterior two-thirds and descends abruptly in
the anterior one-third, but as its height is greater than that of
the normal species so is its descent more conspicuous. The
anterior edge of the cephalon is truncate even in those forms
showing no compression, while in the normal species it is
rounded.*

The sides of the glabella are sub-parallel and do not taper
strongly forward as is usually the case with the normal form.
The spine is about half the length of the cephalon exclusive of
the occipital ring, while in the normal form it is about one-
third.

The very small (young) and very large (old) forms of A.
quadrangularis of Hayward Creek beds are quite similar to this
variety. It thus seems that this variety represents an earlier
(younger) form of the normal (mature) species and that the large
forms of the latter were merely gerontic specimens showing a
return in shape to the characters of youth.

Ptychoparia rogersi Walcott.

One specimen, somewhat crushed laterally, represents the
anterior and most of the central portions of the cephalon. The
anterior head fold is well shown and the glabellar outline.

Another specimen consisting of only the most anterior por-
tion of the glabella and cephalon shows this anterior fold more
faintly.

Paradoxides harlani Green.

There appears to be no difference between our many repre-
sentatives of this species and the normal mid-Cambrian forms
from Hayward Creek. One individual attained an enormous
size, measuring over three inches across the wider anterior
portion of the glabella. Several hypostomes were found, one
two inches wide and attached to the frontal doublure of the
head. Separate pygidia were also quite abundant.

Massachusetts Institute of Technology, Geological Department.

* Bull. Am. Mus. Nat. Hist., i, 147.

Geology and Netural History. G9
SCIENTIFIC INTELLIGENCE.

I. Gronrocy anp Naturau History.

1. The Bermuda Islands. Part IV. Geology and Paleon-
tology; by A. E. VrerrILt. 168 pp., 88 text-cuts, 12 plates.
Trans. Conn. Acad. Science, vol. xii, 1906—7.—Although the text
of this work was printed and a number of copies were distributed
in March, 1906, the complete work, with all the plates, has only |
recently been issued (May, 1907).* This part contains a very
detailed account of the geology of the island, and the changes
they have undergone, by submergence and elevation, erosion, ete.
The rocks are nearly all composed of wind-drifted shell sands
more or less consolidated, but a narrow belt of marine limestones,
formed in shallow water, and containing numerous marine shells
and foraminifera, mostly like those now living there, occurs in
many places a little above sea-level. The origin and nature of
the materials are fully discussed.

The most noteworthy novelty is the division of the rocks into
three distinct geological formations, representing periods parallel
with those of the North American coastal area. The earliest, named
the “ Walsingham” formation, is composed in part of very hard,
compact limestones, of solian origin. In this formation land
snails of many species are found, most of which are now extinct,
including the large forms of Precilozonites Nelsoni, of which
three varieties, all extinct, are recognized and figured. This
formation was deposited during a period of marked elevation,
probably at least 125 feet above the present level when greatest,
as shown by the submerged drainage channels and eroded escarp-
ments, submerged peat deposits, etc. This period is considered
to be preglacial and probably Pliocene. The larger sizes of the
land shells and their abundance indicate a warmer and more
moist climate, with abundant vegetation. This was followed by
a long period of submergence, ending in a depression rather
greater than at present. To this period, named the ‘“ Devon-
shire,” the raised marine limestones, with fossils, are referred, but
the exolian rocks of this age cannot be distinguished from the
later ones. It probably corresponds to the glacial and Champlain
(or Leda-clay) period of New England. The later eolian rocks,
named the “ Paget ” formation, contain in many places an abund-
ance of land shells, nearly all of which are still living on the
islands, though some have undergone some changes, recognizable as
varietal. The phenomena of erosion in all its forms are very fully
discussed and illustrated. The elevated shore cliffs, exposed to
heavy seas, furnish excellent object lessons of this kind. New

* The delay was largely due to a strike in the printing office.

180 Scientific Intelligence.

explanations of the origin of the remarkable structures popularly
called “fossil palmetto stumps” are given. These have been
variously interpreted by different writers. In this article they
are explained as due, at least in most cases, to the solvent action
of carbonated rain-water, localized at first. in depressions of the
surface rock by any one of several causes, but continued down-
ward by the formation of a central core of clay, ete., around
which the water can percolate and evaporate. They are com-
pared to the tubular “sand-pipes” of the English white chalk,
described by Lyell and others.

In the chapter on the paleontology all the land-shells hitherto
found fossil are described and figured, and a list is given of the
marine fossils of the Devonshire formation, with some figures.
Of the latter about ten species are not certainly known to be now
living in Bermuda waters, but they are all West Indian. The
Walsingham formation contains 17 species and 6 named vari-
eties of land shells, of which 9 species and 4 varieties are extinct.
This paper contains a Bibliography and a copious index.

2. The Bermuda Islands. Part V, Section IT, by A. E.
VERRILL. 147 pp., 28 plates, 1 map, and 120 cuts in the text.
Trans. Conn. Acad. Science, vol. xii, 1907.—This part is devoted
entirely to descriptions of the characteristic life of the coral
reefs, particularly the corals, actinie, gorgonians, hydroids, echi-
noderms, and siliceous sponges. The remaining groups are to be
treated in a subsequent section. Nearly all the species are well
illustrated by photographic half-tones, or by drawings. A num-
ber of new species and a new genus of sponges are described, and
two new gorgonians, one representing a new genus. Figures of
the spicules of all the gorgonians and sponges are given. Many
of the figures of the corals, gorgonians, and actinise are from pho-
tographs of the living animals, by A. H. Verrill. Those of the
gorgoniz are remarkably good. Nothing like them has been pub-
lished hitherto. Bibliographies of each group are given, and the
index is very complete. This work is admirably adapted for the
uses of students of tropical marine life, not only at Bermuda, but
in the West Indies and Florida also, for the species are nearly all
found in the West Indies.

3. Maryland Geological Survey, Calvert County; by GEORGE
BuRBaNnk SHATTUCK, Bensamin L. Minter and others. Pp. 227,
pls. xiv, figs. 11. Baltimore, Jan., 1907 (The Johns Hopkins
Press). —This volume is the fourth of a series of reports on the
county resources, and is accompanied by topographic, geologic
and agricultural soil maps on a scale of approximately a mile to
an inch, Calvert County constitutes a peninsula on the west side
of Chesapeake Bay, bounded on its own western side by the
Patuxent River. Its surface is an undulating plain with an extreme
elevation of 180 feet near the northern limits. The geological
formations are Miocene and Pleistocene with some Eocene.

Te

Geology and Natural Mstory. 181

4. Maryland Geological Survey, St. Mary's County ; by
GrorGE BurBank SuHattruck, Bensamin L. MituerR and others.
Pp. 209, pl. xvi, figs. 12. Baltimore, March, 1907 (The Johns
Hopkins Press).—This volume, of the same form as the preceding,
is the fifth of the series of county reports. St. Mary’s County is
the most southern county on the western shore of the Chesapeake,
lying south of Calvert County, and its geology is similar in many
respects to the latter. Both volumes contain full discussions of
the physiography, geology, economic resources, soils, climate,
hydrography, magnetic declination and forests. They are printed
and illustrated in the handsome style customary with the Mary- |
land Survey. os

5. Geological Survey of India.—The first part of Volume
xxxv of the Records of the Geological Survey of India (Cal-
cutta, 1907) contains the report of the Director, Dr. T. H.
Holland, for the year 1906 and two other papers. The former is
a concise summary account of the results obtained by the Survey,
and brings out many facts of interest. We may note an abstract
of the work of L. Leigh Fremor on manganese-bearing rocks of
Vizagapatam. The typical rock of the series, consisting of potash-
feldspar, manganese-garnet, and apatite, having commonly a
granular, medium granitic structure or sometimes assuming a
pegmatitic form, is called kodurite after the Kodur manganese
mine. Types marked by quartz, pyroxene, and biotite are noted.
The name spessart-andradite (shortened to spandite) is suggested
for the typical manganese garnet. It is also shown that the
aluminous laterites, or bauxites of India, which occur on a large
scale, are likely to prove important as a source of aluminium.

A brief statement is made in regard to the Dokachi meteoric -
fall, which is described in full in the same number by Mr. Fremor.
This fall took place on October 22d, 1903, and was accompanied
by some remarkable phenomena, the fire-ball being seen over most
of Bengal and Assam and beyond. Twenty-four fragments have
been collected by the Survey, ranging from 1,571 to 0°73 grms.;
these were obtained along a band running west by south from
Bibandi in the Dacca district to near the east bank of the Ganges
at Kolapara, a distance of six miles, the larger fragments having
fallen near the western end of the line towards which the meteor
was travelling when disruption occurred. The fragments are
nearly all covered with a crust, and on some of the faces of the
smaller fragments the formation of a younger crust shows the
fusion that occurred after disruption. Many more stones were
found in the various villages, the minimum number being estimated
at about 100. Several plates give reproductions of photographs
showing the aspects of prominent specimens.

6. Brief Descriptions of some recently described New Min-
erals._-RUTHERFORDINE—Or, better, rutherfordite (since Shepard’s
name given in 1850 has now no standing in the literature)——is a
uranyl carbonate described by W. Marckwald from the Urnguru
Mts. in German East Africa, and named after Prof. Ernest

182 _ Serentific Intelligence.

Rutherford. It occurs as an alteration-product forming a crust
over masses of pitchblende (U,O, 87:7 p.c., sp. grav. = 8°84).
It has a yellow color and specific gravity of 4°82 ; it is highly
radio-active. An analysis gave :

UOs CO. PbO FeO CaO H.O Gangue

83°8 12°1 1:0 08 ikea O77 |) 038 1008S
For this the composition UO,.CO, is calculated.— Centralbl. Min.,
p- 761, 1906.

HELLANDITE is a silicate of complex composition containing
yttrium and erbium with aluminium and calcium chiefly ; it is
described by W. C. Brogger from the pegmatite veins in the
neighborhood of Krageré, Southern Norway. It occurs in pris-
matic crystals referred to the monoclinic system. The material
was more or less altered ; the freshest had a nut-brown color
with hardness = 5°5 and sp. gravity = 3°70; more altered crys-
tals had a brownish red color with the sp. gravity ranging from
3°55 to 3°33. An analysis of somewhat altered crystals by L.
Andersen-Aars gave :

SiO, Al,O; Fe,03 Mn.0O; Ce.O3 Y.O3 Er.03 Tho, CaO MgO Na,O K.O
23°66 10:12 2°56 5:91 1:01 19°29 15°48 0:62 9°81 0:10 0°23 0:06
H.O 11°75 = 100°55

A relation in form and composition to guarinite is suggested, as
also to dauburite and topaz.—Zeitschr. Avyst., xii, 417.

PoporitE is described by W. Tschirwinsky as a_ phosphate
and carbonate of calcium having the composition 3Ca,P,O..
CaCO,, analogous to apatite. It occurs in the phosphorite of
Government Podolia in South Russia. Minute prismatic hex-
agonal crystals, yellow in color, are noted ; also small spherulitic
forms with fibrous structure. Only a minute quantity of the
crystals was available for analysis ; from this the approximate
results obtained. were:

P.O; 39°04 CaO 51°15 Fe,0; 3°04 CO, 3:90 = 97:14

The evolution of CO, was obtained from perfectly clear and
homogeneous crystals, free from calcite, when treated with
hydrochloric acid. The name carbapatite earlier suggested for
this mineral by P. Tschirwinsky is now withdrawn by him.—
Centralbl. Min., No. 9, 1907, p. 279.

Nepovuire is a hydrated silicate of nickel and magnesium
described by E. Glasser from New Caledonia. It is related to
the similar compounds called garnierite and noumeite, but occurs
in minute hexagonal scales often grouped in vermicular forms
analogous to the chlorites. Analyses lead to the formula
3(Ni,Mg)O.S10,.2H,O, but show a wide variation in the relative
amounts of nickel and magnesia, from NiO 50°7 and MgO 3:00 to
NiO 18:21 and MgO 29°84. Bull. Soc. Frang. Min., xxx, 17,
OMe

GorceixiTE and Hartrire are phosphates identified by E.
Hussak among the ‘“favas” of the diamond-bearing sands of

Geology and Natural History. 183

Minas Geraes, Brazil. The name fava is given locally to spher-
ical or bean-shaped rolled masses of some heavy minerals found
in the diamond washings and regarded as a good indication of
the presence of the gem. They are for the most part brown or
red in color, less often white. Early investigations of them have
been made by Damour, Gorceix and others. Hussak has identi-
fied among the favas, a barium-aluminium phosphate to which
he gives the name gorceixite. ‘The color is light- to dark-brown
and the structure like jasper; hardness = 6; specific gravity
3:04-3:12. An analysis by G. Florence, after the deduction of
S$i0,, Ti0,, Fe,O,, gave:

P.O; Al.03 BaO CaO CeO H.O
24°48 37°68 16°60 3°82 1°67 15°74 = 100

For this the composition BaO.2A1,0,.P,0,.5H,O, is calculated.

The name harttite, after Prof. Fred. Hartt, is given to favas of
a flesh-red to yellow or white color; hardness about 5°5, sp.
gravity 3°14-3°21. Microscopic examination showed a micro-
crystalline structure and suggested hexagonal crystallization.
Analysis by G. Florence gave :

P50; Al,Os3 SO; SrO CaO H.O
21°64 34°40 11°78 17°17 2°19 12°81 = 100

The calculated composition is Sr(Ca)O.2A1,0,.P,0,S0,.5H,0 ;
this shows a close resemblance to svanbergite. Other favas
proved to be phosphates of aluminium and lead closely similar to
plumbogummite.— Min. Petr. Mitth., xxv, 335, 1906.

7. Chiastolite from Bimbowrie, South Australia.—The argil-
laceous schists of Mt. Howden, ten miles north of Bimbowrie,
South Australia, have furnished some remarkable specimens of
chiastolite of unusual size and showing some new features for the
mineral. An account of them was published by C. Anderson in
the Records of the Australian Museum, vol.iv, pp. 298-302. The
locality has been specially developed by Mr. G. R. Howden, who
has furnished specimens to the Australian Museum and also to the
Yale Museum at New Haven.

The chiastolite occurs either embedded in argillaceous schist
or as rolled pebbles with quartz, jasper, etc. Some of the com-
mon forms show a simple cross of nearly pure mineral with also
a light-colored core, surrounded by four dark lines, forming a
rectangle when cut normal and a lozenge when cut oblique to
the axis. These specimens measure up to 24 inches in cross-
section.

A very peculiar form is that called “fish crystals,” the struc-
ture of which is best revealed in weathered specimens. These
have an elongated and flattened form, the length often reaching
five or six inches. They have a central core with projections on
opposite sides (see fig. 1); these show a peculiar series of grooves
and ridges, curved on either side of the axis and in opposite
directions. In some cases, also, the longer arms are slightly bent
in opposite directions. This deformation is well marked when
the grooving is most apparent.

184 Scientific Intelligence.

Other specimens show a complex and beautiful structure in
the cross-section, as suggested in figure 2 (drawings by Mr.
Howden). This variety, which is locally known as howdenite,

2

occurs in square or rhombic faces, often with concave faces, and
enclosed by a green crust from 34th to {th inch in thickness.
The purer portions show many tints of color from white or gray
to yellow, pink, red and purple.

OBITUARY.

Professor ANGELO HetLprty, the naturalist and geologist, died
on July 17 at the age of fifty-four years. A notice is deferred.

TAH

AMERICAN JOURINATE Ole SCTENCE

[FOURTH SERIES]
—_ $+ —___

Arr. XIX.—Plains in Cape Colony ; by Professor E. H. L.
Scnwarz, Rhodes University College, “Grahamstown, Cape
Colony.

Tue southern coastal belt of Cape Colony is divided into
two portions, the western with its lofty and rugged mountain
ranges, the eastern with more gentle topography. Standing
on one of the lofty peaks near the junction of the two zones
an almost bird’s-eye view can be obtained of the lower country
to the east: one sees the land stepped or parcelled out into
narrow shelves with abrupt margins Haeiue the: sea. The
highest of these lies at this point about 2,500 feet and the
width of the terraced portions is 30 miles. Travelling through
this country, say from Port Elizabeth to Bellevue, the highest
point on the railway near here, the terraces are scarcely appar-
ent, as the great fall of the rivers has caused deep gullies to
be cut, and the subsequent denudation has widened them
sufficiently to obscure the details, while the Addo bush, a scrub
forest, covers the more characteristic features ; besides this, in
the Port Elizabeth area, a belt of soft Cretaceous deposits has
been faulted down against the older rocks, and this has allowed
wide alluvial plains to be developed.

Farther east in Pondoland the shelves are far more distinct.
The rock is of uniform hardness throughout, and, supporting
a rich covering of grass, it has been cleared to a large extent
by repeated burning “of all obscuri ing bush and forest.

In the west the shelves are not very apparent except the
extensive 700-foot one all along the coast, but in favorable
spots, especially at the head streams of the oreat rivers where
erosion has had little opportunity of acting, the terraces can
be picked out, and the level certainly carried up to 4,000 feet,
perhaps higher, to the full maximum elevation of the continent,
namely, 6,000 feet.

Am. Jour. Sct.—FourtH Series, Vou. XXIV, No. 141.—SrpremBer, 1907.

15

186 EE. Hf. L. Schwarz—Plains in Cape Colony.

These shelves cannot be other than surf cut: we see them
forming now along the shores by the action of the waves (figs.
1 and 2), the cliffs crumbling with the attack of the breakers,
and the fallen debris washed to and fro with the perpetual
surge that desolates this treacherous coast-line ; the ledge thus
cut is sometimes washed bare, at others covered with what is
really a coral reef, but in which the lime-seer eting organisms
are the red algae, ‘shells and worms. Behind the actual coast
the individual ledges may attain a width of 30 miles as at
Caledon, but near Port Elizabeth they are much narrower,
seldom reaching five miles in width, while the subsidiary
smaller ledges are much less. It is this narrowness that struck

Fia. 1. Arch and cave along the Knysna coast, near Seal Point, Cape

Colony. The rockis quartzite belonging to the Table Mountain Series. The
undermining of the cliffs by the battery of the waves is shown, but the
coast ledge cut by the dragging to and fro of the debris is covered at high
tide. The top of the cliffs belongs to the 700-foot ledge and is here about
600 feet above sea-level.

me while examining the country in the Zitzikamma, a coast
district east of Port Elizabeth, and induced me to revert to
my original idea of their being surf cut, after having tempo-
rarily thought them to be plains of river erosion, “for it is
impossible for a shelf to be cut by a river when the fall from
the mountains behind is so great and the course so short.

Besides this indirect evidence there are undoubted sea-beaches
left on these ledges to attest to their having been cut by the
surf. Over the flats one finds great beach ‘bowlders between
which are wedged the broken shells of Pectunculus and other
marine forms, “mostly of a sub-tropical aspect, foreign to the
shells now living along the coast; elsewhere, as at Sandflats,
foraminiferal sand had collected in large masses and is used

Oe

EI. A. L. Schwarz— Plains in Cape Colony. 187

for building purposes. I am busy collecting material from
these sea-beaches, but the process is a slow one, as in most
cases the shells themselves have been dissolved out leaving
only imperfect casts behind, but the evidence so far is that
the forms represented belong to living species, though these
are now confined to the warmer waters farther north. These
sea-beaches extend from 10 feet above sea-level, or actually at
sea-level, at Port Elizabeth, and can be followed on the ledges
up to 1,300 feet on the Addo Hills; on the higher ones the
calcareous material has all disappeared.*

a H i

!

i

Ti

se
Aer

Fic. 2. The mouth of the Groot River, Zitzikamma, Cape Colony, a little
east of the spot shown in fig. 1. The coast ledge is revealed at low water ;
the ledge behind beiongs to the 700-foot shelf. The rock is quartzite of the
Table Mountain Series: the abrupt change of strike is seen.

Besides these exposed shelves, there is a submerged one like-
wise divided into subsidiary ledges which range from 20 to
60 fathoms, the outermost deepest one lying 90 fathoms below
the ocean surface. The edge of this plain is called the Agul-
has Bank, beyond which is very deep water; it is a zone of
danger to the shipping, as the waves from the Antarctic are
broken here and boil and swirl before sending their crests
towards the land over the shallower sea-bottom. The edge of
the Agulhas plateau continues the southwesterly coast-line-of
the eastern side of South Africa to the longitude of Cape
- Town, and the submerged plain must always be considered as
part and parcel of the present continent. Dr. Gilchrist has
reported bowlders on this plateau 40 miles from land and this
in itself is sufficient evidence of the surface having once been
above the sea; besides this I am unable to conceive of any

* Rogers, A. W., Ann. Rept. Geol. Comm., p. 43. Capetown, 1906.

188 FE. HW. L. Schwarz—Plains in Cape Colony.

natural cause for its existence. Strong currents wash over the
bank and drag the sand to and fro over it, but the sand here
is mostly—I might almost. say entirely—composed of com-
minuted shells, and even if the grains were quartz, the erosive
action of such sand suspended in water would hardly be
greater than that of hailstones on land. Beyond the Agulhas
Bank the ocean deposit is glauconitie.

It has been too much the fashion lately to ascribe every
plane surface on land to the action of rivers, but every plain
must be considered apart, and the causes that led to its forma-
tion examined with no preconceived ideas, for there are many
ways of explaining such surfaces.

(1) A plain may result from the long emergence of a rug-
ged land, when the rocks piled high, in some eases above the
crushing strength of the rocks at the base, will in time flatten
out. Sucha plain is hardly likely to arise on the earth under
present conditions, as denudation works more rapidly than such
action could, but in planets without an atmosphere it might
become operative. At all events, it is a curious coincidence
that in a former paper,* when estimating the maximum height
of a rock mass which could be self-suppor ting, the figure came
out approximately the same as the height of the highest
mountain in the world; that is to say, if my figures are correct,
any mountain thrust ip to a higher altitude than Mount
Everest would slowly subside by mass flowage to a mean of
approximately 30,000 feet.

(2) A plain may arise by an inland sea being filled up, and
then the whole being elevated by block uplift. Supposing the

southern portion of ‘the Caribbéan Sea, or parts of the Yellow.

Sea, were elevated, a natural plain would result. Such a plain
I believe once existed in South Africa. There is an old Per-
mian shore-lne running aprontimately northeast-southwest,
through the Southern Transvaal+; t; on the southward side of

this ARore are found Karroo sediments filling in a wide basin.

whose southern border is now beneath the sea. Towards the
close of the Jurassic period, elevation of the whole of South
Africa took place, and this plain was steadily raised 2,000 feet.
At this time there was a simple watershed extending from
what is now Cape Town to Delagoa Buy, lying parallel to the
old shore line, and from which the rivers came off in or derly
succession, some to the northwest, others to the southeast.
These rivers are still the dominant ones in South Africa, and
are only obscured in the east by a gigantic dam of lavas, which

* An unrecognized agent in the deformation of rocks. Trans. S. A. Phil.
Soc., vol. xiv, p. 887. Capetown, 1903.

+ Schwarz, E. H. L., The Volcanoes of Griqualand East, S. A. Phil. Soc.,

Capetown, vol. xiv, p. 108, 1903; see also Passarge, S., Die Kalahari, p. 63.
Berlin, 1904.

E. H. L. Schwarz— Plains in Cape Colony. 189

was poured out at this time and forms the Drakensburg ; this
thrust back the waters over the main watershed in what is
now called the Orange River.

I have elsewhere* given my reasons for considering the old
Karroo plain, now elevated 6,000 feet above sea-level, to have
been one of original deposition. Similar plains may be formed
in inland lakes as well as in enclosed portions of the ocean, and
I think it should be definitely recognized that such a plain can
be and has been formed by giving such a surface the techni-
cal name of a plain of deposition to distinguish it from a plain of
denudation, qualifying the term by the adjectives marine and
freshwater, according to circumstances. Generally the term
also includes the contingency of block elevation. I believe the
greatest example of this form of plain is that of the pampas of
South America. Ordinary deltas would come under this cate-
gory as well.

3
Kammanassie Mountains Wolve d Wilge Boor Sagers WOE
a
SS SE
ZL ~SS _ZAZZA

g ZZ

Fic. 3. The upper end of the Olifant’s River, Oudtshoorn; showing
the 4000-foot archplain, P. The Cretaceous beds, R.E. and W.E., were
deposited on a plain at this level, were tilted and subsequently cut down to
the same level. Behind arethe folded mountains with a remarkably straight
crest ; on the north, the Swartberg mountains show a similar crest, and these
two suggest that they once were part ofa peneplain at that level; the folded
ranges, however, are far younger than the original plain of deposition. In
the low ground isatemporary peneplain cut in soft Cretaceous beds and due to
the barring of the river further down. T.M.S., Table Mountain beds; B.V.,
Borkkeveld beds (Devonian); R.E., Red (Knon) Conglomerate; W.E.,
White (Enon) conglomerate.

(3) During the extrusion of lavas South Africa sank, to rise
again at the termination to its former level, now 4,000 feet
above the sea, so that there are two clearly marked levels at
this elevation, some cut in older rocks, and some cut in Lower
Cretaceous rocks which were deposited on the older 4,000-foot
level, were folded and tilted, and again eroded or planed down
(fig. 5). Both of these 4,000-foot levels extend far inland and
must have also been represented by coast plains at about the

* Schwarz, 4H. H. L., The Rivers of Cape Colony, Geographical Journal,
pp. 265-279 ; London, 1906; see also Baviaan’s Kloof, a contribution to the

theory of Mountain folds, Addresses and papers, Brit. & S. A. Association
for the Advancement of Science, pp. 56-67. Johannesburg.

190 £. H. L. Schwarz— Plains in Cape Colony.

same elevation. In the long narrow valleys between the moun-
tains the character of the plains becomes that of a double bevel—
clearly a plain of river erosion (fig. 4),so that we must in this
instance distinguish between a plain of marine erosion, a plain
of river erosion and one which has been of so general a nature

that it includes both. A surf-eut shelf will naturally be a little

below that of a river eroded plain, but with long base-levelling,
both will so nearly approach the same plane that they will
become indistinguishable if only moderately old. I have

loosely referred to these great plains as peneplains, and if we can

Fic. 4. Bird’s-eye view of the northern half of the double level in the
Kammanassie River, Oudtshoorn, Cape Colony. The 4000-foot plateau
which opens out towards the sea and becomes a plain of marine denudation,
in the narrow valleys between the mountains becomes a double level of
(river) erosion. The river which runs parallel to the Olifants has reached
the same temporary base level at a much slower rate, owing to the more
resistant rocks traversed, and hence no alluvial plains have formed along its
course.

include a marginal zone of marine denudation the term will
stand, but Lam rather inclined to think that writers have always
excluded marine denudation from their definition, and in such
case we must either coin a new word, such as archplain, or
refer to them simply as plains of denudation.

(4) Peneplains or plains of river erosion have been sufli-
ciently described by Tarr, Davis and others. Temporary pene-
plains may be formed by the barring of rivers. Of these we
have many excellent examples i in South Africa. On the south-
ward side of the main watershed the country is everywhere

ee

E. H. L. Schwarz— Plains in Cape Colony. oe

deeply cut into by kloofs and river gorges, but on the north,
the great falls of the Orange river at Augrabies interposes a
bar of resistant granite to the downward sawing of the river.
As a consequence, behind the falls there are extensive plains,
the rivers are low grade and when they do happen to have
water in them, which is not often, the water spreads out into
great kolks and floors some of which are 3800 and 400 square
miles in extent, for instance, Verneuk Pan. In one place on
the southward side of the main watershed, north of Willow-
more, an effective bar is produced by the rivers being deflected
along the strike of the beds ; where the neighboring rivers, ex-
actly similar in fall and volume, strike the mountains at right
angles, they have sawn through them with apparently no diffi-
culty, but in this particular spot the rivers run parallel to the
strike of the same mountains, the erosion has been insufficient
to allow the rivers behind to cut downwards and a temporary
peneplain has resulted.*

(5) Plains of marine erosion are seldom more than ledges,
and I would suggest that in order to keep the two forms dis-
tinet they should not be referred to as plains. They are cut
by the surt, which undermines the cliffs, and by the off-shore
currents which drag the debris backwards and forwards. Gen-
erally, however, they may be said to be formed between wind
and water, i. e. between high and low waters. Where the
rock is soft, for instance in the case of the Upper Cretaceous
deposits of Eastern Pondoland, the ledge is so rapidly cut and
extends so far that the outermost edge is only exposed at
extremely low tide. This case has puzzled me as to whether
erosion has gone on below the general low-water mark or
whether the submergence is due to slight sinking; remember-
ing, however, the recent sea-beaches round Port Elizabeth
elevated only 10 feet or so above high-water mark, showing
recent emergence, I am inclined to include off-shore currents
as effective in denudation in exceptional cases and where the
rock is soft. In regard to the submarine plains such as those
on which the Maldive Islands stand, I cannot agree with Stan-
ley Gardiner* that they can have been formed by submarine
erosion. Soundings in many places showed that the surface
had been swept by currents clear of all deposit, but as this is
only composed of soft muds, the shells of foraminifera and
such like, even on dry land it would have very little effect in
wasting rocks harder than itself. In hard rock such as we
mostly find on our coast line, the process of cutting shelves is
a slow one and the resulting ledge when elevated above the
sea is a correspondingly narrow one.

* Schwarz, E. H. L., Geological Survey of parts of Prince Albert, Willow-

more and Uniondale, Ann. Rept. Geol. Comm., p. 97; Capetown, 1904 ; see
also Geographical Journal, p. 272, 1906.

192 LE. H. L. Schwarz—Plains in Cape Colony.

I am of opinion that our levels up to 2,500 feet along the
coast are surf-cut, but most of them where they run between
narrow valleys orade into bevels of river erosion, so that we
must not draw too tine a distinction between such surfaces as are
cut by marine and river erosion. As, however, the dominant
feature of all of them is marine erosion, we can legitimately
call them coast ledges (fig. 5).

Dr. Reusch has described the same phenomena in Norway
and uses the term Strandflade or coast plane, but plane cannot

i)

~

(ERS IES = bh
ie ai

ae Sy, SS x i

oT CAEN

Sn al
may SLAY RGR See Nos; Wie

Fie. 5. View overlooking the coastal plateau near Plettenberg Bay, Cape
Colony. The ridges on the left are part of the dissected 1500-foot ledge, and
those beyond belong to the 700-foot ledge. The break from the one to the
other is shown. The crests of the mountains behind rise to the 6000-foot
level, the highest peak being 5497 feet.

in English be nsed for a natural surface, while plain includes
the ideas of both length and breadth and is inadmissible.
Ledge or shelf appears to me to be the only term we can use,
while a collection of these ledges rising step-like on the borders
of a continent might be called a klimakotopedion,—a stepped or
terraced plain. On the Atlantic border of Europe and Amer-
ica the klimakotopedia are submerged.

(6) When a continent has been shaped it retains its outlines
for a considerable period unchanged. South Africa, including
the Agulhas ledge, is sharply bounded on the west by a north
and south coast line, and an almost straight northeasterly one
on the eastern side. Now there isa limit below which denu-
dation as regards the present history of South Africa has never
acted. This I have called the absolute base level of erosion.+
I had in view when suggesting the term the comparison of
the surface features of South Africa and Europe, the latter
with dissected topography, the other with the story of its eleva-
tion still so apparent in the flat tablelands, and I wished to
obtain some quantitative estimate of the relative positions of

* Stanley Gardiner, J., The Indian Ocean, Geographical Journal, vol.
XXvili, pp. 320; London, 1906; this Journal, xvi, 203, 1903.

+Schwarz, E. H. L., Coast Ledges in Southwest of Cape Colony, Quart.
Journal, Geol. Soc., vol. 1xii, p. 84. London, 1906.

E. H. L. Schwarz—Plains in Cape Colony. 193

the two continents in respect to some datum. The absolute
base level of erosion I found to lie submerged some 9,000 feet
in Europe and 1200 feet in South Africa.

Beyond the absolute base level of erosion lie the plains of
the abyss, but we cannot legitimately include these in possible
land forms unless we can prove that the theory of the perma-
nence of ocean basins is an illusionary one.

(7) Finally there is the plain of deposition due to subsidence.
Of these we have many examples along our coast, and to
understand them it must be remembered that our rivers run
out to the sea through rock gates and never form deltas. Eleva-
tion and subsidence are alternate motions in the earth’s crust,
and whereas the evidences for elevation are so marked in South
Africa, those for subsidence are necessarily as plain. In the
mouths of the Knysna, Keurboom, Kowie and Buffalo rivers,
the rock channel has been cut to a hundred or more feet below
present sea-level. In the last cited case the contours of the
rock bottom have been accurately surveyed by Mr. J. J. God-
frey by means of bore-holes, and show a normal V-shaped
river channel with three terraces, the deepest part being 124
feet below low water mark. In this lie accumulations of clay,
sand and shells; the latter were sealed up while yet decompos-
ing and have yielded reservoirs of marsh gas which caused
explosions in the bore-holes when tapped. The actual river
bottom hes now 14 feet below low-water mark.* These sub-
sidence plains of deposit are naturally limited in extent, but
they are sufficiently large in some places to be taken into
account in the present connection. In the Knysna estuary
they form wide salt-meadows which are partially covered at
high water, but with very little work could be reclaimed. +

* Schwarz, E. H. L., The Rock Channel of the Buffalo River, East London,
Records Albany Museum, Grahamstown, vol. ii, p. 1, 1907.

+The Coastal Plateau in George, Knysna, etc., Ann. Rep. Geol. Comm.,
p. 85, map. Capetown, 1906.

194 J. K.and M. A. Phelps—Use of Zine Chloride.

Art. XX.—The Use of Zine Chloride in the Esterification
of Succinic Acid; by 1. K. and M. A. Puetps.

[Contributions from the Kent Chemical Laboratory of Yale Univ.—clxii. }

Iy a former paper* from this laboratory it has been shown
that in the action of succinic acid with ethyl alcohol contain-
ing hydrochloric acid to form diethyl succinie ester, the largest
yield from a known weight of succinic acid is obtained where
maximum dehydration is accomplished. In the work given
here the action of zine chloride in forming diethyl succinic
ester in the mixture of succinic acid and ethy! alcohol both
with and without the addition of hydrochloric acid is shown.

The apparatus figured was used in all of the experiments
recorded here. Tworound bottom flasks of 500°™ capacity, each
provided with inlet and outlet tubes
held in rubber stoppers, were con-
nected as shown. Of these, B, con-
nected to a condenser through a
Hempel column, carried a thermom-
eter, and from flask A an inlet tube
adjusted to dip beneath the lquid
to the same depth as the thermom-
eter. The flask A carried a separa-
ting funnel provided with a drying
tube, as well as the exit tube to the
flask B. In the flask B succinic acid
was heated, by means of a bath of
acid potassium sulphate, in some ex-
periments with a definite amount of
absolute alcohol alone, in others with
the same amount of absolute alcohol
containing hydrochloric acid gas,
while from the flask A gaseous alcohol in most cases, and in a
few cases gaseous alcohol with hydrochloric acid, was driven
over into the mixture in B where esterification took place.
The crude succinic ester left in flask B was freed from impuri-
ties by treating with sodium carbonate solution after first
removing the zine chloride by washing with water. The ester
was freed from carbonate ‘by rinsing with distilled water
containing sodium chloride. The mass of ester carried mechan-
ically with the several wash waters was extracted by shaking
out separately three times with ether. The ether extracts
and the succinic ester were gathered in a 250° distilling flask
fitted in the usual way for a vacuum distillation with a “1008°
distilling flask used as a receiver, and, after fractioning off
low boiling impurities, largely ether, alcohol, and water, was

1

* This Journal, xxiii, 368.

I. K. and M. A. Phelps— Use of Zine Chloride. 195

distilled and collected in the receiver, cooled by a stream of
water striking it constantly, and then weighed.

The pure zine chloride of commerce was freshly fused for
use in the experiments. The alcohol employed was made as
anhydrous as can be obtained by successive treatments with
fresh quicklime. The alcohol containing hydrochloric acid was
charged with the dry gas in the proportion of ten grams to the
liter. The succinic acid used was in most cases the pure acid
of commerce, in the others, pure succinic acid made by recrys-
tallizing the product formed by the hydrolysis of the pure ester
in the presence of nitric acid.

The result obtained in experiment (1) of the table was found
by heating with an acid potassium sulphate bath in the flask
B a known weight of succinic acid with 40° of the total
amount of absolute alcohol used in presence of ten grams of zine
chloride at a temperature of 100° to 110°, while driving into
it the remainder of the alcohol in the form of vapor from
flask A. All vapors from the flask passed through the Hem-
pel column to the condenser. The Hempel column had an
active surface of beads of 10°° in height and a diameter of 2°.
At its lower end it was in connection with a tube 5° in length
and of 0°5™ bore, and had an opening blown in its side 1°5™
from the end. By the use of a column of this construction the
hot vapor is enabled to go upward while the condensed liquid
flows downward readily. It was found that by the use of this
column neither succinic ester nor succini¢ acid distilled in such
amount, if at all, that it could be detected in the liquid distil-
late. The impure ester in flask B, when all alcohol as vapor
from flask A had been passed into it, was cooled and then
poured into a separating funnel containing water with ice, using
a small amount of ether to rinse the ester from the flask, and
the zine chloride was removed as far as possible. After sepa-
rating the ester from this solution any acid impurities were
neutralized with an excess of sodium carbonate in solution, and
the ethereal solution was then washed with distilled water.
The aqueous solutions in which the ester had been washed
were shaken out three times separately with ether.. The ethe-
real extracts were gathered in a 250° distilling flask connected
in the usual way for a vacuum distillation ah a 100™* distil-
ling flask used as a receiver. The low boiling impurities, ether,
alcohol, and water largely, were separated by a vacuum fr action-
ation, the 250° flask being heated in a bath of hot water at
60° finally for fifteen minutes after the manometer registered
15™" and the succinic ester was then distilled and w eighed.

The procedure in case of experiments (2) and (3) was the
same excepting that the alcohol used was charged with hydro-
chloric¢ acid in the proportion of ten grams to the liter ; and in
case of (3), and, also, in case of all other experiments in the

196 L. K. and M. A. Phelps—Use of Zine Chloride.

table where only one gram of zine chloride was used, the rins-

ing with cold water before neutralizing with sodium carbonate
was omitted. In experiments (4) to (10) the 40™* of alcohol
heated with the succinic acid contained 1°25 per cent of hydro-
chlorie acid gas, while absolute alcohol in amounts recorded in
the table distilled into this mixture heated at 100° to 110°.

Succiniec ester

Succinic Aleohol Reaction —— a s

acid ZnCl, with HCl time Theory Found per

No! grm. germ. )cm® | percent, hr) min.) erm. grm. cent

(1) 50 10 200 O 2 30 UB27 66°90 90°8

(2) 50 10 200 1325 Aye 50 EBT 71°25 96°7

(3) 50 1 200 IES) nie 45 HOOT 69°40 94°2
160 0)

(4) 50 10 40 Vie25 4 Wows 69°70 94°6
160 0)

(@) SO 10 40 D257 Pe 1 O30 eho OOceegO iia
60 O

(6) 50 1 40 1:25 24m 20 HET Haya ye iis) 72°]
60 0)

(7) 50 ] 40 1°25 BE DOM Ol, 7 tO OD mame ORG
160 0)

(8) 50 1 40 1°25 i Wrz dual 70°30 95:4
160 0)

(9) 50 1 40 125 te! 45 HOM TES 97°4
160 0)

(10) 50 1 40 1°25 jes 50 73°7 71°88 97°5

From an inspection of the results in the table, it is evident
that in presence of zine chloride to the amount of ten grams
with the proportions of succinic acid and alcohol given in the
table, a fair yield of succinic ester is possible. Introducing
hydrochloric nei and shortening the time of the reaction tends
to increase the yield as shown by (2). Reducing the amount
of zinc chloride present gives satisfactory results, as is clear by
comparing (5) with (9) ‘and (10). Although the amount of
alcohol present in (6) and (7) during the reaction is double that
theoretically required to esti the acid present, it is not
sufficient for the esterification under the conditions of the
experiments. The simplest conditions of all those given in
the experiments where a yield is satisfactory are those under
which (8), (9), and (10) were made.

Hence, it is clear that in presence of zine chloride diethyl
succinic ester is easily obtained in large amount closely approx-
imating that theor tically possible from a known amount of
succinic acid. This is most easily done by heating at a tem-
perature about 100° succinic acid with alcohol containing a
small amount of hydrochloric acid in presence of zine chloride
in small quantity, while gaseous alcohol is driven into the
mixture.

Drushel — Volumetric Estimation of Lanthanum. 197

Arr. XXI.—On the Volumetric Estimation of Lanthanum
as the Oxalate; by W. A. DrusHeEt.

[Contributions from the Kent Chemical Laboratory of Yale Univ.—clxiii. |

Nearty thirty years ago Stolba* stated that cerium, lantha-
num and didymium may be estimated by treating their oxalates
with potassium permanganate in the presence of sulphuric acid,
but gave no experimental evidence in the form of analytical
results. Later this statement of Stolba was contirmed by a
paper from this laboratory+ in which it was shown that cerium
may be estimated by precipitating cerium oxalate with a defi-
nite amount of a standard solution of ammonium oxalate used
in excess. The precipitated cerium oxalate was decomposed
by dilute sulphuric acid and the oxalate estimated by perman-
ganate, and the ammonium oxalate in excess of the amount
required for the precipitation was also estimated by permanga-
nate. By this process the results were checked.

The work to be described was undertaken to determine
the best conditions for the estimation of lanthanum as the
oxalate and also to furnish the desirable experimental data in
support of Stolba’s original statement.

For this work about ten grams of pure ammonium lanthanum
nitrate were prepared by separating the lanthanum and didym-
ium trom the cerium in a kilogram of the mixed sulphates
by Mosander’s chlorine method. The mixture of lanthanum
and didymium chlorides thus obtained, having been shown to
be free from cerium by the hydrogen peroxide test, was con-
verted into a mixture of ammonium lanthanum nitrate and
ammonium didymium nitrate. By fifty recrystallizations about
forty grams of pure ammonium lanthanum nitrate were ob-
tained. This salt was dissolved in eighty cubic centimeters of
water, and the solution throngh a depth of nineteen centime-
ters showed no trace of absorption bands. A few cubic centi-
meters of the solution evaporated to dryness and ignited gave
a pure white oxide. Out of this solution of the double nitrate
about ten grams were recrystallized for the experimental work
of this paper.

The procedure is as follows:

From a neutral lanthanum solution (a 1 per cent ammonium
lanthanum nitrate solution being used in this work) the oxalate

oy n
is precipitated by a measured amount of standard Fig: oxalie¢

acid, or ammonium oxalate after the addition of a few drops of

*Sitzber. d. Bohm. Gesellsch. d. Wissenschaften, iv, July, 1879.
Zeitschr. fiir Anal. Chem. xix, 194.
+ Browning, this Journal, viii, p. 451, 1899.

198 Drushel— Volumetric Estimation of Lanthanum.

acetic acid. The precipitate is thoroughly stirred and allowed
to settle. It is then filtered through a perforated crucible
fitted with an asbestos felt. After thoroughly washing with
water the crucible and precipitate are placed in a beaker with
100 to 300%™* of water and 10 to 30° of (1:4) sulphuric acid.
The contents of the beaker are heated nearly to boiling and at
once titrated to color with standard potassium permanganate.
The filtrate is similarly titrated as a check on the titration of
the precipitate. The lanthanum is calculated as La,O, from
the two titrations. The mean of the two closely agreeing
values thus obtained is taken.

The solution of ammonium lanthanum nitrate used in this
work was standardized by evaporating measured portions to
dryness and carefully igniting the residue to constant weight.

Table of Results.

[MLAS FTeETe ASN Uren ree ee ee eye 138-9]

La.O; taken as

La.Os; found.

the double nitrate Precipitate Filtrate Average Error

gram. gram. gram, gram gram.
Ibs 0°0148 0°0152 0°0144 0°0148 0°0000 +
De 0:0148 0'0149 0°0139 0:0144 0°0004—
3. 0°0296 0°0302 0:0291 0°0296 0.0000+
4. 0°0296 0°0302 0°0293 0°0297 0°000) +
5. 0°0592 0°0599 0°0586 0:05938 O-0001 +
6. 0°0592 0°0598 0°0585 0°0592 0:0000-+
ie 071184 O90 071179 0°1185 0-0001 +
8. 071184 0°1191 01182 O°1187 0°0008 +
ee 0°2368 0:2376 0°2562 0°2369 00001 +
10. 0:0148 0:0149 0°0145 0-0147 0:0001 —
ile 0°0148 00-0150 O-0147 0:0148 -0°0000 +
12. 0°0296 0°0298 0°0293 0°0295 0:0001 —
13. 0°0592 0°0596 0°0589 0°0598 0:0001 +
14, 0°1184 071190 0°1182 O0-1186 0°0002 +
15. 0°1036 071040. 0°1029 0°1055 0:0001—

In experiments 1 to 9 ammonium oxalate was used in making
the precipitation, in 10 to 15 oxalic acid was used.
In conclusion the author acknowledges his indebtedness to
Prof. Philip E. Browning for helpful suggestions during the

progress of the work.

A. J. Lotka—Mode of Growth of Material Aggregates. 199

Arr. XXII.—Studies on the Mode of Growth of Material
Aggregates; by Aurrep J. Lorxa. *

le

Ix a material system in which the physical conditions vary
with the time, certain individual constituent elements may
haye a transitory existence as such, each lasting just so long as
its conditions and those of its neighborhood continue within
certain limits.

Although the “life period” of each individual element may
be thus limited, an aggregate of a number of such individuals
may nevertheless have a prolonged existence, and may even
grow, provided that the variations in the conditions of the sys-
tem do not exceed certain limits, and that by some process or
other new individuals are continually formed as the old are
eliminated.

Aggregates of this kind play an important role in nature,
and a study of their mode of growth or decay would be of
considerable interest.

The general problem is evidently one of extreme complexity,
but some special cases which lend themselves to simple mathe-
matical treatment would seem to be rather instructive.

Let us then consider an aggregate which at time 7) consists
of N, individuals of a specified kind. If N, is the number of
individuals in the aggregate at time @, let us examine the rela-
tion between N, and N,.

In the first place we note, that if

B,= number of individuals added to the aggregate per unit
time at the instant ¢,
and if

D, = number of individuals eliminated therefrom per unit
time at the instant ¢,

then,
aN
re — B,—D,. (1)
t
N= Not f (B,—D,)at. (2)
to

D, is in general a function of N,, its value depending :

1. On the character of the limitation of the “life period”
of the individuals.

2. On the “distribution of ages” among the individuals.

B, may also in some cases be more or less directly dependent
upon N, and upon the “distribution of ages” in the aggregate.

200 A. J. Lotka—Mode of Growth of Material Aggregates.

We will for the present restrict our considerations to cases
in which:

1. The individuals are either all of one class as regards their
general properties, and especially as regards those which affect
the character of the limitation of their “ life period”; or, if they
belong to a number of different classes (e.g., males and females
of a community of living organisms, ete.) then the relative
proportion of individuals of each class among those formed
during any element of time is constant.

2. The“ length of life” of each individual is independent
of the total number of individuals in the ageregate, and of the
distribution of ages among them.

3. Ihe oeneral conditions of the system, in so far as they
affect the “ length of lite” of the individuals (see § 2), are, on
an average, uniform and constant throughout.

The variations in the conditions of the system are of such
fixed type that, when conditions 1, 2 and 38 are satisfied, the
number of individuals sur viving age @ out of any large number
Q counted at the moment of their formation and picked out at
random, can be expressed in the form © p(a), where p(q) is a
(univalent) function containing only a.

Then, if c(@) is such a factor that out of the total N, the
number of individuals whose age lies between the age limits
a and (a+da), is given by N,c(a)da, it readily follows that

Bra i
AG) = N, pla). (3)
ye log PY) a5

o a d p(a) A
— Sh Syne ari aaa da. (5)

We may substitute these values in (2):

t ee) i
ING =eN. fay B, dt +f xf c(a) g Een dt da. (6)

oN oe Bae ou) dt da. (7)

Lastly, given that the average mass of one individual at age
a 1s ma), we have for the total mass M, of the aggregate at
the instant ¢;

Mi wif ela) m(a) da. (8)
=f" B,_. p(a) m(a) da. (9)

A. J. Lotkha—Mode of Growth of Material Aggregates. 201

The history of such an aggregate as we have been consider-
ing may be represented in a system of rectangular coordinates
by plotting as ordinates the values of Nic) corresponding to
the values of @ measured along the Y axis and the values of ¢
measured along the X axis. The surface so obtained may be
ealled the N,c(@) surface ; any section of the same taken at 45°
to the planes of wz and yz is of the form:

z= N,, ¢(0) p(@).
==" D(a).
where 7, is the value of ¢ corresponding to the point at which
the section cuts the X axis.

Special Cases.
A. Let c(a@) be of fixed form.

B,
Then €(O}r = Nie constant = 0 say ;
t

while, by (4) ae = —{" c(a a) ae) 5b)
AN;

= constant = d say ;

hence oe = B,—D,= N,(0—d) =7N say;
N, = Neow (10)

Substituting these values in (3) we have:
e(a) = be~" p(a). (11)

b, d and r are respectively the rates “per head” of forma-
tion, of elimination, and of increase in number of individuals
in the aggregate. From the obvious relation,

Sf ayaa tl (12)

or by (4) and (11), they must satisfy the condition :
1

ac. “ p(a)da Ce)

AM. JOUR. SO Over SERIES, VOL. XXIV, No. 141.—Srepremper, 1907.
4

202 A. J. Lotka—Mode of Growth of Material Aggregates.

era dp(4) 7
ne of « NO ata (14)

i one ae@) da
0 da

= (15)
Hh e '“p(a) da.
0
Substituting (13) in (11) we have
een p(a)
é(@) = (16)

Bie é€ '" p(a) da.
0

For the mass of the aggregate we have:

Moe Nef e(a) m(a) da.
0)

If m(a) as well as c(a@) is independent of ¢

of c(a) m(a) da = constant = m say; (17)
0
then M, = mN;
(18)
== Vi Che

The above is the type toward which tends, for instance, a
population in which the influence of emigr ation and immigra-
tion is negligible, and general conditions are approximately
constant. ‘Equations (10) and (18) then give the number* and
mass of the population as a function of the time (geometric
progression); equations (13), (14) and (15) are three different
forms of the same relation between the birth-rate per head and
the death-rate per head, while equation (16) gives the ‘age
scale.”

A comparison was made between some observed values and
the corresponding figures calculated according to the above
formule. Below are given the results for England and Wales,
1871-1880, which, in spite of very considerable emigration and
immigration, show comparatively close agreement. “For 1881-
_ 1890 the divergence is greater.

* Compare M. Block, Traité théorique et pratique de statistique, 1886, p. 209.

A. J. Lotkha—Mode of Growth of Material Aggregates. 208

England and Wales 1871-1880 (Mean).
Observed* Calculated

Birth-rate per head b 03546 0352
Death-rate per head d 02139 0211
Excess (6—d) =r 01407 (0141)
OP

Age Seale. 1,000 f oe
dy Me pa EN gies

1,000 individuals, in age— O— 5 136 138
groups of 5 and 10 years: 5 — 10 120 116

; Oe a ylo 107 106

oO) 97 97

20s — eZ 89 87

PAD BID) 147 148

35 — 45 113 116

45 — 55 86 87

55 — 65 59 59

65 — 75 33 33

75 — w 15 13

Special Cases.
B. Let B be constant.

Then De —Bf 2) da by (5)
vee ae
== 15}
aN = B—D=0;
dt

i.e., N also is constant, and we have a “stationary condition ”
or “kinetic equilibrium.”

Here evidently b=d
and c(a) == bp(a). (19t)
Furthermore,
i ME == : say, (20)
a p(a)da
0
and Sp) = Ds (21)

*Mean 6} and d from 46th Ann. Rep, Reg. Gen. Births, etc., England and
Wales, p. xxxi; p(a) from Supplt. to 45th Ann. Rep. Reg. Gen. Births, etc.,
England and Wales, pp. vii and viii.

assuming ratio : uenale births = 1:04.
female births
+ Compare Farr, Roy. Soc. Trans., 1859, p. 837.

204 A. J. Lotka—Mode of Growth of Material Aggregates.

/ is evidently the ‘mean length of life” ; its reciprocal then,
by (20), is that value of 6 which just suffices to keep the
ageregate from decreasing under the conditions corresponding
to the particular form of p(@) by which / is given.

Thus, in the case of England and Wales, while according to
the three life tables 1838-54, 1871-80, 1881-90, the mean
length of life has risen from 40-9 to 42°9 and 45-4, the
“equilibrium birth-rate per head” has been correspondingly
reduced from *0245 to ‘0235 and -0220.

a:

Isothermal Monomolecular Reaction.

We will consider a system undergoing the chemical change,

IAG emer men A)

which we will suppose to take place at constant volume and
temperature, and in a homogeneous system.

Let N denote the total number of molecules of <A.

(75 N’ ce ee ce (73 “ce 6¢ ce NY
Also, when ¢ = 0, let IN 0)
N’= Ne

Hence, NeeN = Ni

Let D’ denote the number of molecules of A’ decomposed per min.
(79 D ce ce 73 ce iT4 ce A (73 ce (79
¢ Bf’ 66 oe 6c ¢é 66 SST ANY formed 6¢ 66
6c B 66 14 66 66 66 ce A 66 66 (13

The reaction is monomolecular in either sense. Hence, by
the law of mass action :

Di hNe D = kN:
(x and X’ are the reaction coefticients, )

Then, for the aggregate of N molecules of A. we have:

. B =D! =#'N' =&(N/_N).
= kiN '—(b+W)N.
Integrating :
kl (1—e— (e+e) t
Te aNT es!
N, =N 0 k+k! § (a)

A. J. Lotkha—Mode of Growth of Materral Aggregates. 205

The total mass of the aggregate follows immediately from

the relation:
M, = mN,

where m = absolute molecular weight of A.
Similarly, for the aggregate of N’ molecules of A’ we have:

: hb Bek tk
ee Shem eau (0)

M’, = m'N,’.

Both N, and N,’ approach towards a limiting value as ¢
approaches #, viz:

nok
Ne = Ne ak
| ki
, pees !
MNS ae

These limiting values, of course, represent a state of equilib-
rium, for when they are reached, further increase in ¢ pro-
duces no further increase in N.

In this particular case evidently no tao wledse of the form

dN
of p(a) and c(@) is required in order to determine Tage and

therefore N,, We can, however, conversely deduce the form
of p(a) and c(a).

For, since D depends only upon N, and not at all upon the
previous history of the system,* it is evident that’ the stability
of the molecule is independent of itsage. (Otherwise D would
depend on the distribution of ages in the aggregate, and hence
on its past history.)

Hence, if we pick out a large number 7 out of the total
number N of molecules, such that all these » molecules have
(nearly) the same age, then the fractional rate of decrease
among these will be the same as for the whole aggregate.

We have then ue =— es ily
dt da
Therefore F110 Cmte
i.e, (Ch) —— (c)

In this case we have for mean length of life :

y == (d)

Similarly DG) ee ke . (e)
il

a (f)

* Except in reactions which have a ‘‘ period of induction.”

206 A. J. Lotka—Mode of Growth of Material Aggregates.
Introducing the relation (¢) into (4), we have:

Di= kN f c(a)da.
0

From the nature of ¢(q@) it is evident that

of e(a)da = 1.
0

Hence D, = EN;

which, of course, is not a new result.
The form of ¢(@) follows from the relation :

B =
é(@) = N, p(a) (3)
TiN =
i Ee r
—e¢ N, (by e).
ke + k'e—(k+k’) (t—a)
Lica ie en (by @ and 3B). (7g)

(1—e—@+4)2)

When ¢ = ie., for equilibrium :
CRG) Ce (h)

We cannot determine c’(@) exactly, as we do not know the
ages of the N,’ molecules originally present at time 7=0. By
simply neglecting these N,’ molecules, we can, however, find
an approximate expression which will give c'(@) more and more
nearly as time advances, and the number of survivors of the
original N,’ molecules diminishes.

We thus obtain :

1/7) __ e—(k+k’) (f—a)
k (1 g. ine e uy (i)
het hie @+K
CAG) eee (x)

Equation (z) holds good (approximately) when ¢ is sufficiently
large, for values of @ small as compared with 7.

Equation (#) holds good for all finite values of @.

Equations (/) and (%) correspond to equations (19) and (21)
developed for the general case.

The above conclusions still hold if other substances enter
into the reaction beside A and A’, provided that their concen-
tration is practically constant, (e.g., if they are present in large
eXGess).

As a concrete example, we may, for instance, quote the case
of the reaction :

HCOOH+C,H,OH: <= HCOOC,H,+H,0

C\(@) heme

A. J. Lotka—Mode of Growth of Material Aggregates. 207

which, in the presence of excess of alcohol and water, is mono-
molecular in either sense. According to the values of the
velocity coefficients observed by Kistiakowski,* at 25° C. in a
solution containing 73°2 per cent alcohol, the mean “ length of
life” of a molecule of formic acid is here 7 hrs. 36 mins., while
that of a molecule of ethyl formate is 12 hrs. 6 mins.

Representing the history of the aggregate A (ethyl formate)
by an N,c(a) surface, we should obtain a figure such as that
indicated below.

SA

EEE
NS

K=-00138

'= 00220

We have not the requisite data for constructing the entire
N,c(a) surface for A’; when equilibrium is reached, however,

* Zeitschr, f. phys. Chemie, xxvii, 1898, 250.

208 A. J. Lotkha—Mode of Growth of Material Aqyregates.

all sections given by ¢ = constant are alike both for A and also
for A’, and ‘take the form indicated in fio. 2.

We note that although the zero ordinate—the number of
molecules formed per unit time—is the same for A and A’,
nevertheless molecules of one kind—A say—preponderate in
the system, as shown by the greater area of the curve A; this
is because, in the “struggle for existence,” the stabler (fitter)
molecules A have the advantage, and are, on an average,
‘ longer-lived.”

Viewed in this way, chemical action clearly presents itself as
a case of ‘“ Inorganic Evolution.”

0 500 1000 1500 2000 2900 3000 a”

The physical interpretation of these results would be some-
what as follows:

From the point of view we have taken, we must suppose
that the condition of each molecule at a given instant in
general departs somewhat from the average condition of all
molecules.

For a given molecule of the kind A, there will, in general,
sooner or later, come a moment when the variations in its con-
dition reach a certain limit—we may speak of it as the “limit
of stability ” of the molecules A—at which that molecule of A
ceases to exist as such, and passes into the condition A’.

The number D of molecules which are thus eliminated from
the aggregate A in a unit of time will depend:

1. On the nature of the “limit of stability” of the mole-
cule A.

2. On the “ distribution ” of the variations in the condition of
the molecules in the aggregate.

Any agency which affects either of these factors will, in
general, affect also D, the rate at which molecules are elim-
inated from the aggregate.

Let us consider each of these factors a little more closely.

A. J. Lotka— Mode of Growth of Material Aggregates. 209

I. Limit of Stability of a Molecule of the kind A.

The circumstance which leads to the elimination of a mole-
cule of the kind A is its spontaneous passage from the
condition A to the condition A’.

Now, in order that this passage may take place sponta-
neously, it is necessary :

3

1. That the free energy E of that molecule in the state A
be > E’, its free energy in the state A’.

2. That during the passage of the molecule from the state
A to the state A’, its free energy diminish continually.

A diagram will help to illustrate these conditions.*

Suppose the free energy E of the molecule plotted as ordi-
nate against the time as abscissa during its passage (not neces-
sarily spontaneous) from the state A to the state A’.

Then, if condition (1) is satisfied, the curve E = ¢ (¢); or, as

*TIt is here assumed that, whatever may be the events immediately pre-
ceding the change A A’, during this change the molecule may be regarded as

an isolated system. Without this assumption the conditions for the change
A A’ are more complicated.

210 A. SJ. Lotka—Mode of Growth of Material Aggregates.

we may express it, the “ path” of the change from A to A’
might, for instance, assume any such form as “the above (fig. 3),
for in each case E > E’.

Evidently, however, the change A A’ will not take place
spontanec vusly, except in the case represented by fig. 3a, which
is the only one satisfying also condition (2).

Now, we have no means of ascertaining the character of the
“path” of the change A A’ by direct observation. The
following considerations, however, will show how, from the
properties of the ageregate as a W hole, we can draw inferences
as regards the character of this path for the individual
molecules.

A. Time t'—t required for the change.
q : g

Since /N molecules pass from A to A’ per unit of time, and
each takes on an average (¢’—¢) to complete the change, there
will, at a given instant, be kN(¢’—t) molecules actually under-
going change. The dilution of these in the ageregate will be

y
atleast = kv =)

Now it is a matter of unive ersal observation that in all ordi-
nary chemical reactions* at any rate, this dilution is so great
that the material in the state of transition cannot be detected,
no matter what the value of #, or the methodst+ of analysis, or
the general circumstances. In other words, such a system as
the one we have been considering, can always be fully
described with regard to its composition (within the limits of
experimental error) as consisting of a certain quantity of A
and a certain quantity of A’.

We conclude that in all ordinary cases the time ¢’—¢, during
which the molecule is in the transitional state, is very small.+

B. The form of the path A A’.

In general, the type of the path A A’ will, for each mole-
cule, depend on its condition at any particular instant, and it
may be different for different molecules of the same ageregate.

(a) Let us first of all suppose that for all molecules for
which the condition (1) is satisfied (KE > E’), the path is of
type (a); in this case then the fulfilment of adie (1) will
be sufficient to determine the change A A’, and it follows that
all molecules for which E > E’ will actually be undergoing

*Evidence of the transitional state in tautomeric change, however,
appears to be furnished by the work of Baly and his collaborators,

+ Ordinary methods of analysis, of course, would not reveal the presence
of such transitional bodies. But physical methods should, if they were
sufficiently sensitive. In this connection again we must turn to the work of
Baly for suggestions.

{ But see remark regarding tautomerism on the next page.

A. J. Lotka—Mode of Growth of Material Aggregates. 211

this change. Hence, the aggregate can never contain any
molecules for which E oe vexcept those N&(¢’—7), which are
actually undergoing the change from A to A’; this number
we saw is always negligible for any ordinary chemical change
going on with measurable velocity.

Therefore, A can never, for any length of time, exist except
in equilibrium with A’. This conclusion agrees with the facts
observed in the case of change of physical state unimpeded by
supercooling or superheating, and, perhaps, with certain types
of chemical reaction (e.g., dissociation); but it does not corre-
spond to the characteristics of an ordinary chemical reaction
taking place with measurable velocity.

(8) Next, suppose that the path is of type (6) for all mole-
cules for which E > EK’. In that case no spontaneous change
from A to A’ will take place, although the aggregate may not
be in stable equilibrium with A’; and this condition will con-
tinue indefinitely.

Cases of this kind are: Physical change atrested by super.
heating or supercooling ; also, in practice, every “unstable ”
chemical compound, which may be preserved (practically)
unchanged indefinitely under certain conditions.

(y) Next suppose the path A A’ is of type (a) for some
molecules, of type (6) for others. Then the former will be
actually undergoing change, while the latter, although having
E > E’, will remain in the form of A.

Such an aggregate may, therefore, for an yee ue
length of time, contain molecules for which E > EH’, i.e., it
may have a prolonged existence in a state in which it 1s Hee in
equilibrium with A’, but is under going gradual change towards
that equilibrium.

This description evidently fits the case of the typical chem-
ical time-reaction,* and to its more detailed study we will
return presently, after we have briefly considered some of the
remaining types of the path A A’.

(6) A path of type (ce) or (d) for some of the molecules
would imply that they were more stable in their transitional
state than in the form A or A’.

Such a state of affairs is imaginable—the “stable”? form A,
would then correspond, not to an ordinary chemical compound,
but to a condition of the molecule intermediate between two
compounds. <A case of this kind is perhaps presented by the
phenomenon of tautomerisin.t

* The presence of a maximum in the path of chemical change is indicated
in many cases by the increased reactivity (i.e., increased free energy) of
substances in the “‘ nascent” state. But the mere fact that the transitional
state is evanescent implies such a maximum.

+ In this case then the ‘‘ transitional” state would be prolonged, and we
might expect to be able to detect the material present in that state. In this

connection note the work of Baly and his collaborators. Compare also
G. Oddo, Uber Mesohydrie Chem. Centr. Blatt, 1906, p. 1811.

212 A. J. Lotka—Mode of Growth of Material Aggregates.

The path A A’ might, of course, also be of more compli-
cated form, having a number of maxima and minima. But sueh
a ease does not seem to be of any particular interest in the
present connection.

Let us now return to the case (y), as applied to the system
which we have been considering, viz., that in which the change

A’

SERRE 2
is going on.

A geometrical representation of the conditions in this system,
in accordance with the conception outlined above, may be
obtained as follows:

With our attention fixed first of all upon one particular
molecule, at a time when it is in the condition A’, let a curve
be drawn, whose ordinates represent the values of E’, the free
energy of this molecule, corresponding to the times measured
off as abscissae.

On the same diagram let a second curve be similarly drawn,
such that its ordinate for a given abscissa represents the mini-
mum value E of the free energy which the molecule would
have at that instant, supposing it to be in the state A, and to
retain its actual total energy.

Lastly, on the same diagram, let a curve be drawn, whose
ordinate for a given abscissa represents the least maximum
value E,, thr ough which the free energy of the molecule must
pass under the existing conditions, if the molecule is at the
instant corresponding to the abscissa, transformed from the
state A’ to the state A (limit of stability). A

Now we have no means of directly ascertaining the exactt
form of these three curves for a given molecule, but we do
know something of the gener al aspect which they must
present.

The principal features of interest are indicated in fig. 4, and
are best described by reference to this:

We will suppose that, as shown in the figure, at the moment
¢ = 0, when we begin our observation, EH’ z E. As time goes
on E/and E vary somewhat, and the corresponding curves may
cross at some point ¢, at which E’=E. To the right of this
E’ > E, and the first condition for the transfor smnesiey A’ A is
fulfilled. Following up these two curves further, we may
find them recrossing, but sooner or later we shall come to a
point where, with E < E’, the curve representing E’ will meet
the third curve E,. At this instant t,, the second condition

* For the sake of simplicity it is here assumed that the path which con-
tains this least maximum, contains no other maximum.

+The mean state of the molecule will presumably satisfy the law of
equipartition. See Rayleigh, Phil. Mag. (5), xlix, 98; Kelvin, Phil. Mag. (6),
ii, 1, 7; also W. F. Magie, Science, xxiii, 161, 1906.

A. J. Lotkha—Mode of Growth of Material Aggregates. 2138

also for the spontaneous change from A’ to A, is satisfied, and
this change actually takes place.

Now we can continue our curves, tracing first of all the
value of the free energy during the change from the condition
A’to A. This part of the curve is the “path of the spon-
taneous change, and must, therefore, be of type (a), as shown.
We then continue our set of three curves, very much as before,
except that now the values of E are the actual values, those of
E’ auxiliary (“ caleulated ’’).

The continuation of this figure then shows features precisely
similar to those which we noted in the earlier part, and it is,
therefore, unnecessary to follow up their description any
further. We only note that ¢, denotes the instant at which the
molecule of A is formed, ¢, a point at which the curves E EK’
cross, ¢, the moment when the molecule is eliminated from the
aggregate A.

The interval ¢, ¢, represents the length of life of the mole-
cule A, and this, we saw, has the value (a) for ne—** out of x
molecules counted at the moment of their formation, or, has a

1
mean value h for all molecules.

We can distinguish a number of separate fields in this
diagram.

In the field ¢, ¢, the molecule is in the state A, and is stable
in that condition.

In the field ¢, ¢, the molecule is still in the state A, though
in a meta-stable condition. It does not here pass into the state
A’, for in order to do so, it would have to cover a path of type

B) : Le., the second condition for a spontaneous change is not
fulfilled.

214 A. SJ. Lotka—Mode of Growth of Material Aggregates.

Lastly, the field ¢, ¢, represents the molecule during its spon- ©
taneous transition from the state A to A’.

IT, Distribution of Variations.

The above considerations apply to one individual molecule.
The entire aggregate A will comprise :

1. Molecules whose condition corresponds to the field 4%, ¢,.
These are stable; neither condition (1) nor condition (2) is
satistied for their elimination from the aggregate A.

2. Molecules whose condition corresponds to the field 7, ¢,.
These are in meta-stable condition, but do not change because
condition (2) is not satisfied.

3. Molecules whose condition corresponds to the field ¢, ¢,.
These are undergoing change, and being eliminated from the
agoregate.

The number of molecules in the third class, we saw, _is
always small (at any rate in ordinary cases) and is given by

KN(t'—1).

The distribution of the remaining molecules between the
first and second class is evidently closely connected with reac-
tion velocity and equilibrium. A change in temperature, being
accompanied by a change in the total energy of the system,
must lead to a change in the distribution of the energy in the

system, and so toa change i in the distribution of the variations
in the condition of the molecules. But any agency which
produces such a change will affect D, and hence arises the con-

nection between D and the temper ature.

We still have to consider the influence of agencies which
affect the “limit of stability’ of the molecule.

There is one special case which is of interest here, namely,
that in which, while the values of E and E’ are unaffected for
each molecule, E, is changed—lowered, say. The effect of such
a change will evidently be that, while ‘the total and free energy
of the System as a whole is unchanged for a given composition
(and volume), the rate D will be changed—raised.

We see that this corresponds exactly to the effect of introduc-
ing a catalyser into the system.

‘We may goa step further and ask ourselves in what way the
value of E, may be lowered.

The answer jis that this may take place in one of two ways:

Hither the character of the path A A’ may be changed for
some or for all the molecules; the formation of intermediate
compounds in certain reactions or the production of local dif-

A. J. Lotka—Mode of Growth of Material Aggregates. 215

ferences in concentration, and other unknown effects, may be of
this character.

The effect of the catalyser would in this case be closely analo-
gous to that of “ nuclei, » or“ erystals ” in supercooled systems,
or of gas bubbles in a sperheaien liquid. In these cases change
of state is arrested because in the absence of “ germs” the sys-
tem would have to pass through a stage in ‘Which its free
energy had a value greater than its actual value,* in order
to reach the final lower value. The introduction of the germ
furnishes another path requiring no such maximum value to be
passed, and the change then takes place spontaneously.

But the catalyser might also produce its effect in another
way, namely, by pr oviding mechanism through which a portion
of the energy set free by one molecule during its descent along
the downward limb of the path A A’ is diverted into such a
channel that it raises another molecule up the ascending limb
of that path.¢ This possibility is of some interest because a
large class of natural phenomena, especially in the living world,
depend on an action of this kind. We may only mention here
the phenomenon presented by an organism which obtains its
food by a series of (muscular) efforts, the energy for which is
derived from food previously ingested. The phenomenon as
a whole takes place spontaneously, although for each portion of
food there is astage in the process through. ‘which it passes when
it requires the application of external energy. Another exam-
vle of the same kind taken from the world of mechanies is the
continuous operation of a heat engine in which the working sub-
stance goes through a cycle.

Lastly we note ‘that according to the view developed above,
such agencies as elevation of temperature, or the introduction
of a catalyser, which increase #, will shorten the life of the
molecule, since we found that the probability at the moment
of formation that a given molecule will reach age @ is given
by e

this brings us to the close of our consideration of the system
A + A’. A similar treatment might of course be applied to
more complicated chemical systems, but, although the results
obtained would of course differ in form from those deduced
above, the underlying principles would be the same. Indeed,
the development outlined in this paper appears to promise per-
haps more interesting suggestions in the treatment of aggre-
gates, the elements of which, unlike the molecules in chemical

*Owing to the dependence of vapor-pressure, solubility, or melting-point
on the form of the surface of contact between the two phases.

+ A somewhat similar idea has been expressed by Raschig (Zeitschr. f. ang.
Chem., 1906, p. 1761), who compares the action of the molecules on one another

to that of the consecutive members of a file of dominos set up on end, and
overthrown in a body by the fall of the first in the file.

216 A. SJ. Lotka—Mode of Growth of Material Aggregates.

reaction, are accessible to individual observation ; and it is hoped
on a future occasion to prosecute the work further in this
direction.

For the present, however, we will conclude with a brief sum-
mary of the main points developed so far:

1. We have recognized the problem of chemical dynamies
as a special case of a wider problem: The former is the study
of the laws governing the changes in the distribution of matter
among’ different chemical compounds, as determined mainly by
their chemical character; the latter is the study of the laws
governing the distribution of matter among complexes of any
specified kind, as determined by their general physical character.

The statement of the wider problem may be taken to repre-
sent the quantitative formulation of the problem of evolution
in its most general terms.

2. We have illustrated a statistical method which is suffi-
ciently general in its application to comprise such widely differ-
ent cases as that of the growth of a population under certain
simple conditions, on the one hand, and that of a simple chem-
ical equilibrium reaction, on the other. The fundamental
feature of this method is the splitting up of the characteristics
governing the rate of growth of a material aggregate into two
factors—the one relating to those properties of the system which
determine the formation of new individuals, and the other
relating to those properties of the system which determine the
limitation of the “life period” of the individual constituents.

3. Incidentally we have suggested a physical conception of
the character of chemical action, of the “ passive resistance ”’
which checks its velocity, of catalytic action, of the “nascent”
state, and of tautomerism.

Washington—Catalan Volcanoes and their Rocks, 217

Arr. XXIII.—The Catalan Volcanoes and their Rocks ; by

Henry S. WASHINGTON.

INTRODUCTORY.

Tue volcanoes in the vicinity of Olot in Catalonia seem to
have been first noticed by F. Bolos in 1796,* and later by
Maclure,t and were described in the nineteenth century by
several geologists, among whom may be named Lyell? and
Carez.§ Within the last “few years renewed attention has been
directed to these voleanoes by the Spanish Society of Natural
History, which appointed a commission to undertake a mono-
graphic study of them. A preliminary report of this com-
mission appeared in 1904,| and it is expected that the final
results of its labors will be published in the near future. The
saine year also saw the publication of the two latest papers on
these volcanoes. The one is a small, illustrated pamphlet by
Gelabert,4, who describes the physical features of the district
in a rather popular way, and without any petrographical or
chemical discussion. The other is a brief but very instructive
sketch of the general voleanological features of the vicinity of
Olot by Sapper.**

While the main physical and geological features are thus
well known, the petrographical characters of the rocks have
been comparatively neglected. Recognized in the field by the
earlier observers as basalts, they were studied by Quiroga,tt
who established the presence of feldspar basalts, nephelite
basalts, and limburgites—determinations which my own
observations substantiate. Some of Carez’s and Sapper’s speci-
mens were reported to be feldspar basalts, which is a suft-
ciently accurate designation for many of the rocks in the
absence of chemical analyses.

It was to collect material to supply these deficiencies in our
knowledge, and especially to undertake the chemical investiga-
tion of these presumably interesting rocks, that the writer
visited the region in the summer of 1905 with the aid of a
grant from the Carnegie Institution of Washington. Head-

* Cf. L.M.Vidal, Bull. Soc. Geol. Fr., vol. xxvi, 1898, p. 675.

+Cf. C. Daubeny, Description of Volcanoes, London, 1848, p. 295.

tC. Lyell, Principles of Geology, 1840, vol. iii, p. 185.

<L. Carez, Etude des Terrains Orétacés et Tertiaires du Nord de 1’ Espagne,
Paris, 1881, p. 299.

|S. Calderon, Bol. Soc. Esp. Hist. Nat., p. 330, 1904.

“| J. Gelabert, Los Volcanes extinguidos fae la Provincia de Gerona. San
Feliu de Guixols, 1904, 120 pp., small 8vo, with map.

** K, Sapper, Die Catalonischen Vulkane, Zeits. d. deutsch. geol. Ges.,
vol, liv, P- 240, 1904.

+ CE: - Calderon, Bull. Soc. Geol. Fr. (8), vol. xiv, pp. 18, 114, 1885.
AM. a Sc1.—FourtH SeEries, Vout. XXIV, No. 141.—SepremBer, 1907.

15

218 Washington—Catalun Volcanoes and their Rocks.

quarters were made at both Gerona and Olot, from which
a considerable number of the cones and lava flows were visited,
especial attention being paid to those near Olot. Not much
stress was laid upon the. general geology, the tectonics, and the
stratigraphy, as these were beyond the main scope of the in-
vestigation, and will be fully covered by the forthcoming
rep vort of the commission referred to above. At the same time,

Ry Flay

Lohia. Nera PEastell fulrit

BS. Juanr

s
} Pstany oe

Paoqudd

Bisarecas
° Batet
4 OLOT wg crageet
Sta, Margate
Ptese, Q
SaCot

< “a

LF 6
Bos chs”

deTosca

8
San\ Esteban San Peli
7 Pein
#,,

Las Planas\P.

0%
aol Lago de

Banolas

a B Barolas

Sellen?

ta Pau p

Puig Moner

Og Granollers
P, Ment cal
° ,
mAdrt

| Bana de Boch
B Llora
a

Fi
or

Alen es

Bescano

Sketch Map of the Catalan Volcanoes
1: 400,000

a short sketch of the main physical and topographical features
of the voleanoes themselves will be given here, as they are but
little known in this country and the descriptions may be of
interest.

Study of the region is somewhat hampered by the want of
an accurate and detailed map, a need which the Spanish Com-
mission purposes to fill. The only map available was one of
the Province of Gerona by Marti ‘and Marifio, on a scale 6f
1: 200,000, and published at Barcelona by Barral Hermanos in
1900. No contour lines or other physiographic features,
except the hydrography, are presented in this, even the volea-

Washington—Catalan Volcanoes and their Rocks. 219

noes being omitted, so that the map leaves much to be desired.
It is probably on this that the sketch map mentioned by Sapper
is based, of Which I was unable to obtain a copy, and it also
serves as the basis for the map on the same scale given by
Gelabert, on which the positions of the volcanoes are roughly
indicated. A sketch map of the district is given in fig. i
based on the map mentioned above. Only the more pr ominent
voleanie cones are shown (by small circles), and the railroad,
which extends from Gerona to San Feliu, is omitted.

I would take this opportunity to express my thanks to Sefiores
Calderon and Fernandez Navarro for courtesies and assistance
shown me in Madrid preparatory to my visit to Olot.

ToroGRAPHY AND GEOLOGY.

Prevoleanic Geology.—The voleanoes which form the subject
of the present paper are situated in the modern Province of
Gerona (embraced within the ancient Catalonia), in north-
eastern Spain, between the eastern end of the Pyrenees on the
north and the Sierra de Monseny on the south. The area
occupied by them is part of the Eocene gulf which extended par-
tially across northern Spain from the ‘Gulf of Lyons, the vol-
canic area proper occupying the site of a Pliocene bay, which
formed the last filled and most eastern part of this. The chief
sedimentaries of the district are nummulitic limestones of
Eocene age, with some later conglomerates, and a small patch
of Silurian shales and limestones near Llora, along the Rio Ter.
These earlier rocks are quite variable in dip and strike, and
faulting is common in the Eocene limestones. The fault scarps
thus produced are quite prominent features in the topography,
being often several hundred feet in height. From such obser-
vations as I made, it seemed that the area is one of faulted
blocks, the dips being in diverse directions and usually at
rather low angles, seldom over 25°, though I was able to de-
vote so little time to this phase of the geolog ey that I cannot
speak decisively on the subject.

In the depressions and fault valleys thus formed were laid
down thin deposits of Quaternary sands, clays, and conglom-
erates, which, with the Tertiary strata, underwent very consider-

able erosion prior to the beginning of the voleanic activity.

It would seem that there was some feeble vuleanicity mani-
fested in the Tertiary,* but the era of the volcanoes which
interests us here began about the middle of the Quaternary,
and continued into the present period, as is evident from the
excellent state of preservation of the voleanic cones, though
there is no historic record of any eruptions. The influence of

* Cf. S. Calderon, Bol. Soc. esp. Hist. Nat., 1905, p. 340.

220 = Washington—Catalan Volcanoes and their Rocks.

these upon the preéxisting topography was comparatively slight,
and consisted largely in the building up of cinder cones ‘and
the filling of the river valleys by lava flow s, resulting in the
formation of some lakes and marshes. But, on the whole, the
preéxistent drainage persisted, and the rivers now follow their
old courses, having cut new channels either through or along-
side of the lava flows, coming out into their old beds when the
lava ends. Excellent illustrations of this are to be seen in the
valleys of the Fluvia and of the Amer; in the latter case the
relations of the lava flow to the old drainage, and the abrupt
end of the former being especially well seen “from the railr oad,
a few kilometers below Las Planas.

The Lava Filows.—The igneous ejections are readily refer-
able to two distinct periods, a separate type of eruption dis-
tinguishing each. The earlier seem to have been uniformly
quiet, and formed extensive flows of basalt, which poured out
into the river valleys and other depressions as explained above.
The points of origin of most of these are not readily made out
and are somewhat problematical, owing to their subsequent
burial beneath later ejections and recent deposits. But in
general they seem to have issued from at or near the sites of
scoria cones now in existence. These flows, of which Gelabert
enumerates ten principal ones, varied in length from 5 to 15
kilometers, and seem to have been uniformly quite thick, up to
40 or 50 meters. Of these the most important are: one which
issued from near Santa Margarita and flowed northeasterly
down the valley of the Rio Cor, past Santa Pau and Sellent, a
distance of about 10 kilometers; a second which issued from
near La Garrinada and flowed easterly down the valley of the
Fluvia past Puig Estany and Baguda as far as Castellfullit, a
columnar structure being especially well developed in this ; a
third which flowed from near Bafia de Boch down the valley
of the Riera de Llémana as far as the chapel of Sant Medir
near Llora; and a fourth which issued from near the Puig
Monteal at Adri and flowed southeasterly as far as Domeny,
near Gerona, a distance of about 12 kilometers.

Partially filling the preéxisting river valleys, these flows pre-
sent a characteristic feature of the topography, and when they
have been cut into by subsequent river erosion, or where the
river has cut for itself a channel between the lava and the
adjacent limestone, highly typical and very striking illustra-
tions of columnar structure are revealed. ‘ Of these the best
known is that at Castellfullit, which has been described by
almost every writer who has visited the region. The flow
here is about 50 meters thick, and forms a precipitous tongue
of dark gray, prismatic lava, crested with the little village
above, and with the tur pulent stream below. The upper half

Washington—Catalan Volcanoes and their Rocks. 221

is made up of several series of curved prisms, which Carez
refers to five separate flows, though they seemed to me to be
parts of but one thick flow, the apparent superposition of the
series of columns being due to local variations in the conditions
of cooling. This rests over the greater part of its area ona
thin stratum of loose, scoriaceous material, which les from 18
to 25 meters above the river level. Below this the lava is
decidedly platy, rather than columnar, though even this
assumes in places a distinctly columnar structure, the prisms
being thin and often radiating. The base of this is difficult
to see, owing to the abundant talus, which contains, by the
way, many fragments of Roman glass, but it is distinctly seen
in places to rest upon Quaternary gravels, but little above the
water level. The small * basalt hill” mentioned by Sapper and
marked on his map, which lies a short distance to the east of
the tip of the tongue, across the brook, is evidently an integral
part of the main flow, as it shows the same succession of pris-
matic and platy lavas separated by a thin scoriaceous layer.
In this ease the River Fluvia, and the small tributary from’ the
south, have carved channels between the lava flow and the
limestone. A photograph of this flow is reproduced in fig. 2.

Other excellent examples of basaltic colunins are shown in
the same flow north of La Garrinada, along the small Rio
Ridaura, where the stream has cut into the lava itself, and in
places has formed deep pot-holes, as well as in the first flow
mentioned above near Sellent, and in that from Bafia de Boch,
below the Chapel of Sant Medir near Llora.

An area of rough lava, known as La Malatosquera or the
Bosch de Tosca, extends westwardly from the Fuente de San
Roque near Olot some two kilometers, the breadth being about
half the width. It is a wilderness of the roughest aa lava,
cultivated in spots and cut up by a labyrinth of walls and
narrow paths. Scattered over its surface are many small knolls
and protrusions of lava blocks cemented by finer-grained and
scoriaceous material. These do not seem to be definite cones as
reported by Gelabert, but are rather due to small, local explo-
sions of steam, and the heaping up of solid blocks, such as are
frequently found in extensive rough lava flows. This area,
which overlies a more compact and massive lava flow visible at
the fountain of San Roque, is apparently later than most of
the main flows, but is not connected with any definite cone,
for, though very close to that of Montolivet. it is yet clearly
distinct from this, and is probably to be considered as intermedi-
ate between the two phases of activity.

The Cinder Oones.—Protruding through, and scattered over
the surfaces of these lava flows, and consequently later than
they, and in some cases resting on Quaternary deposits, are

222 = Washington—Catalan Volcanoes and their Rocks.

many well-formed cinder cones, of which Gelabert enumerates
thirty-four. The greater number of these are in the neighbor-
hood of Olot, where they occur very close together, “while
others are met with, but more widely separated, to the east,
near the town of Gerona, but mostly to the west of this. As
far as my observations permitted ine to judge, it would seem
that none of these cones are based immediately upon the
Eocene limestone, but that they rest either upon Quaternary

5)

~

Castellfullit from the north.

deposits or upon the earlier lava flows, in most cases the latter
being true. Furthermore, it would appear that the place of
origin of these cones, as well as of the lava flows, is never near
the upper edge of a fault block, but always in or near the
fault valley, and sometimes close to the face of the fault scarp.
A similar emplacement was observed in the case of some of
the smaller and most recent cinder cones in Sardinia, and such
a situation above or near a well-marked fault is readily con-
ceivable as most favorable to the formation of a volcano.
These small volcanoes are invariably typical cinder cones,

Washington—Catalan Volcanoes and their Rocks, 228

lava beds not entering materially into the structure of most of
those examined by me; though massive flows occur at the base
of some of them, many of which are held to belong to the
earlier eruptive period, ‘and a compact lava stream was “noted at
the summit of La Garrinada. The material of which they are
composed, therefore, is predominantly highly scoriaceous, and
may be either in the form of blocks of very vesicular lava,
seoriz, or very fine lapilli, which may be mingled with the
larger blocks or may make up the cone to the exclusion of
the former.

The largest of these cones is Puig* Moner, near Granollers,
about half way between Gerona and Olot (which I could not
visit), whose height above the surrounding country Gelabert
gives as 950 meters.

But this is exceptionally high, and the great majority have
altitudes above their bases of about 100 meters or so, ranging
from only 30 meters to about 200. They all show well-defined
craters, and many are breached more or less on one side. For
the most part they have but one crater, though La Garrinada
possesses three and Puig Monteal, according to Gelabert,
as many as four. While some of these craters are compara-
tively deep, most of them are quite shallow, and they are
usually overgrown with a more or less dense forest vegetation, -
as are most of the outer slopes, though the bottoms of a few
are given up to cultivation, as at Montsacopa and Santa Mar-
garita. In general the original form is well preserved, as is
seen at Montsacopa and La "Garrinada, though others, as Cruz-
cat, have suffered very considerable denudation, But on the
whole the state of preservation is so good that a comparatively
recent origin must be ascribed to them all.

While detailed descriptions of the voleanie cones visited
would be ont of place here, and are especially unnecessary in
view of the imminent publication of the work of the Spanish
commission, yet a few notes on some of the more important
ones may be welcome, as these volcanoes are comparatively
little known.

The best known are the three in the immediate vicinity of
the town of Olot, namely Montolivet, Montsacopa, and La Gar-
rinada. The first of these (fig. 3), which lies northwest of
Olot, is crescentic in form, having been breached on the north
side, the highest point of the rim, where a fortified watch
tower stands, being 105 meters above the base. The diameter
of the crater is about 300 meter s, the sides being covered with
oaks, and the bottom (which lies about 75 meters below the
summit) being give over to cultivation, while a driven well,

*The Catalan word for these cones is puig, which is the etymological
equivalent of the French puy.

224 = =Washington—Catalan Volcanoes and their Rocks.

which furnishes an abundant supply of water, has been sunk in
the center. This cone is built up of scoriz, and lapilli exelu-
sively, well-defined bedding not being evident. On the south-
west slope these materials are heaped up against and over a low
fault scarp of Oligocene conglomerate, the beds of which dip
at low angles to the southwest, and which separate the products
of this cone from the lower massive flow which is exposed
along the Fluvia at the picturesque Fuente de San Roque, at

3

Montolivet from the east.

the beginning of the lava field of Bosch de Tosca mentioned
above.

The second of these volcanoes, Montsacopa, lies north 52°
east of Montolivet, and just north of Olot, the summits being
about a kilometer apart but the volcanic slopes intermingling,
thongh the juncture is masked by a road and cultivated fields.
The height of its southern rim, on which stands the pilgrimage
chapel of San Frances, above the base I measured with an
aneroid as 66 meters, though Gelabert gives it as 100 meters.
The summit is occupied by a well- defined, circular crater, some

Washington— Catalan Volcanoes and thew Rocks. 225

120 meters across and very shallow, the cultivated bottom
being only 13 meters below the highest, southern rim, and but
5 or 6 below the lowest, northern rim. On the southern flank
is a very shallow and ‘broad, saucer-shaped depression, which
may be the remains of an original second crater, though the
extensive cultivation of this side renders any definite deter-
mination difficult. This cone, which is quite bare of forest
vegetation, is made up entirely of loose scoriz and lapilli,

4

La Garrinada from the south.

though some large blocks of more compact but still highly
vesicular lava are exposed on the top along the south rim.
Like Sapper, I could see no trace of solid lava beds mentioned
by Carez, though I passed around its northern flank, where
Sapper supposed that they might be found. The loose mate-
rial is well exposed in an extensive opening above and just
behind the cemetery of Olot, where the voleanic agglomerate
is extracted for use in making cement and road metal.
Immediately adjoming Montsacopa on the northeast is the

cone of La Garrinada, fig. 4, the height of whose summit,

226 Washington—Catalan Volcanoes and their Locks.

(on the northeast) is 102 meters above the base as measured by
my aneroid, This volcano possesses three well- defined craters.
Of these the southern one is the largest, some 300 meters
across, and also the lowest, the southern rim being almost at
the base of the volcano and scarcely more than two meters
above the crater floor. The northern and highest rim of this
crater is 32 meters above the base and is coincident with the
rim of the second or middle crater, which has a diameter of
about 250 meters. The rim between these two is not very
well defined, but my observations agreed with those of Sapper
in leading to the conclusion that the more southerly and lower
crater was the later. The northern rim of this second crater
forms the highest portion of the cone, and forms a semi-cireu-
lar ridge, the inner walls of which are quite steep. To the
east and west of the northern saddle, between this crater and
the one to be next mentioned, the ridge is formed largely of
massive blocks of compact lava, which are evidently the
remains of a lava flow, though the lower portions of it have
been removed or covered up. On the northern slope of the
voleano is the third crater, which is also the smallest, not more
than 150 meters across. The walls of this are even steeper
than those of the second crater, and the bottom is some 20
meters below the saddle, on which is built a small farmhouse
with a well ot fresh water, this saddle being 80 meters above
the base and so some 20 meters below the highest summit.
Contrary to the opinion of Sapper, it seemed to me that this
northern crater was later than the middle one, and that it was
formed by a small flank explosion, which destroy ed a large
part of the more solid lava flows which must have covered this
slope, leaving a precipitous face along part of the northern
rim of the second crater. There was no apparent way of
deciding as to the relative ages of the most southern and most
northern craters. but the relative states of preservation seemed
to indicate the former as having been formed prior to the latter.
La Garrinada is built up largely of loose scorize and lapilli,
though solid lava, in blocks and probably in flows, enters into
its structure, without, however, affecting seriously its essential
character of a typical cinder cone.

The cones of Estany (A of Sapper’s map), which lies about
two kilometers to the east of La Garrinada on the road to Cas-
tellfullit, and that of Las Bisarocas, abont one kilometer east
of Olot, are both typical cinder cones, the former about 100—
and the latter 90 meters high, each with a well-defined crater,
and composed of loose material, similar to that of the cones
just mentioned, though blocks of compact lava occur at the
summit of Bisarocas and a lava flow was found at the east
base of Estany.

Washington—Catalan Volcanoes and their Rocks. 227

To the southeast of Olot voleanic cones are more numerous,
some twenty being marked by Gelabert. Of these the most
important is Cruzeat, the highest of those around Olot, its
summit being about 180 meters above the base. This is exten-
sively breached on the west side, is not very well preserved,
and is covered with a thick forest growth. Like the preced-
ing, it is built up largely of loose scoriz and lapilli, though
there is also a considerable amount of more or less compact
lava, which seems to be uniformly finely vesicular. This vol-
cano has ejected an immense quantity of bombs, large numbers
of which are to be found on and in the lava- and lapilli-cov-
ered plain to the west of it, north of the road from Olot and
from 4 to 5 kilometers from the town. These are of very
varied shapes, some ovoidal and quite regular, others elongated,
flattened and twisted, and they are all composed of more or
less vesicular lava. According” to Gelabert, who figures many
of them, some contain nuclei ‘ot feldspar or olivine, but these
would seem to be unusual.

Another interesting cone in this vicinity is Santa Margarita,
whose summit is 124 meters above the road between it and
Cruzcat, but about 200 above the lowest part of the base,
which is to the southeast. This likewise shows a well-formed,
circular crater, which is some 300 meters across, and the culti-
vated bottom of which lies 66 meters below the highest point
of the rim, the northeastern part. This voleano is composed
of very loose, fine lapilli, except in the western portion of the
rim, where limestone is exposed, and which evidently formed
part of the original opening produced by a maar-like explo-
sion, the southwest wind which evidently prevailed carrying
the bulk of the loose ejectamenta toward the northeast, as
pointed out by Sapper.

Of the other and smaller cones seen in this vicinity it is
needless to give any details, as they differ in no essential
respects from those already described.

Dikes seem to be very rare in the districts about Olot and
Gerona, only one having been observed by me. This is 25
centimeters wide, cutting tuffs and agglomerates on the road to
Santa Pau, about 3 kilometers from Olot. The rock of this
does not differ mater ially from those of the volcanic cones, as
will be seen later. Gelabert mentions and figures several
dikes which cut the granite of the region around Hostalrich,
20 to 80 kilometers to the south of Gerona.

As will be seen later, the lavas of the flows and of the cones
do not differ mater ially from each other in any respect except
in structure, the former beine uniformly compact and the latter
more usually vesicular, while the chemical and mineralogical
features remain much the same. It is clear, therefore, that,

228 = Washington—Catalan Volcanoes and their Rocks.

while the two may be referred to two periods, yet that the
only essential difference lies in the fact that the eruptions
which produced the cones were the final, explosive effects of
the vuleanicity, the commonly small size of these cones testi-
fying to the comparative feebleness of these last manifestations.

Such a formation of small cinder cones of the Puy type,
and comparatively free from lava flows, during the last, feeble
phases of vulcanicity, is a not uncommon phenomenon* , as
exemplified by the volcanic regions of the Auvergne and the
Hifel. It is of especial interest to note, in this “connection,
that very similar cinder cones, composed of basaltic material
closely allied to those of the Catalan volcanoes, marked the
close ‘of volcanic action in Sardinia and on Pantelleria and
Linosa; the rocks of which are held to be related petrologi-
cally to those of Catalonia,t as will be described in subsequent
papers.

PETROGRAPHY.

In the field the rocks of the Catalan volcanoes would be
unhesitatingly considered basalts, which vary in texture from
compact forms, dark gray to almost black in color, to highly
vesicular, spongy scoriz and lapill, the color of these being
black when fresh, but often yellow or red through surface
decomposition. The more compact forms frequently show
phenocrysts of olivine and augite, both minerals being usually

visible in the same specimen; while phenocrysts of “feldspar
are extremely rare. In some flows, most notably that of the
Baia de Bosch, in the bed of the Llemana near Llora, the lava
contains rounded nodules, varying in size from a few centime-
ters to a decimeter or more, which are made up of small grains
of greenish yellow olivine, Wwith a few grains of magnetite.

Examined microscopically, it is seen that augite in large
amount, and less quantities of olivine and of magnetite, are
present in all these rocks, while hornblende and biotite are
entirely wanting. With these there frequently occurs much
plagioclase feldspar, usually a labradorite, orthoclase not hay-
ing been definitely recognized. In a few of the specimens
this plagioclase is the only salic mineral present, so that these
are true feldspar-basalts, according to current nomenclature.
In other cases a considerable amount of nephelite, or of glass
with probably a nephelitic composition, is present with the feld-
spar, giving rise to nephelite-basanites ; while, again, neither
feldspar nor nephelite is present, a glass base replacing them,
so that the rock may be called a limburgite. From the de-

* Sir A. Geikie, Textbook of Geology, 1903, vol. ii, p. 764.

+ Cf. H. S. Washington, Quart. Jour. Geol. Soc., vol. 1xiii, p. 69, 1907.

Washington—Catalan Volcanoes and their Rocks. 229

scriptions which follow it would appear that the feldspar-basalts
proper differ from the others both chemically and texturally
to such an extent that they may be considered as belonging to
a distinct rock group. On the other hand, it is clear, from
the microscopical as well as the chemical study, that the only

essential difference between the nephelite-basanites and the
limburgites lies in the presence of feldspar and in the erystal-
linity, the chemical composition being about the same and
modal and textural transition forms being common, so that they
may strictly be considered to be merely textural variants of the
samme rock. But for purposes of description a distinction may
be made between the feldspathic and feldspar-free forms.

In terms of the quantitative classification these basalts are
somewhat difficult to classify accurately without chemical anal-
yses. Those that have been analyzed by me, seven in number,
all fall in the salfemane class, and there can be no doubt that
all the others may be safely referred to the same chief division.
As regards the order, the great majority of those analyzed fall
in the sixth, the lendofelic portugare, only one of the analyzed
rocks falling in the perfelic order gallare. But this ordinal posi-
tion is correlated with easily distinguishable microscopical dif-
ferences, so there is little or no hesitation in referring the unan-
alyzed specimens to the one or the other order. For the most
part these rocks are alkalicalcic, the greater part of those anal-
yzed falling in limburgase, with the perfelic one in camptonase,
and a few in the domalkalic monchiquase: but these differences
in rang are not discernible by purely microscopic means. As
regards the subrang they are all dosodic without exception, the
subrangs represented being camptonose, monchiquose, and lim-
burgose, so that we are safe in assuming that the unanalyzed
rocks are likewise dosodic, however uncertain we may be in
regard to their proper rang.

For purposes of description* the camptonose rocks, which are
all referable to one type and which are readily distinguished
from those belonging to the other divisions, will be considered
first. In view of the impossibility of discriminating micro-
scopically between the limburgose and monchiquose rocks,
these will be described together, but considered under the
head of two distinct types, a feldspathic and a non-feldspathiec,
which correspond with the current groups of nephelite-basanites
and limburgites.

The several types will not be named here in the systematic
and definitive manner proposed for the quantitative classitica-
tion, that is, by the use of a root derived from the name of a

* In the following descriptions many terms will be used which have been

recently proposed. Cf. Cross, Iddings, Pirsson, and Washington, Jour.
Geol., vol. xiv, p. 692, 1906.

230 =Washington—Catalan Volcanoes and their Rocks.
g

locality and the suffix —a/, but they will be considered as merely
provisional and applicable only to the rocks described in this
paper, and indicated by the use of a locality name alone, with-
out any suffix. This method is adopted for the reason that, in
the ease of such basaltic rocks, whose modal and textural fea-
tures present no marked peculiarities, but can be duplicated
in many rocks from well-known localities, and some of which
have been thoroughly studied in the earlier days of the science,
it does not seem fitting that the systematic, typal adjectives
should be based upon these little known and as yet imperfectly
described occurrences.
If the quantitative system of classification, or some modifi-
cation of it, is eventually adopted by petr ographers generally,
it will be the task of some future master of the science to
undertake the comparative study of igneous rocks and the
coordination of the various types, the distinctive typal adjec-
tives being bestowed with judicious consideration of all the
facts as to which locality is the most representative, the best
studied, or the longest known. For the present, in the applica-
tion of the quantitative system, it would seem to be wisest,
and the part of moderation, to be conservative in the matter of
systematic typal adjectives, and to bestow them only in the
case of the more unusual magmas, the less common combina-
tions of mode and textures (which may conveniently be called
motexes), or when redescribing in terms of the new classification
some well-known rock which is commonly recognized as ade-
quately representative of some usual motex (combination of
mode and texture). In other cases it may be suggested that
such a simple, substantival form, without the suftix —al, be
employed as is done here, admittedly in a provisional way and
with but a limited and local application. In certain cases, of
course, these provisional type designations may eventually be
rendered entirely systematic and definitive by the addition to
the locality root of the suffix —a/, if the rock described proves
to be worthy of this distinction.

Camptonose ( feldspar-basalt), Castellfullit type.

Megascopic characters.—Color, medium to dark gray: tex-
ture, phanerocrystalline, fine-grained, porphyritic, perpatic :
phenocrysts, 5 per cent or less of the rock volume, 1 to 3™ long,
of black or dark green augite and yellow olivine: groundmass,
very fine-grained, but many very minute, white tables, evi-
dently of ‘feldspar, and black grains, evident with the hand
lens. Some specimens, on weathering, show a mottled surface
of patches of light and dark gray.

Microscopic characters.—Minerals present, labradorite, aug
ite, olivine, and ores. The labradorite, of the average composi-

Washington— Catalan Volcanoes and their Rocks. 231
Y

tion Ab,An,, forms about one-half of the rock, is subhedral to
euhedral, thin tabular, the tables from 0°3 to 0-7" long and
0:02 to 0:05™™ thick, twinned according to the Carlsbad and
albite laws, zonal structure not present. The augite, which
forms about one-quarter of the rock, is very pale g oreenish oray,
with an extinction angle on J(010) of 39°. The phenoer ysts
are subhedral to anhedral, stout prismatic to equant, from 0°5
to 2™™ long, running serially down to the abundant small,
(O;0N6 “to 0-05") anhedra of the groundmass. The olivine
phenocrysts, shghtly larger than the augites, are euhedral to
subhedral, the usual domes and pinacoids being present, almost
equant, often fragmentary, quite colorless and fresh. There
is little or no olivine in the groundmass. Anhedral to subhedral
ore grains, from 0-01 to O-05™™, are common, scattered uni-
formly through the rock. The large amonnt of TiO, shown
by the analysis indicates that it is highly titaniferous, but noth-
ing that could be definitely identified as ilmenite or titanite
was seen. Between these constituents there is a small amount
of a colorless, distinctly anisotropic base, which may be consid-
ered to be probably alkali-feldspar, as ‘the analysis indicates
that a considerable quantity of this must be present. The
texture is essentially intersertal, characterized by the diver-
gently arranged tables of plagiocase, with interstitial granular
augite, olivine, and ore. As the augite is not xenomorphic
towards the other constituents, and is not poikilitic, the texture
cannot be regarded as ophitic in the true sense of this term.
The mottled, “weathered specimens showed no definite peculiar-
ities under the microscope which could be connected with this
appearance,

Mode.—The mode of the specimen from Castellfullit which
was analyzed was estimated by Rosiwal’s method, a large num-
ber of measurements being made on two sections, with the
following results:

Beldspary2 icc sot ate cen ag
TAMU OU Clee praia oe eoe, t eeeeeee 25°8
Olivamemee es yas i Sa aCe en 10°8
Ore gee ta ces oc en mae cree 8°9
BAN ICC ete ies aes egies ee 0°8

100°0

Chemical composition.— A specimen of the upper, prismatic
flow at Castellfullit was analyzed, with the results shown in I
of the table of analyses on page 2 239, which may be regarded as
typical of this type. The silica is distinctly higher than in the
other rocks of the district, but in other respects it does not
differ widely from them. The rather high total oxides of iron,

232 Washington—Catalan Volcanoes and their Rocks.

and especially the great preponderance of ferrous over ferric
oxide, as well as the comparatively large amount of TiO,, are
the chief features to be noted.

Classification.—Vhe norm of this rock is as follows:

Or: e80) (Sale 16585 Class III
Ab __ 24°63 Pe) ta Te Salfemane
An .. 19°18 55°27
NewS 2256. 2256 j F Order 5
Die 9-59) 39°47 1G aes Gallare
OTT 19:88 Wea)
Mites 2418s) k,0+ Na,O Rane 3
nate 12S 9 al AAO Fy oes re ees a (Ny wens

ies iiic8 OR ( CaO’ oe: Camptonase
Ap... 1:00 1:00)

SS KE OW a mae Subrang 4

100°22 INaHOk ms Camptonose
Rest 37

100°59

The relations of norm and mode are clear on comparison of
the figures above with the mode given on a previous page.
The modal augite takes up all the normative diopside, w ith a
little of the normative anorthite, olivine, and magnetite, but
the divergence of the mode from the norm is not a serious one.
As the soda-lime feldspar was determined optically as a labra-
dorite of about the composition Ab,An,, there must be present
about 15 per cent of a soda-orthoclase, of about the composition
Or,Ab,, and this, with 8 or 4 per cent of nephelite, is to be
looked for in the colorless, doubly-refracting base. On the
whole therefore the mode may be considered to be a normative
one, and the rock may be briefly described as a femphyro-
interserti-camptonose. In the current classifications it is a
feldspar-basalt, texturally of the common intersertal type.

Occurrence.—This type of camptonose was observed by me
only in the district around Olot, where it forms extensive
flows with well-developed prismatic structure. Some prom-
inent localities where this type was collected are: the upper
flow at Castellfullit,* a flow of Puig Estany near the 45
kilometer stone on the road from Olot to Castellfullit, a flow
from the Puig Dolors in the Torrent Mal near Castellfullit,-
and the flow in the valley of the Rio Ridaura north of La
Garrinada and also near San Juan las Fonts. A slightly
weathered lava, mottled in light and dark gray, which was
met with as blocks on the south slope of La Garrinada, near
the lowest crater, as well as at Castellfullit, also apparently
belongs to this type.

*My specimens of the lower flow were unfortunately mislaid, but in the

field no difference could be detected between these and those of the upper
flow.

Washington— Catalan Volcanoes and their Rocks. 233

Limburgose and monchiquose (nephelite-basanite), Las Planas
type.

As was mentioned above, no distinction can be made out
either in the field or under the microscope between the
rocks which belong to the two subrangs, so they will be
described together. Furthermore, as no exact dividing line
ean be drawn between the more nearly holoerystalline and
compact rocks, which show evident nephelite, and the highly
scoriaceous ones, in which a glass base replaces the nephelite,
they will be described under one type. This is not strictly
correct, but the abundance of transitional forms, and the com-
paratively unimportant character of the differences between
the extremes, seems to render it the more judicious procedure.

Jegascopic characters.—Compact to highly vesicular and
scoriaceous: very dark gray to black, the latter especially in
the scorize, which weather to yellow and red; perpatic, the
aphanitic groundmass thinly sprinkled with small phenoer 18
of black or dark green augite and light yellow olivine, from
to 5™ in diameter. Nodular masses’ of granular, sellow
olivine present in some cases.

Microscopic characters.—The not very numerous pheno-
erysts of colorless or very pale, greenish gray augite, and color-
less olivine, are similar to those in the ast type, the former
subhedral and stont prismatic or fragmentary, and the latter
more often euhedral. These are common to all the specimens.

The groundmass varies more widely. In all cases small,
thin; tables of plagioclase, usually about Ab,An,, are present,
the laths varying in length in different specimens from 0°5 to
0-057", being uniformly, less broad and thinner in the more
scoriaceous rocks. These tables show a diverse arrangeinent,
and flow structure is seldom seen. With them are usually
present numerous very small anhedral grains of colorless
augite, olivine being rare as a true groundmass constituent.
In other and usually the more scoriaceous specimens, these
small groundmass augite grains are rare or even entirely want-
ing, being replaced by al ageregation of dark, minute, dusty
particles, “which render the base almost opaque in spots. Ore
grains are common and well developed, though very small, in
the more compact specimens; but in the more scoriaceous are
merged in the dusty aggregate just mentioned. The base in
which these lie is often quite holoerystalline, in which case
it is feebly doubly refr ne and is considered to be nephe-
lite, at least in great part. That this attribution is correct is
shown by the vecasional presence in some of the more com-
pact and holocrystalline specimens of rectangular sections,
with parallel extinctions and low birefringence, which are

M. JOUR. ora ee SERIES, VoL. XXIV, No. 141.—SupremBer, 1907.

a! i

234 Washington—Catalan Volcanoes and their Rocks.

indoubtedly nephelite. Im the more scoriaceous material the
base is entirely vitreous, and is apt to be much obscured by
the abundant black dust spoken of above.

The textures offer nothing of special interest, being com-
mon to many such basaltic rocks. The feldspars are not
sufficiently large, or numerous enough to render the rock
ophitic, and in ‘the more compact forms the texture may best
be designated as hy alopilitic, while in the more glassy ones
it is simply vitrophy ric.

Mode.— Owing to the very small size of the groundmass
constituents and the consequent extensive overlapping, as well
as the presence of glass in many cases, no satisfactory estimate
of the mode by Rosiwal’s method was possible. However, by
making certain readjustments from the figures furnished by
the chemical analyses, on the basis of the relations between
norm and mode observed in the Castellfullit rock, the modes
of three of the more compact analyzed specimens were calcu-
lated to be approximately as follows: column I representing
that of the Llora specimen, II that of the rock from the sum-
mit of Garrinada, and IIi that from Las Planas.

iL IT TIT
Labradorite, (Ab, An.) -- ---- 35°0 35°0 38°0
Nephelite sii sare 12-0 10°0 6:0
AUOIt ene, es eis ues Semen. 25°0 30°0 28°0
Qn vinn 4 eae Ee ease ia 15°0 8:0 13-0
ORES 2a Se Be ee ae eS 11°0 15:0 14-0
A atite cers keener nee 2°0 2°0 1-0

100:0 100°0 100°0

Chemical composition. —Analvses of five rocks of this type
will be foundin columns II to VI of the table on page 939.
Of these, the first three pe the more compact forms,
Il and IV being flows and III being a somewhat scoriaceous
lava block, while the last two (V and VI) are scoriaceous lavas
from the recent cones. As compared with the analysis of the
Castellfullit flow, they are distinctly lower in silica, but resem-
ble it in other respects, and each other in all, very closely.
The greatest variation is seen in the relative figures for the
iron ‘oxides, where there seems to be evident an increase of
ferric oxide at the expense of ferrous in the more scoriaceous
lavas of the cones, which may reasonably be referred to the
difference in the conditions which obtained during the extru-
sion and solidification, the possibilities of the oxidation of
ferrous to ferric nde being greater in a vesicular lava than
in a compact flow. The uniformly high figures for T10,,
which vary within very narrow limits, are also to be noted: as
an important feature.

Washington—Catalan Volcanoes and their Rocks. 235

Classification.—The calculated norms of the five rocks
analyzed are given below, the ratios being omitted. It is seen
_ that the specimens from Llora and La Garrinada fall in the
subrang monehiquose (I1I.6.2.4), while those from Las Planas,
Cruzeat, and Montsacopa fall in limburgese (III.6.3.4), the
difference in rang depending on shght differences in the
amounts of the alumina and the alkalies. This is shown in
the norms by the distinctly higher anorthite and lower albite
or nephelite in the last three. It cannot justly be said that
the differences are very significant, and as it seems probable
that the rocks are for the most part in limburgose, they may
be referred to collectively as monchiquose-limburgose.

oa

II Til IV Vv VI
Oe ee ee oi WA 11:12 13°90 15°34
JAN) Dene os eae 10°22 12°58 15°20 8°64 13°86
INU eee 10°84 11°95 14°18 14°73 16°12
Ney 225254 19°92 11:08 6°82 12:07 9°51
Dee ys Oy 23°76 22°02 24°18 21:08
Olin a es NG 8°68 14°10 10°49 9°98
VEG ey eae 4°18 4°64 5°34 4°64 6°73
DG ees a 8 Oi 8°66 9°42 7°75 8°06
EN Dr Se veins 1°68 1°68 1°34 1°34 1°68
vestss222- “08 68 "47 1°53 ts)

100°77 997 70e 100501 99°27 100°66
Symbols ss. Wi1.6.2.4 I116.9.4 1116.3.4 D16.3.4 I1.6.3.4

II. Monchiguose, Llora.

III. Monchiquose, La Garrinada.
IV. Limburgose, Las Planas.

V. Limburgose, Cruzcat.

VI. Limburgose, Montsacopa.

Occurrence.—This seems to be the most abundant rock type
in the district, being represented both in flows and at the
cones. Among the former the following localities may be
mentioned : the flow from Bafia de Boch, below the Chapel of
Sant Medir, near Llora; the end of the extensive flow from
Puig Monteal, collected at Domeny, but not very fresh; and
the extensive flow to the southwest of Olot, as at San Esteban
le Bas, San Feliu de Pallarols, and the quarries in the vailey
of the Amer at Las Planas. Among the cone occurrences may
be named: the compact lavas from Cruzcat and from the sum-
mit of La Garrinada, and the scoriaceous ones from Montolivet,
Montsacopa, Bisaroca, Puig Costa, and others. The lava field
of Bosch de Tosca is also made up largely of this type, and
the dike found 38 kilometers from Olot, on the road to Santa

236 Washington—Catalan Volcanoes and their Rocks.

Pau, is transitional between this and the succeeding type, the
feldspar tables being extremely small and very sparingly
‘present.

Limburgose (limburgite), San Roque type.

Rocks of this type were but rarely met with, and they differ
from the most vitreous of those just described chiefly in the
absence of feldspar. While, therefore, they may be regarded
as probably but a variant of the other type, yet they may be
discussed separately on account of the mineralogical dissimi-
larity.

Megascopic characters.—Compact, dark gray, aphanitic,
almost wholly aphyric, only a few very small nodules of brown-
ish augite being visi ble, though in certain lights the fresh
fracture shows an ill-defined olimmer, due to the presence of
scarcely visible, minute prisms (of augite).

Microscopic characters-—The rock is microporphyritic,
about 10 per cent of microphenocrysts of augite and 5 of
olivine being present. The former are anhedral to subhedr al,
from 0°53 to 1:0", long, stout prismatic to fragmentary, and
of a pale brownish gray color: the latter euhedral to subhe-
dral, 0-2 to 0°5™™ in diameter, colorless. Both carry inelu-
sions of small magnetite grains. The microgroundmass
consists of numerous, subhedral prismoids of pale brownish
augite, 0-1 to 0:03™" long, and fewer magnetite anhedra, of
about the same size, but no olivine, embedded in a colorless,
isotropic base, which is evidently glass. No flow texture is
evident, and the small augite prismoids are too stout and nearly
equant for the texture to be considered typically hyalopilitic.

Mode.— As the type is highly vitreous, the mode is indeter-
minate, and the small size and crowded arrangement of the
eroundmass augites renders any exact estimate of their amount
difficult. Consideration of the norm given below, and study
of the sections, however, indicates that the following figures
will express roughly the mode of the San Roque rock :

WaNIUSA NS) ahalepuenes Se ee hy alison ie; A Sat LOA)
Olavai e222 es ea De aang eo aoe 5:0
Macnetite |. 205i Miieaie ares 10-0
Glass, ith Sie eae On a Oe)

100°0

It may be said of this that the microphenocrysts of augite
constitute about 10 per cent and the small groundmass crystals
about 25, while olivine seems to be present only as micropheno-
crysts, of which certainly not more than 5 per cent are pres-
ent. On referring to the norm given later, it will be seen that,

Washington—Catalan Volcanoes and their Rocks. 237

therefore, the other 10 per cent of normative olivine has prob-
ably taken up silica to form modal augite, thus increasing the
amount of nephelite potentially present in the mode over that
shown by the norm, and this, with the feldspar, is represented
by the glass base.

Chemical composition.—The chemical composition of the
rock of the massive flow exposed at the Fuente de San Roque
is shown in column VII of the table of analyses on page 239.
It does not differ in any essential respect from the others, the
only feature of note being the slightly higher ratio of potash
to soda, though the figures for each of these can be almost
duplicated from the other analyses, and the ratio is but 1:2,
so that the rock is clearly dosodic. Alumina is shghtly higher
than in the other analyses, but ferrous oxide greatly surpasses
ferric, and the amount of titanic oxide is of the same order as
in the rest.

Classification.—The norm of this rock is as follows:

Ore 34 Salas fee Class HI
poate 11°79 + 47°37 Fem —_—s Salfemane
AME = (22524 54°61]
ING een 2a ist 24 Ee aie, Order 6
De eee 79) ) ee leaps: Portugare
| 82°55 | 8
O76 ( }
Mt___. 3°48 : K.,0+Na,0___., Rang 3
e760 fd 08 ; 45°31 1 CAO! Tate Limburgase
Apes 1-68 1°68 |
Rest_. °46 K,O Subrang 4
e Na,O — Mae Limburgose
100°38 F

The high ordinal ratio, which is near the limiting value of
7-00, indicates that the rock is close to the border of the 5th
order, gallare, and that it might be regarded as transitional
and called a camptonose-limburgose. But the simpler term
would seem to be sufficient for the present. In the prevailing
classifications the rock is undoubtedly to be considered as a
limburgite, since no feldspar is present, but only glass as a col-
orless (salic) constituent, along with augite, olivine and ores.
Limburgite is used here in the sense of Rosenbusch* as denot-
ing “the feldspar-free extremes of the rock series of the
trachydolerites, tephrites and basanites, and the leucite-,
nephelite- and melilite-rocks: they represent therefore the
- pyroxenic rocks of the essexite family and belong to the thera-
litic magmas.” It is evident from this that Rosenbusch
intends to denote by the term limburgite a rock derived from
magmas so deficient in silica, and at the same time so distinctly
alkalic in character, that lenads (leucite or nephelite) would

*Rosenbusch, Elemente der Gesteinslehre, 1901, p. 376.

238 Washington—Catalan Volcanoes and their Rocks.

form were the rock holoerystalline. They are thus contrasted
with the otherwise mineralogically similar vitreous and feld-
spar-free basalts proper, to which the older name ‘ magma-
basalt” is more truly applicable, though of doubtful form and
etymology. These are less markedly deficient in silica, or even
with a slight excess of this, and at the same time not as strik-
ingly alk alic in character as the rocks mentioned above, so that
lenads would not be necessarily present, and if so only in very
small amount, were the rock holocrystalline, but there may be
considerable olivine if there is a deticiency in silica.

These magmatic differences between the limburgites and the
augitites on the one hand, and the feldspar-free vitreous basalts
on the other, which are only vaguely recognized in the pre-
vailing classifications, are clearly ‘brought out in the quantita
tive classification. In this the typical ‘examples of limburgites
and augitites, so far as they are represented by good analy Ses,
fall in the lenic orders of salfemane, especially in portugare,
and the greater part of them in the ‘alkalicalcic rang limbur-
gase and the dosodic subrang limburgose : while (it may be
added), the most typical monchiquites, in which a highly sodie
and an analcitic base is abundant, fall in the same order, but in
the domalkalic rang monchiquase and the dosodice subrang
monchiquose. On the other hand, the basalts proper belong
almost exclusively to perfelic orders, germanare and gallare,
and to docalcic rangs and presodic subrangs of these, notably
hessose and auvergnose.*

Occurrence. —Clearly recognizable specimens of this type
were obtained by me only at one locality, the massive flow at
the Fuente de San Roque, about one kilometer west of Olot,
on the south bank of the Fluvia River. Other specimens
which much resemble this, but which contain traces of feld-
spar and are therefore transitional toward the former type,
were obtained near a small wine-shop above the chapel of Sant
Medir, at Llora, and at the dike 3 kilometers from Olot, on the
road to Santa Pau, already mentioned.

General Characters of the District.
Mineralogical characters.—These are so simple and are so
evident from the descriptions which have been given above,
that little need be said of them here. Augite, olivine and
titaniferous magnetite are invariably present, the first always
in very notable amount, and the last two in much smaller
quantities. Hornblende, biotite, and titanite (which last
might be expected in view of the richness of the rocks in tita-
nium), are entirely absent, not a single grain of either of these

having been observed by me in any “of the sections.

*Cf. Iddings, Prof. Paper U. S. Geol. Surv., No. 18, pp. 90, 91, 1905;
and Washington, Prof. Paper U. S. Geol. Surv. , No. 14, PP. 7, 76, 80, 1903.

Washington— Catalan Volcanoes and their Rocks. 239

Of the salic minerals a plagioclase of about the composition
Ab,An, is the most common, though it is quite wanting in
some of the rocks. Alkali feldspar seems to be rare, though
its presence in some amount is demanded by the chemical
composition. It would seem that, in most cases, the condi-
tions did not permit of its crystallization, and that it solidified
asa glass. Nephelite, sometimes in almost euhedral crystals
but usually as an ill-defined base, is more common, but this,
too, is prone to solidify as a glass. The other prominent salic
minerals, quartz, leucite, and the sodalites, are not found in
these rocks.

Chemical characters.—The chemical features of the Catalan
rocks are shown by the analyses in the table annexed, all of
which were made by the writer. Considering the small size of
the district, and the general uniformity in the mineralogical
and textural features, they would seem to be sufficiently numer-
ous to convey a just idea of the true relations.

I II JOU ID W Vi VII
SiO,--- 47°66 44°55 43°64 44°29 44:20 44°82 44°80
ROU 360) 911-4 e ei -1O) 1262 1 13°96). 14-06) 15-51
Fe,0O, Ren 2283 2°81 6°40 3°61 3°19 4°56 2°35
HeO i" 78:44 8°54 D02 8°84 8°41 eu 8°52
MgO ._- 8-19 10°85 9°36 10°06 8°03 8°60 8°83
CaO =) 9°36 TOD 9°52 9:23 Sao) 9°56 9°91
Na OS. 355 i 4°04 3°88 3°25 3°66 3°69 2°99
REO eld 4 2°57 2°18 1°82 2°35 2°30 2°29
BOE Waly 056 0-49 0-21 0°76 OO PUG

0°18

H,O-..- 0°20 0°16 0:09 Onl? 0:05 0:09
€O;==) none, - none none none none none none
LiO, 5 TBS} 4°32 4°55 4°92 4°10 4°25 4°01
Li eee ened n.d. n.d. 0°02 n.d. n.d. n.d.
roe epee ()) 05) 0°70 On 0°57 0°62 0°67 0°68
SO oe. nd. 0°05 n.d. 0°05 n.d. n.d. n.d.
NnOr nid: n.d. n.d. n.d. 0°51 none 0-08
INGOs ned: n.d. n.d. n.d. 0°14 n.d. 0.13
BAOmas nid: n.d. n.d. 0:06 n.d. eG n.d.
SEOsaret Tea: n.d. 0:03 0:04 n.d. n.d. n.d.

100°54 = =—9964 99°60 99°68 99°84 100°13 100°35
Subrang IIJ.5.3.4 II1.6.2.4 I11.6.2.4 II1.6.3.4 111.6.3.4 I11.6.3.4 I11.6.3.4
i. Camptonose (feldspar-basalt). Upper Flow, Castellfullit,
near Olot.
Ii. Monchiquose (nephelite-basanite). Below Chapel Sant
Medir, Llora, near Gerona.
Ill. Monchiquose (nephelite- basanite). La Garrinada, Olot.
LV. ee (nephelite-basanite). Las Planas, South of
lot
VY. Limburgose (nephelite-basanite). Cruzcat, South of Olot.
VI. Limburgose (nephelite-basanite). Montsacopa, Olot.
VII. Limburgose (limburgite). Fuente San Roque, Olot.

240 Washington-—Catalan Volcanoes and their Rocks.

Apart from the silica of the Castellfullit camptonose (feld-
spar-basalt), which is about 3 per cent higher than in any of
the other rocks, all these analyses reveal a remarkable uniform-
ity as regards all the constituents. With the exce ption just
noted, silica is rather low and varies only between 43°64 and
44°82, alumina between 12°48 and 15°51, the total iron oxides
between 10°87 and 12°45, magnesia between 8-03 and 10° 85,
lime between 7:99 and 9: 91, soda between 2:99 and 4: O4, potash
between 1:54 and 2°57, titanium dioxide between 3°83 and 4:92,
and phosphoric pentoxide between 0-45 and 0-74. The only
variation of moment is that in the relative amounts of ferric
and ferrous oxides, the former being between 2°35 and 6°40,
and the latter between 5°32 and 8° 84, the maximum for ferric
oxide and the minimum for ferrous ‘being found in the same
specimen (III), and apart from this the figures being fairly con-
cordant. As noted above, it would seem that the conditions
of extrusion had some influence over these relations, the ratio
of ferrous to ferric oxide being in general higher in the mas-
sive flows and lower in the scoriaceous lava, this being due pre-
sumably to the more favorable conditions for oxidation in the
latter. A somewhat similar state of affairs has been observed
in the basalts of Sardinia, Pantelleria and Linosa, as will be
described in subsequent papers.

Apart from this general uniformity, and the chief charac-
ters of the main constituents, such as ‘the rather low silica and
alumina, the high iron oxides, moderately high magnesia and
lime, and rather high soda (for salfemic rocks) and low potash,
the most striking “feature is the very large amount of tita-
nium dioxide. While most or all of the fioures for this con-
stituent here given can be duplicated in analyses of igneous
rocks from other localities, yet the vast mass of analyses hith-
erto published reveal scarcely a single district where the figures
for titanium dioxide are unifor mily so high as those of Catalo-
nia. Aside from analyses of my own of rocks from Pantelle-
ria and Linosa, and not considering titaniferous ores, the only
ones which show figures surpassing those here given are those
of the melilite-basalts of the Hegau by Grubenmann, which
are all dofemanes and rich in perofskite ; while figures approx-
imately the same as those for Catalonia are seen in analyses of
salfemanes and dofemanes from German Kamerun, Magnet
Cove, Brazil, and some localities in New England.

Of the other minor constituents few call for any special
notice. Phosphoric pentoxide is fairly but not abnormally
high, zirconia seems to be present merely in traces, as is also
true of baryta and strontia, while manganese is present in the
usual small'amounts. Nickel is invariably present apparently,
and was determined in two eases, the figures being rather high

Washington— Catalan Volcanoes and their Rocks. 241

for this rare constituent, while its presence in approximately
similar amounts in the other specimens was indicated by the
greenish color of the filtrate from the ammonia precipitate. [In
this respect, as in the high titanium, these rocks resemble the
basalts of Sardinia, Pantelleria, and Linosa.

Relations to other Districts.

It would seem most natural to connect the igneous rocks of
the Catalonian volcanoes with those of the igneous areas which
extend in a well-defined zone down the east coast of Spain ;
including the Columbretes Islands,* the islands of Mallorca
and Iviza,+ the igneous areas of Murciat and Cabo de Gata,§
and the island of Alboran. | This Conncenon is, indeed,
assumed to be true by the Spanish geologists, as Quiroga and
Calderon,*| who also point out the preponderance of salic rocks
toward the southern end and of more femic ones towards the
northern. While the tectonic relations and those of time, as
most of these eruptions are, apparently, of comparatively
recent geologic date, like the flows and cones of Catalonia,
favor this view, yet it cannot be accepted as definitely proven.

With the exception of the Murcian rocks, of which excellent
analyses by Dittrich are published by Osann, the analyses of
the other localities leave much to be desired both as to com-
pleteness and accuracy. ‘Titanium especially is apt to be neg-
lected in them, and the amount of this constituent is a highly
important feature of the more femic rocks, for the purpose of
comparison with those of Catalonia. In the only analyses in
which it is reported, those of the rocks of the Columbretes and
of Murcia, the amounts are low, the highest figure being 1°60
per cent, in a verite of Fortuna in Murcia, so that there is lit-
tle resemblance in this respect with the Catalan rocks. Fur-
thermore, the Murcian rocks described by Osann are relatively
and absolutely much higher in potash, and leucite even appears
in some of them, while hornblende and biotite are not uncom-
mon at many of the localities south of Catalonia. It would
seem best, therefore, to look upon any connection of the Cata-
lan rocks with those of the more southerly localities with
doubt, and as a matter to be settled in the future by more sat-
isfactory analyses and further investigation.

The possibility may be suggested here of a connection of the
Catalan rocks with those of the central voleanic line of Spain,

=F. Becke, T. M. P. M., vol. xvi,.p. 155, 1896.

+ Cf. S. Calderon, Bull. Soc. Géol. Fran., vol. xiii, p. 112, 1885.

¢ A. Osann, Rosenb. Festschr., p. 263, 1906.

SA. Osann, Z. da. G. G., vol. x1, p. 694, 1888 ; vol. xli, p. 297, 1889 ; vol.
xliii, pp. 323, 688, 1891.

| F. Becke, T. M. P. M., vol. xviii, p. 525, 1899.

4] CE. Calderon, Bol. Soc. Esp. Hist. Nat., p. 335, 1905.

242 Washington—Catalan Volcanoes and their Rocks.

which runs north and south in the provinces of New Castile
and Aragon, and where nephelite-basalts and limburgites are
the prominent types.* It must be said, however, that no anal-
yses have yet been made of these rocks, so far as I am aware
of, so that the question must be left an open one. The matter
of the geologic age does not seem to be decisive one way or
the other, as the Volcanoes of the central line belong to the
Tertiary, while there is reason for thinking that the Catalan
activity began at about the same time, and also the earliest
eruptions at Cabo de Gata antedate the Pliocene.

Looking north, it would seem to be probable that some of
the volcanoes along the southern French littoral, as at Agde
and Montpellier, are connected with those of Catalonia. “Ba-
salts” are known to occur here, thongh no analyses of them
have yet appeared, to the best ‘of my knowledge. Through
the kindness of Prof. Lacroix I have obtained some specimens
of these rocks, and hope in the near future to make some anal-
yses of them, so as to test this point chemically.

Chiefly on account of the character of the basalts, and their
high content in titanium, but on other grounds as well, I have
elsewheret expressed the opinion that the Catalan volcanoes
are connected genetically with those of Sardinia, Pantelleria,
and Linosa, and possibly with eruptions in Tripoli, and even
with the voleanoes of the Great Rift Valley in East Africa.

But the full discussion of the question of this comagmatic
region or petrographic province must be postponed until many
more analyses are made and the various districts have been
described in detail.

Locust, N. J., May 1, 1907.

* Calderon, Bol. Soc. Esp. Hist. Nat., p. 336, 19095.
+H. S. Washington, Q. J. G. S., vol. lxiii, p. 78, 1907.

. Holm—Anemonella thalictroides. 943
I ISS A lla thalictroides 943

Art. XXIV.—Anemonella thalictroides (L.) Spach; an ana-
~ tomical study; by Taro. Horm. (With five figures in text,
drawn from nature by the author.)

Amone the plants that have been changed from one genus to
another Anemonella offers a good example. It was an Anem-
one according to Linneeus, who called it A. thalictroides, and it
was placed as the last species of this genus. According to
Richard it became a Thalictrum: Th. anemonordes, and as such
it was accepted by De Candolle and Gray; Bentham and Hooker
brought it back to Anemone, while Prantl removed it from
there, and placed it once more under Thalictrum, as a mem-
ber of the section Camptonotum. Meanwhile our plant had
been raised to generic rank as Syndesmon by Hoffmannsege,
though only as a name (1832), but a few years afterwards
Spach (1839) described it as a monotypic genus: Anemonella.
As such it was finally accepted by Gray in his manuscript to
the Synoptical Flora.

In several respects our plant is quite remarkable and
possesses some features in common with Anemone (the involu-
cre and the white, petaloid calyx), some others with Zhalic-
trum (the costate achenium with one pendulous ovule), and
finally with Zsopyrum (the habitus). In accordance with our
own observation the ovule of Anemonella has two integuments,
and in this respect it agrees with Zhalictrum, but not with
Anemone, in which only one integument is developed. In
Lecoyer’s st monograph of 7) halictrum, our plant is considered
a close ally of Zh. iuberosum, and enumerated as a member of
the section ‘‘ Brévistaminés”; the same disposition has been
made by Mariet in a paper dealing with the anatomy of the
Ranunculacee.

The anatomical treatment by Marié is, as far as concerns
Anemonella, very incomplete, since only dried material was
studied. For this reason we have examined the plant again,
and the following notes may, thus, be considered as a supple-
ment to the various diagnoses already e given by other authors.
But to draw a comparison between our genus and /sopyrwm
from an anatomical point of view, as attempted by Marié, is at
present not possible ; because Isop yrum is actually an aggre-
gate of several very distinct types, of which L. thalictroides
seems to be the only one, that has been studied, so far, from
satisfactory material. We shall therefore confine ourselves to
treat Anemonella alone, but leave the comparison between

* Bull. Soc. Roy. de Bot. Belgique, vol. xxiv, p. 78, 1885.
+ Ann. d. Se., ser. 6, vol. xx, p. 40, 1885.

244 T. Holin—Anemonella thalictroides.

this and other genera to future students, who might have access
to suitable and more ample material of ‘the family.

We have already described the seedling of Anemonella,*
and from this will be seen that the cotyledons are epigeic, and
that the hypocotyl may not be developed ; since then we have
noticed, however, that a hypocotyl is most frequently present,
and not so very short either. The primary root does not per-
sist for more than about one season, and becomes replaced by
secondary roots, all of which become tuberous at the base, but
not until during the fall; they die off during the next spring,
when new ones become developed. A rhizome of the mature
plant may be seen in the accompanying figure 1, which was
collected in the month of April. In this specimen there is an
old, shrivelled root from the previous year (1), and several
tuber ous, which are active; besides these there is, furthermore,
a slender secondary root (R ), which will become tuberous dur-
ing the fall. Characteristic of these roots is that only the base
is “tuberous, while the apex is very long, filiform and amply
ramified. Sometimes these roots become branched from the
tuberous base, as illustrated in figure 2; in one of these the
tuberous portion is four-cleft, each division being terminated
by a long, filiform apex.

The very slender stem bears two sessile, involucral leaves,
which are opposite and trifoliolate ; the inflorescence is a cyme
with one central and two or a few more lateral flowers. Of
these the terminal is often larger than the others, and has —
generally a larger number of sepals, from six to nine, while
the lateral flowers ver y often have only five. The achenium
has eight very strong ribs (figures 3 and 4), and contains a sin-
gle pendulons ovule with two integuments.

Our plant is vernal, and one of the earliest to bloom; it
inhabits open, especially deciduous woods along the Atlantic
coast from Canada (Ontario) to Florida, westward to Minne-
sota. It is generally associated with Luzula campestris,
Houstonia coerulea, Hypoxis erecta, ete.

Tue INTERNAL STRUCTURE OF THE VEGETATIVE ORGANS.
The roots.

As stated in the preceding pages, the roots of Anemonella
are tuberous, when fully developed. But when we examine
fruiting specimens, we always notice that besides the tuberous
there may be a few others ‘which are slender in their entire
length. These are young and develop from the very short
rhizome, a little above the tuberous. The function of these
slender roots, however, is also to store nutritive matters, and
their increase in thickness commences at a very early stage,

*Mem. Torr. Bot. Club, vol. ii, No. 3, p. 57, 1891.

T. Holm—Anemonella thalictroides. 945

even if they do not become tuberous until at the end of the
season.

\ > \
x i / feos
he g
, Ay \t a
| \ J a
WN 4
‘ \ | D> &) “a J
\ / y
\ i oS Hae i)
fi" 1,
q Yh J, 4
WY /: y

In the old roots we must distinguish between the tuberous
hase and the long, filiform apex. The latter shows only a

246 T. Holu—Anemonella thalictroides.

very slight indication of increase in thickness by a few eell-
divisions inside the leptome. Otherwise the structure is very
primitive. There is an epidermis with numerous hairs, and
four strata of thin-walled cortical parenchyma. Endodermis is
thin-walled, and the pericambium consists of one layer outside
the leptome, but of two outside the proto-hadrome. Four
broad groups of leptome alternate with four rays of ves-
sels, which extend to the center of the stele, no pith being
developed.

In the tuberous portion of this same root the structure is
very different on account of the large increase in thickness.
Epidermis, the cortical parenchyma and endodermis haye been
thrown off, but replaced by a few strata of pericambial cork ;
inside the cork is a large secondary cortex of thin-walled paren-
chyma with narrow, rhombic intercellular spaces. The stele
now represents eight rays of mestome radiating from the
broad central pith. The primordial hadrome is readily to be
seen as four very short rays (H in fig. 5, which represents
only one-half of the stele), while the secondary rays are longer
(H*); outside these secondary rays of vessels are corresponding
strands of leptome (L), thus constituting ordinary collateral
mestome-bundles. The primary leptome is no longer visible,
but three secondary leptomatic strands have become “developed
outside each of the primary and secondary rays of hadrome.
There are thus twenty-four strands of leptome in the tuberous
root-portion, but only four of these border directly on had-
rome (the secondary) ; all the others are isolated, but arranged
very regularly in eight radii, corresponding with the hadro-
matic rays.

A corresponding structure is to be observed in the slender
roots (R in figs. 1 and 2), and as stated above, these roots
become tuberous during the first season. Epidermis and the
primary cortex become ‘thrown off very early, but are replaced
by the endodermis and a few layers of pericambial cork. A
secondary cortex is amply developed at this stage and the stele
shows very distinctly four primary and four secondary rays of
hadrome. Outside the latter are numerous strata of cambium
and groups of secondary leptome. A thin-walled pith occupies
the center of these roots. These roots are secondary, and they
all grow in a horizontal direction; their increase in thickness
depends upon the formation of a secondary cortex from the
pericambium, and upon the development of secondary mes-
tome originating from cambial strata in the shape of arches in-
side the primary leptome; they are not contractile, and possess
no exodermis. It is now interesting to see, that the primary
root of the seedling is vertical, that it shows the same manner
of increase in thickness as the secondary, and that it possesses

T. Holm—Anemonella thalictroides. 947

an exodermis with very distinct foldings on the radial cell-
walls; in other words, the primary root is contractile and at the
same time able to store nutritive matters. It is diarch and con-
tains a broad central pith.

In Picaria ranunculoides the roots are, also, tuberous, but
lack the long, filiform apex observable in Anemonella. More-
over, the swelling of the root in /%caria depends merely upon
the pr esence of a broad primary cortex, while the stele shows
only a few secondary vessels at the very base of the root, in
the immediate vicinity of the overwintering bud. In Tsopy-
rum biternatum the roots are very different, since they attain
a tuberous development in several places on the same root, and
very irregularly.

The flowering stem.

This represents a long internode terminated by the inflor-
escence, and bears only two, opposite involucral leaves. It is
cylindric, glabrous and perfectly smooth. The cuticle is thick,
and the outer cell-walls of epidermis are moderately thickened.
There is no collenchyma, and the cortex consists of three lay-
ers with distinct intercellular spaces surrounding a closed
sheath of thick-walled stereome of about three strata, The
stele contains nine collateral mestome-strands arranged in one
circular band, separated from the stereome by a few layers of
thin-walled parenchyma. The mestome-strands contain cam-
bium between the leptome and hadrome; the vessels do not
show the arrangement in the letter V as is otherwise quite fre-
quently to be observed in this family, A broad, thin-walled
pith occupies the central portion of the stele.

The leaf.

The long and very thin petiole is cylindric and smooth, and
in regard to the internai structure it agrees in most respects
with the stem. However, there are only four large and two
very small mestome- bundles, aud the hadrome shows here the
position of the vessels in the shape of the letter V. But other-
wise the structure is identical with that of the stem. The
petioles of the leaflets show the same structure, but contain
only three mestome-strands.

The leaf-blade is dorsiventral; the cuticle is smooth on both
faces, and the lateral cell-walls of epidermis are undulate, espe-
cially on the lower face. The lumen of epidermis is wider on
the dorsal than on the ventral face, and the outer walls are
moderately thickened; the stomata, which are confined to the
dorsal face, are surrounded by mostly five, ordinary epidermis-
cells; they are level with epidermis, and the air-chamber is
wide, but shallow. The chlorenchyma consists of two layers

248 T. Hola—Anemonella thalictroides.

of high palisades and three open strata of pneumatic tissue.
A few cells of collenchymatic tissue support the leptome side
of the larger veins, which are, furthermore, surrounded by
thin-walled parenchyma-sheaths.

The development of collenchyma in Anemonella is rather
weak so far as concerns the stem and the leaves, and, as we
have seen from the preceding pages, stereome is the only
mechanical tissue that is represented im stem and petioles. A
very firm structure is, on the other hand, to be observed in the
achenium, when fully mature. We find here ridges of collen-
chyma and a ventral epidermis, of which the outer and radial
cell-walls are extremely thickened. There are eight collateral
mestome-strands, one in each rib, and they are surrounded, at
least partly, by thin-walled parenchy ma sheaths. The chloren-
chyma represents a few strata of homogenous tissue with much
chlorophyll.

In bringing together these facts derived from the internal
structure, our genus may be characterized as follows: The tu-
berous portion of the roots contains a broad secondary cortex
and pith ; the hadrome is represented by four short rays alter-
nating with four long ones, of which the latter are secondary ;
isolated strands of leptome occur in rays throughout the see-
ondary cortex, but radiating from the center of the root and
located in the same radii as the eight rays of vessels. The
stem has no collenchyma, and no endodermis, but a closed
sheath of stereome, which surrounds the single circular band
of mestome-strands. The petiole shows the same structure as
the stem. The leaves are bitacial; they possess normal pali-
sade cells, and the larger veins are supported by collenchyma,
besides that they are surrounded by typical parenchyma-
sheaths ; no stereome is developed in the leaves; finally may
be mentioned that the stomata lack subsidiary cells. Anemo-
nella has, thus, a monostelic axis, while Thalictrum has an
astelic. However, the general structure of the North Ameri-
can Ranunculacee is so little know n, that it is impossible at
present to decide whether our genus, from an anatomical view-
point, is more closely related to Anemone than to some of the
other genera, even if the ovule possesses two very distinct in-
teguments, a character which it shares with Zhalictrum.

Brookland, D.C., April, 1907.

EXPLANATION OF FIGURES.

Fies. 1 and 2.—Rhizomes of Anemonella thalictroides; R = young roots,
which will become tuberous during the fall; r—an old tuberous root from
the preceding year; St.= base of Howering stem; L= petiole of leaf; nat-
ural size. Fig. 3.—Achenium. Fig. 4.—Same in cross-section ; both mag-
nified. Fig. 5.—Transverse section of part of tuberous root ; H = primor-
dial hadrome-ray ; H® = secondary hadrome-ray ; L = secondary leptome ;
x 820.

CO. Palache—Mineralogical Notes. 249

Arr. XXV.—MMineralogical Notes; by Cuaries Paracnn.

1. Zoisite Crystals from Chester, Mass.

2. Phenacite as an Alteration Product of Danalite
from Gloucester, Mass.

3. Crystal Form of Chalmersite.

1. Zoisite Crystals from Chester, Mass.

Tue crystals of zoisite here described were found some years
since by Mr. E. L. Cowles of Chester, to whom the writer is
much indebted for the generous loan and gift of material for
study and for information concerning its occurrence.

Mr. Cowles states that the zoisite was obtained from “a vein
from two to five feet in width, located about three miles below
Chester,” which is taken to mean down the valley of the West-
field River from Chester village. The locality is in the town
of Chester and is distinct from the well-known locality in
Huntington.

The zoisite occurs in a rock consisting of a confused aggregate
of tremolite needles and prisms with colorless to pinkish diop-
side in stout prismatic crystals. In parts of the specimens
these minerals constitute the whole rock and again they are
cemented with granular calcite, suggesting that the rock as a
whole is a metamorphosed limestone. The zoisite is confined
to portions of the rock which appear to have once been open
cavities, into which the zoisite as well as tremolite and diopside
erystals projected. These cavities were afterwards infilled,
partly with quartz in which many of the crystals are imbedded,
partly with coarse granular calcite, removal of which with
dilute acid yielded the measured crystals. Irregular small
patches of lustreless granular graphite are also sparingly present
in the rock.

The zoisite crystals are slender prisms up to 8° in length
with lense-shaped cross section, the greatest diameter of which
is about 15°". The prisms are deeply striated in the direction
of their length and commonly show a brilliant cleavage parallel
to the side pinacoid which truncates the thin edges of the len-
ticular prism section. They are white in color, transparent and
glassy in parts, but largely opaque and milk-white owing to the
presence of many flaws and cross-joints which traverse them
in various directions.

Comparatively few of the crystals are terminated and only
one whose termination was complete was detached for
measurement. This crystalis shown in the figure* in about the

*The figure was drawn and the table of angles compiled by Mr. H. E.
Merwin.

Am. Jour. Scere SERIES, VoL. XXIV, No. 141.—SEpremBer, 1907.

250 C. Palache—Mineralogical Notes.

proportions of the original, which measured 1™ in length by
5" in maximum diameter. A number of fragments showing
partial terminations were also measured and a list of the forms
found upon them is presented in the table. The list includes

most of the forms previously observed on zoisite and a number
of new forms of which a few are well established by the data
observed.

In the prism zone the dominant forms are m and 6, present
on every crystal; a, #, g, and / are also frequently well de-
veloped. The remaining prism forms are present only as
narrow line faces and, as the zone is striated, are somewhat
uncertain forms. Where they occur with several faces of each
symmetrically developed on a single crystal and in good agree-
ment with calculated position, they have been accepted as good

forms.

Caleulated
Sym-
bol @
6 0102 002,00!
GE N00- 902100
ees Omeio 2 0
GareO i247
iO S30 CO Bo
§ “S20, 7-33
mp WU@ 6) Als}
yf 120 38 54
& BO) > Bere sly
14i@) — Bile ays}
a OPI OO Oo
z 041 00 OO
af WAL XO WO)
O Witte bie NS
ee ile Om o,4:
Par lsilets Lore 7
Ga alOr 82157
Fee ANO ee Si ei
5 OO) “7A Wo
j 540 63 38
y 221 58 13
A NO aii
/55 WEL DA Bie}
(20 AD Bis}
AC OME 2S
AOR iAa Je
470 42 41
isi = Ie ie
481 38 54
RMON Be sy
TOOL = Bion Byes
TG Bab Te AL, By;
Gp Auseil Bie) aby

C. Palache—Mineralogical Notes.

251

TABLE OF ForMS AND ANGLES OBSERVED ON 7 MEASURED CRYSTALS.

The terminations are chiefly

90°

96
90
90
90
90
90
90
90
90
34
53
28
33
4]
49

90
90
90
90
52
30
55

90
90
90
90
60
73
77
83
85
86

Pp
00'
00
00
00
00
00
00
00
00
00
26
54
57
O04
23

26

00
00
00
00
28
05
56

00
00
00
00
58
37
15
31
12
54

Measured Average Limits (in minutes)

SS (2 ee ee No.
@ p No. of
———J —-— SCO: erys-
co) p + — + — faces tals Quality
OO WO XO OO! NOT ae eS EH al retoxaya
89 56 90 00 CQ) SA LOD as RI AO (State) good
78 24 90 00 SO. 1S se 2 IO Be ate
2,309 9On 00 Te oe Aa ae lore te 8 eae kale
OOO 4 9 OOO Ot sea a erin 2 Oe DOOT:
CCL One JOR FO OR Ge ee ee Om AE DOO
a8 16-90 00 DOs Gin ea ell Cas good
SOLO 90; OOF OSie 4 iitieeyah iain aiktarird te DOOR
Zope onno OnO0 DOs eh nOn Ae OOF
220279 0 a0 Om ene > Oe eee Oho OOU
00 00 34 32 Se ee AD oa BS ate rae
OO OO) Ha HO) Be yar eben ct oahu lane Eee OOO
NO) ORE BS D7 DS Week TL QO) or tay Ab poor
8 O8 “S83 OS Wy eek Bil 414 5 good
Sie. Abs} ETL © ial Wik Qs Be SW Il 4G} good
PAS} WO ZEB) = B35) Wy AGA Me IO) 8} good
New forms well established.
SSO Ss OOO OR aio, 2h ele OO
81 04 90 00 fea PPh etoyasuinerm haan yay 0) poor
HO? 55), “8X0 OO) ee Oe peeaiinta: Be Ik ary he
Gia) Bh QO 00) A Op uenan cine ze ines ni vel poor
BS UBS Bs} Oa? 2 4 4 good
Hib Wes BOS. BO 55 NOR 2G es BB. tehie
ONL exert ILS} hs De aU NDE a Mae Mie Les dees eee
Uncertain and vicinal forms.
79 47 90 00 STAD Sree ave 4 4 poor
CEEEZORSI OO Onin 27 carl itn ep 2a ©. DOOT
74 34 90 00 DQG arose: as Het Oe POOL
Aloe alal. BQ)" X00) LO i en Senos OOK
NGF Ore Ot NO ane yr uinmeeen ty elles DOOR
BH OR 72 AG Sel OA Ae wei ona, LAlT:
By Dy Ae. BO Fe BES) NOB SS a 8} poor
A) 341827 OO wel nero Lian Oo) - POOL
Abe BX) el) Abi ie Oem m2 Oma wae rant
Bias iit a. WY) SA eG OMe Orie © imei poor
characterized by the strong

development of the parallel zone of pyramids, 0, v, and p and

the dome d in the same zone.

Three narrow line faces of

pyramids in this zone new to the mineral were observed, of
which two, A (212) and B(141), seem to be established by the
observations.

The latter is one of the forms mentioned by

252 C. Palache—Mineralogical Notes.

Dana* as found by him on zoisite from Ducktown, Tenn. but
not supported by measurements and therefore excluded from
recent lists of forms.

The form vy (221) is well developed on several crystals with
good faces and is much the best established of the new forms.

On nearly every crystal the edges between brachypinacoid
and terminal planes are replaced by groups of very steep
pyramidal faces which are sometimes curved and reéntrant,
again plane and smooth, affording good reflections. The meas-
ured angles indicate forms of large and complex indices and
the variation in position is so considerable in different crystals
that they can oniy be regarded as vicinal forms. Some of the
measurements: made upon them are however presented in the
table since they are very characteristic for the locality.

The axial ratio calculated from the measurements of thirteen
good faces on a single crystal coincides almost exactly with the
value calculated by Tschermak for zoisite from Ducktown,
Tenn. 7

a One Cc
Tschermak, Ducktown 0°6196 : 1 : 0.8429
Palache, Chester O-61 Oi ales 0r34219

2. Phenasite as an Alteration Product of DPanalite from
Gloucester, Mass.

Tn 1908 an abundance of danalite in small grains and masses
up to 8" diameter was found in a small ballast quarry opened
in granite on the line of the railway, about half way between
Gloucester and Rockport, Mass. On some of the larger masses
of pale to dark pink danalite can be seen faint indications of
octahedral planes. Most of them are quite irregular and much
fissured, showing the beginnings of decomposition, and in one
case the greater portion of such a mass has undergone com-
plete alteration, the resulting products showing that all the
essential constituents of the danalite were retained in the
cavity.

It will be remembered that danalite is a silicate of beryllium
iron, manganese and zinc, containing sulphur. The decompo-
sition products found in this cavity comprise phenacite, spha-
lerite, pyrite, manganiferous siderite, hematite, quartz, chlorite,
albite and kaolin. The fr agmentary condition of the cavity
when found made it impossible to attempt a quantitative esti-
mate of the proportion of the original chemical constituents
preserved in this aggregate. But it is evident that in a quali-
tative sense phenacite represents the beryllium content of the
danalite, sphalerite and pyrite the sulphur, zinc and part of the

* Dana, J. D., Mineralogy, 1877, 290.

C. Palache— Mineralogical Notes. 253

iron, siderite and hematite the manganese and part of the
iron; quartz may represent silica set free in the reaction.
Chlorite is probably derived from biotite, and albite and kaolin
from feldspar present in the granite.
The novelty of this paragenesis for
phenacite and the fact that very few
occurrences of this mineral are known
in New England seemed to make it
worth while to record the following
facts concerning the mineral contents
of this small cavity.
Phenacite.—Phenacite is in sharply
formed yellowish white crystals, short
prismatic or lense-shaped in habit,
implanted on the quartz of the cavity
wall or loose and wholly imbedded in
siderite. The crystals vary in size
iO oe My oriea testa dt am elt ex,
downward and proved to be suffi-
ciently well developed to permit of
measurement, although the faces are
pitted and somewhat dull. The
following forms were determined :
m(1010), @(1120), r(1011), 2(0111),
p (1123), 0 (4223), and «(1322). The
ficure* (fig. 2) shows the prevailing
habit, which much resembles that of the phenacite from Flo-
rissant, Col., described by Penfield. On many erystals, how-
ever, the prism planes are much reduced or entirely wanting,
giving a lense-shaped appearance. to the crystal. The attach-
ment is generally by a side, so that both terminations are seen.
The prism planes are brilliant and give good reflections; the
rhombohedrons 7 and #, which are the dominant terminal
forms, are also of fairly good reflecting quality. Faces near
the center of the crystal are, however, uneven and appear to
have been etched; the planes of the rhombohedron p are
therefore not usually sharp as in the figure but are replaced by
a rounded or irregularly facetted surface. As no new forms
were observed the measurements are not here recorded.
Siderite.—Siderite is the most abundant mineral in the
cavity. It is transparent with pale brown to pinkish color
when freshly fractured, but natural surfaces show an iridescent
or steely tarnish. The larger anhedral grains that border the
cavity reach a diameter up to 3™ and show a brilliant cleav-
age. Most of the siderite isin the form of cellular aggregates
of distinct but rounded erystals of two generations, the older

* The figure was drawn by Mr. R. W. Richards.

254 C. Palache—Mineralogical Notes.

being larger, dark colored and deeply corroded, the later of
a pale pink color, small and with some of their planes still
brilliant. It was " possible to measure these, and the forms
determined, as shown_in the figure (fig. 3) were ¢(0001),
3 r (1011), s (0551), and v(2131). The figure
gives about the proportions of many of
these later erystals, which, being generally
attached by their sides to the older erys-
tals, were doubly terminated. Other erystals
show a broader base and greater relative
development of the form v, giving an equi-
dimensional habit, which is also, so far as
could be judged, the habit of the older
generation of crystals. In view of Schaller’s*
study of siderite in which he questions the
accuracy of the accepted axial ratio, sey-
eral cleavage rhombohedrons which gave
brilliant and single images on the goniometer
were measured. The angle measured, 1011
to 1101 (average of six), was exactly (ioe
which agrees with the accepted value. It is of course recog-
nized that the considerable manganese content of: this siderite
vitiates the comparison with Schaller’s measurements made on
pure siderite, but the fundamental angles of siderite and rho-
dochrosite are identical according to Dana and replacement of
iron by manganese would ther efore affect the angle but little.

Sphatler ite. —Sphalerite is sparingly present as light yellow
transparent grain and imperfect crystals too fragmentary to be
measured satistactorily.

Pyrite.—Pyrite is present only in minute amounts in erys-
tals showing cube and octahedron faces, implanted on siderite.
These crystals, which are exceedingly minute and somewhat
dull, present under the micr oscope an appearance quite unfamil-
jar for this mineral. The cube faces of each erystal are
divided into four equal areas by grooves running from a slight
prominence in the center of each face to the middle of each
edge. The appearance suggests interpenetration twinning, but
this could not be established by measurement.

Quartz.— Quartz crystals of two types are found in the
cavity; (1) simple combinations, m, 7, 2, with dull faces,
attached to the quartz of the cavity walls; (2) fragments or
complete crystals imbedded in siderite. The second type is
glassy and contains chlorite inclusions. The crystals are pris-
matic in habit, often much distorted and highly complex in
development, as shown by the following forms found on the
two measured crystals:

* Siderite and Barite from Maryland, this Journal, xxi, 364, 1906.

C. Palache—Mineralogical Notes. 255

CR ee ae : re et WE
(GOI) 1 (23°0°23-7) 28R Anes 2P2r
alae ae
oom) me GG aE | PG} aE

The forms ¢, &, and &,, which are developed on nearly every
erystal examined, seem to be characteristic for the locality.
The positive rhombohedrons are variable in different vertical
zones on the same crystal but have sharply marked faces which
give good reflections. The prism planes are almost wholly
replaced by the steeper rhombohedral forms.

Albite——Wherever the feldspar of the enclosing granite
reaches the wall of the cavity it is coated with a parallel group
of water-clear albite erystals. The erystals are all albite twins,
consisting in nearly all cases of a single pair of individuals, and
the forms present were determined by measurement to be as
follows :

e (001), 6 (010), m (110), M (110), (180), z2 (130), x (101),
7m (021), 0) (111); and 6 (112).

Hemutite.—Hematite is seen occasionally as lustrous scales
or spangles but more commonly as a red pigment staining the
surface of quartz crystals or mingled with the kaolin that
coats all the minerals in the cavity.

Chlorite.—Green crystals of chlorite of tabular habit, rudely
hexagonal in outline and with rounded edges, are implanted on
quartz and feldspar. .

Kaolin.—Kaolin forms a thin coating on the various miner-
als described above. It is finely scaly under the microscope.
The kaolin is loosely adherent and when removed leaves the
underlying crystals bright and fresh.

3. Crystal Form of Chalmersite.

The original description of chalmersite by Hussak* gave
somewhat meager data concerning the crystallographic develop-
ment of the mineral. The following note results from the
study of five well-developed crystals taken from a specimen
belonging to the Harvard Mineral Collection from the original
locality, the “ Morro Velho” mine, Brazil.

The crystals studied and all those visible on the specimen
from which they were detached are untwinned, unlike Hussak’s
material which is described as commonly in twin or multiple
crystals. The crystals are slender prisms, the largest, measur-

* Hussak, E., Ueber Chalmersit, etc., Centralb. fiir Mineral, 1902, 69.

256 C. Palache—Mineralogical Notes.

ing about 2°™ in length and 0-5™™" diameter, being doubly
terminated, while the smaller ones show but a single termina-
tion. Except for a coarse striation on the brachypinacoid due
to oscillatory combination of that form with a prism, the erys-
tal faces are plane and brilliant and despite small size give
excellent measurements.

The following forms were observed :

c (001), 6 (010), @ (100), m (110), 2 (130), 7 (012), g (O11),
d (021), y (108), p (111), 0 (286), 7 (233), s (263), ¢% (136),
u (19°12).

In the prism zone the forms 6 and 7 are
dominant, often giving a slightly tabular
habit to the prisms. They are striated and
less brilliant than the narrow faces of m,
which are always present The pinacoid a,
present on all but one crystal, gave rather
poor reflections. Of the terminal forms
the brachydome g and the unit pyramid
p are dominant; the base c, while always
present, is broad on but one crystal. The
domes f and y and pyramids o and 7,
while found on several crystals as tiny
facets, are quite subordinate to the fore-
going forms, and the three forms d, ¢ and
uw were each seen but once on the same
erystal.

i

The observed combinations are as follows:

1. ¢, 6, a, m, 1, f, g, p, s. Doubly terminated. Figure 4
(drawn by J. B. Marvin, Jr.).

2. ¢, 6b, m, l, g, p, 8. Commonest type on specimen as a whole.
Figure 5.

3. ¢, b, a, m, 1, f, g, Y, P, 0, 7 s Two measured crystals.
Figure 6.

4. ¢; 0, a, my Ua, G5 dy Y; Py 0, 7,18, it, Uv. Kicurem audrey
in about the natural development.

C. Palache—Mineralogical Notes. Orn

The somewhat complex symbols of the pyramids are sup-
ported by concordant measurements and by the simple and
normal zonal relations existing between them. The choice of
the pyramid 7 as unit would simplify the pyramid symbols
slightly, but there would then be no unit prism and the relation
in form to chalcocite would be obscured.

The table which follows presents the calculated and measured
angles with the range of variation (in minutes) of each form.
The measurements were made on the two-circle goniometer.

Calculated Measured Limits (in minutes) No. of
Sym- SSS OSS =) = = — obser-
bol 0) p @ p @ p vations Quality
COCs 0000, eens 00" 00! abate agers 5 good
6 010 00°00’ 90 00 00°00’ 90 00 0'to + 7 — ‘8 fair
a@ 100 90 00 90 00 89 57 9000 -27to + 7 = 8 fair
m 110 COMUZE 90 00R CO 2 SOOO Roy toro — 10 good
~ 130 SOR 2 9000s SO mice 2000 Fen —9 tor ind — 6 poor
hi Oke 00 00 25 48° 00 038 25 36 Oto +12 -23'to +14’ 3 fair
g 011 0000 48 56° 00 00 43 55: Oto+ 7 -2to+ 3 7 good
d 021 00 00 62 34° 00 11 62 40 = — > Teepe
Y 103 OO OO 2 ig WOO, 2 7 Oto +12 -5to+ 5 6 good
pastel 60 12 62 43 60 10° 62 43 —5 to + 6 -8to + 3 12 verygood
0 236 49 20-36 29 49 19 36 33: —11.to.+12 -4to + 7 7 fair
ry 233 49 20 55 56 49 24 5557 -7to+ 7 -6to+ 8 6 fair
S 2638 30 12 65 51 30 11° 65 51° -8to+ 9 -9to+ 7 9 good
t 136 SOM Z ee ZO 0 S8e e299 20) 22:9) s12 —_— — 1 fair.
2 NON 2a Oe 9) 636222 lk? 33630 = = 1 good

The axial ratio was calculated from the measurements of 50
best faces of 12 forms on 5 crystals. The average values of
the closely accordant results differ but slightly from the ratio
determined by Hussak.

>|
Qu

Opars2 ama
(Hom BE 87 Te 2

Qe

: 0°9637 (Palache).
0°9649 (Hussak).

The angles calculated from this ratio in the form of the
Winkeltabellen of Goldschmidt are given in the table of
measurements above.

As pointed out by Hussak, chalmersite stands in close
relationship to the chalcocite group of minerals both in compo-
sition and form. If we also include in this group the mineral
pyrrhotite, recently shown by the magnetic studies of Kaiser*
to be orthorhombic and pseudohexagonal through twinning, and
if we calculate its ratio from the angle for the pyramid 2021,
using Seligman’s value as the most reliable observation yet

* Kaiser, E., Die Krystallform des Magnetkies, Centralb. fiir Mineral,
1906, 261.

258 C. Palache—Mineralogical Notes.

obtained, the relations of the group appear as in the following
table:

c

ct b
Chalcocite Cu,S O°5822 = 1s0-S770m
Stromeyerite (Ag,Cn),S 0°5822 : 1: 0°9668
Sternbergite Ag,S:Fe,S, 0°5832 : 1 : 0°8391
Chalmersite Cu,S'Fe 8, 0°5725 : 1: 0°9637
Pyrrhotite He,S,4, O:sc60 = 1: 079524

The propriety of placing pyrrhotite in the Chalcocite Group,
long ago suggested by Streng and others, and confirmed by the
discovery of the intermediate chalmersite, seems no longer open
to question with the established orthorhombie nature of pyr-
rhotite. The discordant value for the c-axis of sternbergite
suggests the need of a revision of the crystals of that species.

Harvard University, April, 1907.

Hillebrand and Schaller—Mercury Minerals, ete. 259

Art. XX VI.—The Mercury Minerals from Terlingua, Texas;
Kleinite, Terlinguacte, Eglestonite, Montroydite, C wlomel,
Mercury ;* by W. F. Hittepranp and W. T. SCHALLER.

In the late fall of 1905 the senior author received for identifi-
eation from Mr. H. W. Turner, at that time connected with
one of the mining companies of Terlineua, Texas, specimens of
two minerals from the well-known Terlingua mercury field in
Brewster County. One of these proved to be the unidentified
mineral referred to as No. 5 by Professor A. J. Moses in his
papert on new mercury minerals from that district, namely
terlinguaite, eglestonite and montroydite, the last of these
being mercuric oxide, the others oxychlorides. Preliminary
tests having shown that No. 5 belonged to the so-called mer-
cury-ammonium compounds, hitherto unknown in nature, a
brief announcement of this fact was made in order to secure the
field for as full an investigation of this unique mineral and
its associates as the material on hand and to be obtained might
permit.

This work has been conducted at intervals during the past
18 months and is yet incomplete with ce to the new
mineral. It has, however, extended over soNong a time and
the chances for obtaining more perfect material than that
already available are so slight, that it is deemed inadvisable to
longer delay publication of the results obtained. The full
details of the work herein summarized will be found in a bul-
letin of the U. S. Geological Survey, the appearance of which
will unfortunately be delayed still longer, chiefly on account of
the plates that are to illustrate it and the unavoidable delays
attending publication. Although the present condensation
reproduces the essential points as to the chemistry of the min-
erals, it but touches their crystallographical. side, which,
though of much interest for the great number of forms shown
by most of the minerals, requires too extended treatment for
a resumé of this character. Further, many observations of
interest that can not be detailed here were made upon which
some of the conclusions were based, particularly in studying
kleinite, the mercury-ammonium com pound. For these, as well
as the details of crystallography and association, reference must
be made to the full report.

A few words, however, with reference to their association
as observed by us are necessary in this place. The minerals

* Condensed from a forthcoming bulletin of the U. 8. Geological Survey,
chemistry by Hillebrand, crystallography, etc., by Schaller.

+ This Journal [4], xvi, 253 (1903).

¢{ Science, xxii, 844 (1905); J. Am., Chem. Soc., xxviii, 122 (1906) ; this
Journal [4], xxi, 85 (1906).

260 Hillebvand and Schaller—Mercury

are deposited on a matrix of two kinds, first a soft siliceous-
aluminous, earthy mass, sometimes gray but usually of a pink-
ish color and containing a small amount of calcium carbonate,

and second, a fairly pure layer of calcite with large scalenohe-
dral erystals projecting from the surface. The general
associations of the several minerals are given below, but there
are many exceptions that will be noted in the full report.
Kleinite is found with gypsum, calcite, seldom with barite and
calomel, either loose or on a whitish clayey gangue, only once
or twice accompanied by terlinguaite ; calomel with calcite,
mercury and eglestonite on the pinkish earthy gangue; egles-
tonite with calomel, calcite and mercury on the pink ¢ gvangue or
on calomel ; montroydite with calcite, terlinguaite, ‘and mer-
cury on the calcite layer ; terlinouaite with calcite, montroy-

dite and mercury on the calcite layer.

Several members of this group of minerals are characterized
by a most unusual property, namely, proneness to change color
rapidly on exposure to light. With respect to terlinguaite
and eglestonite this change is of a permanent character and the
result is to impart to the minerals an appearance often quite
different from that they originally possessed. With kleinite
the change is not to a different color but only to a different
shade, and it persists only as long as the exposure itself, the
original color returning in the dark. From published and
privately communicated statements it would seem as if these
minerals, in their earthy forms at least, must be difficult to
distinguish as a rule when first found, by reason of the simi-
larity “of their original colors, all more or less pronowuncedly
yellow.

In addition to the specimens first received from Mr. Turner
many fine ones were donated by Mr. J. H. Hartley, who was
also connected with one of the Terlingua mining companies,
and later Mr. R. M. Wilke, of Berkeley, California, gave kleinite
when more was needed. All was, so far as known to us, from
the properties of the Marfa and Mariposa Mining Co., and
chietly from the Terceiro shaft. Professor A. J. Moses kindly
identified the new mineral with his No. 5 and sent us his orig-
inal measurements of the latter. To these gentlemen, espe-
ally to Mr. Hartley for his most generous liberality, also to Dr.
P.G. Nutting of the Standards ‘Bureau and Prof, B. B. Bolt-
wood of iNew: Haven, who kindly made certain tests, we take
occasion to express our deep sense of obligation.

In the several descriptions that follow we have MOE ERLE E
data already correctly given by Professors Moses and Sachs,*
as well as the new matter gathered by ourselves, in order to
present as complete a record as possible of the minerals

* Sitzb. K. Preuss. Akad. Wiss., 1905, 1091-1094.

Minerals from Terlingua, Texas. 261

described. In general due recognition is given of observa-
tions made by others, though it may not have been possible in
every case.

Kleinite.

Before describing the unique compound referred to in the
foregoing it will be necessary to impose on the reader a little
history. On the day preceding the appearance in Science of
the announcement regarding the new mercury-ammonium
compound there was read at a meeting of the Royal Prussian
Academy of Sciences a paper by Professor A. Sachs, of Bres-
lau, descriptive of an oxychloride of mereury which was
regarded by him as perhaps identical with the No. 5 of Pro-
fessor Moses, and to which he assigned the formula Hg,Cl,O,,
or 3Hg0O. HeCl,, and the name kleinite, after the eminent
mineralogist “Prof. Carl Klein. This paper appeared in print
on Jan. 11, 1906. After reading the announcement of the
mercury-ammonium mineral, said ‘to be also identical with the,
above No. 5 of Professor Moses, Professor Sachs made new
~analyses of material in his possession and obtained results*
agreeing qualitatively in each case and quantitatively in one
with those which had already been obtained in Washington.
His later data appear in the table below:

Sulphur-yellow crystals Orange crystals
c= a =)
Hg at suis 85°29
Cl ate Pe 6°97
SO, 1:05 0°85 2°57
NH; 0°44 1:09 ifs)

He regarded the sulphur-yellow crystals as the purer, and
from the varying values for sulphur and nitrogen ar gued that
these could not be integral components of the mineral. _With-
out committing himself to any view ‘as to the molecular
arrangement, the cen formula was suggested as the most
plausible—He, (CL4ASO,), [O, (NH,),],—which is of the same
type as his original oxychloride, He CLO. duty isi tar trom
being in close agreement with his analytical data and is also
opposed to the chemical behavior of thé mineral as a mer cury-
ammonium compound. The assumption by Professor Sachs
of hydrogen and of oxygen other than that in the SO, radical,
was purely arbitrary, and it may be said here that his surmises
as to the formula of the mineral have not been verified by the
results of our work, nor have we found any certain difference
in composition between the light yellow and the orange crystals.

*Centralbl. Min. Geol. Pal., 1906, 200-202.

262 HTillebrand and Schaller—Mercury

It had been the intention of the senior author to assign to
the mineral a name indicative in some manner of its composi-
tion when this should have been fully established, but to now
substitute for the name kleinite, already in the literature even
though applied to an incorrectly identified species, another
name, no matter how appropriate, would occasion confusion
that it is desirable to avoid; hence the name kleinite is
accepted without reserve.

Physical properties.—Most of the material as received was
in loose crystals or crystal aggregates, to which adhered more
or less of a dull earthy white to reddish foreign matter of a
clay-like character. While many of the crystals were very fine
and brilliant, much of he material was far too impure for the
chemical tests that were contemplated. Even the selected
erystals and aggregates held here and there a little of the
gangue firmly ‘attached and some must also be included in the
cr ystals themselves in an extraordinarily fine state of division,
to judge both from microscopie evidence and from the amount
of non-volatile matter that was left on ignition, which ranged
from about 0°75 to nearly 3 per cent. This gangue interfered
much with the correct determination of the water given off by
the mercury mineral on heating and hence with the establish-
ment of a formula.

A peculiarity not noticed by other observers is that when
exposed to sunlight or even to the diffused light of a room,
after having been in the dark, the crystals, at first canary-) ak
low, almost immediately became much deeper in color,
generally reddish yellow or orange, but that they regained
their original color very soon when again placed.in the dark.
This phenomenon could be reproduced as often as desired. It
was also noticed that the exposed crystals were not all of the
same shade of yellow; there were some that had not changed
at all and others that showed all gradations between almost
colorless and orange, and one crystal was seen with an orange
core and a light outer zone. Professor Sachs also noticed dif-
ferent shades in the same crystal. In powder the color is
sulphur-yellow. One of the first specimens received was com-
posed entirely of very pale yellow, coherent crystal masses
held tog ether by or holding tog ether a reddish ear thy gangue.
The lighter crystals that were picked out for separate analysis
from samples of the loose crystals varied in color, but all were
much lighter than the rest in daylight.

The density was determined on several specimens and found
to average for the orange crystals 7-975 and for the light yel-
low 7:987, but the results are all low because of the attached
or included earthy matter. These figures are much higher
than the 7-441 given by Sachs. The er rystal form is hexagonal ;

Minerals from Terlingua, Texas. 263

c=1-6642 (mean of Schaller’s and Sachs’ values). The habit is
short prismatic, rarely equidimensional. Single crystals seldom
exceed one millimeter in length, but masses: of crystalline
material may exceed one centimeter. Five forms have been
Cae namely, ¢ {0001}, m {1010} O1LUAY p {1011}, and

{1012} (mew). Cleavage i is good geaMel to {0001 { and imper-
fect parallel to {1010}. “Brittle. Luster a ennoenatene ae greasy
on bright surfaces. Hardness apparently slightly over 3 Not
radio-active (Boltwood), but Nutting reported faint ce
of helium on first warming the mineral in a vacuum.

2 TO

The mineral being geometrically hexagonal, a basal section
should remain dark under crossed nicols. But, as described
by Moses, such a section does not remain dark but shows
double refraction, and if thin enough will be seen to be composed
of innumerable individuals, none of which is large enough to
show interference figures. The double refraction is str ong, the
colors being of the third and higher orders. At about 130°
the double refraction begins to decrease, as seen by the
descending colors, until finally it becomes zero and the min-
eral remains dark under crossed nicols. The section now
gives a uniaxial positive interference figure. After cooling,
the section remains dark but after the lapse of many months
is seen to be slowly returning to its doubly refracting condi-
tion. This phenomenon seems to show that kleinite is dimor-
phous and that the uniaxial optical state agreeing with its
outward hexagonal form is stable only above 130° approxi-
mately, while below that temperature its stable condition is
biaxial, probably triclinic. According to this the hexagonal
crystals of kleinite must have been formed at a temperature
not much if any below 180°. As is stated just below, it is at
a point but a few degrees higher than this that the first per-
manent browning of the mineral become visible when it is
heated, and that considerable loss of water has then taken
place. What connection, if any, there may be between these
two phenomena is not known.

Pyrognostic behavior.—When carefully heated in a closed
tube, or better in one through which passes a slow current of

264 LHillebrand and Schaller—Mercury

air, the mineral loses a little water, begins to brown at 135°
150°, and as the temperature rises becomes still darker and
yields more water, but gives no visible sublimate under 260°
even after several hours. Between 260° and 280° mercury
and calomel (not HgCl,) sublime. When most of the calomel
has passed off the residue begins to grow lighter colored, then
yellowish and _ finally becomes nearly white. During the
expulsion of the calomel some gas (Cl?) is evolved in minute
amount that sets free iodine from solution of potassium iodide.
On heating to 400° more of this active gas is evolved, but it is
soon followed or accompanied by one that destroys ‘the color
of the free iodine (SO,?). At 400°-420° appears a further
sublimate less volatile than the calomel. If the test is made
in a closed assay tube this last sublimate and the still unvola-
tilized residue may melt to a dark reddish liquid, which on
cooling solidifies with a yellowish and then white color.
Ammonia turns both sublimates black instantaneously, although
the later one often contains mercuric as well as mercurous_sul-
phate. Most of the nitrogen escapes in the elemental state dur-
ing the formation of the calomel, but not quite all. There is
not the least evidence of the formation of ammonia. If the
heating is carried out in vacuo the evolution of the active gas
is much more marked than at atmospheric pressure and if col-
lected by a pump is seen to be of the color of chlorine. Under
the vacuum this gas does not act on the mereury of the pump,
but fouls that in the collecting tube strongly under atmos-
pheric pressure. The scum on the mercury gives tests for
chlorine. This liberation of free chlorine was at first supposed
to indicate direct union of nitrogen and chlorine in the min-
eral, but since it is also given oft on heating in vacuo a mix-
ture of the artificial compound NHg,Cl.vH,O with a sulphate
of mercury (3HgO. SO, in the test) it is evidently a secondary
reaction between one of the products of breaking up of the radical
SO, and vapor of calomel or of the still undecomposed chlo-
rine constituent of the mineral.

Behavior toward liquid reagents.—Soluble in warm hydro-
chloric as well as nitric acid without deposition of calomel.
Also soluble in sodium sulphide and in ammonium bromide.
The latter liberates as ammonia for every one part of nitro-
gen derived from the mineral itself three parts from the
reagent. Fixed alkalies do not liberate ammonia. Hydrogen
sulphide blackens speedily, ammonia not at all.

Additional data.—Several tests were made by decomposing
the mineral in vacuo under varying conditions of treatment
in order to get evidence as to the presence of either hydrogen
in addition to that which was afforded as water or of oxygen
other than that in the water and the SO, radical. For the

Minerals from Terlingua, Texas. 265

evidence on these points reference must be made to the full
report; suffice it to say here that no certain evidence was
obtained in favor of the presence of one or the other of these
elements other than in the combinations above named, though
the proof is not absolute that there may not be small ‘amounts
of one or both (see also p. 268). If present, the hydrogen is in
such small amount as to be unimportant in the formula of the
mineral as a homogeneous unit, and oxygen must also exist in
all probability as basic oxygen in a minor compile of.a mix-
ture. The summation given on p. 267 is a strong argument
against the presence of considerable percentages of either.

When, however, the mineral is fully decomposed by heat in
presence ‘of lime or sodium carbonate, oxygen is liberated in
quantity. Theoretically the amount should be exactly equiv-
alent to the SO, and Cl, found in a particular sample if the
compound is normal and not basic, and this should afford an
exact means for deciding the question of the presence of oxygen.

“But asa matter of fact the oxygen never did equal the eal-
culated amount by several tenths of one per cent. Of the
various possible explanations to account for the deficiency the
following, based partly on qualitative tests, seems the most prob-
able. The oxygen is liberated partly in an active state and
forms a chloroxy-salt of sodium or else nitrite or nitrate of
sodium in small amount. It would require but little of one of
these salts to bind enough oxygen to account for the observed
deficiency. The evidence favors an oxy-salt of nitrogen in
preference to one of chlorine. It is, of course, assumed that
any such salt was formed by the act of decomposition and did
not preéxist in the mineral.

Analytical methods employed and the results.—( 1) Nitro-
gen. For nitrogen three methods were used: (a) expulsion
as ammonia by sodium sulphide and gravimetric determination
as the chlorplatinic salt; (6) expulsion as ammonia by ammo-
nium bromide in a closed vessel, collection of the ammonia in
titrated oxalic acid and determination of the acid left over;
(c) direct determination as nitrogen gas expelled in vacuo, col-
lected by the aid of a Topler pump, and measured in a gas
burette after freeing from other gases if present.

Numerous data are to be found in the full report relating to
these methods, particularly the last. The second method wa
found to yield about two-tenths per cent more nitrogen than ie
first or than the second when sodium carbonate was used as a
retainer for the chlorine and sulphur, the results by catch
agreed well.

This was probably in at if not altogether due to the
action of the clayey gangue on the ammonium bromide with
liberation of some ammonia (see p. 268). Were it not for

Am. Jour. Sct.—FourtH Seis: Vou. XXIV, No. 141.—SEPTEMBER, 1907.
18

266 Hillebrand and Schaller—Mercury

this effect of the gangue the method would be an excellent
one for obtaining proof as to the presence or absence of basic
oxygen. The nitrogen found by the third method was finally
tested by passing it over hot magnesium, which absorbed
apparently the whole of it. It was also tested spectroscopically.
The analytical results were as follows, calculated to gangue-free
substance, the modes of decomposition employed being indicated
at the heads of the several columns :

NITROGEN PERCENTAGES BY DIFFERENT METHODS.
Gas-volumetric

— _—_———
Naes HCl With Na.CO; Without Naz COs. NH,Br
PP 5\0) 2°37 DOS Oil 2°78
Deon 2299 2°74 2°76
DOs) 2°60 2°67 2°74
2°59 2°58 2°86 BoTKY)
2°43
Av. 2°555 2 a eee
2°54

The single determination after solution of the mineral in
hydrochloric acid and removal of the mercury as sulphide and
of the SO, as the barium salt, is regarded with confidence, as
also those by the sodium sulphide method. The greater vari-
ation among the results by the gas-volumetric tests with sodium
carbonate is to be ascribed to the small amounts of mineral
used—0°25-0°5 gram, the uncertainty in the burette readings
in the upper section of the instrument and the greater chance
for loss or gain during the numerous manipulations. The high
results by the ammonium bromide method have been explained
above. Those by the gas method without sodium carbonate
are not clearly accounted for, but the results obtained in that
way were always less satisfactory by reason of the fouling of
the pump outlet and of the mercury in the collecting tube by
the free chlorine that was formed.

(2) Mercury. Mercury was determined in several ways
almost always in connection with one of the nitrogen deter
minations: (@) As mercury by ignition with sodium carbonate
and once as in organic combustion of mercury compounds.
Most of the determinations were by this method. (6) As mer-
eury by electrolysis from sodium-sulphide solution. (¢) As
the sulphide. The last method usually afforded higher and
probably truer resuits than the first.

(3) Chlorine and sulphur. Since slight loss of chlorine and of
sulphur as sulphide dioxide almost always resulted when the
mineral was heated by itself, these components were determined
in the sodium carbonate employed for the gas-volumetric deter-
mination of nitrogen. A greater variation among the few

Minerals from Terlingua, Texas. 267

chlorine results for the lghter colored crystals was observed
than for the orange-colored ones, which latter afforded excel-
lently agreeing results. But the results for the lighter crystals
were in no case so markedly different from the others as to con-
firm Professor Sachs’ conclusion that there was an appreciable
chemical difference between the crystals of different shades,
and they were in part affected by obvious errors.

(4) Water. None of the water afforded by the mineral is
hygroscopic. About one-half comes off at 135°-150° and the
total that is obtained by careful heating of the mineral by itself
to complete decomposition, using a plug of gold leaf in the exit
of the tube, is not clearly in excess of that obtained after the
manner of organic combustion with copper oxide preceded by
lead chromate and a roll of copper. Some of it comes from the
clayey gangue, but most is beyond question given off by the
mercury mineral. How much belongs to one and how much
to the other it has been impossible to ascertain. The water
determinations constitute the least satisfactory portion of the
analytical results.

In vate different specimens analyzed the gangue ran from
Oonto 3 oe cent in the dehydrated state, as obtained by igni-
tion. In the full report the many analyses are given both as
made and after recalculation to a gangue-free basis for both
orange and light-colored crystals. We reproduce here only the
averaged results for the deeper colored crystals since they were
in best agreement, repeating that although the light-colored
erystals afforded oveater variations in chlorine and sulphur than
the orange ones, this was in part due to obvious errors, and
that the means ron these constituents were slightly higher than
for the orange crystals, rather than far lower as found by Sachs.
To include them ‘would hardly affect the general average.

AVERAGE COMPOSITION OF ORANGE CRYSTALS OF KLEINITE.

elie: SU ek S06 = 200 == 04S = a

Olgas a (SO 2) BMG = He ee

SOees ak yams Sl 596-06 us 00616 mm ai

IS lee ee Os7 = Te = O18. = Tl

EEO cso ee Ie G02 = OST) = One
99°86

In discussing the above ratios it must be borne in mind that
the number for water is of doubtful value, so that it can receive
little attention. The water can not exist in large part even as
hydroxyl, for that would require an amount of basic oxygen
entirely opposed to all the evidence, in which that of the sum-
mation is not of least import. The ratio shows at once that
there is far too little nitrogen for a general formula of the type

268 Hillebrand and Schaller— Mercury

NHg,X.vH,O, in which X represents Cl and SO,. It is even
insufficient for the chlorine alone in such a formula and we are
therefore obliged to consider the possibility of the body being
a mixture. Calomel as a constituent of such a mixture is ex-
cluded from consideration for the reason that but a trace is
indicated on dissolving the mineral in warm hydrochlorie acid
and this appears to be derived from the gangue. Mercuric-
chloride would seem to be excluded by its solubility in water.
As oxychlorides there would naturally be considered first ter-
linguaite and eglestonite, but both are excluded for the reason
that they yield calomel on solution in hydrochloric acid. It
remains to assume a mercuric oxychloride and then an oxysul-
phate, both of which might perhaps be formed simultaneously
with NHg,Cl.zH,O from a mercuric solution containing more
mercury than is needed by the ammonia present to form a com-
pound of this type. If, perchance, the oxychloride were
Hg,Cl,O, or HgO.HgCl,, and allowance were made forit on this
basis in the ratio above given, the ratio of the residuals would
show an oxysulphate with He to SO, as 4 to 3:23 and compo-
sition nearly Hg,O(SO,),, or HgO. 3HeSO,. The calculated
amounts of these salts would demand 0°33. per cent of basic
oxygen, an amount that happens to coincide almost exactly
with that indicated by the nitrogen determinations according
to the ammonium bromide method (p. 265), which as will be
remembered gave more than was obtained by other methods.
While the excess of ammonia obtained by the ammonium bro-
mide method is with considerable reason believed to be in part
due to the action of the aluminous gangue on the reagent, cal-
culation shows that it can not well have all originated thus.
If not, the only alternative, assuming that the tests were rea-
sonably correct, is that an oxy-salt of mereury must have given
rise to some at least of that excess of ammonia. The chief ob-
jection to accepting this alternative is the yolume of oxygen cor-
responding to the required weight percentage, which if wholly
given off as oxygen gas on heating the mineral itself ought not
to have escaped detection and. approximate measurement.
Still, it is conceivable that in this method of operating, the basic
oxygen might not all escape as gas. The evidence as to the
presence of a little basic oxygen is ‘conflicting, and further spec-
ulation would be profitless in the present state of our knowledge
regarding this remarkable mineral. To assume that the nitro-
gen is very low and should be 2°885 per cent or exactly equiv-
alent to the chlorine, is opposed by much evidence, including
the fact that the sulphate would then be strongly acid.

It is regrettable that the long labor has resulted in nothing
more definite than the fixing of this mineral as the first nat-
urally occurring member of the so-called mercury-ammonium

Minerals from Terlingua, Texas. 269

compounds and the untenableness of the possibility suggested
by Professor Sachs. The question as to the structure of these
mercury-ammonium bodies, whether they belong to one or
other of the several types that have been suggested for them,
is outside the scope of this investigation.

Montroydite.

Orthorhombic-holohedral; @ : 6 : c=0:6375 : 1 : 11977
(Schaller). Fifty-six forms observed, 45 new. Two erystal
habits with all intermediate oradations : (1) Prismatic, flexible,
dark red needles, commonly 150", occasionally 2$™ long, by less
than 1™™ thick, often partially “erayish from what appears to be
a thin coating ot some other (presumably mercury ) mineral, or
minute and or ange in several forms, notably wormlike and (2)
nearly equidimensional erystals of a few millimeters diameter.
The larger needles occur also in curiously twisted and curved
shapes, the minute or ange ones In irregular rounded and loosely
coherent masses. There are also hollow, irregularly shaped
and bubble-like formations that in their interior resemble geodes,
being lined with or nearly filled with one or both of the above
mentioned types of crystals. A somewhat different form of
bubble is found between large calcite crystals, smooth and
somewhat glistening exteriorly, gray-black and partially filled
with a spongy mass of crystalline material that is commonly
very dark in color. Precise descriptions of these and other
modes of occurrence are difficult to give in few words.

Color, dark red to yellow-brown or orange-brown. Streak
yellow-brown. Transparent to translucent. Cleavage, perfect
{010;. Hardness, 2-3, less than 2 (Moses). Brittle, also sec-
tile, but the long needles, extremely flexible, can be rolled
around a thin réd. Density not determined because of inability
to separate completely from free mercury enough for a satis-
factory test.

Completely volatile without fusing, yielding in a closed tube
a sublimate of mercury only. Slowly blackened by hydrogen
sulphide, but not equally over all surfaces.

Since the oxygen as given for montroydite by Professor
Moses was assumed by difference, a direct determination was
made by dissociating the mineral in vacuo, collecting, measur-
ing and testing the gas evolved. The result was to confirm the
identification of Professor Moses :

ANALYSIS OF MONTROYDITE.

Theory Found in
HgO 0°2215 g

os 92259 92°74 weighed as metal
Oe ial 7°49 calculated from the volume.

100°00 = 160°23

270 Lillebvand and Schaller—Mereury
Terlinguaite.
Monoclinic prismatic (holohedrai), @ : 6 : c=1:6050 : 1 :

9:0245 (Schaller), B=74° 23’. Of the 133 3 forms observed 102
arenew. Orystals often extended in one direction and also equi-
dimensional. The largest crystal measured 16x44", though
faces over a centimeter broad are sometimes to be seen. Also
occurs in powdery form impregnating the earthy gangue, to
judge from the greenish color of some specimens of ore, and
perhaps in a similar state admixed with eglestonite, in which
case its identification is at least difficult. Much confusion
seems to exist as to the original color of terlinguaite before it has
been exposed to light. Moses writes of it as “ sulphur-yellow
with a slight ereenish tinge, very slowly darkening on exposure

to an olive green,” but ‘Mr. Hartley, in reply to our inquiry,
wrote that the terlinewaite crystals were green ere they were
touched by the sunlight, but that most of the terlinguaite occurs
as a yellow powder changing to green. Sometimes brown
crystals are seen and occasionally the green and brown colors

appear in the same crystal. When brown they are difficult to
distinguish at sight from eglestonite in one of its. transitional
color stages. Some of our earthy specimens that, were yellow
at first turned greenish on exposure and presumably contained
terlinguaite. Beautiful spots of emerald-green reflected light
appear when the crystals are examined with a lens as they
occur on the specimen. If at times originally yellow the min-
eral is not in that state always distinguishable from kleinite,
and perhaps not from eglestonite or even from the orange- red
form of montroydite.

Streak yellow, turning greenish gray in light. Transparent
or nearly so. Luster, | rilliant adamantine. “Cleavage, pertect
‘101%. Brittle to subsectile. Hardness 2-3. Density, 8°725
(Moses).

The effect of heating crystals in a closed tube differs some-
_ what according as this is done slowly or quickly. When
quickly done there is violent decrepitation, continuing till the
mineral has volatilized, the substance turning red- brown or
almost vermilion in color (orange-yellow cold) and much of
the resulting powder being projected up onto the sublimate of
calomel and mereury above the assay. Eventually there is
complete volatilization. With slow heat decrepitation is hardly
noticeable. With the first burst of calomel sublimate there
appears a little mercury, but then only calomel so long as there
is any chlorine left in the residue. Sometimes at the last,
when the flame is removed, brilliant short red needles of mer-
curic oxide form on the warm glass by recombination of some
of the mercury vapor and oxygen.

Minerals from Terlingua, Texas. 21.

Tn vacuo the color changes of the crystals as the heat increases
are more marked, these being, after first appearance of a sub-
limate, red, black (without loss of luster), red-brown, orange-
brown and dull. Before becoming completely orange-brown
some faces are olive-green. When orange-brown the only
visible sublimate is calomel and no trace of oxygen has been
evolved. The residue then seems to be mercuric oxide, upon
the decomposition of which partial recombination of its con-
stituents occurs, to judge from the deposition on the warm
glass near by of a slight orange-brown sublimate.

Hydrogen sulphide blackens the edges of a crystal, but fur-
ther action is very slow ; ammonia blackens only after some time.
Thesecond test serves to distinguish the mineral from eglestonite,
which is at once blackened by ammonia. This difference in
behavior of terlinguaite and eglestonite is in line with their
chemical difference, the former being mercuric-mercurous, the
latter wholly mercurous. Hydrochloric and nitric acids decom-
pose terlinguaite with separation of calomel. The hydrochloric
filtrate yields much bivalent mercury. Cold acetic acid slowly
decomposes the mineral when in powder, also with separation
of calomel, and in the filtrate hydrochloric acid produces no fur-
ther precipitate, or but a very faint one. LEeglestonite under
similar treatment yields a heavy calomel precipitate, the filtrate
trom which is free from mercury.

Moses’ empirical formula was confirmed by analyses in
which the oxygen was measured directly and found to be wholly
absorbed by ‘phosphorus, thus showing its freedom from
nitrogen.

ANALYSES OF TERLINGUAITE CALCULATED TO GANGUE-FREE SUBSTANCE.

Theory Ratio

Hg.ClO I IT HOE of III

Hg Sh ae ee 88°65 88°92 88°3] 88'61% 2°00

Ce eres. 7°85 7°83 1:00

OF mee er 3°50 B30 7/65) 1°06
100-00 100°19

The high oxygen found in IIT is due probably more to error
in measuring so small a volume as 2°67" than to the little
montr oydite “that was present.

The only artificial mercuric-mereurous oxychloride hitherto
preparedt has the formula of terlinguaite.

Liglestonite.

Tsometric-holohedral. Crystals small, equidimensional, usually
about and under one millimeter in diameter. They show two

* Mean of I and II.
+ Fischer, T., and yon Wartenberg, H. Chem. Zeit., xxix, 308 (1905).

272 LTillebrand and Schaller—Mercury

habits, one determined by the dev elopment of the rhombie
dodecahedron and not particularly rich in forms, the other
determined by development of the octahedron and with abun-
dant forms. Of the 21 forms observed 17 arenew. Whether
eglestonite can be distinguished by its color in the mine or soon
after removal therefrom from terlinguaite, with which it is
sometimes closely associated, or from kleinite or the orange
montroydite, we are unable to say, but from the confusion that
existed in the minds of those who sent us our specimens It seems
that it must be at least difficult to do so. The first specimens
received were dark brownish and of dodecahedral habit, others
were of octahedral habit and light brownish yellow. These
last became darker in time. If sufficiently exposed the erys-
tals turn black, but without losing their luster, as noticed by
Moses. Streak yellow, tarning black. Luster adamantine to
resinous. Transparent to translucent. Brittle. Cleavage lack-
ing. Fracture uneyen and apparently sometimes conchoidal.
Hardness, 2—3 (Moses). Density, 8-237 (Moses) ; not determined
by us for the same reason as “with terlinguaite, ditheulty of
freeing perfectiv from mercury enough material for a satisfac-
tory test.

When heated in a closed tube comports itself in almost every
respect like terlinguaite. The residue, after expulsion of the
calomel, seems to be mercuric oxide as with terlinguaite,
formed in this case, however, from mercurous oxide at the
expense of half the mercury of the latter, a reaction which
accords with the observation that no oxygen escapes till all the
calomel and some mercury have sublimed.

Hydrogen sulphide acts very much as upon terlinguaite and
ammonia blackens at once, the latter reaction serving as aready
distinguishing test between the two minerals. Hy drochlorie
and nitric acids decompose it with separation of calomel. The
hydrochloric acid filtrate contains no mercury. Cold dilute
acetic acid acts more quickly on the powder of eglestonite than
on that of terlinguaite, calomel is left and from the filtrate
much more can be obtained by hydrochloric acid. The final
filtrate is free from mercury. These tests, confirmed by the
analysis, show clearly the mercurous nature of the compound,
the first authentic instance of a mercurous oxychloride, native
or artificial.

Analysis did not confirm the empirical formula He,Cl,O,,
deduced from J. 8. McCord’s analyses in the paper by Professor
Moses, a formula which, in fact, is invalidated by the qualita-
tive data above given, since it calls for mercuric as well as mer-
curous mercury. The analyses were made in the main as for
terlinguaite, with the exception that tue chlorine and mercury
in the sublimate were each time determined, the separation

Minerals from Terlingua, Texas. 273

being effected by sodium hydroxide, and the little mereury
that goes into solution with the chlorine by this operation being
recovered. Calculated to gangue-free substance the results
were as follows:

ANALYSES OF EGLESTONITE CALCULATED TO GANGUE-FREE SUBSTANCE,

Theory I II Tit
Heg,.Cl.O 0:1195¢. At. ratio 0:1008g. At. ratio 0°1198¢. At. ratio
leer sae 90°21 88°33 4°11 88°94 3°86 89°73 3°99
Clea 99 8:32 2°18 8°23 2°02 8°12 2°08
OBE oS 1°80 Tene 1 Tose bor | 1°80 i
100°00 98°37 99°G1 99°65

In all cases the mercury is probably low, and calomel was
present to a slight extent in sample I at least. It is quite prob-
able that the oxygen was less accurately determined than the
chlorine, but the effect of low mercury and the presence of cal-
omel are better brought out by the ratio based on oxygen than
on chlorine as unity. The formula plainly indicated is Hg,Cl,O,
or Hg,O.2HeCl, one that is in full agreement with the quali.
tative behavior of the mineral. The variations in the analytical
data reported by Moses are so wide that the excellent agree-
ment of his averages with the formula Hg,Cl,0, can be due
only to a balancing of large errors. The oxygen values of his
table were indirectly determined and are affected by the errors
involved in other determinations, which inspection shows were
large.

The ammonium-bromide method, used with kleinite for deter-
mining nitroyen, might probably be employed successfully
with eglestonite and terlinguaite and any other compounds of
similar kind for the indirect but very accurate determination
of the basic oxygen in them.

Two specimens of what was supposed to be eglestonite were
analyzed, but with results indicative of a mixture of terlinguaite
with calomel, though the appearance of both was decidedly
against this. Posstbly they represent a new species.

Calomel.

The crystals of calomel often reach a large size, some being
1$™ in diameter. They are very rich in forms, a total of 30
having been observed, of which 10 are new. The crystals are
frequently twinned, twinning plane e{011'. The formula for
calomel being well established, no chemical work was done on
this mineral.

* 1:90 by loss in wt, of ign. tube,

274 Hillebrand and Schaller—Mercury Minerals, ete.

Native Mercury.

Native mereury occurs abundantly on nearly all of the speci-
mens, except those of kleinite, on which we have not seen it.
It is usually present as elobular irregular masses associated di-
rectly with the other minerals. Globules often project from
small cavities on the surfaces of crystals of terlinguaite, egles-
tonite and montroydite, and are sometimes to be seen in the j in-
terior of crystals of terlinguaite and montroydite, notably the
latter. While much of it is in the form of a “fairly pure liquid,
a good deal is mixed with powdery oxychlorides as a sort of
stiff paste having a gray or greenish color and irregularly associ-
ated with the plainly erystallized oxychlorides and montroydite.
The proximate determination of this gray or greenish mass is
almost imposstble. Some of the mass has a yellowish, almost
metallic sheen, which is perhaps largely an iridescent effect.

Summary.

Kleimite, as announced in 1905, belongs to the sc-called mer-
cury-ammonium compounds, but no probable formula can be
deduced from the analyses. It may be a mixture of a mereury-
ammonium chloride in great preponderance, NHg,Cl4H,0,
with an oxychloride and sulphate or oxysulphate oe mereur \

Terlinguaite is a mercuric-mercurous oxychloride, HeO.HgCl,
the formula of Moses being confirmed and the mixed nature
ascertained by tests.

Eglestonite is a mereurous oxychloride, Hg,O.2HgCl, the
first authentic instance of such a compound, “either artificial
or native, and not Hg,Cl,O, as believed by Moses.

Montroydite is mercuric oxide, as supposed by Moses and
proven now by direct determination of its oxygen content.

Laboratory U.S. Geological Survey, Washington, D. C., July.

Kunz and Washington—Forms of Arkansas Diamonds. 275

Arr. XXVII.— Note on the Forms of Arkansas Diamonds ;
by G. F. Kunz and H. 8S. Wasurneron.

Dramonps have recently been discovered in the peridotite
stock of Murfreesboro, Pike County, Arkansas, the first having
been found Aug. 1st, 1906, and about 140 in all up to the date
of writing. <A full description will be published later, but some
preliminary notes on the forms and colors of the stones may
be of interest.

The most commonly occurring forms are distorted hexocta-
hedrons, most frequently elongated, but occasionally flattened,
and with always much rounded faces. The symbols of the
hexoctahedrons have not yet been determined, but more than
one form seems to occur, and in some cases ‘they are merely
vicinal to the octahedral face. A flattened, trigonal, lenticular
form of spinel-twin, similar to fig. 4 in Dana’s System (page 3),
is rather common, especially among the yellow stones. Another
form, also appar ently twinned on Bie is shown by the largest
diamond yet found, which in form, color, and effect somewhat
resembles the well known “butterfly twins” of calcite from
Cumberland. It is crescentic in shape, fattened in one direc-
tion of the twinning plane and elongated obliquely to it on
either side.

There are a few regular and undistorted octahedrons, which
show slightly rounded faces, though the center of these is
usually flat. The edges are often replaced by the dodecahedron,
and they are commonly further rounded by trigonal trisocta-
hedrons, with hexoctahedrons near the apices. The octohedral
faces are commonly marked by small, shallow, triangular pits,
which are most abundant toward the center ; while a few show
small, low, parallel ridges, the direction of ‘which bisects the
face angle. No cubes were seen, and tetrahexahedral and
dodecahedral forms seem to be equally rare, except as replac-
ing octahedral edges and angles.

‘White stones are common, the white being exceptionally
pure and free from all tints of color, and ustially absolutely
pellucid, like the so-called “river stones” of South Africa.
A considerable number are tinted brown or are a rich cinna-
mon-brown, while others are yellow, some of these being of a
pure lemon color. Some of the stones are gray, and a number
are very dark, almost black, bort; but none of the hard,
amorphous carbonado seems ‘to occur, though careful search
was made for it, which should have re evealed it had it been
present.

The largest diamond weighs 64 carats, and is 15°5™™ long,
7°75 high, and 4:2 thick, of an absolutely pure, pellucid white,

276 Kunz and Washington—forms of Arkansas Diamonds.

free from inclusions. Its form was noted above. Another
stone weighs 64 carats, several about 5, 4, or 3, and from this
they run down to ;'; carat, the average being probably about
one carat.

Inclusions are present in some of the stones, especially in
the gray ones, and are apparently mostly of hematite, and in
one case of what appear to be rutile needles. Several of the
diamonds show rough and irregular indentations, as if they
had formed or been attached to surfaces that were not smooth ;
while a very large proportion, possibly one half, are fractured.

While most of the diamonds have been found on the sur-
face of the igneous area, a few have been found among the
concentrates derived from washing the decomposed peridotite,
which much resembles that of Kimberley, and one diamond
was found embedded in the decomposed peridotite itself. The
portion of this that is visible is 9°1™™ long and 3"™™ wide, and
the stone is apparently a distorted and elongated hexoctahedron,
with much rounded faces, grayish in color. The evidence
seems conclusive that the diamonds found are derived from
the peridotite ; and, if so, this is evidently the first occurrence
of diamonds in place on either the North or South American
Continent.

New York, Aug. 12, 1907.

C. Barus—Method for the Observation of Coronas. 277

Arr. XXVIII.—On a Method for the Observation of
Coronas*; by C. Barus.

1. Character of the method.—In the usual practical experi-
ments with the large coronas of cloudy condensation (the
largest types having angular diameter of nearly 60°), the small
round source of light is placed in the equatorial (vertical) plane
of the fog chamber and remote from it. The eye and goniom-
eter are put as near it as possible, whenever sharp vision is
essential. The diffracted rays in such cases come from the fog
particles a, 5, ¢, at the ends of the chamber F, as in figure la,
and are liable to be seriously distorted by the refraction of the

SS Ss A

Jig. /

glass walls. The limit will be reached sooner or later, in which
the fog particles to which the diffractions are due lie at or
beyond the ends of the fog chamber, after which the features
essential to the measurement will no longer appear. Moreover
one eye only can be used in the measurements.

It occurred to me, therefore, to invert the phenomenon by
using two sources, which may be moved symmetrically towards
or from the equatorial plane, as in figure 16, and to observe
the contact in this plane of the two identical coronas produced.
In this way the oblique refractions are diminished as far as
possible, the diffracting particles a, 6, c¢ lie in the middle,
coronas of all sizes are observable, both eyes are available for
observation, increasing sharpness of vision and lessening the
eye strain. The contact method is in itself more sensitive,

*Extract from a forthcoming report of the Carnegie Institution of
Washington.

278 CO. Barus— Method for the Observation of* Coronas.

seeing that the eyes may be placed all but in contact with the
H08 chamber.

Apparatus.— Figure 1b and d, the latter drawn to scale,
ae a general disposition of the apparatus. S’ and S” are
the two circular sources of light lying in the same horizontal,
and movable in opposite directions in equal amounts, at the
control of the observer sitting near at the fog chamber /.
S’ and S” are therefore always symmetrical with? respect to the
vertical plane S/2. The diffraction of rays due to the fog
particles a, b,c, in /, produce coronas seen at nn’ and n/n”
and the lamps S. , S” have been adjusted at a distance S, $0
that the selected annuli of the coronas are in contact at 2’. The
angular radii of the coronas, marked @ or shaded in the diagram,
are nearly equal 2/2 tan @ = S, where 7 is the distance of the
axis of the fog chamber from the track S in parallel with it.

On a double track at S, the two carriages for the lamps,
S’,S” are moved with sprocket and chain or in a similar
manner and provided with a scale stretched between them,
reading to centimeters. This is a lath of wood about 3 meters
long with one end fastened at 8’ , the other free, while the scale
moves across an index at 8”. A pole at #, with the end in the
observer’s hand, moves the two central sprockets and at the
same time serves for the measurement of 2, should this
accidentally vary.

Errors. — The figure shows clearly that the angle of
diffraction corresponding to the fog particles nearer and far-
ther from the eye will not be the same and that this effect will
vanish as the coronas are smaller, as the diameter or thickness
of the fog chamber is less, and as the distance from the
source is ‘ereater. Shghtly different annuli overlap; but the
effect is much less here than in the case of a single source,
where the active fog particles le oblique to the axis. See
figure 1@ and fizure 1b at a, b, c. In practice this effect is
probal bly negligible if the dimensions of apparatus and disposi-
tion of parts are properly chosen, particularly so since the fog
particles themselves are not usually so nearly of a size as to
imply less overlapping. In fact the true corona, if large or
even of moderate size, is seen but for an instant immediately
after exhaustion. It therefore shrinks rapidly, as may be
gathered from the following incidental data given in figure 2
and obtained with fog particles originally about -00002 in
diameter, belonging to the large yellow- blue corona.

The coronas shrink as the fog particles increase in number
and decrease in size at an accelerated rate. This shrinkage
begins as soon as the corona appears. The initial rates must
be estimated at a decrement of number greater than 1-4 per
cent per second, supposing that no water is added from other

O. Barus— Method for the Observation of Coronas. 279

sources than the evaporation of smaller particles. In 100
seconds about 80 per cent have escaped. The case is much
more serious for larger coronas, so that these are characteristic-
ally fleeting and must be observed at once. It may not be
impossible that rapidity of evaporation itself sets a limit to the
largest coronas producible. The nuclei, however, are not all
lost, as a rule. They.occur, at least in part, as water nuclei
and are available for the next coronas, if not removed.

It follows then that for these cases the method of subsidence
is not applicable, as the corona changes totally before measur-

able subsidence is recorded. Hence an instantaneous proce-
dure like the former goniometer method or the present
method is alone available.

4. Data.—Onmitting the tables, I have constructed in figs. 3
and 4 results obtained by the method* of successive equal
exhaustion with phosphorus nuclei in dust-free air, leaving out
the initial fogs. It is seen at once that large coronal diameters
are actually measurable; a result which was not possible
hitherto. Reduced to the goniometer method (figure 1¢ shows
the notation) the present results may be written "12S = s’ for
small coronas; but for large coronas, if 26 is the angular
diameter, S = 2F tan 0, s=2rsin 6 or S = 8°38/V1—s*/r",
§ = 128/V14+8°/R?; R= 250™, r= 30. The letters on
the curves refer to the colors of the discs and first annuli of
the coronas.

*This Journal, xiii, pp. 81-94, 1902.

280 OC. Barus—Method for the Observation of Coronas.

Since the elementary diffraction equation may be put

sin 6 = 1°22 X/d
for the first minimum
S
2A aS

and S would therefore appear to be less immediately adapted
for the equation than s. It does not follow, however, that this
s and the one observed at the goniometer work are the same.
In fact they are not, the latter being larger for reasons involved
in the more recondite theory of the experiment, or else due to
irregular refractions at the remote ends of the chamber.

5. Remarks on the results.— W ithout entering at length into
the details of the subject, it is clear that if ds =a for normal
coronas, where d is the diameter of particles and a an optical
constant, s = a(mn/6m)!/?, where m is the quantity of water
precipitated per cm. and 2 the number of nuclei per cm. on
which this water falls. I will adda few results showing the
relation of the s computed in this way and the observed S or
the s reduced from S.

Figure 3 contains data both for S, 12S =’ and s and leads
to a curious consequence. The computed chords of the coro-
nas, s=a(mn/6m)'” are here not proportional to s=2rsin 8,
but to s=2/ tan 6, where 20 is the angular diameter of the
coronas. This implies a diffraction equation reading tan @
= 1:220/d.

In figure 38 san? is laid off as the abscissas and ‘12sa tan 6
and “12S\/ 1-S'f*a sin @, as the ordinates. If we confine our
attention to values within s = 14, where the readings are more
certain, and where there is less accentuated overlapping of
coronas, the graph .12S oscillates between two straight lines as
the coronas change from the red to the green ty pes. The
slopes of these lines are respectively as 1:08 = 73A,/a and
‘99 = 73X,/a, whence 2,=-000047 and A, = 000048. These
should be blue and violet minima.

The figure shows moreover that compared with the graph
for 12 S=60 tan 6, the curve for sin @ is in series 1 quite out
of the question, as already specified. The results for other
independent series, 2, 3, are given in figure 4, in the same way.
Curiously enough, series 2 and 3, which should be identical with
1, fail to coincide with it in the region of higher coronas. In
these series the graph sa@sin @ (not shown) would more nearly
express the results, though the Bgucoataut is far from satisfac-
tory.

/ “~y ay /
= 1229. /d or S= 2-44 RG / /1=(2-44n/

C. Barus—Method for the Observation of Coronas. 281

It is exceedingly difficult to account for this difference of
behavior. Asthe abscissas s = a(nm/6m)?, or san/, and the
ordinates, s (observed), are independent of each other, the
equality of the two values of s will in a measure check the
work apart from the constant @ which determines 7,. This is
actually the case for the lower series of coronas below s = 10,
which are used in practice. It is the observational value of
the aperture of the given coronas which varies. Very proba-
bly mixed coronas are being observed. To this must be added
the subjective error or personal equation which enters into the
determination of contacts. Finally the tendency of a corona
to shrink at once after the formation of droplets, makes it diffi-
eult to catch the time at which coronas should be observed
soon enough. Under other circumstances there is even lable
to be an oscillation of the coronal aperture in the lapse of
time. All these difficulties are accentuated as the coronas
become larger, for here not only are the droplets more volatile,
but the coronas overlap, and there is an unlooked for tendency
for them to flatten at the point of contact. The dark rings
are liable to invade the bright.

The green coronas in the series 1, 2, 8 show the following
average values, the diameter d being given in cms. :

Ser. Computed Observed Computed ' Observed
— PSS a ar = = AS =
$3 So $3 Se 10d; 10°d, 10°d, 10d.

1 8 i6 8 14 400 200 400 230

2 8 14 8 15 4.00 230 400 210

3 8 13 8 13 400 250 400 260

Mean values are thus:

O, ==" (0) 10°d, = 400
ish 14-3 10°d, = 230
agreeing pretty well with my earlier data, where
6. = (teal 10°d, = 398
Gea ies 10°d, == DONS) a

The subscripts here show the cycle to which the corona
belongs, the 2d and 3d being producible here. The first cycle
is very rarely obtainable ; the fourth and fifth occur best in
experiments with atmospheric air, not filtered, the coronas
here being very crowded.

Brown University, Providence, R. I.

Am, Jour. Sct.—FourtH Series, Vou. XXIV, No. 141.—SreprempBer, 1907.
19

282 Scientific Intelligence.

S CLE NAT ELC iNeS Bia GaneNe Osha
MisceELLANEous ScIENTIFIC INTELLIGENCE.

1. Las Formaciones Volednicas de la Provincia de Gerona ;
by 8S. CatpERon, M. Cazurro, and L. Frernanprez-Navarro.
Mem. R. Soe. Esp. Hist. Nat., vol. iv, 1907, pp. 159-489, 10 plates,
3 maps.—This is the report mentioned in the paper by the reviewer
in the present number, and was received shortly after the final
proofs had been returned. It isan excellent and detailed account
of the voleanoes of Catalonia and will long remain the standard
work on the subject. Cazurro gives a brief history and bibliog-
raphy of the region, after which Calderon describes the general
topographical and geological features of Catalonia and its vol-
canoes, and their relations to others in Spain. <A list of Cata-
lonian earthquakes is given by Cazurro, and the same writer
describes in great detail the physical features of each voleano
and flow, of which he enumerates more than fifty. Fernandez-
Navarro describes the petrography of the rocks with much detail,
finding them to be dominantly feldspar-basalts and nephelite-
basalts, with some limburgites, tachylites, diabases, and others in
very small amount. The “work is illustrated by numerous cuts,
with colored plates of rock sections, and the maps add much to
the value. The report is of great interest and is highly credita-
ble to the Society and to the Commission to noe the investiga-
tion was entrusted. . WASHINGTON.

2. I Vulcani Attivi della Terra ; by G. Manca Pp. 421,
Milano, 1907 (U. Hoepli).—This volume gives a general account
of the active volcanoes of the globe and, while intended primarily
for Italian readers, it contains much matter of interest to foreign
geologists. The author makes full use of his great knowledge
and long study of Vesuvius, but also avails himself largely of
others elsewhere. The structure and varied phenomena of active
volcanoes are described systematically, the examples and illustra-
tions being well chosen and widely scattered, and brief accounts
are given of the most noteworthy eruptions. In a descriptive
catalogue the author enumerates 415 active volcanoes. The causes
of vulcanicity are briefly discussed and the author seems to favor
the theory of Stiibel. The volume is well and abundantly
illustrated, H. S. W.

3. Die Mineralien der Siidnorwegischen Granitpegmatitgdnge.
IT. Niobate, Tantalate, Titanate and Titanonibate ; von W. C.
Broceer. Pp. 162, with 8 tables. Kristiania, 1906 (Jacob
Dybwad).—The author’s earliest studies on the minerals and
rocks of Norway, and the extraordinary richness of the results
which he has obtained, is so well known by all interested that it
need not be discussed here. Mineralogists will rejoice that he
has undertaken a comprehensive work describing the miner-

Miscellaneous Scientific Intelligence. 283

als occurring in the pegmatite veins of southern Norway. The
first part has now been issued, which includes some of the rarest
and most complex species,—namiely, those belonging to the niob-
ates, tantalates, titanates, and titano-niobates. The descriptions
here given leave nothing to be desired and are accompanied by a
series of plates, giving figures of the crystals of many species ;
there is also a map of the portion of Norway involved, showing
the position of the localities.

4. Das Problem der Schwingungserzeugung ; von Dr. H.
BarkKHavsEeN. Pp.iv+113. Leipzig, 1907 (8. Hirzel).—The cen-
tral idea of this book is that the production of oscillating electric
currents may be regarded as due to the variation of resistance,
inductance, or capacity ina circuit containing asteady E.M.F. In
this way the author classifies all alternating current apparatus
from the ordinary generator to the singing are. He points out
analogies between different forms of apparatus, and deduces some
general conditions for stability and efficiency. The last section
deals with mechanical oscillations and their relations to the elec-
trical problem. H. A. B.

5. Lehigh University, Astronomical Papers. Volume T.
Part I. Results of Observations with the Zenith Telescope of
the Sayre Astronomical Observatory from September 11, 1904,
to September 1, 1905 ; by Joun H. OcBurn. Pp. 46 and Table.
Published by the University, South Bethlehem, Pa., 1907.—This
paper embraces the results of the first year of latitude observa-
tions with the new zenith telescope, the purchase of which was
provided for in 1903 by Mr. R. H. Sayre with the general specifi-
cation that it should be equal to the best in existence. The
description of the telescope, its location, and the scheme of work
are outlined in the preface. The instrument is located on the
same meridian and 67:08 feet north of the instrument with which
Professor C. L. Doolittle carried on his well-known variation of
latitude work from 1876 to 1895.

The residuals from the adjusted latitudes show that the mean
probable error of a single observation was 0”:121. The constant of
aberration derived from the whole series for 1904-5 is 20”°4645 +
0”-01098. The extreme variation of latitude shown is from
40° 36’ 237-713 on Feb. 21, 1905, to 40° 36’ 247-007 on Aug. 24,
1905. Plotting the final latitudes obtained at intervals of from
three to five weeks and drawing a smooth curve between the
points, it is seen that the final latitudes mostly lie not more than
0”-01 from the curve and in no case more than 0”:04 from it.
When it is considered that these figures amount on the surface
of the earth to but one foot and four feet respectively, the accu-
racy with which modern astronomy can follow the wanderings
of the earth’s axis within its body is perceived. The author is
to be congratulated on the success of the arduous work of obser-
vation and reduction performed without assistance and in addi-
tion to the usual duties of college instruction. fo 18h

bo
(2)
wee

Scientific Intelligence.

OBITUARY.

ANGELO HEILpPRIN, naturalist, explorer, and well-known writer
and lecturer on geological and geographical subjects, died in
New York City July i 7th, after a long illness, in the 55th year
of hisage. A native of Hungary and a descendant of a culti-
vated and intellectual family, he was brought to this country
while yet a small child by his father and received here his early
education. In 1876 he went abroad and, in London, Paris,
Geneva and other places, studied subjects chiefly of a geological
and biological nature. In London he was awarded the Forbes medal
for proficiency in geological work. In 1879 he returned to this
country and the following year was appointed professor of inver-
tebrate paleontology and geology at the Academy of Natural
Sciences in Philadelphia and in 1883 executive curator of its
museum, a post he held until 1892. He also filled the chair of
geology in the Wagner Free Institute of Science, 1885 to 1890.
During the last three years of his life he was lecturer on physical
geography i in the Sheftield Scientific School of Yale University.

Throughout his life, as opportunity offered, he engaged in
active exploration i in various parts of the world, making special
studies of a geological and geographical character. Thus in 1886
he explored Lake Okeechobee, Florida; in 1888 the Bermuda
Islands ; in 1890 parts of South and Central America and Mexico,
especially Yucatan, devoting attention particularly to the vol-
canoes. In 1891 he became interested in Peary’s work and spent
two summers in Greenland, leading the relief expedition in 1892.
In 1896 he travelled in Morocco and Algeria studying the Atlas
Mts. He also spent two summers in northern Alaska. His best
known work is perhaps that connected with the eruptions of Pelée
on Martinique in 1902, where he made a perilous ascent to see
the voleano in activity. The results of his observations made
during his travels he published in a number of books and maga-
zine articles, and he also lectured extensively upon them. He was
also a contributor to encyclopedic literature and wrote several
popular works on geology. With his brother he edited Lippin-
cott’s New Gazetteer of the World.

During his life Heilprin accomplished a large amount of work
and while this work was chiefly concerned with the popular
rather than with the profounder aspects of science, it was none
the less extensive and useful on this account. Men who have his
ability to seize and present the essentials of scientific research and
progress to the public in a clear and lucid manner by word and
pen are uncommon, and this added to his energy and enthusiasm
for exploration rendered him aman whose death, while still in
the prime of life, is a serious loss in his chosen field of work. His
wide knowledge, intellectual acquirements and varied talents, as
well as his agreeable and friendly disposition, gained for him a
large circle of friends, by whom he will be greatly missed. 1. v. P.

Professor James Merritt Sarrorp, of Vanderbilt University,
for many years state geologist of Tennessee, died on July 3 at
the age of eighty-five years.

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES. ]

—_—_0¢e—___

Art. XXIX.—On the Corpuscular Rays produced in differ-
ent Metals by Réntgen Rays; by OC. D. Cooxsry.

Saenac* has shown that when Rontgen rays fall on a sub-
stance this substance gives off rays very like the Rontgen rays
producing them. Barklat has investigated these rays and found
that those from gases and light solids were almost identical
in penetrating power with the primary rays producing them,
while those from heavy solids such as lead, zine, copper, etc.,
are much less penetrating than the primary. He also finds
that the rays from gases and light solids vary in penetrating
power as the primary varies, but those from the heavy
solids vary very little fora large variation in the penetrating
power of the primary. All these secondary rays are probably
pulses in the ether like Rontgen pulses, though the secondary
pulses may vary greatly in thickness from the | primary produc-
ing them; heavy metals giving out much thicker pulses than
the primary. Hereafter in referring to this type of rays we
shall speak of it as ‘Secondary isoumgen Rays.”

Beside this type of rays, Townsend,t Dorn,§ and Curie and
Sagnac|| have shown that heavy metals, when Rontgen rays fall
on ‘them, give out a very absorbable type of radiation, consisting
in part of negatively charged particles. The velocity of these
particles is of the order of that of the cathode particles in a highly
exhausted tube. Bestelmeyer§] has investigated by a photo-
graphic method the corpuscular rays produced when Roéntgen
rays fall on a plrioarety splice and found that they have veloci-
ties from 07195 to 0°327 of the velocity of light. He shows

* Annales de Chimie et de Physique, xxii, p. 493, 1901.
+ Phil. Mag., xi, p. 812, 1906.

¢ Proc. Camb. Phil. Soc., x, p. 217, 1899.

S$ Abhand. d. naturf. Ges. zu Halle, xxii, p. 39, 1900.

| Journal de Physique (4), i, p. 39, 1902.
§| Ann. der Physik, xxii, p. 429, 1907.

Am. Jour. Sci.—FourtuH Series, Vou. XXIV, No. 142.—Ocrtoser, 1907.
20

286 Cooksey—Corpuscular Rays produced in

that the velocity does not depend at all on the intensity, but
does depend on ‘the penetrating power of the primary Réntgen
rays producing them; the harder or more penetrating primary
rays producing higher velocities in the corpuscular rays. From
some results published by Prof Bumstead* on the heating
effects produced when Roéntgen rays are absorbed by different
metals, I was led to investiyate this last type of rays, which

MIA 2PIS

Lead Boy

we shall refer to as “Corpuscular Secondary Rays.” with a view
to determining the relative number produced in different
metals for equal absorptions of the same beam of Rontgen
rays.

Apparatus.

It was decided to investigate these corpuscular secondary
rays by means of the ionization which they produce, but owing
to the great ease with which they are absorbed it was found
that the thinnest aluminium which could be obtained would
stop all but a few of these rays. It was necessary, therefore, to

* This Journal, xxi, Jan., 1906.

different Metals by Lontgen Rays. 287

let the metal to be examined form one end of the ionization
chamber, and let the primary beam of Réntgen rays pass
through the ionization chamber and fall on the metal. Then
all the corpuscles coming from the metal would pass into
the ionization chamber, producing ions in the air, and if the
ionization due to the Rontgen rays was known, that due to the
corpuscular secondary rays could be deduced. But it was
found that the ionization due to the secondary rays was such
a small fraction of that due to the primary, that this method
was not accurate.

The apparatus as finally used and shown in fig. 1 was
designed to eliminate the ionization due to the primary rays.
Two ionization chambers, “A” and “ B,” were used, consisting
of hollow brass cylinders about 5™ long and about 10 an:
diameter, each encased in thick lead. The cylinders were
mounted on vuleanite insulators, “f’,’ with their axes in a
horizontal plane and parallel to each other and about 14™
apart. Above the cylinder “A” and supported on it was a
gold leaf electroscope consisting of a lead box, “a,” with win-
dows for observing the gold ‘leat, and a narrow strip of gold
leaf hung from the upper end of a vertical brass TOG. saz
The brass rod, “b,” passed through the bottom of the electro-
scope box into the interior of Tak ” being insulated from “ A ”
and “a” by vulcanite rings, «* £ and its lower end termina-
ting in a ring, “e,” about 4™ in diameter, made of fine brass
wire. This 1 ring was placed concentric with the axis of “A”
and ina vertical plane parallel to and midway between the
ends of “A? Al brass rod, “¢,/* insulated from “A” and
Be by vulcanite: rings, £", 7 veonmected Se? to a: similar
ring, “¢’,” similarly placed with respect t0.. du;

The Ronicen ray bulb was about 60° in front of the two
ionization chambers with its target midway between their axes,
so that the beam would pass through each cylinder longitudi-
nally. The ends of A” and “ B” on the side toward the tube
were covered with very thin aluminium leaf, and the rear end
of “B” was covered in a similar manner. The rear end of
“A” was left open, but provided with clamps by means of
which disks of the metal to be tested could be clamped over
this end.

Both cylinders stood ina thick lead box, open at the rear,
and provided with windows in the side for observing the elec-
troscope. In the front of the box were two cireular openings,
“W” and “ W’,” corresponding respectively with the ends of
“A” and “B.” Both openings were just small enough to
prevent the beam of Réntgen rays from falling on the walls
of either cylinder. Lead screens with different sized holes

288 Cooksey— Corpuscular Rays produced in

bored in them could be hung in front of ‘“ W” to control the
amount of rays entering ch? A shutter, “d,”’ placed over
“W’” and controlled by ascrew ‘‘e,” gave a very fine adjust-
ment to the size of this opening.

In use the lead box, chamber “ A,” and electroscope case
were all connected to earth ; the chamber “ B” was kept at a
constant potential of about 400 volts, and the gold leaf, rings
“ec” and ‘“¢’,’ and connections were br ought initially to about
200 volts. “This made a difference of potential between “ A”
and “¢” of 200 volts, and between “B” and “c’” of minus
200 ie The fall of the gold leaf due to the jomieane in
the chamber ‘A’ could be read with a microscope having a
micrometer eyepiece. If the Réntgen rays were turned on, and
the size of the opening, “‘ W’,” properly adjusted so that there
should be the same amount of ionization in “B” as in “ A,”
the flow of electricity from ‘“¢” to “ A” would be just equal
and opposite to the ‘flow on “B” to “e’,” and there would
be no motion of the gold leaf. In this way the ionization due
to the Réntgen rays and any natural leak could be eliminated.

In order to test a given metal, a disk of the metal was made
to fit over the rear end of the chamber, “A.” This disk was
then clamped over the end with a thin disk of aluminium
interposed between it and the primary rays. (Aluminium was
used because it gives off hardly any of the corpuscular second-
ary rays, and there do not enough of the secondary Rontgen
rays come off from a thin sheet. to produce any appreciable
ionization.) The Roéntgen tube was then excited, and the
shutter, “d,” adjusted fill there was no motion of the gold
leaf. The aluminium screen was then removed, and the time
taken for the gold leaf to fall over a given number of scale
divisions measured. The fall of the cold leaf was then due to
the ionization produced by the secondary rays alone from the
metal.

Preliminary experiments showed that an appreciable amount
of the ionization in the chamber, ‘* A,” was due to secondary
rays of the Réntgen type, and it was therefore necessary to
find what proportion of the ionization was due to corpuscular
secondary rays, and what due to secondary Réntgen rays. For
this end the absorption of the secondary rays in thin sheets
of aluminium leaf was measured: ‘The primary rays were
first eliminated by balancing the apparatus with a sheet of
aluminium over the metal about 1°" thick, sufficient to stop
all the secondary rays from the metal, but ‘not thick enough
to give off any appreciable secondary rays itself that would be
absorbed in the ionization chamber. The ionization was then
measured which was due to the secondary rays from the
bare metal and to those which got through successive sheets of

different Metals by Léntgen Ruys. 289

aluminium leaf about :001™™ thick. Curves were then plotted,
having for abscissae the thickness of aluminium in millimeters,
and for ordinates the ratios of intensity of ionization after
absorption to the intensity before absorption.

Since the penetrating power of the secondary rays depends
upon that of the primary rays, it was necessary to keep the
Rontgen ray tube very constant, and to have some standard of the
penetrating power of the primary rays. A self-adjusting tube
was used which was found to be fairly constant. The stand-
ard of penetrating power or hardness of the primary rays was
arbitrarily taken as the ratio of the ionization produced by the
rays after passing through asheet of aluminium 0°82" thick to
that produced by the unobstructed rays. This ionization was
measured in the chamber, ‘‘ A,” after connecting the chamber,
“ B,” to earth, and entirely closing the opening, “W’.” The
hardness of the primary rays was varied by varying the length
of the spark gap in the regulator, and by filtering out the soft
rays by a thin iron screen.

Absorption curves were plotted for the secondary rays from
lead, zine, nickel, copper, tin, and silver for two different de-
grees of hardness of the primary rays. In the case of lead four
curves were plotted for four different hardnesses of the primary
rays. The curves are represented by the full lines passing
through the circles and crosses. In every case the circles rep-
resent rays produced by soft, and the crosses rays by hard,
primary rays. The following tables give the data from which
the curves were plotted. Tables 1 and 2 refer to lead, 3 to zine,
4 to copper, 5 to nickel, 6 to tin, and 7 to silver.

TABLE I,

= Hardness=0 62 — -———-—Hardness=90:89-——_—.
I Il III IV ine Ii’ TV'
0) 1:00 1°00 0) 1°00 1°00 O

001 0°29 0°22 — —0°66 0°49 0°45 —0°35

"002 0°15 0:07 —1'15 0°24 0:19 —0°72

"003 0°10 0:02 —1°70 Orll 0°06 —1°22

"004 0°09 0:01 — 2°00 0:08 0:03 —1°52

"008 0°07 ens So 0°03 Mii ay

TABLE II,

— —Hardness=0°61 ——, —-——Hardness=0'95- s
ih II III IV 1 Tit’ Iv'
®) 1°00 1°00 0) 1°00 1°00 0)

“001 0°32 0°24 —0°62 0°44 0°40 —0.40

"002 0°19 O'll —0°96 Ors 0°21 —0'68

"003 0°12 0°03 —1°52 0°12 O-1l —0°96

"004 0°10 0°02 —1°70 0:09 0:04 —1°40

"008 0°07 oe a 0°02 dis ae

290 . Cooksey—Corpuscular Rays produced in

I II II
O 1°00 1:00
‘001 0°55 0°31
002 0°45 0°16
008 O37 0°05
“004 0°38 0:03
008 0°29 ee
"012 0:26 ee
= —Hardness=0'68
I II III
0 1:00 1:00
"001 0°42 0°23
"002 0.33 OrlL
"003 0°29 0°07
"004 0°29 0:04
“008 0°23 ae
"012 0°22 ae
———-———Hardness= 0 °64——
I II Til
0) 1-00 1:00
‘001 0°45 0:24
“002 0°40 O-ll
7003 0°33 0°05
"004 0°30 0:02
“008 0°24 oe
012 0°19 ec
—— — Hardness=0'59
I II JOE
0) 1-00 1:00
“001 0°41 0°39
“002 0:18 0°15
"003 O'1ll 0°05
“004 0:04 0:02
“008 0:04 ae
"012 cae ies
———_—— Hardness=0°60——
I II III
O 1°00 1:00
001 0:44 0°37
°002 0°17 0°10
"003 0-14 0:04
*004 0712 0°02
008 Orll a
012 0:09 re

— Hardness=0 66

TABLE III.

IY

TABLE IV.
SS

ING

TABLE V.

=)

TVG

TABLE VI.

IV

TABLE VII.

IV
0)
—0°43
—1:00
—1:40
—1°70

——-Hardness=0°97-——_—_,
1 Til’ TV’
1°00 1°00 0)
0°49 0°41 —0.39
0°33 0°19 —0°72
0:20 0:08 —1°'10
0°21 0-04 —1°40
Oa ea As
0:09 es ees
———-Hardness=0°78= —
Il’ Ill’ Tv’
1°00 1:00 0)
0°52 0:44 —0°'36
0:29 0:19 —0°72
19 0°07 —1°'15
0°23 0:03 —1'52
0-10 sy a
———-—Hardness= 0°82-——_—_,
la’ IIT’ Ny"
1:00 1°00 0)
0°33 0°32 —()'"49
0°24 0°14 —0°85
0°13 0°04 —1:40
0°10 0-01 —?2-00
0°09 ao ae
————Hardness=0:°95-———_,
Il’ IDoy Vad
1°00 - 1:00 @)
0°50 0°43 —0°37
0°29 O19 —— (P72
0°19 0-08 —1°10
0°14 0:03 —1°52
Orll aes ras
0-10 te au
—-—— Hardness = 0'90-——_—
100 III’ IV’
1°00 1°00 0)
0°52 0°41 —0°39
0°26 0°16 —0°80
0°20 0:07 —1°15
0°17 0°04 —1°40
0°13 Boks ie

Ow2

different Metals by Léntgen Lays. 291

Column I of the tables gives the thickness of aluminium in
millimeters; columns II and II’ give the intensity of the second-
ary rays after passing through ‘the corresponding thickness of
alumininm for the two hardnesses respectively, the intensity of
the rays from the bare plate being always taken as unity.

ae were
Sa ae
IN oo ieee mat

era

Y,

[rt E066

O01 .002 O03 .00F J/2
Ii KI of AlenCAtin

iby
¥ To.

Tete nect

SY,

001 .002 .003 .004 008
m tH. of Aluminium.

i.

Intensity Yr.

BON as 77
LS
Se eee

L,
ie

Le ory,

-OOZ O02 .003 .00 008
infn. of Aluminium.

Pe ee ae

Aes

aS — ———————

S|) a
C0. 708, O17

Y 2002 003 .004
inm.of Alumi ini

To

TInt id
M7 Sse

1001 .002 .003 .004

008
min of Aluminium.

294 Cooksey—Corpuscular Rays produced in

In all the curves there is a very sudden drop up to the first
three or four sheets of aluminium, and then the curves fall off
more slowly. The first part of the curve up to the first three
or four sheets of aluminium is undoubtedly due, for the most
part, to the corpuscular secondary rays, while beyond this point
the corpuscular rays are practically all absorbed and there
remain only the secondary rays of the Rontgen type. As has
been said before, these corpuscles come off with velocities com-
parable with the velocities of the cathode particles in a highly
exhausted tube; say between 10° and 10°°™*/sec. Seitz™

Ol O02 003 00F 008 O12
nn. of Aduminiurm.

has investigated the absorption of cathode rays driven by 20,000
volts, corresponding to about 9x LO*™s/sec., and finds that they
will penetrate aluminium about -002"" thick, the fraction
which gets through being about 0°16. A little further on we
shall show that this is about the order of absorption for the rays
constituting the first part of the curves under discussion. It
is this agreement which leads us to believe that the first part
of the curves is largely due to the corpuscular secondary rays.

With the exception of zine and nickel the first part of all
the curves due to the secondary rays produced by the hard
primary rays lies above those produced by the soft primary rays.
This would be expected from the results of Bestelmeyer, before
mentioned, where the corpuscles produced by harder primary
rays had higher velocities than those produced by the softer

* Annalen der Physik, xii, p. 860, 1903.

different Metals by Rontgen Rays. 295

primary rays. The fact that in the case of lead and copper the
curves cross each other, and that in the case of zine and nickel
the harder lies below ae softer, would seem to show that for
these metals the harder primary rays produced less penetrating
secondary rays. But there are several reasons to account for
this. In the first place, for zinc, nickel, and copper a compar-
atively large proportion of the secondary rays consists of those
of the Réntgen type which are much more penetrating than the
corpuscular rays, and in the case of lead there is an appreciable
quantity of these rays present. Owing to the shortness of the
ionization chamber, ‘A,’ a considerable proportion of the
secondary Roéntgen rays doubtless passed entirely through the
chamber, producing very few ions. If now the hard primary
rays produce secondary rays of the Réntgen type more penetra-
ting than those produced by the softer primary rays, a larger
proportion of these penetrating secondary rays would pass
through the chamber, without producing i ions, than of the less
penetrating ones. Thus the less penetrating rays would produce
more ionization, and the curve corresponding to them would he
higher than the curve corresponding to the more penetrating
secondar y rays. This reasoning would not apply to the corpus-
cular secondary rays, since the most penetrating of these are
absorbed by a few millimeters of air. But in all but two cases
the part of the curve corresponding to the corpuscular rays
shows greater penetration when the corpuscles are produced
by hard primary rays than by soft. In the two exceptions,
namely zine and nickel, there are so many of the Réntgen type
of secondary rays present that they would obscure the effect
of the corpuscular type, and the above reasoning would again
apply.

In order to show more clearly the variation of penetrating
power of the corpuscular secondary rays for different hardnesses
of the primary, I have subtracted from the whole curve the
amount, shown by the broken straight lines, which seems to be
about the proportion of secondary. Leontg gen rays present; and
again plotted the curves after reducing the initial intensity to
unity. These are shown by the broken curved lines. The
intensities are given in columns III and III’ of tables 1 to 7.
In every case the curve corresponding to hard primary rays
‘lies above that for the soft, except in nickel and copper, where
they cross very near the ‘last point. This is probably due to
not allowing enough for the secondary Rontgen rays.

If we consider the amount of the rays left unabsorbed after
passing through -002™™ of aluminium, as shown by these latter
curves, it is found to be between 0- 10 and 0:20 of the total
amount of corpuscular rays coming off. This agrees very well
with the absorptions found by Seitz for cathode rays of high

296 Cooksey—Corpuscular Rays produced in

velocity, and leads us to believe, as mentioned above, that we
-are dealing with corpuscular secondary rays of velocities com-
parable with those measured by Seitz.

In subtracting the secondary Rontgen rays, where the points
lie badly off the curve, the curve was taken in preference to
the point. Attention should be called to the points for zine
and copper corresponding to the intensity of the rays after pass-
ing through four sheets of aluminium. These points lie as
hich or higher than the corresponding ones for three sheets,
and it is due in all probability to an error in measuring the
thickness of the fourth sheet of aluminium. A different sheet
was used for the other metals, but I have not had time to repeat
the two curves in question.

Anattempt was made to see if the absorption of the corpuscular
rays followed an exponcnia) law. To this end the common
logarithms of the points (given in columns IT and III’ of the
tables) were plotted. The origin was taken at the point on the
axis of ordinates corresponding to unit intensity of the second-
ary rays, and the logarithms of the intensity plotted down,
each tenth division for the intensity being equal to two tenths
for the logarithms. The values of the Jogar ithms are given
in columns IV and LV’ of the tables.

For the more penetrating primary rays the points all lie
pretty well on the straight line passing through the origin.
This would seem to show that the corpuscular secondary rays
produced by hard primary rays were very homogeneous in regard
to velocities and were absorbed according to an exponential
law. When we come to consider the soft rays we find that if
we neglect the origin, the points lie fairly well on a straight,
line which cuts the axis of ordinates below the origin. This in
connection with the absorption curves for the cor puscular rays
leads us to the conclusion that the soft primary rays give rise
to corpuscular rays between 70 per cent and 80 per cent of
which are slow moving corpuscles which are almost all absorbed
in the first -001™ of aluminium, and the rest faster moving
corpuscles which are absorbed accor ding to an exponential law ;
but in ali cases, except nickel and copper, these corpuscles are
slower moving than those produced by the hard primary rays.
These exceptions, I am inclined to think, are again due to not
taking sufficient account of the secondary Rontgen rays present.” »

In the case of tin and silver the pots for the rays produced
by the soft primary rays lie on a straight line passing through
the origin, showing that the corpuscular rays from these metals
are very homogeneous even though produced by soft primary
rays.

*The writer hopes to investigate this question further by using a shorter
ionization chamber.

different Metals by Rontgen Rays. 297

TaBLeE VIII.
0 001 002 003 004 -008 012
i 3°4 6°8 1°8 2°6 3°8 4°8 5:2
II 3°8 7:8 2-0 2°8 3°4 5:0 5:4
Il 40 7-6 2-0 3:2 4*4 5°8 6°6
Average 3°73 7°40 1:93 2°86 3°86 5:2 57
Ratio 1°00 0°50 0:29 0°19 Olt O-ll 0°10

Table 8 has been inserted to give an idea of the actual obser-
vations as taken. It is for tin for a hardness of the primary
rays of 0°95. The numbers in the top row give the thickness
of aluminium in millimeters. The next three rows give the three
sets of observations of the time in seconds for the gold leaf to fall
over a constant number of scale divisions; these values being
inversely proportional to the intensity of ionization. The gold
leaf moved so rapidly for the unabsorbed rays and so slowly
after the rays had been partially absorbed that the same num-
ber of scale divisions could not be used for all thicknesses of
aluminium. The times in the first two columns are for a fall
of forty scale divisions, while in the remaining five columns
the times are for a fall of five scale divisions from the same
zero point. A separate observation showed that for a constant
intensity of ionization the fall for forty divisions was to the
fall for five in the ratio of 6-7 to 1. The last row gives the
ratio of intensity of ionization to the initial intensity after
reducing the times of fall to the same number of scale divi-
sions. It was found that owing to asymmetry of. the appa-
ratus a very small change of hardness in the primary rays
would disturb the balance enough to account for a large part
of the errors, though great care was taken to see that the ioni-
zation due to the primary rays was properly balanced out before
each separate observation. To show the magnitude of this error,
suppose that the time observed for the secondary rays to pro-
duce a fall of five scale divisions was five seconds, it was some-
times found that after an observation the balance had been
disturbed sufficiently for the primary rays alone to produce a
fall one way or the other of five divisions in eighty seconds,
which would tend to make the fall observed for the secondary
rays too large or two small by about 6 per cent.

After this investigation of the character of the secondary
rays coming off from different metals, the way was open to
obtain some idea of the relative number of cor puscles produced
in different metals for equal absorptions of the primary Rént-
gen rays. The method of procedure was nearly the same as
that already described. The absorption curves showed that
essentially all the corpuscles were stopped in the first ‘003"™™
to 004" of aluminium, while practically all the secondary

298 Cooksey— Corpuscular Rays produced in

Roéntgen rays were able to get through. In order to compare
two metals, say lead and zine, the lead was covered with about
-004"™ of aluminium and clamped to the chamber, “A”. The
Roéntgen rays were then turned on and the leak balanced as
before ; but this time not only was the ionization due to the prim-
ary 1 rays eliminated, but also that due to all the secondary Rént--
gen rays which got through the aluminium. After the system
was balanced, the aluminium was removed, and the ionization
measured which was due to all the copuscular secondary rays
which escaped from the lead. Then the zine was tested in the
same way, care being taken to keep the intensity and hardness
of the primary rays constant throughout the comparison. This
method was pursued with all the metals ; lead was taken as the
standard, and all the other metals compared to it separately
in order that the intensity and hardness of the primary rays
might be the same throughout the comparison. Hardnesses
were used as nearly as possible the same as those used in the
first eaten

It might be well here, by the way of parenthesis, to give
some idea of the hardness of the tube used. The softest rays
obtainable were those equal to a hardness of about O-4 in the
arbitrary scale, the vacuum being down to about the point
where Rontgen rays begin, the green fluorescence on the walls
of the tube just beginning to show. It was found impossible
to keep the tube as ‘constant as necessary with so low a vacuum,
the vacuum changing very rapidly at this point; therefore a
hardness of about 0°6 was the softest used, and this could be
kept very constant. A thin iron screen was always used when
hard rays were wanted, but this does not really make the rays
harder, only cutting out the softer pulses in the beam, and
making the proportion of hard pulses greater. As well as using
the iron screen, the tube was run for a while with a longer
spark gap in the regulator and with higher potential from the
coil. This did make the rays harder, as the cathode particles
producing them traveled faster in the tube.

To resume: the above comparison of different metals gave
only the relative ionization due to the corpuscles which escaped
from the metals, but not the number of corpuscles produced
in the metal by the primary Roéntgen rays. This latter can
be roughly calculated as follows:

The intensity of the primary rays after passing through a
thickness # of the metal, if we assume that for thin layers the
rays are absorbed exponentially, i is I,e-*«, where 2 is the coef-
ficient of absorption of the primary rays in the metal. The
number of corpuscles produced per cm. of path of the primary
rays is proportional to the intensity of the primary. The num-
ber produced ina layer of the metal between w and x + dz is pro-

different Metals by Rontgen Rays. 299

portional to I,e-*«dx. If k’ be the coefficient of absorption of
the corpuscular rays in the metal, the number which emerge
from this layer will be proportional to I,e-*«e-*'«dx, and if we
eall the total number of corpuscles emerging from the metal

N’, we have that
CHL), ae
Ne Life Ce op
I
ioe

Nae
If we let the total number of corpuscles generated in the metal

be N, then
ope
Ne If "dee
J
oa

r
Therefore

No Xt Ni
x

oc (+5) N’

Since the corpuscular rays are much more easily absorbed than
!

or

the Réntgen rays producing them, the quantity x will be very

large compared to unity, and we can say
No —N’
eX

N’ is not, however, directly proportional to the intensity of the
ionization measured in the ionization chamber. ‘To find what
relation the ionization measured bears to N’ we must take into
consideration the absorption of the corpuscles in air, and the
number of ions produced by a corpuscle per cm. of path for
different velocities. For the range of velocities encountered
in these experiments, however, I think we are justified in
assuming that the corpuscles from all the metals produce the
same specific ionization, for the following reason.

Durack * has measured the specific ionization produced by
Lenard rays having velocities of about 4x10° ems./sec., and that
produced by the deflectible Becquerel rays having velocities of
from 24x10" to 2°8x10"° cms./sec. In the first case he finds
the specific ionization equal to about 0-4 and, in the second, to
about 0-17 ions per corpuscle per cm. of path; a fall of about

* Phil. Mag. (6) iv, 29, 1902; and v, 550, 1903.

300 Cooksey— Corpuscular Rays produced in

57 per cent for an increase of velocity of about 78 per cent.
Considering the slight variation in velocity with which the
corpuscles come off from the different metals in our experi-
ments, as shown by the variation in absorption, it seems safe
to assume that the specific ionization due to the corpuscles
from the different metals under consideration is approximately
the same for all the metals.

To find the relation between N’ and the ionization measured
let m be the number of ions produced by one corpuscle per em.
of path, and let K be the coefficient of absorption in air for the
corpuscles from any metal. Then the number of ions pro-

duced in a layer of air between x and w+dw is n N’e-Kadz,
and the total number of ions produced is

2K; nN’!
nif 6. ae
‘ K

This is proportional to the quantity measured, say Q. There-
fore
K Q -

iD

N’'ec

In comparing the number of corpuscles produced in any
metal with the number produced in lead, we have:

The ratio of the number of corpuscles produced in any metal
to the number produced in lead for equal absorptions and equal
intensities of the primary rays is equal to

k’
enn) SiN

Ree =
a, Q, IN;

The coefficient of absorption for the corpuscular rays in pass-
ing through any substance is proportional to the density of the
substance. We have found the absorption of the corpuscular
rays in aluminium; let the coefficient of this be called %. Let
p be the density of any substance, and p’ the density of alumin-
ium; then the above reduces to

N A, ke p Q
N, A k*, Pi Q,

The values of X were measured for the different metals used.
The values of & for the different metals are proportional to
the slopes of the straight lines plotted with the curves in the
cases where the lines pass through the origin. In the case of
the secondary rays produced by the soft primary rays, where
the straight lines do not pass through the origin, the following
method for finding & was used. The slope of the straight line

different Metals by [ontgen Rays. 301

passing through the origin and the first point was multiplied
by the percentage of rays absorbed by the first layer of alumin-
ium, and the slope of the line passing through the first and
remaining points was multiplied by the percentage of rays
remaining unabsorbed by the first layer, and the sum of the
two numbers thus found divided by 100. This seems to me
to give the fairest value of & when the rays are not homoge-

neous.

TABLE IX.

Metal Hard k Be an k ki Q N’ N
pi a ky k? Q: N/ Nz
Lead eran : 1:00 1:00 1:00 1:00 1°00 1:00 1:00
Silver 0°52_.__- A Caepen 210) O18) O61 0:52 10 S058
sc 0:-90eue 3°8 2-09) 03) 1-10" 0:77 0279" 1-60
Nickell 10:59. 553 5 a 09s! 0787: 10:36)2 0/33) (0135
‘6 0°94._.._4:9 oe lh 71-3011 1-70 .0:30 | 0-46.0:70
Tin 0:59" DOT ee ils OU 8051) 0.64005) 0.28
«“ O95. oe 3°6 1:0 0:97 0°94 0-75 0°73 0:46
Tine). 0:66. .2 AD ogg oy 016) 0:58 10-41) 0:31 1. 0/22
«“ O80 3s 7 1:9 1:00 1:00 0°35 0:35 0°42
Copper 0°63____- Bioeng 3) |) 0.93). 0 8%) 039770 36 1 0:35
‘s GQ ne 3:7 16 1:00 1:00 0°33 0°33 0:42

aed

Table 9 gives the values of the ratios of the different quanti-
ties in the formula. The absolute values of 4, given in the
second column, are not the true coefficients of absorption, but
the absorption per 0°01"™" of aluminium calculated from the
common logarithms and not the natural logarithms; but as &
comes into the formula only as a ratio, the other constants
eancel. The numbers after the metals in the first column
denote the hardness of the primary rays used on the metal and
on lead when the two were compared. The seventh column
gives the ratio of intensity of ionization produced by the cor-
puscular secondary rays from the different metals to that
which was due to the rays from lead, as measured experiment-
ally for the different hardnesses given. The numbers in the
eighth column are the ratios of the number of corpuscles
which get out of the different metals to those which get out
of lead. The numbers in the last column of the table repre-
sent the ratios of the total number of corpuscles produced in
the metals to the number produced in lead for equal absorptions
otf the primary rays of equal intensity, as calculated from the
formula.

It will be seen at once that, in all cases, this ratio is larger
for the harder primary rays, in fact in silver the harder prim-
ary rays produce more corpuscles than in lead. For both hard

Am. Jour. Sct.—FourtH SERIES, VoL. XXIV, No. 142.—OcrosEr, 1907.
21

302 Cooksey—Corpuscular Rays produced in

and soft primary rays fewer corpuscles are produced, the less
dense the metal. In the case of copper and nickel, which have

the same density, the ratios are the same in the case of the soft
primary rays; ‘for the hard primary rays the ratio is much
smaller for copper than for nickel, but it will be noticed that
the hardness of the primary rays used on copper was consider-
ably less than for the rays used on nickel. In the case of tin and
zine, which have nearly the same density, and on which primary
rays of nearly the same hardness were used, we find the corre-
sponding ratios nearly equal. In the case of silver, which is only
slightly /ess dense than lead, there are much fewer corpuscles
produced than in the lead by the soft primary rays, but the
hard primary rays seem to produce more in the silver than in
the lead.

It will be noticed in the case of nickel, zinc, and copper for
the hard primary rays, and in silver tor both soft and hard rays,
that N/N, is larger than N’/N,': This means that of the
total number of corpuscles pr oduced a less proportion escapes
from these metals than from lead, which would not be
expected since the corpuscles should be more absorbed in lead,
owing to its greater density, than in the other metals. But

reference to the values of = the ratio of the absorption of the

primary rays in lead, to that in the other metals, will show that
the hard, and in the case of silver also the soft, primary rays
are absorbed to a much greater extent in the lead than ‘in the
above mentioned metals; while the corpuscular rays coming
from lead for hard primary rays possess nearly the same pene-
trating power as those from zine and copper, and greater
penetrating power than those from silver and nickel, as can
be seen from the values of 4/%,. Therefore for equal absorp-
tions of the primary rays the corpuscles must be produced
much nearer to the surface in lead than in the other metals,
and consequently a greater proportion of the number produced
will escape from the lead than from the other metals, notwith-
standing the fact that the corpuscles are more easily absorbed
in lead owing to its greater density.

In the case of the soft primary rays, except for silver and
nickel, the value of N/N, is less than the value of N’/N/’.
But the soft primary rays are absorbed nearly as much by the

other metals as by lead, and therefore the corpuscles will come
from as near the surface in one as in the other, and, the den-
sity of lead being the greatest, more will get out of ‘the other
metals in proportion to the number produced than out of lead.
In the case of silver we see that the absorption of the primary
rays in lead is again much the greater. Nickel is the only
other exception to the above, while zine, which is more trans-

different Metals by Rontgen Rays. 303

parent to the primary rays in comparison to lead than nickel
is, follows the rule; but a comparison of the absorption coeffi-
cients for the corpuscular rays will show that the corpuscles
from lead, due to soft primary rays, are nearly as penetrating as
those from nickel and much less penetrating than those from
zine. This should easily account for the exception.

Summary.

The results from these experiments may be summarized as
follows :—

1. The velocities found from the absorption in aluminium
of the corpuscular secondary rays produced when Réntgen
rays fall on heavy metals are in good agreement with the veloc-
ities found by other investigators, for all the metals considered.

9, The more penetrating primary rays produce more pene-
trating corpuscular secondary rays.

3. For the more penetrating primary rays used the corpus-
cular rays from all the metals tested, with the exception of
nickel, possessed nearly equal penetrating power; while for
the less penetrating primary rays the penetration of the cor-
puscular rays varied considerably for the different metals, those
from lead being the least penetrating.

4. The corpuscular rays produced by hard primary rays are
very homogeneous and are absorbed according to an exponen-
tial law. Those produced by soft primary rays are not homo-
geneous, except for tin and silver, but consist of a large number
of corpuscles which are absorbed in the first 001" of alumin-
ium, and a smaller number which are not so easily absorbed.

5. In all the metals tested, with one exception, the ratio of
the number of corpuscles produced in the metal to the number
produced in lead for equal absorptions of the primary rays
was greater for metals of greater density. In silver, however,
the hard primary rays pr oduced more corpuscles than in lead,
which has the greater density.

6. The ratio N’/N, depends partly. on the penetrating
power of the corpuscular rays and toa great extent on the
ratio of the absorption coefticients of the primary rays in lead
and the metal compared with it. Where the primary rays are
absorbed to a much Sa extent in lead than in the other
metal, the corpuscles will be produced much nearer to the sur-
face in lead, and consequently more wiil escape from the lead,
if the corpuscles from the lead are not too much less penetra-
ting than those from the other metal. A good instance of
this is silver, where more corpuscles are produced than in lead
by the har d. primary rays, but yet more get out of the lead
than the silver.

304 Cooksey—Corpuscular Rays, ete.

If the energy of the corpuscular secondary rays is due, as
Prof. J. J. Thomson* has suggested, to an explosion of the atom
and not directly to the energy in the primary pulse, it is hard
to explain why the more penetrating primary rays give rise to
corpuscles of higher velocity than the softer rays do; while
the intensity of the primary rays makes no difference in their
velocity. The author has been unable to explain this variation
in velocity with the nature of the primary pulse, but hopes
that further experiments, with modifications in the apparatus,
may throw more light on the subject of these corpuscular rays.

It gives me great pleasure to close by expressing my thanks
for the valuable advice received from Prof. H. A. Bumstead,
at whose suggestion this work was begun.

Sheffield Scientific School of Yale University,
New Haven, July, 1907.

* Conduction of Electricity through Gases, 2d ed, p- 820.

Verrill— Hereules Beetles from Dominica Island. 305

Art. XX X.— Description of a New Species or Sub-species of
Hercules Beetles from Dominica Island, B. W. L., with
notes on the habits and larve of the common species and other
beetles. Brief Contributions to Zoology from the Museum
of Yale Umversity, No. LX VII; by A. Hyatt Verritu.

Waite collecting a very large series of Dynastes Hercules,
consisting of sever al hundred specimens, during the past two
years in Dominiea, my attention has been frequently attracted
by the wide variation in coloration and markings exhibited by
this species. As a general rule, however, the shade of ground-
color, as well as the extent, shape and number of markings,
vary so interminably and so grade into one another that it is
impossible to separate any distinct color varieties which are
constant. Moreover, the ground-color of the elytra is almost
always of the same general color—a sort of straw or yellowish
olive—only variable in depth of color, or at the most, varying
from olive-ochre to greenish olive.

A few specimens have been obtained, however, which are
so very distinct from all others and so remarkable in their
variation from the normal color of Hercules that I have
deemed it advisable to describe them as a new and distinct
species or subspecies. In fact, if coloration alone constitutes
a specific character among the coleoptera, then this Dynastes
should be considered as a new species peculiar to Dominica.

DywastEs ARGENTATA, nov. Silvery Hercules Beetle. Male.
Size, general shape, processes of thorax and head, ete., indis-
tinguishable from Dynastes hercules. Head, thorax, lower
parts, legs, etc., deep jet-black, instead of dark brown or
purplish as in Hercules. Elytra pale plumbous, silver gray
or grayish white, with a silvery metallic sheen. Edges of
elytra all around, as well as a few large, circular scattered
spots, intense jet- black. Hair on the posterior extremity of
the abdomen, tibize, along edges of ventral segments and dorsal
segments of thorax, pale silvery gray or whitish. Velvet hairs
on lower surface of thoracic process deep orange brown or
ferruginous when viewed from below, but silvery white or
pale creamy white when seen from above or viewed in profile.
Hair on ventral surface of head and thorax deep ferruginous
or vandyke brown.

Length (type), including thoracic process, 6 inches. Length
of thoracic process, 3 inches.

Habitat, interior mountain ranges of Dominica Island,
Antilles.

306 =Verrill—LHercules Beetles from Dominica Island.

The species formerly described by me* as Dynastes tricornis
(fig. 1) should be referred to the allied genus Strategus, of.
which it is probably the largest known species. Its habits are
similar to those of Dynastes.

Fie. 1. Strategus tricornis, about nat. size. Phot. A. H. Ver.

Note on the habits and larve of the common Hercules beetle
(Dynastes Hercules), and other beetles.

The Hercules beetle is common in the forests of inland
mountains, where it is found in bunches of a dozen or more,
clinging to the branches and trunks of the “ La Glui” tree,
on the sap or gum of which it feeds. As they usually
congregate on the branches at a considerable elevation, they
are seldom seen by the casual observers. In order to secure
specimens it is usually necessary to shoot them, thus spoiling
a large proportion for scientific Pu Occasionally it is
possible to secure a regular supply of beetles by cutting down
a tree and scoring the trunks so that the sap runs freely. If
any beetles are in the vicinity they are soon attracted to the
felled tree and can be picked off by hand. Lights seem to
have no attraction for the Hercules beetles of Dominica. The
larva (fig. 2) is a huge grub, in general appearance resembling

* This Journal, vol. xxi, p. 317, 1906.

Verrill— Hercules Beetles from Dominica Island. 307

a gigantic New England June-beetle larva. It feeds on dead
and rotten wood of the La Glui tree, and requires several
years to reach maturity. The pupa case is formed from rotten
wood and the tough fibers of the bark. The imago when first
emerged has the thoracic and occipital appendages but partly
developed. They rapidly increase in size, however, and by
the time the elytra and other horny parts are hard, the appen-
dages have attained their full size.

oo
ow

Fia..2. Larva of Dynastes Hercules. Nat. size.
Fie. 8. Gru-gru Worm, the larva of Palm Weevil. 14 nat. size.

Of the two other new species, Dynastes vulcan and D.
Lagaii (loc. cit., fig. 2), the latter is bv far the rarest, only four
specimens having been obtained by the writer during three
years residence on the island.

D. vulcan (fig. 1, 6) is rarer than D. Hercules (fig. 1, a),
but far more common than D. Lagai, while of D. argentata
only one specimen is known.

Another noteworthy beetle of the island is the great Palm-
Weevil (?hynchophorus palmarum). It is of interest mainly
because the larvee are eaten by the natives. These larve
(fig. 3), which are locally called “ Gru-gru worms,” feed on
the decaying wood of the mountain Cabbage Palm and Gru-
* gru Palm. They attain a length of 2°15 to 3 inches when fully
grown. ‘They are fat, firm, legless grubs, and are considered
a great delicacy by the natives, as well as by many foreigners
who, having sufficiently overcome their natural repugnance,

308 Verrill—Hercules Beetles from Dominica Island.

have ventured to taste them. The larvee, after having been
spitted on the slender midribs of palm leaves, are roasted over
hot coals. Treated in this manner they pop open like roasted
chestnuts, and taste much the same. From personal experience
the writer can testify to the fact that they are excellent eating.

Of the other large coleoptera, Philemus didymus, a large
black Rhinoceros-beetle, and Weleus unicornis are the most
likely to be observed. The larvee of Phzlemus live in decaying
wood and other vegetable mnatter. The adult beetles have the
peculiar habit of rolling balls of rotten wood and leaves, in
which the eggs are deposited, much in the manner of the
northern “ Tumble-dung Beetles.”

The Weleus larvee live in rotten wood exclusively, and the
beetles at certain seasons become exceedingly abundant. This
beetle is remarkable for the peculiar loose-appearing manner
in which the head and thorax are joined to the bedy. Even
when living, the thorax wobbles about as if br oken and
entirely beyond the control of the insect.

A large undescribed species of Stag-beetles was also obtained
by the writer,

Most of the remaining beetles are small or so scarce as to be
but seldom met with, but special mention should be made of
the huge “ Fireflies” of the Elater group. These are abund-
ant in the forests of the interior, and to one who has never
seen them, the stories of their brilliancy seem incredible. One
of these insects flying about an ordinary room renders most
objects clearly visible at night, and fine print can be easily
read by holding one near the page. The light emanating from
these Firebeetles is continuous and issues from various por-
tions of the body, especially from between the abdominal seg-
ments and from underneath the elytra.

C. Barus— Successive Cycles of Coronas. 309

Arr. XXXI.— On the Successive Cycles of Coronas, by
C. Barus.

1. Introductory.—Using the method of successive identical
exhaustions or withdrawal of nuclei, by which the nucleations
left in the fog chamber should decrease in geometrical progres-
sion® from the high initial value given to dust-free air by the
emanations of phosphorus, I have recently attempted to gain
an estimate of the trustworthiness of the results, by using drops
of pressure (67 from atmospheric p) of widely different ‘values.
These range from one extreme 6p = 10 to the other 6p = 20™.
Leaving the data, for which there is no room here, for publica-
tion elsewhere,+ it appears that if the diameter of particle d
of a given type and order of corona, worked out from the vari-
ables and constants of the experiment, agree reasonably well,
the general method of reduction is assured.

The whole of the work was done in time series. Thus the
exhaustion cock between the vacuum and fog chambers was
kept open five seconds after exhaustion, only, and the time
between two successive exhaustions was one minnte. Other
intervals were also tested without modifying the results. The
time during which the cock remained open is of special impor-
tance, for the exchange of air between the two chambers
depends upon it, and hence also the rate at which nuclei are
removed per exhaustion. These details were determined in
preliminary experiments. The isothermal pressure in the fog
chamber after five seconds of communication between the
chambers, obviously differs in marked degree from the isother-
mal pressure if the chambers are permanently left open in
communication.

Additional correction was made for subsidence of fog par-
ticles during the short time, in all about fifteen seconds, needed
for the observation of the coronas. They were then at once
dispelled by opening the filter cock, evaporating the fog par-
ticles and restoring the nuclei left behind to their original size.
Other losses, as for instance the evaporation loss which oceurs
when fog particles are precipitated on ions or on the vapor
nuclei of dust-free air, or the time loss (decay), were found to
be negligible. The amount of water precipitated, however,
varies both with the adiabatic drop of pressure and with initial
temperature, all of which was carefully allowed for. ©The
ratio of sections of the fog chamber and the exhaust pipe was
about 6/1, sufficient to reach the highest coronas. In fact the
large red type of the first series was certainly obtained.

* This Journal, xiii, pp. 81-94, 1902.
+ To be given in a forthcoming Carnegie publication.

310 C. Barus—Successive Cycles of Coronas.

. The violet and green coronas.—As has been suggested,
ae ‘object of the series of experiments made at very low
exhaustions (drop of pressure 6p = 10™) and compared with a
series for high exhaustions (69 = 20°5°") was an estimation of
the impor tance of the time effect and of the convective effect
in causing loss of nuclei. If the latter series be reduced to
the former by modifying the constants in terms of pressure
and temperature, the coincidence of the graphs was found to
be complete. This indicates that the method of reduction is
reliable.

To determine in how far the results themselves are trust-
worthy, it will be necessary to find the computed values in the
different independent series of measurements, of the diameter
of the fog particles producing a given corona. For this pur-
pose the violet (v) and green coronas (g) are suitable, the red
(7) less"so. Where are three of the green coronas, ‘the two
upper being very brilliant. In a former paper the diameters
of particles were estimated as d = -000460™ for the middle
green corona. Ratios of 4, 38, 2 were usually apparent, the
data being multiples of a ‘diameter something larger than
d = 00015, the corona for which is not producible i in the fog
chamber. In the present experiments the values of the diam-
eter of fog particle d and s, the aperture of the green coronas,
still show considerable fluctuation ; but the data approach
closely to a common mean, remembering that the color itself
has a certain latitude, necessarily, and wide differences of
exhaustion are involved in the tests. The ratio 2, 3, 4 of
diameters of fog particles is not as well suggested in the pres-
eat results as in the former, while the absolute sizes themselves
are throughout smaller. It is nevertheless convenient to retain
these ratios for the division of coronas into successive cycles.
If these may be considered as beginning with deep red and
ending with violet, the following cycles may be postulated, the
subscripts denoting the order:

®,,(@="00011) v,, d= 00019 vd = 00033 vd —=-0004n
Gad — 0003) sons 23m ge 40 4, Ben
PA — 000 NG) vrs Sey Vee 64°"

Only the red and crimson of the first series are certainly
observable in any apparatus known to me. Their aperture,
5/30, is about 40 degrees, their rings diffuse and their disc filmy,
so that in a small appar atus they ‘night be mistaken for clear
air. The second series is producible and vivid throughout,
and the same is even more true of the third. The fourth is
already closely packed, while the fifth and subsequent series
merge into each other too rapidly for separation.

C. Barus—Successive Cycles of Coronas. 311

Cycles 3 and 4 were obtained in great number in my work
with atmospheric nucleation. Selecting some twenty or more
cases, the mean ratio 1/s,:1/s, = 146 ::206 =d,:d,. Hence
the ratio of 3:4 is very well sustained. The voniometer dis-
tance from the fog chamber was nearly a meter in this case.
Hence it is probable that the cyele 1 is actually the first occur-
ring, although the smallest active particles (violet) must exceed
‘0001™ in diameter. The same terminal conditions are sug-
gested by the axial colors of the steam jet. It seems curious
that the diffraction phenomenon should begin with particles of
the order of three times the wave length of light.

Using the method of contact of coronas from two sources
described elsewhere,* the ratio of diameters of the four first
series is again very ay as 1, 2, 8, 4, for the green coronas,
for instance.

3. Wilson’st data and conelusions.—The table contains
Wilson’s exhaustions (v,/v)* at 18° to 19° C. and the corre-
sponding disc colors, as | interpret them. It also contains the
Estimate of the diameter of nuclei d and their number n per cm? corre-

sponding to Wilson’s colors of cloudy condensation in wet dust-free air.
Temp. 18°-19°.

ALO Bis Dise iO Perea Bes 1Om® OY
v/v Op/p color ds, 2 dla, 4 Na, 2 Ne, 1
1°410 384 green 40 23 ae "900
1°413 386 as ve ies oo Bip
1°419 390 violet 33 19 27 1°5
1°420 390 6 ape Ep pat ia
1°426 394 red 32 16 32 2°5
1:469 418 greenish white 23 2 "90 7:0

equivalent relative drop of pressure 6p/p used above. From
these and the colors of diameters of fog particles (d) may be
estimated, very roughly of course, provided the cyele in which
these colors lie is known ; hence dz. refers to the probable
case of the occurrence of pane third and second ser ies, 724 to
the very improbable case of the occurrence of the second and
first series. If the values m be found for the corresponding
temperature and expansions (6p/p), the nucleations v3. and
ma, respectively follow. Wilson gives but a single cycle
between green coronas. There are ae such series and three
detinite large green coronas producible, and I shall assume
that the very vivid upper one is meant. The first cycle is not
producible by any means known to me, except in its lower
red coronas. Ignoring mo, and taking 739, the data are dis-

* This Journal, xxiv, p. 277, 1907.
+ Phil. Trans. Royal Soc., vol. 189, pp. 265, 1897; cf. p. 285.

312 C. Barus essive Cycles of Coronas.

tributed similarly to my own so far as the slope of the curves
for the vapor nucleation of dust-free air is concerned.

There is another way in which the estimate in question may
be made. Let the nucleations corresponding to the colors be
taken and reduction be made for the different drops of pres-
sure in question. This is merely a corroboration of the method
of computation. The evincidence is as close as may be ex-
pected, seeing that the methods of approach are widely. differ-
ent and that the nucleation varies as the cube of the inverse
diameter of particle.

Wilson’s views of the nature of the phenomena are quite
different and lead to enormous nucleations, even as compared
with the improbable m2;. He says (le: p. 3801) “When all
diffraction colors disappear eed the fog appears white from all
points of view, as it does when (the expansion) ¥,/v, amounts
to about 1:44, we cannot be far wrong in assuming that the
diameter of the drops does not exceed one wave leneth i in the
brightest part of the spectrum, that is, about 5x10-* centim.
That the absence of color is not due to the inequality of the
drops is evident from the fact that the colors are at their
brightest when v,/v, is only slightly less than 1-44, and from
the perfect regularity of the color changes up to this point.

Taking the diameter of the drops as 5x107-° centim., we
obtain for the volume of each drop about 6X10~™ cub. ‘cen-
tim., or its mass is 6X 10-" erm.

Now, we have seen that when the expansion is such as pro-
duces the sensitive tint (when v,/v,= 1°42), the quantity of
water which separates out is about 7:6 107° grm. in each cub.
centim. With greater expansion rather more must separate
out. We therefore obtain as an inferior mit the number of
drops, when v,/v, = 1:44, 76 X107°/6X10-" = 10° per cub.
centim.”

In my data the smallest green coronas correspond to a
diameter of particles of about d, = 00052, the next to
d, = (00040, the next to d, = -00023; the first (which © have
not been able to produce by any means whatever, however
large the nuclei) should correspond to d, = 00013", and even
this calls for particles nearly three times as large as Wilson’s
estimate, 00005. Ina small tube but 2™ in “diameter, like
Wilson’s test tube apparatus, it is improbable that the d, green
corona, which is about twenty-seven degrees in angular diame-
war, could look otherwise than greenish white, whereas the
filmy dise of the large crimson coronas (the largest produci-
ble d, = 00016 with an angular diameter of about 39 degrees)
would be regarded colorless. I shall venture to believe there-
fore that Wilson’s large greenish-white coronas corresponded
to about -9X10° rather than to 10° nuclei per cubic centimeter,
that the maximum nucleation would not exceed 10’ even if
colors of the unapproachable first order were produced.

DL. Randall—Behavior of Molybdic Acid. 313

Art. XXXII.— The Behavior of Molybdic Acid in the Zinc
Reductor; by D. I.. Ranpatt.

[Contributions from the Kent Chemical Laboratory of Yale Univ.—clxiv. ]

TuereE has been a difference of opinion in regard to the
degree of reduction obtained when molybdic trioxide, in sul-
phuric acid solution, is passed through the column of amalga-
mated zinc as applied i in the Jones reductor. Jones,* who first
determined molybdenum by this method, considered that the
reduction goes to the condition represented by the formula
IMox@: the same degree of reduction that Wernket obtained
with zine and sulphuric acid in a closed flask. Doolittlet and
Eavenson found that, by varying the strength of the acid and the
speed at which the ‘molybdenum was passed through, differ-
ent degrees of reduction might be obtained. They calculated
no formula corresponding to the degree of reduction obtained
by them which was lower than that represented by Mo,,O,,.
W. A. Noyes and Frohman,§ by taking pains to replace the air in
the reductor flask by carbon dioxide, were able to obtain a
reduction to the form of Mo,O,. Blair} :
ever, were unable to press the reduction below the condition
represented by the symbol Mo,,O,,. Miller] and Frank in
repeating the experiments of Blair and Whitfield obtained in
general the same results, though by taking extraordinary pre-
cantions they were able to get a reduction to a point a little
below midway between the conditions represented by the sym-
bols Mo,,O,, and Mo,O,.

The ease with which the reduced molybdenum solution is
oxidized by the air has been noted and suggests the possibility
of charging the reductor flask with some oxidizer unaffected
by air, which shall anticipate the oxidizing effects of the air as
the reduced molybdenum compound comes through the redue-
tor and register the oxidation.

In the first experiments an excess of a standard solution of
potassium permanganate was used in the receiver. But it was
found, by running blanks, that somewhat more than the theoret-
ical amount of permanganate was used up. Table I shows the
results of experiments made by charging the reductor flask
with 20™° of approximately tenth normal standard potassium
permanganate, passing 300%" of warm dilute (2°5 per cent)
sulphuric acid through the reductor, adding 20° of standard
ferrous sulphate solution and titrating with permanganate to
color.

* American Institute Mining Engineers, xviii, 705.

+ Zeitschr. Anal. Chem., xiv, 1. ¢ Jour. Amer. Chem. Soce., xvi, 284.

3 Jour. Amer. Chem. Soc., xvi, 5538. ~—‘|| Ibid., xvii, 747.
{| Ibid., xxv, 919.

314 D. L. Randall—Behavior of Molybdic Acid.

TABLE I.
KMn0, KMnO,;
FeSO, used theory Difference
em’. em’, em’, em’,
20 24°45 23°25 1°20
20 23°88 23°25 63
20 25°40 23°25 2°15
20 24°90 23°25 1°65

This excessive reducing action on the permanganate must be
due either tc impurities in the zine, to small particles of zine
worked through the asbestos layer at the bottom of the reduc-
tor, or possibly to the hydrogen formed in the reductor. That
the last mentioned possibility is a sufficient cause of the effects
obtained was shown by passing hydrogen, formed by the action
of hydrochloric acid on zine in a Kipp : generator and washed
with water and caustic potash, through 23° standard perman-
ganate diluted with 300° of hot dilute (2°5 per cent) sul-
phuric acid, adding 20°™* of a solution of standard ferrous
sulphate and titrating back with permanganate. The results

obtained by such action of hydrogen during fifteen minutes is
shown in Table II.

TABLE it
KMn0O, KMn0O, Reduced by
FeSO, required theory hydrogen
cm?, em, em’, em?,
20 24°6 2301 11259
20 Daley Zeal 1°6
20 23°6 Zi Sill O55

Having seen that potassium permanganate could not be used
in the reductor flask, ferric alum was next tried as ‘an oxidizer.
The only objection to the use of this salt, that at the usual
dilution of 500-600" an excess of ferric alum colors the solu-
tion strongly and makes the end point in the titration some-
what difficult to recognize, is overcome by the addition of a
few centimeters of phosphoric acid which renders the solution
colorless.

For the work with the ferric oxidizer the C. P. ammonium
molybdate of commerce was recrystallized and found to contain
81:55 per cent of molybdenum trioxide by treating weighed
portions ina platinum crucible with nitric acid and evaporating
to dryness on the steam bath and igniting gently and weighing
the molybdic anhydride formed. The ferric iron solution for
use in the reductor flask was prepared by dissolving 100 grams
of the crystalized salt in one liter of water. The column of
zine in the reductor was 36™ in length. The procedure was
to charge the reductor flask with 20 or 30% of the iron solu-
tion and 4° of phosphoric acid and to pass through the redue-

D, L. Randall— Behavior of Molybdic Acid. 315

tor in succession 100° of hot dilute sulphuric acid (2°5 per
cent), the ammonium molybdate dissolved in 10°™* of water and
_acidified with 100°* of the hot dilute acid (2°5 per cent),
200" of the hot dilute acid and finally 100°™° of hot water.
The solution was titrated while still hot with approximately
tenth normal permanganate standardized against ammonium
oxalate.

The molybdenum as it passed through the lower part of the
reductor was green, but on coming in contact with the ferric
salt was changed to a bright red color. The results in the fol-
lowing table are calculated on the assumption that the molyb-
denum was reduced to the form of Mo,O,. The close
agreement with theory indicates that the reductor is able to
reduce molybdenum to that form and that the reduction does
not stop at a point represented by the formula Mo,,O,, or
MoO

TaBueE III.

Ammonium

molybdate Iron KMnO, MoO;
taken sol. H;PO, used found theory Error
grams. em’, cme. em’, grms. grms.
"2000 20 4 30°38 1628 1631 — "0003
.2000 20 4 30.45 "1632 "1631 +.0001
*2000 20 4 30°35 1626 1631 --"0005
*3000 30 4 45°65 "2447 "2447 +:0000
“3000 30 4 45°80 °2455 "2447 +0008
*3000 30 4 45°73 "2451 - :2447 +0004
“3000 30 4 45°76 °24538 "2447 +0006
°3000 30 4 45°68 "2448 "2447 + -0001
°3000 30 aa 45°64 °2446 "2447 — 0001
*3000 30 A A Seri "2449 "2447 +:°0002

It having been shown that molybdenum may be successfully
and uniformly reduced to the condition of Mo,O, by anticipa-
ting any previous action of air and registering the oxidation
by ‘the action of a ferric salt, the method was ‘then applied to
the determination of phosphorus in ammonium phospho-
molybdate precipitated in the usual manner.

For this work a solution of microcosmic salt was made up
1:2 grams to the liter and the phosphorus content determined
by evaporating 50° in a platinum crucible and igniting to form
the sodium metaphosphate.

With the exception of using an equivalent amount of ammo-
nium molybdate instead of molybdie anhydride, the method
given by Blair* was used for making up the moly bdate reagent.
According to this method, 100 grains of pure molybdie anhy-
dride and 400™* cold distilled water are thoroughly mixed and
the mixture is treated with 80™ of strong ammonia (0°90
sp. gr.). When the solution is complete, the liquid is filtered

* Chem. Analysis of Iron, vi ed., p. 97.

316 D. L. Randall—Behavior of Molybdic Acid.

slowly and with constant stirring into a mixture of 400
of strong nitric acid (1°42 sp. gr.) and 600° distilled water.
Then 50 milligrams of microcosmic salt dissolved in a little
water are added, the mixture is agitated thoroughly, the pre-
cipitate is allowed to settle for twenty-four hours and the solu-
tion is filtered before using.

In order to have iron present during the precipitation of the
phospho-molybdate, a solution of ferric nitrate was prepared
and used in each experiment. A solution of the C. P.
ferric nitrate of commerce, Fe(NO,),9H,O, was made up to
contain 150 grams to the liter, with enough nitric acid added
to give the solution an amber color. To free the iron solution
from traces of phosphorus, 400° of the molybdenum solution
were added to the ferric nitrate and the whole well shaken and
allowed to stand for twenty-four hours.

To measured portions of the phosphorus solution, in a 300°"
Erlenmeyer flask, were added 150°™* of the iron solution fil-
tered and warmed to 35° C. The flask was stoppered
with a rubber stopper and shaken for five minutes. The
precipitated phospho-molybdate was allowed to settle, and was
then filtered on asbestos in a perforated crucible and washed
with a solution of ammonium acid sulphate (15°™* ammonia,
25°™* sulphuric acid, 1 liter water). The Erlenmeyer was
washed out with a solution of 20° of water and 5°™ of
ammonia, and this was poured on the asbestos in the crucible.
The molybdenum solution was acidified with 10°™* of strong sul-
phuric acid and was passed through the reductor into the ferric
alum solution. The molybdenum solution was preceded by
100°" of hot water and followed by 200° of the hot dilute acid
and 100°™ of water. The reduced solution was titrated imme-
diately with approximately tenth normal permanganate. In
the following table are shown results thus obtained in precipi-
tating phosphorus as the phospho-molybdate, passing the acidic
solution through the reductor into a solution of ferric alum,
and titrating the ferrous salt which is formed.

TABLE IV. 4

P taken P found Error

grms. grms. grms.
"003645 "003673 + °000028
003645 "003697 + :000052
003645 "003638 —:000007
"003645 "003726 +:000081
003645 "003630 — 000015
003645 003661 +:000016

These results are calculated on the assumption that the ammo-
nium phospho-molybdate contains phosphorus and molybde-
num in the proportion given by the symbol (NH,),12Mo00,PO,,
and that the reduction proceeds to the condition represented
by the symbol Mo,O,.

Am. Jour. Sci., Vol. XXIV, 1907.

Plate |

4

FE. Wright—Measurement of the Optic Axial Angle. 317

Art. XX XIII.—The Measurement of the Optic Awial Angle
of Minerals in the Thin Section, by Frep Kucrne Wricur.
(With Plates I and II.)

Durine the past few decades the development of the petro-
graphic microscope and petrographic methods has been exceed-
ingly rapid and has now attained a state of such high efficiency
that nearly every optic constant of a rock mineral in the thin
section can be ascertained with ease and some degree of cer-
tainty. In many instances, in fact, the petrographer has at pres-
ent at his disposal two and even more methods for the determina-
tion of a single optical constant under the microscope. The
results obtained by these different methods are, however, often
not of the same value, nor are all methods equally applicable to
a given mineral section. Com parative and critical studies of the
relative merits and accuracy of such methods under different
conditions are not abundant in petrographic literature. It has
therefore seemed to the writer desirable that such a study
should be undertaken, and the microscopic examination of
the artificial mineral ageregates produced in the Geophysical
Laboratory has furnished a favorable opportunity. It has been
found necessary in this work touse most of the available pet-
rographic microscopic methods and to test them under different
conditions.

Minerals are determined under the microscope by means of
their crystallographic and optic properties; the more accurately
such properties or constants can be ascertained for any given
mineral, the more reliable and satisfactory is the determination.
Of the several optical constants which are thus made use of in
the identification of minerals, the optic axial angle, and with
it the optical character, whether positive or negative, are among
the most important. Different methods have been suggested,
from time to time, by means of which the optic axial angle
of any given biaxial mineral can be measured with greater or
less accuracy and more or less rapidly ; certain of these methods
are of general application and can be employed on all mineral
sections, while others can be used only on sections of fixed
orientation. In microscopic work, however, all possible sections
are to be encountered, and in each par ticular case, that method
is the best and should be selected which not only gives the
most accurate results, but also requires the least time. This is
particularly true of Ge fine-grained preparations where
it is frequently difficult to obtain results of any value whatever.
In the following pages, the methods for measuring optic axial

Am, Jour. Sci.—FourtTH SERIES, Vou. XXIV, No. 142.—Ocrosrr, 1907.
29

318 FE. Wright—Measurement of the Optic Axial Angle

angles of minerals in thin section are to be described briefly,
together with several novel methods which have been found
serviceable, and a summary given of the results of comparative

tests of all these methods on mineral sections of known and
measured optic axial angles. As many of these methods have
not been described in English, it has seemed desirable to intro-
duce the general discussion by a mention of their underlying
principles “and an outline of their mode of application to differ-
ent sections.

wy) MII INAS

Le HIN \\

G
eZ A ff

WAAL WA
\\RBCAELRRURR ERE
QA
WAAAY

LT ETL LL,
HTL LL
LH HI YY SA
ITLL LL.
DQITYELLLLYLLLY YY
LI LELLLAL LL LLG
DELL LLL JHE
[i] EEA ELED ELE
ae CILLLILMLE ha
LY WL LL ip
BOT Le
hE s

Sy
SESS

Ws
ot ee,

AV TT LAE
SME Sriranee

Several of the methods are in part graphical methods and
involve the use of stereographic, orthographic or gnomonic
projection plats. (Plate I and fig. 1.)* A few paragraphs,

* Plate Lis a photolithographic production of the stereographic projection
plat by Professor G. Wulff (Zeitschr. f. Kryst., xxxvi, 14-15, 1902). Both
great circle and small circle ares are located two degrees apart in the pro-
jection plat, the great circle arcs having a radius R, fora given angle p, Ri=
aa r being the radius of the projection plat; the small circle arcs, a
diameter, Re=r cot p. In the projection plat, these arcs have been con-

of Minerals in the Thin Section. 319

therefore, on these different methods of projection, their
essential characteristics, and their application to microscopic
work, are inserted at this point because of the important role
they play in the theoretical consideration and elucidation of
the optical methods to be given below.

The spacial relations between the optic and morphologic
properties of a crystal are frequently complicated and extremely
difficult of correct conception and description without the aid
of special models or projections of the same. Both observation
and theory have shown that for any crystal the optical phe-
nomena which it presents can be ascertained for a particular
color of light, both in direction and length, by reference to an
ellipsoid, either triaxial ellipsoid or ellipsoid of rotation, or
sphere in the limiting case. Having once determined "the
exact position and character of this ellipsoid within the crystal
and the relative lengths of its axes, it is possible to figure
mathematically and to represent ovaphically the optical proper-
ties of the crystal for light waves of the specified length
transmitted in any given direction. In the general treatment
of the optical properties of crystals, the source of light is
considered to be located at a central point within the crystal
and the light waves to emerge in all directions from that central
point along the radii of an inscribed sphere of unit radius. In
space any radius can be represented accurately by its intersec-
tion with the surface of the unit sphere, and like any point on
the earth’s surface, its position can be fixed accurately by two
angles equivalent to those of latitude and longitude.

structed with great accuracy and measurements can be made directly by their
use with errors of less than }°. Having given the stereographic and ortho-
graphic plats of Plate I and fig. 1, the process of plotting angles from them
graphically and measuring angular distances is rendered easy ‘and certain by
the use of tracing paper, as first suggested by Professor Wulff. With the
plat as a base, the measured angles, obtained from the microscope, are
plotted directly on tracing paper from the projection plat underneath, and
the various necessary operations of passing great circles through given points
and measuring angular distances, etc., accomplished by revolving the sheet
of tracing paper about the center of the projection plat as a pivot, until the
required great or small circle of the projection plat beneath is found. The
circle is then sketched with sharp pencil point on tracing paper and is suffi-
ciently accurate for practical purposes, the slight errors produced by using
tracing paper and stereographic plat being of the same order of magnitude as
those of observation and therefore negligible. The amount of time and energy,
moreover, expended in this method of free-hand sketching, is much less than
that necessary for constructing accurately the required small and great
circles. In actual practice it has been found expedient to draw with colored
ink, on both stereographic and orthographic projection plats, radii and circles
2° apart, and thus greatly facilitate the ease with which observed data can be
plotted. To incorporate these radii and circles in the original plats did not
seem advisable because of the great complexity of lines which arises when
radii, circles and great and small projection ares are all superimposed in
black ink. By the use of colored ink, no confusion arises and the radii and
circles can be added at any time with little effort by the aid of rule and
compass.

320) FE Wright—Measurement of the Optic Axial Angle

As a general rule, it is neither convenient nor feasible to
work with actual models in the study of optical phenomena.
Several different types of projection have been devised to
overcome this difficulty by representing the spacial relations
ona plane. In all cases the relation of the object to its pro-
jection plot is one of definite construction, and is dependent
on the method of projection adopted.

In optical work, any one of three different methods of pro-
jection, the gnomonie, the orthographic, and the stereographie,
may be used, each of which possesses certain favorable and
certain unfavorable features. In each projection, the points
on the sphere are pictured in a fixed plane.

In the gnomonic projection,* the plane of projection is the
horizontal tangent plane passing through the crest of the unit
sphere. (fig. 2.) The point of intersection. of the radius

PLANE oF -
7
BTERECARAPAIC AND ORTHOGRAPHIC PROJECTIONS

SeHU

fe)

In this figure, the point D is the gnomonic projection point of the
point P on the sphere, or cf the direction CP in the erystal. The distance
MD is evidently r. tan p, r being the radius of the sphere and p the angle
MCD. Similarly, the distance CE in the stereographic projection isr tan
- and CF in the orthographic projection, r sin p.
drawn through any given point on the sphere with the tangent
plane is the projection point for that direction. In this pro-
jection, the great circles of the sphere become straight lines in
the plat, and are small circles, hyperbolae. In the following
pages, however, the gnomonic projection will not be used to
any extent and will not therefore be described here in greater
detail. It is particularly well suited to crystallogr aphie work,

*For a comparative study of projections, see V. Goldschmidt, Ueber
Projektion und graphische Krystallberechnung, Berlin, 1887.

of Minerals in the Thin Section. 321

since by its use all crystal faces are reduced to points and zones
to straight lines.

In the orthographic (also called orthogonal or parallel or
ocular) projection, the eye of the observer is supposed to be at
an infinite distance above the plane of projection and to look
directly down upon the sphere. The lines of sight are then
parallel and the points on the sphere are vertically above their
projection points on the central diametral plane (figs. 2 and 3).
In this projection, great circles appear as ellipses and small
circles as straight lines. This projection is especially impor-

)
Vv
ia

N

In fig. 3 the point P of the sphere, which is located in this case by
the intersection of the great circle ATP and the small circle DPK,
in the orthographic projection F and is there located by the ellipse AHF,
the orthographic projection of ATP, and the straight line DFL, the
projection of DPK. F is also the point of intersection of the diametral
plane CGB with the line PF, normal through P to that plane.

tant since all interference phenomena observed in convergent
polarized light under the microscope appear to the eye of the
observer as they would were the actual interference phenomena
plotted in this projection. The serious drawback to its general
application in optical work hes in the rapid decrease of its
sensitiveness to differences in angular distances near its outer
margin. The interference figures below are all represented in
orthographic projection, although their construction was
accomplished in part by using the stereographic projection.
In fig. 1, the ellipses represent great circles with a common
diameter of intersection and drawn at intervals of 2° apart;
while the straight lines are the projections of small circles, also
2° apart and corresponding to small circles of latitude. As on
the sphere itself, the angular distance between any two points
in projection can be found by passing through the two points
the common great circle (ellipse in projection) and counting
directly the distance in degrees by means of the small circles.

322 F. EB. Wright—Measurement of the Optic Axial Angle

The actual modus operandi of this and the stereographie pro-
jection will appear more clearly later when actual data of
observation are plotted.

In the stereographic projection,* the eye of the observer is
considered at the lower extremity of the vertical radius, the
lines of sight being directed toward the points of intersection
of given radii with the surface of the sphere. The intersection
of the line of sight for a given radius or direction of light wave
propagation with the central horizontal plane determines then
the stereographic projection point of that radius. (Plate I
and figs. 2 and 4.)

In this figure the point P of the sphere, corresponding to the direc-
tion CP within the crystal and located in this instance by the great circle
ATP and the small circle DPK, becomes E in the stereographic projection
plat and is there located at the intersection of the great circle arc AHE,
the stereographic projection of ATP, and by the small circle arc DEL, the
stereographic projection of DPK. LE is also the point of intersection of the
line OP with the horizontal diametral plane CGB.

The stereographic projection is unique in that all circles,
whether great or small, appear in the projection as circles
instead of ellipses, as might be supposed at first thought. The
angle, moreover, which two great circles make with each other,
is preserved unaltered in the projection. The projection is
thus angle-true. In Plate I, the portions of great circles of
the upper half of the sphere are represented by the circular
ares of which the horizontal radius is the limiting case, and
the small circles by the ares of which the vertical radius is the
limiting case.

Biot’s or Fresnel’s rule-—In several of the optic methods to
be deseribed, frequent use is made of a rule first formulated

*S. L. Penfield, this Journal (4), xi, 1,115; E. Fedorow, Zeitschr. Kryst.,

vols. xxvi, xxvii, xxix, and G. Wulff, Zeitschr. Kryst., xxi, 249, 1893; xxxvi,
14-15, 1902, have given complete descriptions of the stereographie projection.

of Minerals in the Thin Section. Bho

by Biot* by which the directions of extinction for any section
of a birefracting mineral can be found. Some ten years before
Biot announced this general rule, Malust had found that the
light waves emerging ; from a calcite rhomb were plane polar-
ized and that for any given section of calcite the lines of
extinction were parallel and at right angles with the trace of
the plane containing the optic axis and the normal to the sec-
tion ; in other words, the orthogonal projection of the optic
axis on any given section of a uniaxial mineral determines its
lines of extinction which are parallel with and normal to this
projection line. By modifying the wording of this rule
slightly, it is possible, as Biot proved experimentally and Fres-
nelt demonstrated theoretically, to make it of general appli-
cation to all birefracting substances; thus, the directions of
extinction of a biaxial mineral§ cut after any section are par-
allel to the traces, on that section. of the planes bisecting the
angles between the two planes c¢ containing the normal to the
section and the optic axes (or optic binormals); in other words,
the lines bisecting the angles between the lines of orthogonal
projection of the optic binormals on any given section of a
biaxial mineral are the directions of extinction for that section
for the particular color of hight employed. It should be noted
that this rule apples to optic phenomena within the crystal
itself, and that for oblique incidence of light, as in convergent
polarized hight, the apparent angles observed in air must be
reduced to true angles by means of the average refractive

index of the mineral in question; by the formula sm V=
sin E

ome
and 8 the mean refractive index of the mineral.

The above methods of projection, together with the Biot-
Fresnel rule of construction for finding the directions of
extinction on any plane, constitute the basis on which several
of the methods described below rest. With practice, these
projections become working models for the observer and aid
him very materially in grasping and picturing the optical
phenomena presented by different minerals under varying
conditions.

* Biot, J. B., Mém. de l’Acad. de l’Inst. de France, iii, 228, 1820.

+ Malus, E. L., Théorie de la double réfraction de la lumiére dans les sub-
stances cristallisées, Mém. prés & l’Inst. Sc. math. et phys., ii, 303, 1811.
Sane Second Mémoire sur la double réfraction, Pogg. Ann., xxiii,
t Bs)

§'The uniaxial minerals may for the moment be considered the limiting
case of biaxial minerals for which 2V = 0.

, EK being the observed angle, V the required true angle,

B24 FE. Wright—Measurement of the Optic Axial Angle

Part J. THEORETICAL.

In this part of the paper, the attempt has been made to
describe briefly the methods, at present known to the writer,
for measuring the optic axial angle of mineral sections under
the microscope and to emphasize particularly their modus
operandi and the principles on which they are founded. In
the preparation of these descriptions constant use has been
made of the Mikroskopische Physiographie, I, 1, by Rosen-
busch and Wiilfing, and of the original papers in which the
methods were first ‘mentioned, especially those by Becke and
Fedorow. Several of the text-fig ures below have, in fact, been
derived, with shght modifications, from these sources.

Convergent Polarized Light.

Optical phenomena in general can be observed and be meas-
ured by use of either convergent or plane polarized light. The
study of optic axial interference figures has, however, until
recently, been effected solely in convergent polarized light,
and the methods applicable thereto are better known and more
fully developed than those for plane polarized light. We shall,
therefore, begin with the methods for measuring the optic
axial angle of minerals in conve ergent polarized light.

There are several different lens combinations which can be
used to advantage for obtaining and observing interference
ficures under the microscope in convergent polarized hght;
and of these, the one suggested by Amici® in 1830 has been
found to be the best suited for optical measurements. With
this method, the primary interference image, which is formed
directly above the high power objective, is magnified and
reproduced as a secondary image in the upper part of the
microscope tube, where it can be observed either with magni-
fying glass or ocular with cross hair and micrometer scale.
The small Amici-Bertrand lens by means of which this change
of microscope to conoscope is effected must be inserted at such
a point between ocular and objective that the secondary inter-
ference image observed through the ocular is as sharp and
clear as possible.

Both theory and observation show, however, that all parts of
the interference figure thus formed are not in perfect focust
at one and the same time. Fig. 5 illustrates the path of a light
beam from condenser lens to eye of observer. From this
figure, it is evident that the surface of focus for light waves
entering in all possible directions is not a plane but roughly

* Ann. Chim. Phys. (8), xii, 114, 1844.
+ Rosenbusch-Wilfing, Mikroskop. Phys., I, 1, 215, 1904.

of Minerals in the Thin Section. 325

a sphere. It is necessary, therefore, to focus sharply on
points about midway between the center and margin of the

Fic. 5. In this figure, C represents the condenser lens system; O that
of the objective, B, the Bertrand-Amici lens, P, the mineral plate and Oc,
the ocular of the Ramsden-Czapski type. The primary image EF (d) is
reproduced as a secondary image at E’F’ (d’) and later still more magnified
by the ocular at E"F"” (D), where it is observed by the eye back of the small
aperture S. In the writer’s microscope with Fuess No. 9 objective, of
equivalent focal length, 2°7™", and numerical aperture 0°97, the diameter
of the primary interference figure measures about 5™™ and this is magnified
very considerably by the Bertrand lens and by the ocular in the secondary
interference figure as shown in the figure. By choosing a Bertrand lens of
suitable focal length and inserting the same at the proper point between
objective and ocular, it is possible to increase the magnification of the
secondary interference figure and thus to enhance the accuracy of the
measurements, providing the interference figure is sufficiently sharp and
distinct to permit magnification to such an extent.

326 FB Wright— Measurement of the Optic Axial Angle

field (F of the figure) and to adjust the lenses for these points
in order that the remainder of the field be as distinct as possi-
ble under the circumstances.

In order to minimize the error due to the shght parallax of

rays near the margin of the field, a small stop, 8, may be
placed above the ocular as shown in the figure.

The exact position of any point in a given interference
figure can be determined with considerable accuracy by using
either a micrometer ocular with fixed scales or movable screw
micrometers, or by means of a projection by camera lucida
on paper. The actual position in space of any point thus fixed
in the projection can then be stated if the relation between
projection and object be known.

In 1882, Mallard* proved theoretically that the interference
figure observed in the microscope is approximately an ortho-
grapbie projection of the optic phenomena in space. To take
the simplest case, let both condenser lens and objective consist
each of a simple hemispherical plane convex lens, C and O,
fig. 5. The rays which in the objective lens are parallel come
to a focus at F, where they can be viewed directly by the
unaided eye. As suming the radius of the lens to be 7, the

radius R (MF, fig. 5) of the focal circle is evidently ——_, where

nm is the refractive index of the objective lens glass. It should
be noted that under the nye seone the optic phenomena
are observed as they appear in the objective itself, i. e., modi-
fied by their refraction in the glass. The distance

ER = (de Phasing so Min (1)

Replacing x EMF by its corresponding angle in air, E, the
relation between which is expressed by sin E = n. sin X

EMF, equation (1) can be written
Me = @ = a sin E. (2)

By fixing rigidly the relative positions of objective, Amici-
Bertrand lens and ocular (fig. 5), the distance d@ for all angular
values E, which can be measured in the microscopic ‘field,
bears a constant relation to D of the secondary interference
figure

D=kd (3)
This distance D can be measured accurately with the movable

screw micrometer ocular. Substituting D for d, equation (2)
becomes

*E. Mallard, Bull. Soc. Min. Fr., v., 77--87, 1882.

eo
S
“I

of Minerals in the Thin Section.

R :
D = Rae ay Es - Sin K,
an equation which can be simplified to
Dy Ke sin, B (4)

by replacing an expression consisting entirely of con-

One
stants, by K, the Mallard constant of the microscope for the
given fixed position of microscopic lens system. The constant
K can be determined once for all by ascertaining the D for
any known angle KE. The method usually employed to deter-
mine this constant K is to measure, on a section of a biaxial
mineral of known optic axial angle and cut normal to the acute
bisectrix, the distances between the apices of the dark hyper-
bolic bats of the interference figure in the diagonal position ;
this distance is 2D, and the optic axial angle measured is 2E.

This method, however, does not furnish a check on the value
of K, thus obtained, and its validity verified for other angles E,
unless many similar sections of different biaxial minerals be
taken and the K of the microscope for each angle E be ascer-
tained. The formula, D = K sin E, of Mallard assumes that
the loci of the focal points of waves entering in all directions
le on a perfect spherical surface, an asstmption which actual
microscopic objective lens systems in the strict sense of the
word do not fulfill.

To test the validity of the formula for the microscopic field
of a given objective (No. 9, Fuess, with Amici-Bertrand lens,
system 7—9, and movable screw micrometer ocular with Rams-
den ocular), two polished plates of calcite were used, the one
cut after 0001 and the second at an angle of 23° 17’ with 0001
as measured on a two-cireled goniometer, the adjustment on the
goniometer having been accomplished by means of fresh cleav-
age faces along the edge of the plate. These plates were placed
successively on a car efully adjusted Fedorow-Fuess universal
stage (large model) and the positions of the optic axis meas-
ured on the micrometer ocular screws as the stage was turned
about its horizontal axis from degree to degree on both sides
from the horizontal position. By means of the two plates, a
continuous set of readings was obtained for D for angles E
ranging from 0° to 47°. These values were ascer tained for
both scales (horizontal and vertical) of the double screw microm-
eter ocular described on page 336, and are listed in Table I.

The values for D listed under the heading “ calculated” in
this table were figured by Mallard’s formula by taking the
average value of K for both the horizontal and vertical scales
of the double micrometer ocular; for the vertical micrometer

328 FL Ef. Wright—Measurement of the Optic Axial Angle

TABLE 1.
= 2h | ee (o> se Se nit
Observed Caleulated : Observed Calculated
E Dy Dh Dv Dh EK Dv Dh Dy Dh
Oho 10 0 0 0 OA Wish 186 BE i
1 8°5 8 sete ats Ta SVico ml O25 = cE reere
2 16°5 16 Lee. cies 8 19475 199: dpe aaa
3 23 ZI ny eR aves 9 199 206°5 Bee ae
4. 29 30 Seaeee Beaks 30°) 205s5 2025) QO Galeries
5 36°5 a 3) 35°9 36°8 Wy PANS ULSI) ae eae ewes
6 44 44 BEAL jer ea es DW DAY) 225 spleens ees
7 52 51 pars a Ye oy 5 230°5 IG x jagetien
8 58°5 (Cts) Rye ieee Benet AL DSO | 837/05 pee Bee a
9 65 65°5 ae sash fonts 35 1 2386°5)) QA3 on 238 Crammer
Op aeeuzalicts UBD 71°6 73°4 6 243 250°5 nee ech
1 78°5 80 weer Ee 7 249 256°5 Aor aye
2 85 87°5 ieee panded 8 254 261°5 BRS te Sanayi
3 92°5 94°5 Beja etAnel 9 259 266°5 Zee pee
A 1008s 102 esp MER 40° 264 Qe, 965—27 1:
Lo elOncore, OO. MO GsfelOOss iL AS) rik eee
GUMS sonal ley BLS as ag Do Wiss 283 aya Ene
Top MAO SAQA 5 er Dihe Be ene By) DA Wot isis) Jie Bit as
Shel 2 os ommelisilich Moh a! aa ae TAS VSG 29 3no AS) Gaye
By Me} 138 gees pas Ea 45 290 299 291-5 § VI8e%
20m aSOrbu es 144-5 4cOnul 44e5 6 295 303°5 ates aye eis
LPM GLa ee Oi: pa NS Spits: es 00) 308 eS Aa seas
2 154 158°5 sheen mae 8 ete ata Se Bettye ope
68) S1GO NGA 25 wale ne We eset pe OR soln a ee ee
4 166 7 aera gee lee. FOnp ent area Se OGG. - BLBAD
Diss US} 179 WG 2 WAST)

scale, K, = 421°3; for the horizontal, K, = 422°4. The curves
of fig. 6 express these relations graphically, the zero line of each
succeeding curve being one division above the preceding curve,
to prevent confusion. For the objective in question (No. 9,
Fuess), and with the precautions observed to avoid parallax by
placing a fine stop diaphragm S (fig. 5) above the Ramsden
ocular, the agreement between theory and practice is remark-
able.* The screw of the horizontal (H, fig. 11, p. 336) micrometer

*In the Mikroskopische Physiographie I, 1, p. 330, by Rosenbusch and
Wiilfing, the latter gives a series of measurements with an a-monobrom-
naphthaline immersion objective of R. Winkel and finds disagreements

as high as 8° between observed and calculated values, as indicated in the
following table :

H D K lala
Aragonite 252264 2-22 11°33” 0°325 1°623 10°59
Mascotte eee eaten 24°43 0:700 1°674 24°14
Moy ata Alu ene 1 UG oe 39°09 1:075 1°705 (39°05)

Calcite aa eee ee 60°51 1:590 1821 68°50’

The angles under H are half the axial angles for these minerals obtained

of Minerals in the Thin Section. 329

scale registered -005"™" for each division on the head, while the
vertical (V, fig. 11) scale screw, which was constructed at a
later period and on a different lathe, was a trifle coarser and
registered a slightly greater movement for one division on its
head. For this reason the values K, and K, are slightly

HH

—
1S 20° =o : AGE ay

Fic. 6. Curve I of this figure contains the observed values of D, for the
vertical (V) micrometer screw of the double micrometer ocular (fig. 11, p. 336)
for the different values of E from, 0-47° ; the values of D in Curve II were
calculated from Mallard’s formula for the same micrometer screw scale ;
Curve III in like manner was plotted from observed values of Dy, the
horizontal micrometer screw for values of E from 0-47°;. while Curve
IV was plotted from values of Dy calculated from Mallard’s formula,
Di ke, isinvk.

from measurements on an optic axial angle apparatus while the angles under
H; were calculated by Mallard’s formula on the assumption that the K
obtained for topaz (1°075) is valid for all angles. The differences between
observation and calculation are large and indicate that the determination of
the positions of optic axes near the periphery of the field is less accurate
than that for more centrally located points; on comparison of this series of
results with those obtained by Fuess, No. 9 objective, it is evident that
objectives vary considerably in this particular and that for accurate work
the constants K of each microscopic objective and lens system should be
determined for a number of directions, either by using minerals of known
optic axial angles or a uniaxial mineral as calcite in conjunction with the
universal stage.

330 FE. EB. Wright—Measurement of the Optic Axial Angle

different. On an average, a movement of 6 divisions or ‘03™™
corresponded to one degree, so that with this method of special
refinement, the probable error for E remains at least + 10’,
and in wide hyperbolic bars, differences of 1° and over should
be expected.

If only a single screw micrometer ocular be used, the section
should be cut very nearly normal to the acute bisectri ix, other-
wise the values become much less certain. With a double
screw micrometer ocular, however, this error can be eliminated
directly and equally wood values obtained on sections only
approximately normal to the acute bisectrix, as will be shown
later (page 336).

In place of solving the above equations D=K sin E and
sin E=8 sin V by logarithms, it is possible to use a graphical
method which is sufficiently accurate for the purpose and which
Fedorow appears to have been the first to use.* An accurate
drawing (Plate IT) is made once for all which serves for all
possible angles and all refractive mineral indices to be encoun-

D
tered. To solve the equation D=K sin E, or sin E=)—, draw

the circle with radius K (Plate II, preferably in colored ink);
the intersection of the ordinate D with this circle makes then

the angle E in degrees, as is evident from the right angled tri-
sin ie sin V

1 123

angle. To solve the equation sin E= 8 sin V, or

find the intersection of radins E® with the circle for the given
refractive index and pass horizontally from this poit to inter-
section with outer circle of drawing, which point indicates V

° >)
in degrees.
Examples.
(ll) 2 Ke 54-0 D=211
Intersection of ordinate D with K-cirele is at radius 238°.
(2) e402 Be a65

Pass along radius 42° to intersection with circle labelled
(Seale 6 5: and then horizontally to outer circleand read V = 24°.+

* Fedorow, Zeitschr. f. Kryst. xxvi, 225-261, 1896. F. E. Wright, this
Journal, xx, 287, 1905.
+The drawing of Plate II can also be used to solve the birefringence for-

mula of Biot, - = sin a sin a, which is very approximately correct and

ye
has been used frequently in optical work and in which y'— a’ denotes the
measure of birefringence for any given section of a mineral, y — a, that of
the maximum birefringence of he mineral, a, and a, the angles included
between the normal to the section and the two optie axes respectively.
This formula can be solved graphically at once by noting the length of the
ordinate of the point of intersection of the radius a, with that circle whose
radius is equal to the ordinate of the point of intersection of the radius q%
with the outer circle. This graphical solution gives directly the relative
birefringence of the section in per cent of the absolute birefringence (y — @)
as represented by the radius of the outer circle.

<7

of Minerals in the Thin Section. B31

Mallard’s method for measuring the optic axial angle is one
of the most satisfactory of the microscopic methods and if
available sections are at hand which show the required phe-
nomena, Mallard’s method should be adopted without question,
especially if the measurements can be made with a double
screw micrometer ocular. The limits of error of measure-
ments of 2V by the Mallard method should not exceed 1°-2°
on clear interference figures.

Methods of F. Becke. ane place of the single screw microme-
ter ocular which in itself is of very limited application, F.
Becke* has substituted a graphical method in which the
observed optical phenomena are projected by a camera lucida
on a revolving drawing table fixed in position relative to the
microscope. “Accurate drawings of the interference phenomena
are thus prepared and serve in place of the actual interference
figure. This method has been fruitful in its results and with
practice the necessary manipulative skill can be acquired to
obtain trustworthy axial angle values. The accuracy of the
method is dependent on several factors—the accuracy with
which the drawing is prepared, the exactness with which the
drawing table is centered, and the care with which measure-
ments are made on the finished dr awing.

The actual field of the projection ‘does not measure over
25™™ in diameter, and a difference of 1° of E corresponds to a
difference in D of about °25™™, a distance which is easily meas-
urable. With unusually sharp axial bars and nice adjustment
of the optical system, it is theoretically possible to obtain an
accuracy of about ah 4°; in practice, however, a greater
accuracy than 1—2° cannot be claimed for the method.

The writer has not seen the revolving drawing table described
by Professor Becke, and has used in his work a small revolving
disk D, oraduated in degrees and supported by an arm E
attached to the microscope stand by the collar F, as shown in
fie. 7. This device was constructed in the mechanical work-
shop of the Geophysical Laboratory, after specifications fur-
nishéd by the writer. The results obtained with it have proved
satisfactory and the manipulation with the same convenient.

Having once fixed the position of this table so that its axis
of revolution coincides after reflection in the camera lucida
with the optical axis of the microscope and is also at the
proper distance from the eye for distinct vision, its constant K,
corresponding to the K of the microscope in the formula
D=K sin E, must be determined by one of the methods
described above.

With the drawing of an interference figure thus properly
prepared, itis possible to determine the angular distance,—polar

* FW, Becke, Tschermak’s Min. petr. Mittheil., xiv, 563, 1894 ; xvi, 180, 1896.

332 FL EL Wright—Measurement of the Optic Awial Angle

angle p and longitude angle ¢, of any point in the projection,—
and to plot the same in ‘stereographic or orthographic or gno-
monic projection and thus to measure the angular distance
between any two points, as those between optic axes occurring
in the field of vision.

~

Fie. 7. In this figure, C is the camera lucida: D, revolving disk of
drawing table graduated into degrees and supported by the arm HE, which in
turn fits into the collar F clamped to the stand of the microscope at the
proper distance from C; R, axial angle reflector; B, Bertrand-Amici lens; 8,
sliding stop diaphragm; L, upper lens of condenser system fitted in brass
ring in horizontal circle H; of universal stage (page 343); N, brass cylinder
extension to hold lower portion M of condenser lens system (page 342).

In a recent article,* Professor Becke has described an
ingenious method by which any section, in which only one
optic axis appears in the field, can be used for the measurement

*B. Becke, Tschermak’s Min. petr. Mitth., xxiv, 35-44, 1905.

Am. Jour. Sci., Vol. XXIV, 1907.

Plate ||

5.0}}2.0)

30)

Lol

co

BD

of Minerals in the Thin Section.

of the optic axial angle, although the values obtained are only
close approximations to the true value of the 2V. He utilizes
the fact that sections of biaxial minerals, cut approximately
normal to an optic¢ axis, exhibit, in convergent polarized light,
dark axial bars which resemble hyperbolas in the diagonal
position and whose degree of curvature is dependent on the
optic axial angle 2V. For any given position of the stage,
the points along the dark bar of the interference figure corre-
spond to those directions of hight wave propagation in space
whose planes of vibration coincide with the principal plane of
the lower nicol (polarizer) and for which the extinction angle
is zero.

To measure graphically the optic axial angle of a given min-
eral from the degree of curvature of its dark axial bar (zero
isogyre) on a section about normal to an optic axis by this
method, the axial bar is first drawn when in a position parallel
to the horizontal cross hair (fig. 8, the straight line A, C in this
position being the trace of the plane of optic axes) ; the micro-
scope stage and drawing table are
then revolved in the same direction
about some convenient angle 30° or
45° and the axial bar drawn in the
new position (A, P of fig. 8).*
These drawings are repeated after
revolution of the microscopic stage
or drawing table alone through 180°
(Caw Ce pial ge).

The point P in the projection is
any convenient point on the dark
curve or zero isogyre, and is there-
fore a direction in the crystal in its
given position relative to the nicols along which light waves
are propagated without changing their original plane of vibra-
tion. The plane of vibration for the point P is thus known,
and the law of Biot can be applied directly, to find by con-
struction the second optic axis A,,.

A convenient form of construction is shown in fig. 9, the
details of which are the same in every case. After plotting
the observed points P and A, on tracing paper above the pro-
jection plat, the great circle A,’CA,’, of which P is the pole,
is first found by revolving the tracing paper about the center
O until P coincides with the vertical diameter of the under-

*In his work the writer has found it more convenient and accurate to

revolve the nicols instead of the stages, which remain stationary except for
revolutions of 180°.

Am. Jour. Sct.—FourtH Series, Von. XXIV, No. 142.—Ocroser, 1907.

9
28

334 F. EL Wright—Measurement of the Optic Axial Angle

lying plat, and then finding and sketching that great circle
aglibee intersection C with the vertical diameter is at 90° from

P (to be counted from P, each of the great circles of the
projection being 2° apart). Similarly, the great circle A,A,,
containing A, and the extremities of the horizontal Giameioe
DE, is located and drawn. The great circle which indicates
the plane of vibration for P, can be found by either one of
two methods: (@) it is that great circle containing P and
tangent at P to the small circle, GPL (ig. 9), which parallels

Fig. 9. In this figure the points P and A, are the two points obtained from
the original drawing. The plane of vibration for the point P in its position
of darkness is indicated by the great circle PC tangent at P to the small cir-
cle GPL, which is parallel to FOC, the trace of the plane of vibration of the
polarizer when the zero isogyre passes through P. Further details of con-
struction are added under fig. 10.

the trace GOF in projection of the principal plane of the
lower nicol; (6) it is that great circle passing through P and
the intersection ,C of the polar great circle A,’A,’ C with the
trace of the principal plane F OC of the lower nicol.* In
actual work, however, it is not necessary to draw this great
circle PC, since the point C is the point sought and deter-
mines at once the direction of extinction for the given section.
The simplified construction is illustrated in fig. 10, where C is

*In his paper, Professor Becke determines this great circle as the one
which is tangent at P to the straight line parallel to the trace in projection
of the principal section of the lower nicol. This method does not facilitate
the finding of the great circle in any degree, and introduces an error, as
Professor Becke himself recognizes, which decreases the accuracy of the
method to just that extent. It seems advisable, therefore, to use one of the
new methods described above which are theoretically correct and equally
simple.

ae

of Minerals in the Thin Section. 335

the intersection of the great circle A,’A,’, and the diameter
OC, the trace of the plane of vibration of the lower nicol.

Having thus determined the point A,’ and C, the pr ojection
of the second optic axis A, is found by making A,’C = AC
(Biot’s law). The intersection of the ereat circle PA,’ with
the plane of optic axes A,A, determines then the position of
A,, and the angle A,A, in projection is 2V, the angle between
the optic axes.

The actual time consumed in this operation is not great, and
the values obtained are approximately correct. The objection

Fic. 10. In this figure the operations of actual construction are given
which are required to measure A,A» from the data in the drawing. The
points A, and P are first located accurately in the drawing, reduced to
angular values and plotted directly on tracing paper in stereographic pro-
jection ; the great circle A,’A,’ polar to P,, and the horizontal great circles
A, A, are then sketched; the point A,’ is the intersection of the great circle,
containing P and A, with great circle A,’A»', while the line OC is the trace
of the plane of vibration of the lower nicol- (polarizer) as it appears on the
drawing after the revolution of 30° or 45°; the angle MOF indicating directly
the angle of revolution. By definition Nee C is equal to A,'C, and the inter-
section A, of the great circle PA,’ with the great circle A, A». determines Ag.
the second optic binormal, the angle A,A» being the desired optic axial
angle.

to its use lies chiefly in the subjective factor involved, namely,
the skill required in drawing accurately the phenomena
observed, and also in the nice adjustment of all parts of the
instr ument.

The location of the optic axis A, is at all times more accu-
rate and trustworthy than that of P, owing to the indistinct-
ness and width of the black axial bars near the margin of the
field in consequence, chiefly, of elliptic polarization.

336 FL EL Wright— Measurement of the Optic Awial Angle

New Methods with Double Screw Micrometer Ocular.—
In seeking for more accurate and at the same time simpler
methods than those of Professor Becke described above, the
writer has substituted in place of the usual single screw
micrometer ocular, with a movement in one direction only, a
double screw micrometer ocular with movements in two ee
tions normal to each other. By the use of this double screw
micrometer ocular, which was constructed in the workshop of
the Geophysical Laboratory (fig. 11)* it is possible to determine

11

the position of any point in the interference figure accurately
by means of two micrometer screw readings, which corre-
spond to rectilinear codrdinates in the orthogonal projection
and small circle codrdinates in the stereographic projection.
By means of the constant K of the microscope for each of
these micrometer movements, Ky and Ky, which must have
been determined previously by means of known angle values
(table 1, page 328), each of these readings is reduced as usual
to its angle value for the crystal by means of K and the aver-
age refractive index of the crystal,

D

KB

Having given the interference figure from a section of a
biaxial mineral, cut so that one axial bar is visible, the course

sin V =

* The double screw micrometer ocular is fitted to microscope of fig. 7
The two movements, H horizontal and V vertical, are effected by fine
micrometer screws, reading accurately to ‘005"". The construction of this
ocular is similar to that of the single screw micromoter oculars with the
exception that here two screws with corresponding movements are used in
place of the single screw. O = Ramsden ocular; 8, small stop aperture to
reduce errors of parallax.

of Minerals in the Thin Section. 337

of procedure in measuring the optic axial by means of the
double screw micrometer ocular consists in : (a) revolving the
microscopic stage until the dark axial bar is parallel to the
horizontal cross hair of the ocular; (6) moving the horizontal
eross hair by means of the vertical micrometer screw V until
it coincides precisely with the center of the dark axial line
(fig. 12, A,C,); (¢) revolving the nicols (not the stage as may
be done in the Becke method) about a suitable angle (usually
30 or 45°), the exact position of the optic axis A, then being
the intersection of the axial bar with the horizontal cross hair
(fig. 12, A,C, with A,P,); (@) moving the vertical cross hair by
means of the horizontal micro-
meter screw until it coincides with
thisintersection and recording both
vertical and horizontal micrometer
screw readings; (¢) the stage is then
revolved about an angle “of 180°,
and similar readings for A, taken
in its new position, A,’. This
last step is necessary in order to
locate accurately the center of the
field O. The position of A, is
thus fixed accurately and can be
plotted directly, after proper re-
duction to true angles within the
crystal, in stereographic (small circles) or orthographic (co-
ordinates) projection. Any point P of the dark curved axial
bar can now be determined by two micrometer readings
(codrdinates from the center) and, after reduction to angles
within the crystal, plotted in the projection. From this point
on the method does not differ from the foregoing methods.
The optic axial angle is determined by construction from the
projection plat thus obtained.

In plotting the angles corresponding to the codrdinate read-
ings of the double screw micrometer ocular, it should be noted
that these angles apply to small circles, the angle for each
micrometer screw indicating the position, from the center of
the projection, of a certain horizontal or vertical small circle.
The intersection of the horizontal and vertical small circles thus
obtained from the two micrometer screw readings for a par-
ticular point of the interference figure determines the location
of that point in the projection.

Measurements with the double screw micrometer ocular are
more delicate, and therefore more accurate, than those with
the Becke drawing table, and the values for 2V_ correspond-
ingly more trustwor thy. This method is of general application
to all sections cut in such a way that either one or both optic

12

388 EF. EL Wright—Measurement of the Optic Axial Angle

axes appear in the microscopic field. By using projection
plats, either stereographie or orthographic, results of a fair
degree of accuracy can be obtained in a very few minutes.

On sections in which both optic axes of the interference
figure are visible, the exact position of A, and A, can be meas-
ured directly, and after plotting, the value 2V obtained from
the projection by direct reading. Such values should be
AGOUENE) HO Se IL,

To form an idea of the relative degree of curvature of the
axial bar in the diagonal position for sections of biaxial miner-
als cut at different angles with an optic axis | (binormal) and for
the optic axial angles 2V = 0°, 15°, 30°, 45°, 60°, 75° and 90°,
ae writer has constructed by gr aphical methods the following

gures :

Fie. 15. This figure illustrates the positions in the interference figure of
the dark curves of no extinction (axial bars) as they would appear in the
field were observations made in air on a series of biaxial minerals having a
mean refractive index of 1:60 and the optic angles indicated, and cut normal
to one of the optic axes (binormals). From the figure it is evident that the
radius of curvature of the axial bar increases with the optic axial angle, so
that for 2V = 90°, the axial bar is practically a straight line.

Fie. 14. The conditions of construction for this figure were similar to
those of fig. 12, except that the section is considered cut at an angle of
2 = 10°, ~=10° (small circle codrdinates from the center) with one of the
optic binormals.

In these figures, the dark lines represent the curves of 0°
extinction (zero isogyres or axial bars) of the interference fig-
ure in orthographic projection, the plane of the optic axes
making angles of about 45° with the plane of vibration of
light waves from the lower nicol.

of Minerals in the Thin Section. 339

The curves of figs. 18, 14 and 15 are constructed for min-
erals with a refractive index 8 = 1,600, while in those of figs.
16-19 orthographic projections of the actual positions of the
directions of 0° extinction within the erystal are given (refrac-
tive index of mineral and medium in which the phenomena
are observed, being considered equal). It is evident from the
figures that the differences in curvature of the axial bars for
the different values of 2V are sufficient to warrant their use in
measuring optic axial angles approximately. The accuracy of
the method depends on the accuracy with which the points A,
and P (fig. 12) can be determined. The positions most favor-

Fie. 15. This figure differs in construction from figs. 12 and 138 only in
the fact that the section is considered cut at an angle 4 = 20°, w = 20° with
one of the optic binormals.

able for these points are located one-half to two-thirds the
distance from center of field to its margin. Near the center
of the field, the errors of construction increase rapidly, while
near the margin, errors due to imperfections in the lenses and
to elliptical polarization tend to modify the interference fig-
ures, and decrease the accuracy to be attained.

The actual diameter of the field covered by the micrometer
screw movements of the writer’s microscope measures about
600 micrometer screw divisions. The distance covered by the
extremes of the curves for 0° and 90° is less than 200 divisions,
or on an average, about 2 divisions for one degree. Taking
into consideration the indistinctness and width of the axial
bars, it is easily possible to make an error of three or four divi-
sions of the micrometer scale in these readings, so that an

340 FE. Wright—Measurement of the Optic Axial Angle

Fies. 16, 17,18, 19. In figs. 16-19 the axial curves (zero isogyres) are
constructed under the assumption that the mean refractive index of both
the mineral and the medium in which its interference phenomena are
observed, are identical ; in short, an orthographic projection of the phenom-
ena as they appear within the crystal is given. In fig. 16, the section is nor-
mal to an optic axis; in fig. 17, the section makes angles, 2 =5°, n=6°
(small circle codrdinates) with the optic binormal ; in fig. 18, the angles are
A = 12°, w=18°, while in fig. 19, the section makes angles A = 20°, uw = 20°
with the optic axis ; the angle of revolution of the stage and consequent new
position of the trace of principal plane of lower nicol in each case is indi-
eated by the arrows. The area included by the inner circle of fig. 16 indicates
the relative field of vision of ordinary microscopes.

of Minerals in the Thin Section. 341

accuracy of more than =: 2° cannot be claimed for this
~ method. With the drawing table, this probable error increases
about + 5° under the same circumstances.

In place of the expensive double screw micrometer ocular,
the writer has had constructed a simpler, although slightly less
accurate ocular, consisting of a Ramsden ocular with tinely
divided cross section scale in its focal plane (arrangement simi-
lar to that in ocular of Czapski). (Ocular of fig. 5.) After
insertion of the Amici—Bertrand lens, the secondary image of
the interference figure is brought to focus in the focal plane of
the ocular, where the location of any point can be read off
directly in codrdinates, which in turn are to be reduced, just
as the readings of the double micrometer screw ocular, to angle
directions within the crystal, and then plotted in suitable
projection. The codrdinate scale employed im this ocular is
a photographic reproduction on thin glass of a greatly reduced
drawing.

Klein’s lens, which was first described by Becke, can also be
changed to fit the new conditions by simply introducing the
above fine cross-grating or codrdinate micrometer scale in ‘place
of the single micrometer scale.

By the use of such oculars with fine coérdinate scales, one
has the entire field of the interference figure under control,
and, by use of projection plats, can readily measure ‘optic axial
angles on all sections which are so cut that one optic axis at
least is in the field. If two optic axes appear within the field
of vision, their positions can be read at once from the coor-
dinate scale of the ocular and after proper reduction plotted in
stereographic or orthographic projection where their angular
distance can be determined directly.

Michel-Lévy Method —¥or sections normal to an acute
bisectrix of a mineral with large optic axial angles, Michel-
Lévy has suggested a method which, although interesting
theoretically, is not of great practical value, owing to the
indistinctness of the phenomena to be observed. His method
consists in reading the angle of revolution of the stage necessary
to bring the interference figure from the crossed position to
that in which the emerging axial bars of the interference figure
are tangent to a given circle.* Actual practice with “this
method has shown that it is not sufficiently accurate and of
such general application as to warrant a more detailed descrip-
tion at this point.

Method with Universal Stage.—In practice, it frequently
happens that a given section is not favorably cut to show the

* Michel-Lévy, et Lacroix, Les Mineraux d. Roches, 94-95, 1888. For a

modification and simplification of his formula, see F. E. Wright, this Jour-
nal, xxii, 289, 1905.

342 EE. Wright—Measurement of the Optic Axial Angle

optical phenomena to the best advantage, and that by tilting it
a certain angle, the interference figures can be improved mate-
rially. This is particularly the case with fine-grained artificial
preparations where, although individual crystals and cleavage
ean frequently be obtained, they do not rest in the section in
the most advantageous position. Such crystals and crystal
plates can be tilted either by means of an axial angle appara-
tus for the microscope, as that described by Bertrand* many
years ago, or by use of the glass hemisphere of Schroeder yan
der Kolk,+ or by the new upper condenser lens of ten Siethoff.t
The last two methods are qualitative methods only, while that
of Bertrand, although quantitative, permits of revolutions only
in one plane. To supply the want of a universal condenser
lens on which angular movements can be accomplished and
measured in any direction, the writer has modified the Fedorow-
Fuess universal stage by having a brass disk, L, constructed in the
workshop of the Geophysical ‘Laborator y to fit in the Fedorow-
Fuess stage (large model) in place of the inner ring bearing the
glass with cross hair (see fig. 7, page 332). Into ‘this ring the
upper lens of the ten Siethoff condenser lens system is inserted.
The partially bevelled upper surface of this condenser lens has
a radius of 1:°5™™, and permits, even with a No. 9 Fuess objec-
tive, angular movements of about 30° on either side of the nor-
mal. By means of a proper cylinder, N, of brass (fig. 7, page
332) resting on the cylinder containing the lower nicol, the
remaining lenses of the condenser system are raised to the
required distance from the upper lens. This type has proved
extremely useful for work with artificial preparations, since by
its use sections may be so placed that the most favorable meas-
urements possible can be accomplished with the double microm-
eter ocular, and in certain cases even, where the optic axial
angle is small, the same can be measured directly by means of
the universal stage angles in convergent polarized light.
Although the ‘measurements accomplished by the universal
stage methods of Fedorow are made in parallel polarized light
and with low power objectives, the same objectives can be used
for weakly convergent polarized light with Bertrand lens and
the position of the optic axes thus determined if it be possible
to bring them within the field of vision and provided they are
sufficiently distinct for accurate location under these conditions.
For general work, however, with thin sections in convergent
polarized light, the methods requiring the double screw microm-
eter ocular are the most accurate and easy of application.

*K. Bertrand, Bull. Soc. Min. Fr., iii, 97-100, 1880.

+ Schroeder van der Kolk, Zeitschr. f. wiss. Mikroskopie, viii, 459-461, 1891,
and xii, 188-189, 1895.

+E, G. H. ten Siethoff, Centralblatt f. Min , 657, 1903.

of Minerals in the Thin Section. 343

The methods for determination of the plagioclase feldspars,
rich in lime, which have been developed by F. Becke* by using
Klein’s lens and Czapski ocular, also by means of revolving
drawing table, can be applied directly to the two screw microm-
eter ocular and more accurate data obtained by its use.

ParautEL Pouarizep Licut.

The introduetion of the universal stage methods by Fedorow
in 1893+ and succeeding years placed a powerful instrument

Fic. 20. In its present form the universal stage comprises. when attached
to the microscope stage, five graduated circles; Hi, the horizontal circle of
the microscope stage, He, the large horizontal circle of the universal stage,
with Hs, the inner and thin section bearing circle, V,, the large vertical cir-
cle, and V2, an inner circle consisting of two segments V2, and V2, and
placed to measure the angles of revolution of the inner disk H; about a hori-
zontal axis. Two glass hemispheres (A, being the upper) are usually
employed with the stage to increase the angle of vision of the microscopic
field.
of attack in the hands of petrologists. With his methods, it
is now possible to obtain the optic properties of mineral sec-
tions which before were considered practically useless. The
universal stage (fig. 20) can be attached securely to any suitable
petrographic microscope ; parallel polarized light only is used.
By means of horizontal and vertical axes of revolution, a erys-
tal section can be brought to any given position and revolved
about any axis for optical examination.

In plane polarized light an optic axis is recognized by the
fact that when placed parallel to the axis of the microscope it
remains uniformly dark during a complete revolution about

*F. Becke, Tschermak’s Min. petr. Mitth., xiv, 415-442, 1895.

+E. Fedorow, Zeitschr. Kryst., xxi, 574-678, 1893; xxii, 229-268, 1894 ;
xxv, 225-261, 1895; xxvii, 337-398, 1897; xxix, 604-858, 1898.

B44 FE EL Wrioht—Measurement of the Optic Axial Angle

that axis. By plotting these directions graphically in projec-
tion, and by determining extinction angles in given zones, it is
possible not only to measure the optic axial angle, but also to
determine the exact position of the optic axes with reference
to the crystal plate, even though it may happen that neither
optic axis appears within the field of vision.

The values for 2V thus obtained on different sections, how-
ever, are not all of the same order of exactness, as will appear
later in the more detailed discussion of the different sections.
It should be noted that in the Fedorow methods, as in the pre-
ceding, the measured angles are reduced by means of the ayver-
age refractive index 8 to true angles within the mineral before
plotting in stereographie projection. Here also the combina-
tion of tracing paper with stereographic projection plat as a
base, as suggested by Wulff, is to be recommended as the best
and most efficient scheme for obtaining results rapidly and
accurately.

In these methods the rule of construction of Biot-Fresnel,
that the planes of vibration of light waves propagated in any
given direction bisect the angles ‘between the two planes con-
taining one of the two optic axes and the given direction, is
used constantly, since the two factors, on which the universal
stage methods are practically based, are the directions of the

optic axes, as they may be determined directly, and extinction
angles for certain zones and directions. Fedorow has also
shown how it is possible with his methods to measure the
refractive indices and also the birefringence approximately of
a mineral from any section. These methods are not, however,
germane to the purpose of this paper, and will not be discussed
further. It may be stated that, although the methods of Fedo-
row involve the use of a stereographic projection plat and are
in part graphical in nature, they are not difficult of applica-
tion and often furnish results where other methods fail. In
ordinary microscopic work, it frequently happens that one
method will yield more accurate data in a shorter time than a
second, and that particular method should then be chosen in
preference to all others.

In general, the Fedorow methods are indirect methods and
frequently involve a large expenditure of time to complete the
observations on a single plate. For these reasons chiefly,
petrologists have not adopted them so rapidly and generally as
might have been anticipated, particularly as the old tested
methods accomplish about what is desired by the busy petrolo-
gist who uses the microscope simply as a means to an end—to
aid him in interpreting geological phenomena and relations.

When attached to the microscope, the Fedorow-Fuess stage
(fig. 20) possesses, when in the 0° (primary) position, three

of Minerals in the Thin Section. B45

horizontal circles, H, (microscope stage), H, and H,, cach circle
graduated into deorees with verniers attached to Jol enockdales
each of these circles is revolvable about a vertical axis; the
horizontal axes of revolution and equivalent vertical circles are
V, and V, also divided into degrees and V, with vernier
attached. On the original stage “described by Fedorow and
made by Fuess, the par rtial scales V,are wanting and have been
attached by the writer. These sca ales were constructed in the
workshop of the Geophysical Laboratory, and have been found
of practical service in several methods, especiaily those involv-
ing the principal sections of the triaxial ellipsoid of any min-
eral (page 353). Each partial scale of V, is accurately divided
and carefully adjusted to the instrument. When not in use,
tha seale segments of V, can be inclined to a horizontal posi-
tion V., and are then entirely out of the way. Measurements
given below will be referred to this modification of the Fedo-
row-Kuess universal stage.

To increase the angle of vision of the field, two glass hemi-
spheres, A, and A, (in fig. 20 A, only appears, A, being hidden
Py H,), are usually employed ; between these the preparation

s placed, either cedar wood oil or glycerine being used to
Eek the same together and to reduce the effects of total reflec-
tion. For gener al work with the universal stage,it is advisable
to follow the suggestion of Fedorow and use special circular
(2% diameter) object glasses on which to mount the prepara-
tions in place of the or rdinary rectangular (26 x 46") thin sec-
tion object glasses.

With the universal stage of this type, it is possible not only
to bring a crystal section in any given position, but also to
revolve that section about any axis; in short, by its use one
has control over all possible directions and zones or axes of
revolution of a crystal.

The Determination of the Crystal System of a Given Min-
eral by Means of the Universal Stage Method.—The fact that
the universal stage allows the observer to study the different
effects of a given mineral section on light waves transmitted
through it in different directions, enables him to determine at
once the crystal system to which the erystal belongs. This is
accomplished most readily by means of extinction angles along
certain directions, since the term extinction angle ‘implies a
definite relation between a given crystallographic and a given
optical direction in any mineral. ‘These relations vary “with
the erystal system of the mineral, and in fact are such definite
functions of the same that, as Brewster was the first to show,
it is possible from extinction angles alone to determine defi-
nitely the crystal system of a given mineral. Briefly, an iso-

* Brewster, D., Phil. Trans., 1814, 187-218; 1818, 199-272.

346 FL EL Wright— Measurement of the Optic Axial Angle

metric mineral is isotropic for all directions of light wave
propagation. Uniaxial minerals (hexagonal ‘and tetragonal)
appear isotropic for light waves passing along the principal
crystallographic axis. or all other directions, they are aniso-
tropic, but even then can generally be distinguished from
biaxial minerals at once by the fact that every section of a
uniaxial mineral contains the @ ellipsoidal axis, parallel with
and normal to which it extinguishes. If the section be placed,
therefore, in the position of darkness between crossed nicols
and be revolved about a horizontal axis, V,, it will continue to
remain dark, if the ellipsoidal axis » coincides with the axis
of revolution, while if the ellipsoidal axis m be normal to the
latter, the crystal will exhibit interference colors of polariza-
tion on revolution except for sections of the prism zone. Biax-
jal minerals, on the other hand, do not in general remain dark
for either axis revolution, and only do so for sections in the
principal zones of the optical ellipsoid. Biaxial minerals
show, moreover, two directions of apparent isotropism, those
of the optic axes or optic binormals. To trace out the rela-
tions obtaining for orthorhombic, monoclinic and triclinic min-
erals and their distinguishing featur es, is not a difficult matter,
but one for which space is not here available. They are in
effect those which are used for the same purpose with ordinary
methods.

The Accurate Determination of the Position of an Optic
Axis when in the Field of Vision.—Although the underlying
principles of determination by means of the universal stage
are the same for all sections of a mineral, it has been found
by experience that for certain sections, certain courses of pro-
cedure for measuring the optic axial angles are best adapted to
produce the best results. Fedorow has divided the possible
sections of any biaxial mineral into four convenient groups,
each of which has its special characteristics and to each of
which certain methods are best suited. The relative positions
of the optic axes to and in the field of vision have been made
the criteria for distinguishing these different gr Oups 5 thus, 1 in
group (1) both optic axes are within the field of vision; (2)
one optic axis is within the field of vision and makes an angle
of less than 20° with the normal to the section; the second
optic axis cannot be brought within the field of vision by any
revolution of the stage ; (3) one optic axis only appears in the
field and makes an angle of over 20° with the normal to the
section, the second optic axis lies entirely outside the field;
(+) both optic axes lie outside of the microscopic field, the sec-
tion in question being cut more or less nearly per pendicular to
the optic normal, or about parallel to the plane of the optic
axes; or about normal to the obtuse bisectrix of a mineral
with small optic axial angle.

of Minerals in the Thin Section. B47

In case one or both optic binormals of a biaxial mineral sec-
tion can be bronght by revolution to coincidence with the axis
of the microscope, it is necessary to determine these angles of
revolution with the greatest possible accuracy. In all cases,
an approximate determination is first effected by revolving the
section about V, and H, until it is dark and remains dark dur-
ing a complete revolution of the microscope stage H,. In
weakly convergent polarized light the optic axis can be seen
in the center of the field. In ordinary microscopes, where
absolutely plane parallel polarized light cannot be obtained, the
section in such a position will not be perfectly dark, owing to

Fic. 21. In this figure, the method for locating the position of the optic
axes by means of optical curves is illustrated. The figures 0°, 20°, 30° and
45° opposite the curves indicate the angles which the plane of vibration of
the polarizer at the time of observation made with the plane of symmetry of
the microscope,

internal conical refraction, but will preserve the same degree
of slight uniform illumination for all positions of the micro-
scope stage.

More accurate determinations of the position of an optic
axis ean then be made by means of extinction angles along
definite directions, which, when plotted in projection, give rise
to curves all of which pass through the optic axis. The aver-
age point of intersection of a set of such curves is then the
true position of the optic axis in projection. (Fig. 21.)

Such curves have been called optical curves by Fedorow and
are obtained most readily by first placing the crossed nicols in

348 EE. Wright—Measurement of the Optic Axial Angle

any given but fixed position, then turning H, through angles
of 5° respectively, and for each position of H, determining
the angle of inclination about V, for which the section is in
the darkest position (0° extinction) (tig. 21); the same results
can also be attained by first turning the preparation about V, a
specified angle and then about H, until darkness ensues. By
this method, those directions in the crystal are obtained (after
proper reduction of observed angles to crystal directions by
means of the refractive index) for which the extinction is zero
for a given position of the nicols. The curve uniting these
directions in projection is the optical curve for the particular
position of the nicols to the axes of revolution. Analogous
curves for cther and different positions of the nicols are to be
obtained and plotted in similar manner. All such curves pass
through the optic axes and also the center of the projection.
Their points of intersection in the projection determine,
therefore, with considerable accuracy, the exact position of the
optic axis or of both axes, in case both axes can be brought
within the field of vision. Since such optical curves are
intended solely to increase the accuracy of the determination
of the positions of the optic axes, their approximate positions
being known from the preliminary determination, It is neces-
sary, In actual practice, to take readings of H, only 5° or 10°
on either side of the approximately correct position of the
optic axis as determined by the preliminary direct observations.
Conv ore positions of the nicols for optical curves are at 0°,
45°, 15° and 30° from the V, axis of revolution. If both
optic axes appear within the microscopic field of vision, the
most satisfactory method of measuring the optic axial angle by
means of the universal stage is to determine the exact position
of each axis by the above method and to plat the same in stere-
ographie projection, in which the angle can be measured
directly by graphical methods rather than by calculation, from
the cosine formula, cos 2V=cos V,,° cos V,, + sin V,,q sin V a
cos (H,,, — H,,) in which 2V designates the optic axial angle ;
V ets the readings for the one optic axis; and V,,,H,,,
those for the second.

The results obtained by the use of optical curves can be
checked and veritied by several of the methods described
below, which are of general application and can readily be
applied to this special case.

Fedorow has shown that in actual practice with minerals of
weak to medium birefringence, the errors can be disregarded
which are due to the reduction of all observed angles by means
of the average refractive index of the crystal in place of the
true refractive indices for each given direction ; and likewise
those errors which may arise from shght deviations in planes

of Minerals im the Thin Section. 349

of vibration (extinction angles) due to refraction from steeply
inclined plates and consequent elliptical polarization, are small
quantities of a low order of magnitude and can be disregarded
in general.

The method of measuring the optic axial angle by means of opti-

22 238

Fie. 22. In the method illustrated by this figure, the visible optic axis
A, is brought to coincide with the plane DCO and the extinction angle DOE
measured while the stage is in the horizontal position. A» must lie then in
the plane OB, the angle BOE having been made by construction equal to
DOE. The section is then revolved about OM (axis V,) until the extinction
angle becomes 45°, in which case the plane OM contains A», since A, has
remained during this revolution in the plane DCO ; on turning the specimen
back to its original position, the line OM becomes the great circle CA.M and
the intersection of this great circle with the plane OB fixes A, definitely in
the projection. In practice, the great circle CA.M need not be drawn, since
on placing the tracing over the plat it is only necessary to find that small
circle A.A.’, the arc of which intercepted between OB and OM is equal to the
angle of revolution.

Fie. 23, The general method of extinction curves shown in this figure is
applicable to all sections in which one optic axis A, can be brought to coin-
cidence with the axis of the microscope. After the determination of the exact
position of A, by means of optical curves the specimen is revolved about
H; until A, coincides with the plane NO normal to the axis V, of the uni-
versal stage. The extinction angle MOK of the specimen in horizontal posi-
tion is then determined ; by construction EOA, is made equal to MOE ; the
specimen is then revolved about V, a convenient angle (apparent angle
observed to be reduced to true angle), and the new extinction angle MOE’
ascertained. In the new position, the optic axis is contained in the plane
OA,.’, angle E’OA,’ having been made equal to MOE’. The exact position of
A» is then determined on the drawing on tracing paper by noting the small
circle of the underlying projection plat, whose are A.A.’ intercepted between
OA, and OA,’ is equal to the angle of revolution. This determination can
be checked by drawing the great circle CF, which marks the position which
the plane OA,’ would assume were the specimen turned back to its original
position. In practice the position of A» is determined for different angles
of revolution about V; and the mean position of all determinations taken as
the most probable and correct location of Ag. :

Am. Jour. Sct.—Fourtu Series, Vout. XXIV, No. 142.—OcrtoseEr, 1907.

2

350 FE. Wright—Measurement of the Optic Axial Angle

cal curves can be used only when both optic axes appear within
the field of vision. In other cases, other methods are to be
employed which involve either the measurement of extinction
angles in zones or the determinations of the position of the
principal planes of the ellipsoid, these latter to be plotted in
appropriate projection. In most cases, however, one optic axis

ean be determined directly by optical curves, w hile the second
optic axis makes a large angle with the normal to the section,
and must be determined indirectly. A simple but compara-
tively accurate method to accomplish this consists in first turn-
ing the stage about H, until the known optic axis comes to lie
in the plane normal to the axis V,(OCD, fig. 22), and in deter-
mining the extinction angle (EOD) when the stage is in hori-
zontal position and also at such an inclination about V, that
the extinction angle is 45°; this can be recognized most ‘read-
ily by placing the nicols in ‘the 45° position and then revolving
the preparation about V, until darkness ensues. By thus
ascertaining the angle of revolution necessary to attain the
required 45° extinction angle, the great circle CA,M is fixed
with reference to the horizontal diameter, the plane in which
the unknown optic axis A,’ must rest when the extinction angle
is 45°. The intersection A, of the great circle CA,M with
the radius OB drawn at an angle, with the vertical line, of
twice the angle of extinction (EOD) for the plate in the hori-
zontal position, fixes the position of the second optic axis in the
projection. This method, however, is not always applicable
owing to the indistinctness of extinction phenomena in steeply
inclined sections (effect of elliptical polarization), and a second
method of extinction curves, of which the above is only a
special case, can be used to advantage. Having first placed
the known optic axis in the plane normal to the axis V, as in
the above method, measure the extinction angles for different
inclinations of the stage about V, (the angles, as usual, to be
reduced to real angles within the er ystal by means of its aver-
age refractive index), and plot these directions of extinction in
stereographie projection. (Fig. 23.) Under these conditions
the radii, which make an angle with the vertical diameter
OM, equal to twice the extinction angles, are evidently the
planes containing the second optic axis A, whose exact location
can be readily found by noting for two given radii, as OA, and
OA,’, the small circle, whose are A,A,’ intercepted ‘between the
radii is equal to the angle of revolution of the stage. In prac-
tice it is advisable to repeat the determinations of the extine-
tion angles and to take as angles of inclination those equivalent
to 0°, 10°, 20°, 30°, 40° and 45° in the erystal on both sides of
the normal to the section.

of Minerals in the Thin Section. 351 |

In actual work with this method, it happens occasionally
that the determination of the location of A, is not accurate
because of the acute angle between the radius and the small
circle A,A,’. In such cases the writer has been able to apply
with favorable results one of the two following new methods,
which, like the preceding method, are based on the measure-
ment of extinction angles for different angles of inclination
about one of the horizontal axes of revolution of the univer-
sal stage. The new circle V, may render hereby valuable
assistance.

In the first of the new methods, the horizontal position of
the section is exactly that of the above method (fig. 24); A,,

24 20

having been previously located accurately, is brought to coinci-
dence with ON, and the extinction angle of the specimen in
the horizontal position ascertained ; and then instead of being
revolved about the horizonal axis V, (the line OL in projec-
tion), it is revolved about V, (or ON in the projection) as an
axis,* a given angle (apparent angle in air equivalent to true
angle in crystal). A, travels during the revolution of stage to
A,’ in the projection, the direction of extinction wanders from
OE to OE’ and the plane OA, containing A,, from OA, to
OA,’, the angle E’OA,’ being by construction = E’/OA,’.
By recording the angle of revolution of the stage about ON
(V,) required to bring the section to its new position, it is not
difficult to find in the projection that small circle, parallel to
OL, whose are A,A,’ intercepted by the lines OA, and OA,’
is equal to the above angle of revolution and thus to locate A,.

* The same effect can be produced by revolving the specimen 90° about
H; and about Vi as an axis.

352 FL EL Wright—Measurement of the Optic Axial Angle

To insure accuracy, this measurement should be repeated for
several different angles of revolution and A. determined in
each case. As in the first method, the great circle CF, indi-
eating the original position of the plane OA,’, can be con-
structed and should pass through A, on the line OA,.

The second new method differs from the first only in the
fact that instead of placing the optic axis A, in the plane OE
(fig. 25), and then measuring the extinction angle of the section
in the horizontal position, the actual direction of extinction
OE is brought to coincidence with the axis of revolution of
the universal stage (V, or V,); the section is then revolved a

Fie. 26. In this figure, the great circles aB’y. aBy' and a’ By of the stereo-
graphic projection denote the traces of the principal planes of the optical
ellipsoid within the crystal. They are fixed in position by determining the
positions of H; and V. for which the section remains dark: for all positions
of inclination about the horizontal axis V, (V. being normal to V,); the
lines O86’, Oy’ and Oa’ are thus fixed both in direction and length and also
the great circles a/’y, a3y' and a’Gy, the planes of symmetry of the ellipsoid,
the intersections a, § and y of which are in turn the ellipsoidal axes.

given angle about this axis and from the extinction angles the
lines OA, and OA,’ determined whose are is equal to the
angle of revolution. The point A, is then the desired direction
of the second optic axis.

In both new methods the determination can be varied by
inclining the specimen first about V, as an axis and then deter-
mining a series of extinction angles for different angles of
inclination about V, (V, in this case being normal to V,) and

of Minerals in the Thin Section. 353

thus locating A, afresh with each extinction. By establishing
a set of observations about V, for each new position of V, it is
possible to extend the number of observations imdefinitely and
thus to locate A, with the greatest possible accuracy. In fact,
the position of A, in the projection is immaterial so long as
this position be definitely known with respect to the axes of
revolution (V, and V,), since with A,’ located at any point in
the projection it is still possible to locate A, by means of extine-
tion angles for different angles of inclination about V, and V,.
This method, involving the use of both V, and V.,, is therefore
a method of general application and is capable of furnishing
reliable data on all sections so cut that one optic axis at least
falls within the field of vision.

Still another method which furnishes trustworthy results and
is of general application, consists in determining first the posi-
tions of the planes of symmetry and the axes of the ellipsoid
within the crystal. (Fig. 26.) In this method, practically all of
the graduated circles of the stage are brought into play, since
not ouly must extinction angles be observed, but also the section
revolved about the ellipsoidal axes and the exact position of
each axis noted. The method of procedure consists in first
placing the stage in the zero (primary) position, H,, H,, H,,
and V, in zero position, and V, normal to V,; the section havy-
ing any orientation and position. The section is then inclined
about V, until darkness between crossed nicols ensues ; if this
be not the case, it is turned about H, a small angle, and the
attempt made a second time, and so on until at a definite angle
of inclination about V, darkness is observed. The preparation
is then revolved about V,, and if by chance the correct position
be obtained, darkness will continue for every angle of inclina-
tion about V,. This is usually not the case, and by repeated
trial that position of H,, H, is to be found for which the prepa-
ration remains dark for every angle of revolution about V,.
The angle of inclination V, and the directive angle H, deter-
mine then the position of one of the planes of symmetry of the
ellipsoid within the crystal, e. g., the plane af’y of fig. 26, this
being fixed by the line O8’; in similar fashion the planes ay’
and a’®y are located and plotted in the stereographic projec-
tion. This method of locating the planes of symmetry of the
ellipsoid within the crystal is comparatively rapid and sensitive,
and a fair degree of accuracy can be attained by its use.
The new circles V, (fig. 20) which were attached in the Geo-
physical Laboratory to the large Fedorow-Fuess universal stage,
have proved extremely serviceable and time savers in this
method.

Having once determined the position of either a or y by this
method, and that of one optic axis A, by optical curves, the

354 FE. Wright—Measurement of the Optic Avial Angle

position of second optic axis A, is readily obtained, since the
angle A,a or A,y is by definition equal to A,a resp. A,y.

After some practice, the exact relative positions of H,, H,
ean be found without difficulty for which darkness remains for
all angles of inclination about V,. To insure accuracy, how-
ever, the fact of remaining dark should be scrutinized very
sharply, since the correct position is not always that of absolute
darkness but rather that for which the same degree of dark-
ness or intensity of uniform lighting obtains throughout.

From the complete deter mination by this method of the
positions of a, 8 and y, which should be mutually 90° apart,
Fedorow has shown that the average refractive index of the
mineral can be derived approximately, although the determina-
tion is not of sufficient accuracy to be of great practical value.

By this method of determining the positions of the principal
sections of the ellipsoid, the distinction between uniaxial and
biaxial minerals is greatly facilitated and the general problem
solved for all possible sections. In case the position of neither
optic axis can be determined directly, both optic axes lying
outside the field of vision, the methods for measuring the optic
axial are based solely on the determination of extinction angles
along certain directions, and are of such a nature that OY their
use only ver y rough approximations to the true value of 2V can
be obtained, errors of + 10° and over being easily possible
within the range of possibility. Fedorow has suggested one
principal method applicable to such cases and the writer has
had oceasion to use several others. They are not so satisfac-
tory, however, as the above methods, and are not of equal
practical value. For the sake of completeness, they are
deseribed briefly in fine type below.

Section nearly perpendicular to the optic normal B.

In case the section of a mineral is so cut that it makes an angle
of 30° or less with the plane of the optic axes, neither optic axis
appearing, in consequence, within the field of vision, the above
method places the observer in a position to measure the optic
axial without even seeing either optic axis. The exact position
of 8 can first be determined by this method, and then brought to
coincidence with the microscopic axis, in which case the plane
of the optic axis is horizontal. In this position the circles V, and
H, are free and the section can be revolved about V, and extine-
tion angles determined on H,. (Figs. 27 and 28.)

Since the exact positions of a and y have been determined and
the two optic axes make equal angles with these bisectrices, it is
possible by trial to bring one of the optic axes A, to coincidence
with the normal to V, (fig. 27), and to test the accuracy of its
position by means of extinction curves for different inclinations

— of Minerals in the Thin Section. 355

of the section about V,. Thus let a be the acute bisectrix (fig. 27)
and assume that one optic axis A, coincides precisely with the
normal to axis V,; A,is then the second optic axis and angle
A, fa equal to angle A,Ba, and A,@a is the extinction angle. On
revolving, now, the section about V,, the optic axial point A, is
brought to A’, and the extinction angle A’, GE for the new posi-
tion of the section should bisect exactly the angle A’ BA’,. If
this be not the case and the extinction angle be too large or too
small, the section should be revolved about H, either counter
clock wise (A’, to A,) or clock wise, A’, to A,, through a small
angle and a new set of measurements made, until after repeated
trials the corrected position is to be found for which observation
and construction agree precisely. ‘The angle A, Ga is then half the
desired optic axial angle.

In certain cases this method of placing the one optic axis A, in
the plane normal to the axis of revolution V,_ has been found
unsatisfactory, and a new method used which consists in first bring-
ing by trial the one optic axis to coincidence with the axis of revolu-
tion and then measuring the extinction angles for different angles
of inclination about V, (or V,) and testing the results of observa-
tion and construction until they coincide. The method is shown
in hg. 28 and is so similar to the foregoing method (fig. 27) that
further description is unnecessary.

Hor a section nearly normal to the obtuse bisectria of a mineral,
both optic axes lie again outside the field of vision and the optic
normal £ cannot be brought to coincidence with the axis of the
microscope. ‘The above methods do not apply, therefore, and new
ones are required to meet the new conditions and of these the
following has been found practicable by the writer.

Place the universal stage in the primary position, the axis of
V, normal to that of V, and the circles H,, H,, and H,, all in the
horizontal position ; determine the exact position of the obtuse

356 FE. Wright—Measurement of the Optic Axial Angle

bisectrix (a or y, as the case may be) by the method of principal
ellipsoidal planes (page 353), and bring it to coincidence with
the axis of the microscope, the plane of the optic axes being then
parallel to the vertical cross hair. (Fig. 29.) Revolve section
some convenient angle about axis V, and then about V, (as shown
in fig. 29), also through any suitable angle. Measure accurately

Fic. 29. To use the method indicated by this figure. turn the section so that
its obtuse bisectrix coincides with the axis of the microscope (center of the pro-
jection plat), and the optic axial plane is parallel to the vertical cross hair ;
turn preparation about axis Ve a convenient angle (reduce to true crystal
angle equivalent to apparent angle observed in air or glass), and then about
axis V, (normal to V2) any suitable angle. and measure extinction angle of
section in its new position. Plat these data in stereographic projection and
find those two points A"; and A”. equidistant from the obtuse bisectrix
and contained in the plane of optic axes in its new position, for which
the observed line of extinction OE bisects the angle included between OA",
and OA"»,

the extinction angle of the section in its new position. Plot data
in stereographic projection after proper reduction of observed
angles to true crystal angles; and find those two points A”, and
A”, contained in the optic axial plane and equidistant from the
obtuse bisectrix a”, which are so located that the observed extine-
tion angle OE, bisects the angle A”,OA”, (fig. 29). The angle
A” A”, is then the desired optic axial angle, 2 V.

With the universal stage in its present form it is not always .
possible to execute the movements indicated in the above method,
since when the obtuse bisectrix is brought to coincide with the
axis of the microscope, the axis of V, is in general no longer
horizontal and the revolution about V, is therefore along an in-
clined axis. In plotting the observed data, this fact should be
carefully noted, otherwise errors may occur and nullify the results.

of Minerals in the Thin Section. B57

With the universal stage, it is thus possible to measure the
optic axial angle of any grain of any transparent birefracting
substance and to distinguish the biaxial and uniaxial minerals.
The degree of accuracy of this measurement, however, is not
of the same order of magnitude for all sections, but differs
very materially with different sections. Asa matter of experi-
ence it has been found that the most accurate results can be
obtained on sections in which both optic axes appear within
the field of vision; that good results can be had from sections
which show only one optic axis within the field, while for
sections in which neither optic axis appears within the field the
determination is uncertain and at best only a rough approxima-
tion.

To summarize briefly the different methods best applicable
to the four different possible cases cited above :

(1) The optic axes are both within the field of vision and
inclined between 15°-55° with the normal to the section.
Determine approximate position of the two optic axes by bring-
ing each one, by means of H, and V,, into the vertical position.

Determine position of each more accurately by means of
optical curves in projection and check by means of extinction
curves and exact location of principal planes of ellipsoid,
especially the plane containing the optic axes.

(2) Section is nearly normal to an optic axis; one optic axis
A, inclined less than 20° to section normal. Place stage in
horizontal position,—H, and H, in horizontal position and V,
normal to V,—turn Es and incline about V, until optic axis
coincides with the axis of the microscope; then revolve about
V, and turn H, until darkness is attained, and thus determine
plane of optic axes and £’. Incline V, back to 0° position,
revolve about H, until the optic axis coincides with plane nor-
mal to V, and determine extinction curve, the intersection of
which with plane of optic axes in pr ojection fixes the position
of the second optic axis accurately. Check by determining a
and y both from projection and observation ; also by extinction
curve for revolution about V,.

(3) One optic axis inclined between 20°-55° within the er vs-
tal to the normal of the section, the second entirely out of the
field of vision. Determine visible optic axis by optical curves
and second optic axis by means of extinction curves, both about
V, and V,. Verify results by determination of a, Band ¥.

(4) Both optic axes are entirely without the field of vision,
1. €., are inclined at an angle of more than 65° in air with nor-
mals to the section. In such instancesthe location of the optic
axes is accomplished by means of extinction angles alone and
the values obtained are not accurate, since a slight error of 1°
in the: determination of the extinction angle may affect the

358 FE. EL Wright—Measurement of the Optic Axial Angle

value of optic axial angle up to 30°. For accurate work, there-
fore, such sections are Wok little value in general at the present
time for measuring the optic axial angle by the universal
stage methods. In case, however, the section be about normal
to the obtuse bisectrix, the measurement of the optie axial
angle is much more certain and satisfactory.

‘As noted previously, experience has shown that the best
and most rapid method of projection is that of Wulff, who uses
an accurate stereographic or orthographic plat as a base and
tracing paper on which to sketch the great circles and to
execute the actual measurements.

Since the accurate measurement of the optic axial angle can
be accomplished only on sections in which at least one optic
axis is within the field of vision, it is of interest to note the
probable relative frequency of occurrence of such sections in a
rock section. The microscopic field of the universal stage
fitted with glass segments includes an angle of about 60°, and
the area on the surface of the unit sphere thus covered for a

biaxial crystal is evidently s = 47.2 (1—cos ¢) = 47.4 sin* —

2h being the angle of vision of the field reduced to the true
value within the crystal; if the observed angle 2W be used,
the average refractive index of the mineral 8 and that of

the glass segments » should be introduced into the formula
2

Tie ae .

B sin? 3" The probability, P,, that a section show-
ing an optic axis is evidently measured by the relative surfaces
s to 8, the surface of the sphere itself :

Se Ge
2
i 47.4 sin os ae Te ioe W
eS SAP GUN neon Sa
S Aor 2 B 2
In case the areas covered by the two optic axes overlap, the
formula should be changed, as Césaro has shown,* to

a 2 sin V tan V
P= 4 sin? & — _( are cos (= — 0s are cos ( -))

7 sin } tan }

in which 2V denotes the angle between the optic axes.
Assuming an average refractive index of 1:65 for ordinary
biaxial minerals, and 1°52 for the glass hemispheres, the pr ob-
ability of encountering a proper section ranges under these
conditions from 4 to ats in unaxial crystals, to 8 to 10 in
biaxial crystals for which the fields for the optic axes do not
overlap. The degree of probability is high and one should be
able to find suitable sections in every slide for the measure-

ment of the optic axial of each mineral present.
* G. Césaro, Mem. de l’Acad. Roy. d. Sci. d. Belgique, liv, 1895.

§ = 47.2

of Minerals in the Thin Section. 359

Fedorow* has also shown how it is possible to measure the
birefringence y — B and 8 — a by use of the universal stage
and the Fedorow mica-comparator and thus to ascertain the optic
axial angle from the approximate formula cos’ ua = a either
by graphical means or by calculation.

Lanet has also used the birefringence of different sections
as a rough measure for the optic axial angle, but his methods
are even less exact than those of Fedorow and can only give
first approximations to the true optic axial angle of a given
mineral. In eases of parallel intergrowths of different amphi-
boles and pyroxenes they have, nevertheless, rendered valuable
service. Both his methods and the one of Fedorow will not,
however, be discussed further in this paper.

Extinction angles of faces in zones whose axes lie in the plane
of the optic binormals.

This method is particularly adapted to monoclinic minerals,
as amphiboles and pyroxenes, and may be of service to secure
a rough estimation of the optic axial angle of such a mineral.
The underlying principle of this method'is again the rule of
Biot-Fresnel (page 322), and mathematical formulae suitable for
its solution have been developed by Michel-Lévy,t Césaro,§
Harker,| Lane,4| Daly,** and others. These formulae show that
for the exact determination of the optic axial angle, the method
of extinction angles on different faces in the same zone is not
well adapted to optic axial determinations, especially when the
optic axial angle of the mineral is small. In certain cases, it
is possible to express this relation, as Lane has shown, in a
slightly different form which is better adapted for measure-
ments. Lane’s method, as applied to the pyroxenes and amphi-
boles, consists in measuring the angle between the clinopinacoid
and that face of the prism zone which shows the same extinction
angle. For this case, in which the plane of the optic binormals
contains the zonal axis, the formule of Césaro and Michel-Lévy
reduce to the form,

(tan A + tan pw) cos v

tan 2% = =
1 — tan X tan p cos’ v

(1)
r and w being angles between the zonal axis and the two optic

* Feodorow, EK. von, Zeitschr. f. Kryst., xxv, 349-356, 1896.

+ A. C. Lane, this Journal (3), xliii, 79, 1892.

t Michel-Lévy et Fouquée, Minéralogie Micrographique, p. 368.

<A. Cesaro, Mem. de l’Acad. Roy. d. Sci. d. Belg., liv, og, 18995.

| Harker, A., Miner. Magazine, x, 239, 1894.

“| See Daly, Proc. Amer. Acad. Arts and Sci., xxxiv, 314, 1899.
** Daly, R. A., Proc. Amer. Acad. Arts and Sci., xxxiv, 314-323, 1899.

oT
He

2V

10
20
30
40
50
60
70
80

360 EF. EE. Wright—Measurement of the Optic Axial Angle

binormals respectively; and v the angle between a plane of
the zone and plane of optic binormals. For the plane con-
taining the optic axes v = 0 and the extinction angle becomes

tanA + tan pw
1 — tan A tan be

For that plane which shows the same extinction angle, the
equation

tan 2)

tandA + tanw _(tanr + tanp) cosv 5
1—tanAtanw 1 — tandA tan pcos’ v (2)

or
cosv = — cotrA. cot pm -(3)
is evidently valid. In order that equation (8) be valid,
cot A cot w must be less than unity, or
A + p> 90°.
But the values 2V, the optic axial angle, and 2x, twice the

extinction angle on ‘the plane of optic binormals are related to
r and w by the equations

Snbstituting these values in (3), we find

eet ile: cos 2a + cos 2 V (4)
— cos2a2—cos2V~

or
cos v — 1

cos 2 V = cos 2 2, ———_—_—_—_ (5)
KOS OS ih

In table 2, the values of v are given for different optic axial
angles (2V) and different extinction angles (x), the extinction
angle being considered taken invariably to the acute bisectrix of
the optic binormal angle. It is evident from this table that for
small optic axial angles this method has no practical value for
even rough measurements. The larger the axial angle, how-
ever, the more sensitive the method becomes.

TABLE II
50 ay) 60 65 70 75 80 85
40 35 30 25 20 15 10 5

134-97 118°58’ 109°03"* 102°08! “O7o10" 97237 Ole ano Oman

133, 29% 117 47° 107 47° 100 49%" 95 50-492 10 9050 0RiaSS
131 45 115 42 105 32 98 31-93-3190 00) Sit A Omare
129-05) s1t2 30-102 508 94°01 — 90 00) 986: 29) 84 a0eeRS2
125 04 107 47 97 10 9000785259)" 81529 Om eee
118 58 100 49 90 00 82: 50 9°71 )2) 7428)" V2 aim
109 03 90 00 oe 12, 13) 6%.30 2164 V8") G2 hommmol
90 00 70 57 61 02 54 56 5055 48 15 46 31 465

of Minerals in the Thin Section. B61

Measurement of the optic axial angle on the total refractometer.

Pulfrich,* Soret,+ Viola,t Cornu§ and Wallérant| have shown
that it is possible on a single section of a biaxial or uniaxial
mineral to determine not only the three principal refractive
indices a, 8 and ¥, but also, by observing the planes of polari-
zation of each wave corresponding respectively to aB,By and
ya, to determine accurately the relative position of the princi-
pal planes of the ellipsoid to the given section; and from the
accurate refractive indices thus ascertained to fioure the optic
axial angle with great exactness. These methods, however,
require specially eround and polished sections and are not, in
general, microscopic methods, although the total refractometer
of Wallérant is attached directly to the microscope and is em-
ployed on thin uncovered and polished sections of rocks.
Unfortunately, the writer has had practically no opportunity
to work with the total refractometer of Wallérant, and is, there-
fore, not in a position to judge personally of its fitness for optic
axial angle determinations. Viola and others have shown that
on the Abbe total refractometer results of great accuracy and
certitude can be obtained rapidly and without difficulty. The
mineral plates should measure then 1°¢™" or over to furnish
sufficiently intense reflexion signals for nice adjustment and
measurement.

Measurements.

So much space has been devoted above to the theoretical con-
siderations and descriptions of methods that in this section only
a part of the available observational data can be enumerated
and a brief résumé of the results presented. Enough data will
be offered, however, to indicate certain inferences bearing on
the relative accuracy and applicability of the different methods
under test.

Different minerals, as.aragonite, topaz, muscovite, ete., were
first chosen and oriented sections cut to show the different
phenomena required by the several methods. The correct
optic axial angle for each mineral was then measured in sodium
hight on a Wiilfing- Fuess axial angle apparatus, the angle
obtained thereby being adopted as the standard of comparison
for all methods. For each mineral a series of measurements
_ of the optic axial angle for different sections and by the differ-
ent methods was taken and the relative degree of accuracy of
each method judged, not only by the results obtained, but also

* Pulfrich, C., Das Total er eee Leipzig, 1890.

+ Soret, Zeitschr. Kryst., xv, 43, 1899

¢ Viola, C., Zeitschr. Kryst., Xxxi, 40 48, 1889 ; xxxvi, 245-251, 1902.
8 Cornu, Compt. Rend., ¢xxxiii, 125 Bull. Soc. Min., xxv, 7.

| Wallérant, F., Bull. Soe. Min., xx, i2 2-26, 1898.

362 Ff. EB. Wright—Measurement of the Optic Axial Angle

by the factors on which the method itself is dependent and
their relative exactness under the conditions of observation,

Measurements with Axial Angle Apparatus.

The optie axial angles obtained in sodium light on the
Wiilfing axial angle apparatus varied slightly and the average
of five determinations of each angle is given below :

Topaz, Willard Co., Utah.
AD AaB eA ea GO ez
Aragonite, Bilin, Bohemia.
Yass OS) ING = Aer
Muscovite (a)
Ps 7h AO
Muscovite (6)
OH = 59> 49%.

Measurements with the Becke drawing-table.—To economize
space, the results are given below in their reduced form ready
for plotting directly im projection, the angle @ denoting the
longitude from the horizontal K—W line of the projection and
p the polar distance ; A, as usual denotes the visible axial point
and P, any point on the dark axial bar.

Topaz.
(a) ¢ p
PA ee ners 0° 5°°0
IPacnet nh teres +65 20 °5

In projecting these angles and performing the requisite
mechanical operations, the optic axial angle thus determined on
this section was 2V = 62°°5. For a second section the values
were :

p
AQ eee ee 3°°8 2V == 70°
poe ene =U79 19 “6
For a third section
g p
ON eres One 8°°8 2V = 63°
Tei Wied: Ree —58 23

The average of these three values is 65°°2.

Aragonite.—In aragonite the optic axial angle is so small
that both axial bars A, and A, are visible and the direct deter-
mination of 2V,.,, should in all cases be accurate within one
degree. The birefri ingence is so strong, however, that the meas-
urements involving a point P, or P, on the dark axial bar and
consequent introduction of the refractive index 8 for that
point, may be decidedly incorrect. The use of the refractive

of Minerals in the Thin Section. 363

index 8 presupposes only slight differences between the refrac-
tive indices of the mineral in order that the errors thus caused
may not be too large.

(1) ¢ p
Aue ar —80° Wo De i ines aia
ASO —130 23 QIN ial
Bea +15 18 4

(2) :
SNE ee eta —10 Si Vii =s8y3
Aas tee OE Sol pek A NO sO.ene Vet enol
1 Eek Sate A A: +492 21

(3)
ARS sitet aT Woe uae — al Bcc 0)
DNS Sa SL +85 TO OPO Vi eae 6,
Pee eS 2b in PON 4

(4)
SAN ees URES —56° 1
AS ci) Se at —140 Si oh 2) Veguage = (
Ree eatas +29 23524 a ON a6
JEgaa ee De +158 26 Ou QIN Ve 23

The values for 2V,,,, do not differ over 1° from the true
value, while those for 2V,,,, differ as much as 5° from the true
value.

Muscovite. (b)

¢ p
JG Pepe OF =n BG 2S tile?
Ae AO 85156,

Double Screw Micrometer Ocular.—The data given below
appear also only in corrected form ready for plotting in projec-
tion, the actual scale readings having been reduced to equivalent
angles in air and these in turn figured to true angles within
the crystal, by means of the refractive index 8. The errors
observed above in aragonite sections because of strong birefrin-
gence apply equally well here. In the following tables H.

indieates the horizontal and V the vertical micrometer screw of
the ocular.

Topaz.
(1)

(2)

364 FE Wright—Measurement of the Optic Axial Angle

Aragonite.

(1) H
Ds egestas es lee 9°°8 2Vi, = 82:0
AG ear —10 °8 9-8 NR 19), 5
IP ae eeay Up ye Bs DV = 14
Exe Pea —17 5 6 °8
9
(2) Pee sitet)
eR AOS PX0
yaNapo ec —13 -20 20 ON eal Tee
| eee e a 8d 5 3 DE a
| eis, Bees —23 5 3 DES = OF
Oe te
Rete Alara 08 1°8
ANG cave Nea —11 °5 8 Vile
Eee ees RD GB ME 3 ONG =e I
Muscovite. (a)
(1) a H Ms
SN ER aya 3ore5 0° Qi vile
Ava sues aes —35 °5 0
(1) d
eG ie 0° 36° Oreo
MEARS Doo US 0 —36
Muscovite. (6)
H V
OD eee 30°°1 0° O60 422
Ag mee —30 1 0
Aue Segments 0° 30° PAD 5X0)
eas ae 0 —80

Measurements with the Fedorow-Fuess universal stage.—
The angles given below were read directly on the different
circles of the universal stage and before plotting in projection
require reduction to true crystal angles by means of the refrac-
tive index 8 of the mineral and w (=1: 5239) of the glass
hemispheres used. The letters H,, H,, H, and V,, V, desig-
nate the different circles of the universal stage (see ite 11) on
which the angles were read. The angles after the letter N
designate the angle made by the principal plane of the lower
nicol with that of the micr oscope.

Topaz. Section after OO1 (acute bisectrix).—A direct
preliminary determination of the position of the optic axes in
parallel polarized light was first made and the approximate
location of each axis determined. These values were later
corrected by means of optical curves.

of Minerals in the Thin Section. 365

Direct preliminary determination.

Ai, He Hs Vi Ve
Ne yee MSOs O Okeram2O4 cere One o aiintee Via 6. Ol4cO
A. BB ont eae 66 66 66 36° eRe
Corrected by method of optical curves.
= N = 30° NWA oe
H, He Hs Ve Vi Vi Vv;
— = oe SS (Se =)
Ay Ao Ay As Ay Ae
NSO mee Oeer O44 a la enor Oona) MOUanl— SOLD moo mii)
£6 85 « &¢ 35 —36 36 —38 33°09 —3d71°5
G 90 se uo 35°65 —36 84.°— 37 34 —36°5
06 95 ee ve 34 —34°5 3. —36 34 — oil;
OS 100 s¢ “ 33 —345 33 —34 B35) —37

After proper reduction of these angles, the corrected angle,
obtained directly from the ster eographic plat, is 2V = 66°°5.

Topaz. Section nearly normal to an optic axis —The deter-
mination in this case can be most readily accomplished by first
locating A, accurately by optical curves and then fixing the
position of A, in projection by means of the principal ellipsoi-
dal planes.

Optical curves for Jae

ING Oe ay N= 305 meNG = 40s

Ay He Hs Ve Vi Vi Vi
180° 80° 225°5 —1 4°5 4°5 5

<4 85 ce 6“ 5 5 55

ce 90 ce 73 6°5 6 5

(13 95 4 (19 U7 6 5

ce 100 (5 ce e 5 4

After reduction to true angles, the position of A, in projec:

on was found to be: H,, 180°, H,, 90°, H,, 225°°5, ve 6° and V.,

12] Lhe Ss: ellipsoidal plane was located by: El OPS 18155
0°, H,, 315°, V, —, V., 26°, while for the ay ellipsoidal plane,
the ‘readings ‘Were: BL, TSO et 0 pee 2 OO Nias Visser:
The optic axial angle thus determined in projection plat is
2V=54°. In such cases, where the section is nearly normal to
an optic axis, the method of extinction curves is not of practical
value, owing to the difficulty of determining extinction angles
with requisite accuracy.

Topaz. Section nearly normal to obtuse bisectrix.—Optie
axial angle was found by first locating the principal ellipsoidal
planes af and ay and then measuring the extinction angle of
the section when a coincided with the microscope axis and

Am. Jour. Scr.—Fourra SERIES, VoL. XXIV, No. 142.—OcrosEr, 1907.

At

366 FL EL Wright—Measurement of the Optic Axial Angle

after revolution of the section from that position through
known angles about V, and V,
For the ellipsoidal planes the readings were:

Ay Hy Hs Vi V2
af plane 180° 90° 230555 = 0°
ay plane* : rs 326° ixi —17°5
For a in coincidence with the microscope axis, the readings
were found from the projection to be: H,, 185°, H,, 90°, H.,
BIOr A We 0 aN Dee vAviters tlie reeclGtinn ‘about Ve and

V,, the angles recorded were + Ele SOR oa ele OO aaeelee 326°,
V,, 40°, Woes De

‘From these rales 2V was measured in projection and found
to be 66°.

By direct observation of the optic axis, the same angle was
also obtained. This method may in favorable instances give
reliable results, but in general it cannot be considered an accu-
rate method, owing to the undue influence in projection of
slight deviations of the extinction angle on the value of the
optic axial angle.

Second Section:

lal He Hs Vi Vo
af plane 180° 90° 140° an + 5°
ay plane oe 5 234 at One
Optic axis A, WG G 140 4° a)

The optic axis A, was determined by direct readings. After
proper reduction to true angles the value 2V =63° was obtained
from the projection plat.

Topaz. Section about perpendicular to the optic normal.—
In this instance the principal elliposoidal planes were first
determined and ellipsoidal axis 6 brought to coincide with the
microscope axis and the extinction angle measured in that posi-
tion. By trial that position of H, was found for which A, coin-
cided with the principal plane of the lower nicol, and the optic
axial angle thus ascertained by measuring the extinction angles
of the section in different positions of V , and comparing “the
data of observation with those obtained by graphical methods
from the projection plat on the assumption that A, did actually
coincide with the principal plane of the lower nicol. In like
manner, the section was revolved about V, and extinetion
angles measured until theory and observation furnished identi-
eal results.

The principal ellipsoidal planes of the section were deter-
mined by the readings:

A, H, Hs Vi V2

By plane 180° 90° 333° be +16°
‘a plane ce ss Dares is — 1
Ba }

of Minerals in the Thin Section. 367

For the different positions of H,, the extinction angles for a
given angle of revolution about V, were:

laly H, Hs Vi Vo
145° 123° 333° ik? 31°
145°5 122 ‘6 “é Ge
144:5 124 “e <6 ee
144°3 123°5 ih s cs

On plotting these values in projection, it was found that 2V
was about 64°—67°, but a more decisive result was not attain-
alles. Ahe Boas js not accurate and can only furnish very
rough approximations.

In the second method, which involves revolution about an
axis normal to that of the above, the values observed were:

lek lel 18ke » V, Vo
148°°5 33° Sass Ips 0 lh
148 °5 34 “ “< “
149 32 ue “ oe

and from these angles, 2V was found to lie between 64 and 68°.

The determination cannot be termed satisfactory and this
method, like the above, can furnish only rough approximations
to the true values of 2V.

. Summary.

(1) The optic axial angle of minerals in the thin section
can be determined under the microscope in either convergent
or parallel polarized light.

(a) In convergent polarized hight, methods for the meas-
urement of the optic axial angle are ‘available for all sections
in which at least one optic axis appears within the field of
vision. Of these the method requiring the use of the Becke
drawing table is of general application and furnishes results of
a fair de; 9
about + 1° if both optic axes be visible, and + 5° if only one
optic axis be visible. More accurate and somewhat simpler in
manipulation and of the same general application is the method
involving the new double screw micrometer ocular, described
above. ‘This ocular combined with the method of projection
of Professor Wulff, is a general extension of the Mallard
method, and, like the Becke method, utilizes the rule of Boit and
Fresnel which defines the planes of vibration for any direction
of wave propagation. With this ocular the probable errors of
determination on sharp interference figures should not exceed
1° if both optic axes are visible, nor BO He only one optice axis
appears in the field.

368 FE. EL Wright—Measurement of the Optic Axial Angle

(6) In parallel polarized light, the methods involving the
Fedorow-Fuess universal stage are ‘used and furnish satisfactory
results, provided the position of one optic axis can be deter-
mined directly. If both optic axes are outside of the field of
vision, the results obtained are usually unsatisfactory and inac-
curate. Theoretically, it is possible to measure the optic axial
angle of any biaxial transparent mineral on any section by
means of the universal stage. If both optie axes appear within
the field of vision, the error of determination should not exceed
1°, and if only one of the optic axes be visible, the accuracy
may decrease to + 5°. The exact location of a visible optic
axis is assisted somewhat by use of the method of optical
curves. Having once fixed the location of one optic axis, that
of the second is determined by the method of extinction curves.
If both optie axes lie entirely outside of the field, special
methods must be resorted to, but in general without marked
success, owing to the great difference in the value of 2V caused
by a very slight deviation in the measured extinction angle.

The range ‘of the field of vision of the universal stage is
greater than that of any possible interference figure; the ‘Fed-
orow universal stage methods are, therefore, applicable to a
greater number of sections than the methods with convergent
polarized light and may furnish results on sections otherwise
useless for ordinary methods. Both experience and _ theory
show that for all these methods the accuracy of the determina-
tion varies considerably with the section and mineral in ques-
tion. The most accurate results can be obtained on sections for
which both optic axes appear within the field of vision; less
accurate but still satisfactory measurements can be made when
only one optic axis appears, particularly when it is situated
about midway from the center to the margin of the field.

For convergent polarized light, the general extension of the
Mallard method by means of the new double screw micrometer
ocular is the most satisfactory and accurate method available,
but good results can be had by use of the Becke drawing table.

The readings in both methods require to be reduced to. equiv-
alent crystal ‘angles and plotted in stereographic projection.
The optical axial angle is then measured upon the plot by
graphical methods.

(2) For the purpose of plotting and measuring observed and
calculated angles, the stereographic projection is without doubt
best adapted to optical work in general. The stereographic
plot of Plate [ is a photographie reproduction of the accurate
drawing by Professor Wulff published in the Zeitschrift fiir
Kr ystallographie,

(3) Since the interference figures are roughly orthographic
projections of the phenomena in space, an accurate ortho-

of Minerals in the Thin Section. 369

graphic plat (fig. 1) has been added with the standard stereo-
eraphie plat to serve as a base for future plotting of similar
phenomena.

(4) Plate II is intended to serve as a graphical base by
which to solve Mallard’s formula, K sin E=D; also n, sin a,=
n, sin a,; also y’—a’=(y—a) sin a, sin a,, a, and a, being the
angles included between the given direction of wave propaga-
tion and the two optic axes respectively.

(5) The Mallard formula and method were tested by a new
method and the agreement of the formula with fact for the
special objective used and the particular precautions observed
found to be remarkable. It was evident that for each micro-
scopic objective similar tests should be made at intervals across
the entire field in order to insure accuracy and certainty in the
results obtained.

(6) A disk-shaped type of the Becke drawing table was con-
structed in the Geophysical Laboratory and found satisfactory
in practice.

(7) An improvement was made in the Fedorow-Fuess uni-
versal stage, consisting in the addition of two hinged graduated
circles on which to read the inclinations of the second vertical
circle V,, and found to be of service in several methods.

(8) A new form of condenser lens system which combines
the advantages of the ten Siethoff qualitative adjustable conden-
ser system with the exact movement of the universal stage, was
also described and applied to the examination of minute min-
eral sections, especially of artificial preparations.

(9) A set of accurate drawings of the position of the dark
axial bar of the interference figure in convergent polarized
light for sections cut at various angles with one optic axis but
always so that the optic axis is still visible, has been prepared,
and the theoretically probable limits of error of determinations
of the optic axial angle by the different methods and for the
different sections established graphically.

(10) In the course of the investigation, several methods,
based solely on extinction angles for different faces, were tried,
but without exception they were discarded because ‘of the diffi-
culties in the measurement of the extinction angle and the
undue influence of small differences in extinction angle on the

value of the optic axial angle.

Geophysical Laboratory,
Carnegie Institution of Washington, May 15, 1907.

370 Boltwood—Note on a New Radio-Active Element.

Arr. XXXIV.—Wote on a New Radio-Active Element ;

by Brrrram B. Botrwoop.

In an earlier paper* the results of some experiments were
described which indicated the separation of the parent of
-radium from a solution of a uranium mineral. Some pure
thorium nitrate was added to a solution obtained by treating a
kilogram of carnotite ore with dilute hydrochloric acid and,
after the removal of the substances precipitated by hydrogen
sulphide, the thorium was precipitated as oxalate. The oxa-
lates were converted into nitrates, the precipitation with oxalic
acid was repeated and the substances were converted into
chlorides. Measurements of the amount of radium emanation
produced by the solution of these chlorides showed that in a
period of 193 days the amount of radium present had more
than doubled, and it was therefore evident that the process
described had separated the immediate parent of radium from
the uranium mineral.

From a number of earlier experiments I had found that
after this treatment the thorium salt contains a radio-active
body which retains its activity without apparent alteration in
the course of several years. ‘As it was easily proved that this
substance was not radium, uranium or polonium, it was there-
fore assumed to be actinium, Debierne+ having stated that
the chemical properties of actinium are similar to those of
thorium. Moreover, it had been found that small amounts of
an emapation which completely lost its activity in less than
half a minute were evolved from the oxides of the thorium
treated in this manner. I therefore suggested that actinium
was the parent of radium and the intermediate product between
uranium and radium.

Rutherford, using a commercial preparation of actinium,
has recently obtained resultst which prove that the immediate
parent of radium is distinct from actinium itself although it
is present in his actinium preparation. He states that the
parent substance can be separated from actinium by precipita-
tion with ammonium sulphide.

For the past ten months I have been continuing my exper-
iments with the object of determining definitely the radio-active
properties and chemical behavior of the radium parent. As
sources of material I have used carnotite, Joachimsthal pitch-
blende, gummite, uranophane and a specimen of very pure
uraninite from North Carolina. ;

In confirmation of Rutherford’s statement it has been found

* This Journal, xxii, 537, 1906. { Nature, Ixxvi, 126, 1907.
+C. R., exxx, 906, 1900.

Boltwood—Note on a New Radio-Active Hlement. 371

that the rate of production of radium in solutions of the parent
is not affected appreciably by the presence of radio-actinium
and its products. Continued observations of the growth of
radium in my original solution indicate that the rate of produc-
tion of radium has been constant, within the limits of experi-
mental error, for a period of over 500 days. Using one of my
own preparations, | have been unable to repeat the separation
of the radium parent from actinium by the ammonium sulphide
treatment which Rutherford has described. With pure, freshly-
prepared ammonium sulphide no separation could be detected.
The radium parent can, however, be quite completely separated
fron actinium by precipitation with sodium thiosulphate, under
the conditions usual for the precipitation of thorium. As am-
monium sulphide readily changes into ammonium thiosulphate,
it would appear probable that “the separation noticed by Ruth-
erford was due to the latter compound.

An interesting and important relation has been observed be-
tween the growth of radium and the activity of the substances
other than thorium in my solutions containing the radium
parent. This proportionality is most striking in those solutions
containing the more completely puritied salts. More significant
still is the fact that this radio-active constituent does not appear
to possess any of the characteristic properties af the recognized
radio-active elements. Less than half a gram of thorium oxide
containing an amount of this new body having an activity
about equal to that of five grams of uraninm did not produce
sufficient actinium emanation to permit its detection in a sensi-
tive electroscope, although under the conditions of experiment
the thorium emanation evolved could be detected and measured
without difficulty. e

That the active substance is not actinium was also demon-
strated by the fact that from a solution over five months old,
containing about 3 grams of thorium and a quantity of the new
substance with an activity equal to that of about 35 grams of
uranium, no active substances other than thorium products could
be separ ated by precipitation of the earths with ammonia, by the
formation of finely divided sulphur from sodium thiosulphate
or by the precipitation of considerable quantities of barium
sulphate in the solution. The first process should have sepa-
rated actinium X and the last two should have separated radio-
actinium had these products been present.

The behavior of the oxides obtained by strongly igniting the
hydroxides precipitated by ammonia from a solution similar to
the above is also significant. The activities of these oxides
remain nearly constant for long periods, showing only a slight
initial rise corresponding to the formation of thorium’ X in the
thorium present. No rise corresponding to the formation of

372 Boltwood—Note on a New Radio-Active Klement.

actinium X can be observed, but if actinium were present a
separation of this product would be expected.

The most conclusive proof that the substance described is
a new radio-active element is furnished by the properties of
its a-radiation. The a-rays which it emits are much more
readily absorbed by aluminium than the a-rays from polonium,
with which it has been directly compared. Their range
in air as determined by the scintillation method appears
to be less than 3 oun which is less than the range of
the a-particle from any other known radio-active element. The
new substance also gives out a S-radiation which is less pene-
trating and more easily absorbed than that from uranium, the
value found for the coefficient of absor ption being about 1°8
aluminium.

Experiments which have been carried out with a view to
obtaining a quantitative separation of this new element from
small quantities of very pure uraninite have given results which
are in good agreement with one another and which indicate
that the activity of the new element in equilibrium with radium
is about 0°8 of the activity of the radium itself with which itis
associated. This is about the value to be expected if the new
substance is intermediate between uranium and radium when
the range of a-particles in air is taken into consideration. It
is very likely that this radio-active element is present in Debier-
ne’s actinium preparations and in some of Giesel’s “emanium”

compounds which have been put on the market by the Chinin-
fabrik, Braunschweig, especially in the former, and its presence
may perhaps explain the confusion which has resulted from
Debierne’s earlier assertions that actinium accompanied thorium
as opposed to Giesel’s positive statements to the contrary.*

Strong evidence has, therefore, been obtained of the existence
in uranium minerals of a new radio-active element, which emits
both a and 8 radiations, which produces no emanation and
which resembles thorium in its chemical properties. It is with-
out doubt a disintegration product of uranium and is in all prob-
ability the immediate parent of radium. The name ‘“Tonium”
is proposed for this new substance, a name derived from the
word ‘ion’. This name is believed to be appropriate because
of the ionizing action which it possesses in common with the
other elements which emit a-radiations.

Further experiments are in progress which it is hoped will
afford additional information as to the properties and chemical
behavior of this new body.

Sloane Laboratory, Yale University,
New Haven, Conn., Sept. 21, 1907.

* Chem. Berichte, xl, 3011, 1907.

Chemistry and Physics. 373

SCIBNTIFIC (INTELLIGENCE.

I. CHEMISTRY AND PHYSICS.

1. The Action of Ozone upon Metallic Silver and Mercury.—
The blackening of bright silver by ozone is generally given as one
of its characteristic reactions, particularly in distinguishing it
from hydrogen peroxide. It would seem, therefore, that the
reaction should take place easily under any conditions. However,
Mawncuor and KampscHuLtE have found that when a piece of
silver is held at the mouth of a Siemens’ ozone tube, there is
little blackening when the ozone is dry, and even when the silver
is moistened the reaction is uncertain and not characteristic.
They have found, on the other hand, that if the silver is heated
to near redness and then is exposed to ozone as it cools, a beautiful
coating is quickly obtained, but the reaction becomes indistinct
when the metal has cooled to a temperature still considerably
above that of the room. By a series of experiments at known
temperatures it was shown that the reaction with very dilute
ozone is scarcely appreciable at 100°, but becomes more distinct
as the temperature rises, until it becomes best and very beautiful
at 220-240°. ‘The intensity of the reaction decreases gradually at
still higher temperatures, on account of the decomposition by
heat of the silver oxide, and it does not occur at all at 450°.
The experiments just described were carried out with carefully
dried gas, but it was found that moist ozone gives the same
results. The handbooks of chemistry say, on the contrary, that
completely dry ozone does not oxidize dry silver, The interest-
ing fact was observed that a great number of substances adher-
ing to the surface of silver, even in exceedingly minute quantity,
impart to it the property of being instantly blackened by ozone,
even in the cold. For instance, silver polished with emery paper,
or etched with nitric acid and carefully wasbed and dried, shows
this property. Many metallic oxides give this catalytic effect.
Curiously enough, certain silver surfaces which were susceptible
to the action of ozone in the cold lost this property after stand-
ing for some time at ordinary temperature. In the case of mer-
cury it was found that the temperature of greatest action of
ozone is 170°. Similarly to silver, the action is gradually less
below and above this temperature. — Berichte, x1, 2891.

H. L. W.

2. The Separation of Tellurium from the Heavy Metals.
BRauNER and Kuzma have made a useful contribution to the
analytical chemistry of tellurium. They find that when this ele-
ment is precipitated in the customary manner by means of sul-
phur dioxide, other metals, such as mercury, lead, bismuth,
copper, cadmium, arsenic, antimony, and thallium, tend to come

B74 Scientific Intelligence.

down with the tellurium to some extent. They find this to be
the case particularly with copper. As a means of separating tel-
lurium from some of these metals they convert the tellurium into
telluric acid and then precipitate the other metals as sulphides,
leaving the telluric acid in solution, since hydrogen sulphide
does not act upon this compound at all until after a long time.
Their method in detail is as follows: The tellurium contaminated
with a heavy metal (Cu, Bi, Sb) is collected on a Gooch filter, dis-
solved in nitric acid, the solution is evaporated, the residue is
dissolved in potassium hydroxide (1:5), and in a well covered
vessel (Erlenmeyer flask) is oxidized by the gradual addition of
ammonium persulphate (about 4—6*). After the excess of the
latter has been removed by boiling, the liquid is acidified with
sulphuric acid, and to the cold liquid 100° of hydrogen sulphide
water are added. The excess of hydrogen sulphide is then
removed by means of a rapid stream of carbon dioxide. The
sulphide is then filtered off and brought into a weighable form
by an appropriate method. The telluric acid may be reduced to
tellurous acid by adding hydrochloric acid and boiling down to
25°, as has been shown by Gooch and his co-workers, “and then
the tellurium may be precipitated by means of sulphur dioxide
water. The authors collect this precipitate on a Gooch filter and
dry it at 120° in a stream of carbon dioxide.— Berichte, xl, 3362.
H. i. We

3. The Rays from Thorium Products.—The assumption by
Haun of the existence of a product, mesothorium, intermediate
between thorium and radiothorium, was noticed in the July num-
ber of this Journal. Hahn now considers the existence of meso-
thorium as fully established by further experiments. He has
found also that mesothorium, at first considered to be rayless,
gives off B-rays, and he reaches the conclusion, in an indirect
manner, that thorium itself gives off a-rays of a peculiar kind.
He gives the following list of the radio-active thorium products
and the kind of rays given off by each:

Phorum er ae a eae a-Tays.
Meso thon tums 02 as sey ee ene (3 daca
Radiologe ea eee eS
SPORE NEE ee ee oie eee niet a-—“S
JOfoee hae yHWoy ey ee Se oe ee a- ‘
Thorram SAGs. + tes Bees see eee slowly 8-rays.
Thorium B )
} libra ae NA CIN Reece: PARTS
Thorium C | a, B-y-tays
— Berichte, xl, 3304. Fenn wa

4. The Decomposition of Gaseous Hydrocarbons by Ignition
with Powdered Aluminium.—Ktsnerzow has found that the
hydrocarbons, methane, ethane, ethylene and acetylene, are com-
pletely decomposed when they are passed several times through
a tube containing aluminium powder and heated to the melting

Chemistry and Physics. B75
peint of the aluminium. In each case pure hydrogen was pro-
duced in the theoretical amount, while the carbon was partly
deposited in the free state on the surface of the metal, and partly
formed a carbide. he latter by the action of water gave off
methane which was contaminated with other hydrocarbons.—
Berichte, xi, 2871. ives Wes

5. Studies on the Mode of Growth of Material Aggregates.
II (Addendum) Distribution of Variations ;* by A. J. Lorna.
—It was pointed out in the body of the paper, under the above
heading, that the distribution of the molecules amongst the three
classes: stable, metastable and transitional, must bear a close
relation to reaction velocity and equilibrium, and the influence of
temperature on the same.

A perusal of Arrhenius’ paper on the influence of temperature
on reaction velocity, published in the Zeitschr. f. phys. Chem,
1889, iv, 226-234,+ suggests that his “inactive ” and “ active ”
molecules correspond to the “stable” and “metastable” mole-
cules of our presentation. We may then directly apply Arrhe-
nius’ theory.

If we denote

by N the total number of molecules of A

by N, the number of stable molecules of A

by N,, the number of metastable molecules of A

by N, the number of transitional molecules (of A—~+ B)
then, by Arrhenius,

N, = &N, (1)
where is very small and follows Van’t Hoff’s relation
we oe
Re oeret
Gr = Ro €

On the other hand we have
N, = &N (t'—?¢)
= kN, (t’—7@) very nearly.

Hence, ING NG Ne es te (Cie)

This proportion then expresses the distribution of the molecules
amongst the three classes at a given temperature T, in those
cases in which Arrhenius’ theory is applicable, viz: in reactions
whose velocity coefficient varies with the temperature according
to the relation :

* See this Journal, Sept., 1907, p. 214.
+Compare also Arrhenius, loc. cit., 1899, v. xxviii, p. 317, et seq. and
Rothmund, loc. cit., 1896, v. xx, pp. 168-179.

376 crentific Intelligence.

Similar considerations apply to the molecules of A’. The
relation between the temperature and the equilibrium constant
then follows immediately from

ky
in)

6. On a Method for the Observation of Coronas; by C.
Barvus.—The following corrections are called for in the above
article, as printed in the September number (pp. 277-281).

K, —

§3, line 17; for ‘00002 read -0002°.

$4, line 16; for (2°44X/d* read (2°44 /d)’.

$5, lines 15, 16, 30, 34; for a read “ varies as.”

$5, line 16 ; for S12)4/ WeSulea read: <12).S)/\/ es Sayers
§5, line 61; for o, read s,.

$5, line 13 ; for tan read tan 0.

Il. Gronroey.

1. United States Geological Survey.—Recent publications of
the U. 8. Geological Survey are included in the following list
(continued from p. 83) :

For1o, No. 150. Devils Tower Folio, Wyoming; by N. H.
Darton and C. C. O?Harra. Pp. 9, with 5 charts.

PRoFEssIoNAL Paper, No. 53. Geology and Water Resources
of the Bighorn Basin, Wyoming ; by Cassius A. Fisuer. Pp.
72, with 16 plates and 1 figure.

Butietins.—No. 300. Economic Geology of the Amity
Quadrangle, Eastern Washington County, Pennsylvania; by
Freperick G. Crapp. Pp. 145, with 8 plates and 7 figures.

No. 304. Oil and Gas Fields of Green County, Pa. ; by Ratru
W. Stone and F. G. Crapr. Pp. 110, with 3 plates and 17
figures.

No. 308. A Geologic Reconnaissance in Southwestern Nevada
and Eastern California; by Sypnry.H. Batt. Pp. 218, with 3
plates and 17 figures.

No. 311. The Green Schists and associated Granites and
Porphyries of Rhode Island ; by B. K. Emerson and Josrpn H.
Perry. Pp. 74, with 2 plates and 6 figures.

No. 312. The Interaction between Minerals and Water Solu-
tions, with Special Reference to Geologic Phenomena; by
EKuaeene C. Sutrivan. Pp. 69.

No. 817. Preliminary Report on the Santa Maria Oil District,
Santa Barbara County, California; by Ratpa ARrNoup and
Rosert ANDERSON. Pp. 69, with 2 plates and one figure.

No. 318. Geology of Oil and Gas Fields in Steubenville,
Burgettstown and Claysville Quadrangles, Ohio, West Virginia
and Pennsylvania; by W. T. Griswotp and M. J. Munn. Pp.
196, with 13 plates.

No. 320. The Downtown District of Leadville, Colorado; by

Geology. 377

S. F. Emmons and J. D. Irvine: Pp. 72, with 7 plates and 5
figures.

W ATER-SUPPLY AND IRRIGATION PapERs—No. 190. Under-
ground Waters of Coastal Plain of Texas ; 5 Dy Tuomas U. Tay tor.
Pp. 73, with 3 plates.

No. 195. Underground Waters of Miccoutsl their Geology
and Utilization ; by EK. M. SuHeparp. Pp. 224, with 6 plates and
6 figures.

No. 197. Water Resources of Georgia ; by B. M. Hati and
M. R. Haty. Pp. 342, with 1 plate.

No. 199. Underground Water in Sanpete and Central Sevier
Valleys, Utah; by G. B. Ricaarpson. Pp. 63, with 6 plates
and 5 figures.

No. 201. Surface Water Supply of New England, 1906
(Atlantic Coast of New England drainage) ; by H. K. Barrows.
Pp. 120, with 5 plates and 2 figures.

No. 203. Surface Water Supply of Middle Atlantic Water,
1906 (Susquehanna, Gunpowder, Patapsco, Potomac, James,
Roanoke and Yadkin river drainages) ; by N. C. Grover. Pp.
100, with 4 plates and 2 figures.

No. 204. Surface Water Supply of Southern Atlantic and
Eastern Gulf States, 1906 (Santee, Savannah, Ogeechee and
Altamaha rivers and eastern Gulf of Mexico drainages) ; by M. R.
Hatt. Pp. 110, with 5 plates and 2 figures.

No. 206. Surface Water Supply of Great Lakes and St.
Lawrence River Drainages, 1906 ; 5 BY. H. K. Barrows and A. H.
Horton. Fp. 98, with 3 plates ; and 2 figures.

No. 208, Surface Water Supply of Missouri River Drainage,
1906; by R. FotitansBer, R. I. MEexer and J. EH. Srewarr.
Pp. 190, with 5 plates and 2 figures.

9. Carnivora Jrom the Tertiary y Formations of the John Day
Region ; by Joun C. Merriam. Univ. of Calif. publications.
Bull. of the Dept. of Geol., vol. v, No. 1, pp. 1-64, pls. 1-6.—
This valuable paper is based upon a collection made by the Uni-
versity of California parties in 1899 and 1900 in the John Day:
Valley of eastern Oregon, supplemented by further collections
made during the fall of 1900, in the region of Crooked River
and Logan Butte, south of the John Day Basin. The illustrations
are reproduced from the first rough proofs, as the original plates
and drawings were destroyed in the great fire of San Francisco.

The sequence of formations in the John Day region is as
follows :—

John Day River terraces Quaternary
Rattlesnake formation Pliocene

Masceall formation Miocene

Columbia Lava formation Miocene

John Day series Miocene to Oligocene

Upper John Day

Middle John Day

Lower John Day

Clarno formation EKocene
Upper Clarno

Lower Clarno

378 Scientific Intelligence.

The John Day and Mascall formations are almost entirely ash
and volcanic tuff in various forms. The mammal remains are
from the John Day, Mascall, Rattlesnake, and Terrace deposits.
The John Day consists for the most part of evenly stratified
beds containing a characteristic dry land fauna; the higher
strata, however, are crossbedded and contain fresh-water types.

The lower Mascall is composed of fresh-water sediment con-
taining fresh-water fishes, molluses, and fossil plants, while the
upper portion consists of evenly stratified ash beds like those of
the John Day.

The carnivora of the Tertiary faunas of the John Day region
are known through numerous types, but the actual number of
specimens is not lar ge and may be counted among the rarities.
It is probable that the fauna is still only imperfectly represented
in the collections. Carnivora are known as vet only from the
John Day and Mascall formations of this series, the distribution
of species being as follows:

Middle John Day 11 species, 23 specimens.
Upper John Day 6 6(?) species, 12 specimens.
Mascall 2 species, 2 specimens.

There is also a considerable number of species of which the
geological range or occurrence is unknown.

Canide,—TVhere is a remarkable variety of canid types com-
pared with those of other formations in America, nearly all from
the John Day beds; as but twe of the eighteen species are from
the Maseall. Most of the types exhibit primitive characters
though much variation in structure is shown, and in some cases
differentation has led to the development of considerably spec-
ialized forms.

Compared with the canids of other Tertiary formations in
America, the John Day dogs represent a stage of evolution
which does not correspond to that of any other formation. Of
the nine generic types but one, Cynodictis is considered identical
with a White River genus; the type is, however, much more
advanced than the White River species. ‘The genera have all
advanced along lines of specialization laid down in the White
River epoch.

The Mascall is separated from the John Day by at least one
period of erosion and by the epoch of the accumulation of the
Columbia Lava. The relationships of the Maseall Canide and
the stratigraphic relations taken together indicate that the epoch
of the Mascall beds is not far from that of the Deep River.
The position of the John Day below these beds puts it into a
division much earlier than the Loup Fork.

Felide.—Though fairly well-known from skulls and teeth,
the John Day cats have, as a whole, presented some of the most
puzzling features of this fauna. The most common and best
known forms included in the genera Archelurus and Mimravus
have been generally considered as representing the most primi-

coo)
tive division of the machzrodont group of the Felide. In the

Geology. 379

White River beds, held to be older than the John Day, there
appeared to be among the felines no forms so primitive as these.
As the other elements of the John Day fauna are nearly all more
advanced than the corresponding forms of the White River, the
evidence regarding the age of the beds which is furnished by
these cats seemed to contradict that of the remainder of the
fauna. he persistence of the primitive running type of feline
seems due to the fact that the country was in the main open and
ill suited to the development of the larger, slower animals upon
which the more specialized saber-toothed cats preyed.

The nine species of Felidze described are entirely confined to
the John Day, none being known from the Mascall. In the
Loup Fork, however, the species of true Felis and of Machero-
dus represent a more "advanced stage of development and a closer
approximation to the recent fauna than is found in either the
John Day or the White River.

Conclusions.—Taken together the Canide and Felide of the
John Day represent a stage of evolution somewhat more advanced
than that reached in the White River, and less advanced than
that of the Loup Fork. Compared with the known faunas of
Europe, they appear to be not older than the Middle Oligocene
of Fontainbleu, and not as young as the Middle Miocene of
Sansan. Bie St aby

3. A Lower Miocene Fauna from South Dakota; by W. D.
Matruew. Bull. Amer. Mus. Nat. Hist., vol. xxii, art. 1x, pp.
169-219.—in this bulletin Dr. Matthew announces the discovery
of a fossil fauna which links the latest of the White River with
the earliest of the so-called Loup Fork faune of the western
plains. This gap had been filled in part by the John Day of
Oregon, but this is much more nearly allied to the White River
than to the Loup Fork.

Matthew and Gidley have given the name Rosebud Beds to
the Lower Miocene formation of South Dakota lying between the
White River-and Loup Fork. These beds are divided by a
white, flinty, calcareous layer lying about half way up, into an
upper and a lower series, each with its characteristic animal
forms.

In the Lower Rosebud Matthew has identified five new species
of Carnivora, nine species of Rodentia, of which six are new, three
species of Perissodactyla, and six of Artiodactyla, one of which he
describes as new. ‘The Upper Rosebud fauna is almost entirely
distinct, few species passing through. It contains four new species
of Carnivor a, one of which is the type of a new genus, a new genus
and species of Insectivora, four new species of Rodentia and one
form representing an undetermined genus, five species of Artio-
dactyla of which three are new, and one new genus and at least
three species of Perissodactyla. The Rosebud fauna is derived
from the John Day, there being but one immigrant, the antiloca-
prid Blastomeryx. ‘The species are in advance of those of the
John Day, though the great majority can be referred to John
Day genera.

380 Scientific Intelligence.

A comparison with the Middle and Upper Miocene faune is
much more difficult on account of our imperfect knowledge-of so
many of the species. These appear, however, to be a further
outgrowth of the Rosebud, but contain new elements which can-
not be derived from this source, such as the Proboscidea, the
Pecora (modern ruminants), Protohippine (horses with long-
crowned, cemented teeth and reduced lateral metapodials but
retaining a vestigial pollex), and probably certain Carnivora
(Lutrinee, ete.). Aside from these foreign elements of the later
Miocene, the Rosebud fauna presents two stages in the evolution
of the Miocene fauna fairly intermediate between the John Day
and the Deep River—Pawnee Creek beds ; the remainder are suf-
ficiently more primitive for generic separation or represent phyla
which have not survived. If the John Day represents the Upper
Oligocene of Europe and the Deep River-Pawnee Creek the Mid-
dle Miocene, the Rosebud represents an earlier and a later phase
of the Lower Miocene. :

The discovery of these intermediate stages will enable us to
clear up the relations of most of the Oligocene and Upper Mio-
cene genera and to trace the descent of the various phyla and
subphyla much more exactly than has hitherto been possible.
The more elaborate studies and extensive collections of the past
few years in the American Tertiaries have shown that the simple
phyletic series, based upon more fragmentary and imperfect data
than are now available, are true only in a general and approxi-
mate way. Recent phylogenetic study has tended quite as much
to negative as to positive results—to break up accepted phyla as
to reinforce them by more complete knowledge of the genera.
It is peculiarly satisfactory, therefore, to find a fauna which is
intermediate between two stages hitherto disconnected, and ena-
bles us to perceive the exact relationship between genera which
could until now be connected only in a general or provisional
way. ‘The preliminary results here presented are very incomplete
and various additions and modifications may be needed when the
collections are more completely prepared and studied. R. s. L.

4. Points of the Skeleton of the Arab Horse; by H. F.
Ossorn. Bull. Amer. Mus. Nat. Hist., vol. xxiii, art. xili, pp.
259-263.—In this brief article Professor Osborn discusses the
distinctive features of the Arabian horse as shown in the skeleton
of the horse ‘‘ Nimr” recently mounted at the American Museum
of Natural History. It is interesting to compare these points
with those of the Arab mare, “Esnea,” the skeleton of which is
preserved in the Yale University Museum. “Esnea” was a pure
bred Arabian, imported from Damascus by Mr. John W. Garrett
in 1852. Some of the points are as follows :

(1.) Arab horses possess but five lumbar vertebre. This is
true of “ Nimr,” of “Lexington” in the U. 8. National Museum,
and also of a thoroughbred in the British Museum. “ Esnea,”
however, has six lumbars, the last three being céossified.

(2.) The characteristic elevation of the tail due to the upturned
sacral and anterior caudal vertebre together with the remarkable

Geology. 381

horizontal position of the pelvis “ Esnea” would show were the
skeleton properly mounted. As it is, the back is so highly arched
that the zygapophyses are pulled apart.

(3.) The short tail is a distinctive feature, “ Nimr” having six-
teen vertebre compared with the eighteen of a draught horse,
while “Esnea” has but. thirteen and the tail is seemingly
complete.

(4.) The fourth character, that of a complete shaft to the ulna,
“ Ksnea” does not show, for with her the shaft is discontinuous
for about 65™™, although a fractured end implies that the actual
break in the continuity of the bone may have been less.

(5.) The Arab skeleton is noted for the great density of the
bone. This is not especially true of “Nimr.” Whether or not
it is true of ‘“ Hsnea” has not been ascertained.

(6.) The skull of “ Nimr” has a large brain case, prominent
orbits, a broad forehead, and a “dish profile.” That of “ Esnea”
agrees except that the profile is not dished.

(7.) Lhe development of the sagittal crest in “ Nimr,” as well
as the fosse for the insertion of the masseteric muscles in the
angular region of the jaw, are probably more exaggerated than
in ‘*Ksnea” as a sexual character. The slender, tapering jaw is
characteristic of both individuals.

(8.) ‘‘Esnea” shows the slight depression in the malar region
in front of and below the eyes, to which Lydekker calls attention:
as characteristic of the skulls of several thoroughbreds. This
““Nimr” does not show.

_ Finally, Osborn says: “Altogether in my opinion these osteo-
logical characters justify the separation of the Arab as a distinct
species (Hquus africanus Sanson), of distinct origin and from
wild ancestors very different from those of the northern horse.”’

Royse ke

5. Hiszeit und Urgeschichte der Menschen; von Dr. Hans
Pouriag. Pp. 141. Leipzig, 1907 (Quelle & Meyer).—This is
a valuable little book recently published by the well-known
authority on the Pleistocene faunze of Europe.

6. Physikalische Kristallographie vom Standpunkt der
Strukturtheorie ; von Ernst Somerretpr. Pp. vi, 131, mit 122
| ee im Text und auf eingeheften Tafeln. Leipzig, 1907
(Chr. Herm. Tauchnitz).—This compact volume will be found
useful and interesting by those desirous of obtaining a knowledge
of the modern theory of molecular structure and the crystal-
lographic and physical relations connected with it. The matter
is presented after the manner of Sohncke, and is liberally illus-
trated by photographs from models, which will be helpful to the
student. The application of the theory of structure to the
different aspects of crystallographic physics, as the etching figures,
the phenomena of rotatory polarization, etc., are well presented i in
the latter part of the work.

Am. Jour. Sc1.—FourtH Series, Vou. XXIV, No. 142.—Ocrosrr, 1907,
26

382 Scientific Intelligence.

III. MisceELLANEOUS SCIENTIFIC INTELLIGENCE.

1. Carnegie Institution of Washington.—Recent publications
of the Carnegie Institution are noted in the following list (see
earlier, p. 87, July, 1907:

No. 5. A General Catalogue of Double Stars within 121° of
the North Pole; by 8S. W. Burnaam. Part I. The Catalogue,
4to, pp. lv. Part II. Notes to the Catalogue. Pp. vill, 259-
1086.

No. 54. Research in China. Volume One in Two Parts: Part
Two. Petrography and Zoology; by Exior BLackWELDER.
Syllabary of Chinese Sounds; by FrrepRicH Hirru. 4to, pp. iv,
367-528.

No. 62. Condensation of Vapor as induced by Nuclei and
Ions; by Cart Barus. 8vo, pp. v, 164, with 55 tables and 66
figures.

No. 64. Variation and Correlation in the Crayfish, with Special
Reference to the influence of Differentiation and Homology of
Parts; by Raymonp Peart and A. B. CLawson. 8vo, pp. 70,
with 32 tables.

No. 68. Further Researches on North American Acridiide ;
by ALBERT P. Morse. 8vo, pp. 54, with 9 plates.

- No. 71. Atlas of Absorption Spectra; by H. 8. Usxer and R.
W. Woop. 4to, pp. 59, with 26 plates (102 figures).

No. 72. Investigation of Inequalities in the Motion of the
Moon produced by the Action of the Planets; by Simon NEw-
comp, assisted by Frank E: Koss. 4to, pp. v, 160, with 49 tables.

No. 84. The Proteins of the Wheat Kernel; by Tuomas B.
OsBORNE. 8vo, pp. 119.

2. A Laboratory Manual of Invertebrate Zoilogy ; by Git-
MAN A. Drew, Ph.D. Pp. xii, 201. - Philadelphia, 1907 (W. B.
Saunders Co.).—The manual is based on laboratory directions
which are the result of the experience of the last six years in
teaching the class in general zodlogy at the Marine Biological
Laboratory of Woods Hole, Mass. In addition to the particularly
satisfactory directions for dissecting a large number of inverte-
brate types, special emphasis is placed upon such facts as lead
the student “to an appreciation of adaptation.” B. W. K.

OBITUARY.

Dr. Witsur Oxin Atwater, Professor of Chemistry in Wes-
leyan University, Middletown, Ct., since 1873, died on Sept. 22 at
the age of sixty-three years.

THE

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES.]

Art. XXXV.—On the Electric Arc between Metallic Elec-
trodes;* by W. G. Capy and H. D. Arnorp.

First Parrer.—Jntroduction,

§1. Up to the present time most investigations on the electric
discharge between metals have been confined either to the
glow discharge,—chiefly at low gas pressures,—or else to the
are discharge at relatively high current densities, where a pro-
nounced volatilization of both electrodes takes place. No
systematic examination of the transition from one of these
forms of discharge to the other, for various metals, seems to
have been carried out. The present paper has to do with this
transitional region, haying regard particularly to the phenomena
observed with the electric arc at relatively small currents.t+

The starting-point of the investigation was the observation
made by one of the writers, that the iron are at a certain crit-
ical value of current undergoes an abrupt change somewhat
similar to the well-known “hissing point” of the carbon are.
The similarity was so strong that in our preliminary reports we
used the terms “quiet state,” “ hissing state,” and ‘ hissing
point” to denote the phenomena observed. More recent obser-
vations have shown that the effect is not to be compared to the
hissing point of the carbon are, but that it is a different phenom-
enon, casual reference to which has been made by various
observers in the past. Thus Maiselt notes that the iron are

* This investigation is being carried on with the aid of a grant of $200 from
the Elizabeth Thompson Science Fund, grateful acknowledgment of which
the writers desire to make here.

+ Brief reports on these experiments have appeared in Nature, Ixxiv, 443,
1906 ; Electrician, lviii, 816, 1907; Phys. Rev., xxiv, 381 and 446, 1907.

¢ Phys. Zeitschr., v, 550, 1904.

Am, JouR. Soo — owas SERIES, VoL. XXIV, No. 143.—NovemBer, 1907.

A

384 Cady and Arnold—FElectrie Are.

ceases to hiss at a current somewhere below two amperes;
Steinmetz* mentions an unlooked-for change in voltage at
about one ampere ; and as we shall show later, “the darkening of
the arc in the neighborhood of the anodet and ‘the appearance of
striations in the are,{ which several experimenters haye noted,
may also be referred to the same common cause. Among the
names mentioned, Stark and Cassuto seem to have been the
only ones who recognized the nature of the phenomenon.

Hlence it seems better to use the term cr itical point in reter-
ring to the effect described in this paper, and to substitute for
e quiet state” and ‘‘ hissing state” the terms first stage and
second stage respectively.

I. The Critical Point of the Iron Are.

§2. Apparatus.—Among the metals that have been tried in
air at atmospheric pressure, iron shows the critical point by far
the most easily. No essential difference in any of the phenom-
ena was noticed whether the are was horizontal or vertical,
anode above or below. The effects here described are, however,
best obtained with the anode down, probably because the
anode, which as will be seen is the seat of the critical point, is
then free of disturbing influences from the heated vapors.

Most of the observations on the iron are were obtained with
a lamp constructed for the purpose, in which each terminal
could be raised and lowered by means of a rack and pinion.
Horizontal adjustments permitted the alignment of the ter-
minals, which were inclosed in a wooden box with glass sides,
to protect the are against air currents. Current and voltage
were measured by means of Weston instruments. The direct
current mains yielded a supply at either 120 or 240 volts, which
eould be raised to 475 volts by connecting a storage battery
and small dynamo in series.

The appearance and length of the are were observed by pro-
jecting an image, magnified about ten times, by means of a
small lens, onto a mirror, from which it was brought to a focus
on a vertical paper scale graduated in millimeters. The scale
was mounted close beside the are, thus enabling one observer
to control the are and record lengths, while a second observer
recorded voltage and current.

The following method of determining length of are also
gave good results in cases where it was desirable not to darken
the room. <A telescope was focused on the image of the are

* Tr, Elec. Cong. St. Louis, vol. ii, p. 710, 1904.

+ Matteucci, C. R., xxix, 263, 1849; Arons, Ann. Phys., i, 700, 1900; Stark,
Phys. Zeitschr., v, 81, 1904; Stark and Cassuto, Phys. Zeitschr., v, 264, 1904.

tGassiot, Pogg. Ann., cxix, 133, 1868; dela Rue and Miller, Phil. Tr.,
elxxi, 65, 1879 ; Child, Phys. Rey., xx, 369, 1905.

Cady and Arnold—EHlectric Are. 385

as partially reflected from a piece of plate glass. Back of
the glass, at the same distance as the arc, was a vertical paper
seale, brightly illuminated. The are was "then seen superposed
upon the scale.

Beside German silver rheostats, a large carbonate of soda
resistance was used to secure a finer regulation of current.

83. The Iron Are in free Air.—In order to make the
transition from the second to the first stage it is of prime
importance that the are shall have settled into a normal state.
When the arc has just been struck, with fresh electrodes, oxida-
tion immediately commences, resulting in the formation of a
mass of molten magnetic oxide on each terminal. Owing to the
comparatively rapid consumption of the cathode, the molten
oxide here presents a flat or even concave surface; at the anode
on the other hand the oxide forms a rounded elobule of con-
tinually increasing size, due to the transportation of matter
from the cathode. It was found best to touch the terminals
together from time to time, thereby transferring a part of the:
positive globule to the cathode and causing the discharge to
take place between convex surfaces. When cold, the oxide
tips are easily detached from the iron electrodes, leaving gener-
ally a blunt point of iron at the anode and a slight concavity
at the cathode.

Thus the normal iron arc in free air is in reality one between
terminals of magnetic oxide of iron. A test for iron nitride in
the globules gave negative results. Air-holes of considerable
extent are often found in the cooled positive globule. The
origin of these is not certain, but an examination of the globules
indicates that air finds an entrance at points around the edge
between the globule and the metal, making its way to the
neighborhood “of the are, where owing to the highly liquid state
of the globule it emerges, leaving a cavity behind. The same
is true of copper (see X18) and of iron when the are has been
burned in nitrogen. When the vapors from the iron are con-
dense, brown oxide forms ina copious powder. Most of these
details are doubtless familiar to all who have worked with the
iron are, but it seems worth while to refer to them in view of
the important relation borne by the oxidation to the critical
point.

S4. In free air, with iron terminals from three to six milli-
meters in diameter, the arc is generally on the first stage as
long as the current is below one ampere. The are proper has
then a bluish-white color at the negative end, which gradually
shades into dark blue as the anode is approached. At the
anode is a layer that seems almost completely dark. A thick
yellowish-green mantle surrounds the are, apparently due to
the oxidation of the iron in the arc itself. This mantle seems

4

386 Cady and Arnold—Electric Arc.

to avoid touching the anode. Outside the mantle is a thin
second mantle or ‘shell, springing from the cathode and prob-
ably due to oxidation os the negative globule.

The are starts from an intensely bright spot on the cathode,
but the anode end is very dim and diffuse. As the current is
increased by diminishing the external resistance, the various
parts of the are ¢ brighten i in intensity, and the fennel voltage

decreases in the usual manner. Fig. 5 (p. 403) shows a group

Fic. 1. The Ivon Arc. The positive terminal is below in each ease.
Right: Ist stage, 1 = 4:6™™, l= 1°3 amp., E= 74°5 volt.
Left: 2d stage, 1=4°5™™,I1=1°3 amp., E = 61:0 volt.

of characteristic curves for iron, for various lengths of are.
The curve for 8™" may be taken as a typical case. Here
the part J/A is on the first stage, with large external resist-
ance. In this, as in the other diagrams, full lines represent
those regions where the discharge is stable, broken lines those
where no steady discharge can be maintained.
When the current reaches a value a little over one ampere,

a very small decrease in external resistance causes an abrupt
change to take place. The potential difference drops about
12 volts, the current increases slightly, a distinct hissing
sound is heard, and the blue-white light of the cathode extends
nearly uniformly to the anode, where a small bright spot

, A ie

Cady and Arnold — Electric PAG: 387

appears. At the same time the mantle and shell surrounding
the are increase greatly in brightness.

This sudden change i is represented in fig. 5 by the broken
line ae Between the points A and B no readings can be
taken. As the current is still further increased, the portion
BW on oe second stage is obtained. This is, of course, the
stage on which the iron are is usually operated for spectro-
scopic or other purposes, and it constitutes the final state of
the are.

The energy expended at the beginning of the second stage
is always less than that at the end of the first stage, owing to
the great decrease in voltage, together with the fact that the
increase of current is limited by the external resistance.

The photographs, fig. 1, give a fair idea of the appearance
of the arc, though it was “impossible to reproduce the sur-
rounding mantle. When the are is viewed directly through
dark olasses, the change in appearance between the two stages
is much more striking than in the photograph. The bright
spot seen on the front surface of the positive globule in each
photograph is the reflection of the light from the are.

If one begins taking observations on the second stage and
decreases the cur rent, it is usually found that the change back
to the first stage does not take place along the line BA, but
that a smaller value of current can be reached. Thus in the
observations shown in fig. 5, as the external resistance was
gradually increased, the break took place along the line CD.
The value of the current for the critical points A and C is
subject to variation, but in general it may be said that the
longer the are, the wider is the parallelogram. In the obser-
vations shown in fig. 5, no difference was observed between
the lines corresponding to 48 and CD for lengths of 1™™
and 2™™, though even here on other occasions wide dif-
ferences have been found. Thus at least for long ares there
may exist two values of voltage for the same current. ‘As
would be expected, the lines BA and CD eut the axis of ordi-
nates at a point corresponding to the supply voltage. These
characteristic curves are discussed more fully in $30.

$5. The first step was to see how far these two stages are
analogous to the quiet and hissing states of the carbon are.
itive experiments described by Mrs. Ayrton* were repeated,
with the following results: When the are is on the second
stage, the first stage can be precipitated by (1) decreasing the
current, (2) drawing the electrodes farther Lee or (3) length-
ening the are by an air-blast or by a magnet. Also, as with
the carbon are, the largest current that will maintain the
first stage is gr eater the longer the are.

*The Electric Arc, p. 279.

388 Cady and Arnold—EHlectric Are.

$6. The Drop in Potential at the Anode.—Myrs. Ayrton*
found for the carbon are, that of the change in voltage in
passing through the hissing point, about two-thirds occurs at
the anode, the rest in the gas. There are serious difficulties
in the way of carrying out this test with iron, for although
the iron are when undisturbed burns steadily enough, still the
approach of an exploring pencil, especially when the current
is small, causes the are to be repelled and to burn with great
irregularity. Indeed, the sources of error in the use of
exploring electrodes mentioned by Pollakt in connection with
the mercury are and by Child§ seem to be particularly danger-
ous with the metallic are in free air.

Carbon was first tried as an exploring electrode, but it was
soon abandoned for fear of error from the vapor which it
introduced into the are. It did, indeed, change the are volt-
age by about five volts.

Then a rod of iron was employed. <A small diameter could
not be used, as the rod then melted too rapidly. Platinum
melted almost as readily as iron. When an iron pencil
2:5" thick was introduced into the arc, a globule of oxide
formed on its extremity, which was so Tes ge im comparison
with the length of are that any thought of accurate explora-
tion of the tield was out of the question. Still, it was possible
to show at least qualitatively by means of a quadrant electrom-
eter that the change in voltage when passing through the eriti-
cal point was chiefly, if not “wholly, on the anode side of the
exploring electrode.

When the are is formed between a cathode of carbon or
copper and an iron anode, the critical point occurs practically
the same as between iron terminals. But when iron is the
cathode and carbon the anode, no sudden change takes place
until the hissing point for carbon is reached. With an iron
cathode and co pper anode no critical point occurs (unless it be
the critical point for copper, which is discussed farther on)
until a deposit of iron has formed on the copper; then the
normal change for iron takes place.

The Iron Arc in Nitrogen.

§7. Pursuing the analogy with the carbon are, the question
arose whether oxidation plays a part im producing the observed
effect. Some preliminary tests in which oxygen had been
excluded from the are as far as possible indicated that either a
very small trace of oxygen is still capable of precipitating the

*The Electric Are, p. 288.

+ Child observed the positive and negative drop in the case of the metallic

arc, but he used a current of six amperes. See Phys. Rev., xii, 149, 1901.
+ Ann. Phys., xix, 218, 1906. § Phys. Rev., xXiv, "498, 1907.

Cady and Arnold—Klectric Are. 389

change in the iron are, or else it is possible to obtain both first
and second stages between terminals of magnetic oxide of iron
even when no free oxygen is present in the surr ounding gas.”

The question was settled conclusively by starting an are
between carefully cleaned iron terminals in pure nitrogen.

The apparatus shown in fig. 2 was used for this purpose. The
terminals A and A’ were located inside a glass bulb B15 in
diameter. They were made from iron rod 4°5™" thick, and the

Fic. 2. Apparatus for the are in nitrogen.

ends from which the discharge took place were turned down
to a diameter of 8°". They extended into the bulb through
vertical glass tubes about 80 long, which were provided with
ground jointsat Dand #. The glass tubes dipped into vessels
of mercury, which rose to barometric height when the air was
exhausted. It was possible to raise and lower each electrode
independently, as well as to rotate it about a vertical axis in
order to secure exact alignment. The rod A’ was jointed
at F. The short portion beyond i” was bent outward by a
suitable device when A was to be taken from the bulb or intro-
* When the iron are was formed under a bell-jar containing nitrogen in
which only a small trace of free oxygen can have been present, a surpris-
ingly large quantity of nitric acid was produced. Traces of moisture or

other impurities from organic substances in the jar may have contributed
toward this.

390 Cady and Arnold—Electrie Are.

duced into it. The curvature of the rod was such that it
could just be slipped in through the neck of the bulb. Z was a
side outlet leading to the manometer, Geryk pump and nitro-
gen generating and purifying apparatus.

This construction was chosen in order to exclude from the
are all substances that might introduce traces of oxygen, and
at the same time to permit as much freedom as possible in manip-
ulating the electrodes. Great precautions were taken in puri-
fying ‘and drying the nitrogen, which was generated chemically.
The bulb was carefully heated to drive off occluded gases.
Several sets of observations were made, in each case the ¢ appa-
ratus having been exhausted and filled with nitrogen several
times. In some eases a trace of oxidation was noticed at the
start, which soon ceased, leaving the tips of the electrodes bright
and silvery in appearance.

In agreement with an observation of Arons* we found that
the black metallic dust from the are became deposited more
densely on the negative than on the positive electrode. This
would seem to indicate that the dust particles on leaving the
are carry a positive charge. It is not what one would expect
in view of the fact that it is chiefly the negative electrode that
disintegrates.

§8. The following results were noted. As long as a trace of
black oxide remained on the ,electrodes both stages could be
obtained, in nitrogen as wellasinair. But when the dischar ee
took place between terminals of metallic iron, it presented the
appearance of the second rather than the first stage, even when
the current was reduced until the are went out. Only in the
last instant before the are ceased, when it was flickering irreg-
ularly, was there any indication of the appearance of the first
stage. This alone is enough to show that oxidation is not the
cause of the second stage in the iron arc.

$9. Observations in a Partial Vacuwm.—We shall show in
423 that in the case of certain metals the critical point can easily
be obtained at reduced gas pressures, although it does not appear
in free air. With unoxidized iron and platinum it is almost
impossible to get the first stage at any pressure. With a sup-
ply of 475 volts, and iron terminals in nitrogen, using the
apparatus of fig. 2,a glow discharge could be obtained at any
pressure, but when the current was sufficiently increased the
glow persisted in changing to the second stage of the arc, pro-
vided the pressure was high encugh to make the formation of
an are possible.

A careful examination was made at various pressures, but no
stable first stage could be found. Ata pressure of 60 and
with very short arc there was what seemed to be an exceed-

* Ann. Phys., i, 700, 1900.

Cady and Arnold—Klectric Are. 391

ingly unstable first stage, but it could not be established with
certainty.

A very unstable region was found at pressures between 10 and
30. Here the discharge changed back and forth from glow
to are so rapidly that it looked like both simultaneously.

The Critical Point due to Vaporization.

§10. After it had thus been proved that the difference between
the first and second stage of the iron are cannot be attributed
to oxidation, the thought suggested itself that perhaps in the
first stage we have, even in free air without artificial cooling,
an anode that is below the temperature of vaporization. The
observation that the positive globule appears in a more agitated
condition on the second stage, as well as the glow-like appear-
ance of the are in the neighborhood of the anode on the first
stage, tend toe strengthen this view. This means, of course, that
on the first stage ‘the dischar ge has the character of an are at
the cathode, but of a glow at the anode.

Such a possibility was mentioned several years ago by Stark,*
and more recently Stark and Cassutot have carried out observa-
tions with an are in which the anode was not even incandescent.
The mercury are with iron anode belongs also to this class of
phenomena. What made the case less obvious with the iron
are was, first, the incandescent state of the positive globule on
the first stage, and second, that any characteristic of the glow
discharge was hardly to be anticipated with ordinary electrodes
in free air with a supply voltage as low as 110.

§11. Several methods of testing this hypothesis at once
suggested themselves. The first was, to test for vaporization
by measuring the loss im weight of the positive globule on the
two stages. Even to the eye it was apparent that the positive
globule grew somewhat more rapidly on the first than on the
second stage.

In order to avoid the growth of the globule from oxidation
of the anode or from matter tr ansported across from the
cathode, a small globule of magnetic oxide was placed as anode
in a cavity in the end of an electric light carbon, while another
carbon served as cathode. The are was allowed to burn long
enough for the globule to absorb all the carbon it could, and
the globule was taken out and weighed. The loss in weight
was then observed after the are had burned a certain number
of minutes on the second stage, then again for the first stage.

Table I contains a summary of the results :

TABLE I,
Stage Current Voltage Are Length Watts Loss per Min.
Ist. 115 61 Cae 70°1 0-82ms
2d. 1:4 48 2a 67°72 Doone

*Die Elektr. in Gasen, 1902, p. 154. + Phys. Zeitschr., v, 264, 1904.

392 Cady and Arnold—EKlectric Are.

Thus although more energy is expended on the first than on
the second stage, still the loss in weight of the positive globule
is seven times as great on the second as on the first. The loss
on the second stage, according to this theory, is due to vapor-
ization of the globule; that on the first, in so far as it cannot
be accounted for by errors, is due in part perhaps to ionic
bombardment, and in part to local heating from the are itself.

Strictly speaking, the energy considered should have been that
expended at the surface of the anode, and not that for the whole
are. Still, since the change in voltage occurs chiefly at the
anode, it follows that if the true anode drop were known and
multiplied by the value of the current, the ratio of the energies
for the first and second stage would be even greater than is
indicated in Table I, thus making the argument even more
conclusive.

$12. Spectroscopic Hvidence.—The are was formed between
a carbon cathode and iron anode; its image was then projected
onto the slit of a spectroscope. On the first stage, the spectrum
of iron was faintly visible, owing to traces of iron in the
carbon and probably to a small amount of iron set free from
the anode, as indicated in the preceding paragraph. At the
instant when on decreasing the external resistance the critical
point was reached, the iron spectrum flashed out with great
brilliancy. The same is true of the are between a carbon
cathode and silver anode ($24). This indicates that even if
vaporization is taking place on the first stage, it is exceed-
ingly small in comparison with that on the second,

iy Evidence from Temperature of Positive Globule.—
A series of tests was made with a thermo-electric junction
imbedded in the positive globule. A junction consisting of a
platinum wire 0°137™ in diameter and a wire of platinum and 10
per cent rhodium 0°25™™ in diameter, held by an arm that could
be controlled by slow-motion screws, was thrust into the globule
after the arc had been formed. It was connected to a d’Arson- -
val cea ea teh, and scale deflections were noted as the are
current gradually increased. According to the vaporization
theory, the positive globule on the first stage, with the excep-
tion perhaps of a small region immediately under the are, is at
a temperature below the boiling point. ‘At the critical point,
copious vaporization suddenly sets in, and from then on the
temperature of the globule would remain practically constant
if it were not for losses by conduction and radiation. Owing
to these losses, and to the diminution of energy on passing
through the critical point, changes in the mean temperature of
the elobule cannot be taken as a sure indication of its physical
state in the neighborhood of the are.

It was found, however, in spite of the great decrease in

Electric Are. 393

energy at the beginning of the second stage, that the galvan-
ometer deflection decreased but little. In so far as the junction
can be considered as giving the mean temperature of the
globule, this means that the globule temperature is not propor-
tional alone to the energy expended, but tends to become con-
stant when the second stage is reached.

$14. Heating the Positive Globule-—Anything that raises
the temperature of the positive globule should cause the criti-
eal point to be reached at a smaller current. This we have
found to be the case. With thin electrodes, which conduct
the heat away less readily, the critical point comes at a slightly
smaller current than with thick electrodes. Also, the smaller
the positive globule, the sooner is the critical point reached.
Thus with the same anode, we have recorded within a few
minutes values of the critical current from 1:1 to 1:7 amperes,
depending on the size and curvature of the positive globule.

The experiment was also tried of heating the positive termi-
nal, an iron rod 3™™ in diameter, by means of a blast lamp. A
large mica shield fitted closely around the rod about a centi-
meter below the are, to protect the latter from the direct effects
of the flame. When. the positive terminal. was heated to
redness, the critical current was 1-4 amp. Without artificial
heating, the critical current was 1-7 amp. The globule was of
approximately the same size in each case, and the blast lamp
was kept burning throughout the experiment in order that the
nature of the gas surrounding the are might be constant.

Stark and Cassuto (1. ¢.) have shown that heating the anode
causes the positive drop to increase slightly, owing ‘probably to
increased thermal electromotive-force. This, by increasing the
expenditure of energy at the anode, would also tend to cause
the critical point to come at a smaller current.

When it was attempted to cool the anode by surrounding it
‘with snow, anomalous results were obtained, which we found
to be due to the effect of the water vapor. For example, when
steam was allowed to flow freely into the inclosure surrounding
the are, the break from the first to the second stage took place
at a much higher voltage and smaller current than when the air
was not thus saturated with moisture. The difference—amount-
ing to about ten volts in the case of the potential difference—is
too great to be accounted for by change in the heat conductivity
of the air due to the presence of moisture. It is more likely
that the water vapor, which as the experiments of Merritt and
Stewart* indicate becomes highly ionized in the are, gives to
the are quite different characteristic curves from those prevail-
ing in ordinary air.

* Phys. Rev., vii, 147, 1898.

394 Cady and Arnold—Electric Are.

Il. Observations with Various Metals.

$15. Change from Glow to Are with Tron and Platinum
Terminals.—Mention has already been made ($9) of the glow
discharge in nitrogen between iron terminals with a supply
e.m.f. of 475 volts. This has been observed in free air with
terminals of iron, copper, and platinum, but only when the
electrodes were nearly free from oxidation. In each ease,
almost the whole supply e.m.f. was used up in the are. The
effect could only be obtained with an are less than 1°5™™ long.

I 2 AMP.

Fie. 3. Change from glow to are in free air. Heavy lines, iron.
Light lines, platinum.

The space between the electrodes looked entirely dark, but the
discharge remained in this state only a few seconds, for the
cathode rapidly became incandescent, causing the glow to
change to an are, with the accompanying drop in voltage and
increase in current. |

With iron electrodes on which there were only small parti-
cles of oxide, the are at small currents burned irregularly, and
at the positive end it sprang wildly from one particle to
another, while its negative end was concentrated on one parti-
ele. If the wandering of the are caused the negative end to
leave its minute globule of oxide, the discharge instantly
changed toa glow. Then as soon as the whole cathode had
become incandescent, the are formed again, concentrating itself

Cady and Arnold—Electrie Are. 395

upon an oxide particle as before. This may be due simply to
the loose particles having a higher temperature than the elec-
trode itself, but it may indicate also that the oxide of iron,
like that of calcium, emits negative corpuscles more readily than
does the metal itself. When the electrodes have become com-
pletely tipped with oxide, no glow discharge can be obtained.*
A similar spontaneous change back and forth between glow
and are was also observed with terminals of copper, and of
platinum.

§16. Fig. 3 shows the limiting values of current and volt-
age for the change from glow to are, with platinum and iron
electrodes. The length of are is about 1™™. In the case of
platinum, AZ is the line along which the change takes place
in going from glow to are. The are current can then be
decreased so that the change back to glow is along CD. This
is analogous to the “parallelograms” referred to in $4, and is
similar to an observation of Kaufmann’s.t No such overlap-
ping was detected in the case of iron. The platinum arc is
very irregular, but as nearly as can be judged it is on the
second stage.

When the glow between iron electrodes changes along A’ B’
to an are, the latter is on the first stage. Further decrease in
external resistance results farther on in the change to the sec-
ond stage, as shown in fig. 38. Points based on observations
made during this experiment are represented by circles. The
lines representing the first and second stages of the iron are in
air are based on other data.

The close agreement of the curve for the platinum are with
the second stage of the iron are is noticeable. It looks as if
the higher boiling point of platinum allowed the glow to per-
sist until the current was greater than is possible with iron
electrodes, so that when the change to are occurred, the initial
are current as determined by the slope of the line AB was
too great for the first stage to exist. In other words, cathode
and anode begin to volatilize at the same instant. If a higher
supply e.m.f. had been available, it is quite probable that the
first stage could have been maintained with the platinum are.
With the are between platinum terminals in nitrogen, no better
success in reaching the first stage was obtained, even at reduced
gas pressures.

After the platinum electrodes had cooled, an examination
under the microscope revealed a system of grooves or striae on
their surfaces, somewhat like those mentioned by Arons.t{ A
similar structure, apparently crystalline in nature, can be seen

* Stark (Phys. Zeitschr., vy, 81, 1904) found that the glow did not change to

an are when the electrodes were clean.
+ Ann. Phys., ii, 158, 1900. tTbid., i, 705, 1900.

396 Cady and Arnold—Electric Are.

sometimes on the surfaces of the iron globules. After cool-
ing, the platimum rod which had served as cathode looked ©
perfectly bright at the tip, but the platinum anode showed a
discoloration, apparently the same as that which de la Rive*
took for a sign of oxidation.

$17. MMagnetite-Copper.—The two electrodes belonging to
a magnetite arc lamp were used. When the magnetite was
positive, the critical point was observed just as with the iron
are. The large size of the electrodes made observations at
small currents very irregular.

§18, Copper.—After some difficulty, the critical point in
free air was found. For comparison with iron, the observa-
tions have been plotted in fig. 5. They were obtained with a
475 volt supply, length of are 4™™, diameter of electrodes
6™™. In fig. 5 only the “parallelogram” mnop correspond-
ing to ABCD for iron is shown, so that pm is a part of the
characteristic curve on the first, no on the second stage. The
cause of the parallelograms for copper and iron is explained
in $80. The change in appearance of the copper are between
the first and the second stage in air is in a general way similar
to that for iron, though less pronounced.

The arc takes place between molten globules of copper
oxide, which shrink considerably on cooling. Air cavities are
sometimes found inside these globules, probably of the same
nature as those in the case of iron.

$19. Aluminium and Zinc.—The only observations were
in free air. The formation of oxides made the arcs so
unsteady that no accurate observations could be made. Over
a wide range of current no evidence of anything other than
the second stage could be found.

§20. Mercury, Cadmium and Tin.—Arcs were formed in
air between these metals as anodes and carbon as cathode.
The discharge was very unsteady, and no certain indication of
the first stage was found. This is not surprising, owing to the
low boiling points of the metals.

§21. Bismuth.—A piece of bismuth was placed as anode
in a brass receptacle, with a carbon cathode. Much oxide
formed, but it became reduced in the neighborhood of the are.
The are burned normally on the second stage, but as the cur-
rent was decreased, the bright spot on the anode disappeared
and the first stage seemed to have set in. Unfortunately the
are could not be kept burning steadily on this stage, and no
readings could be taken owing to the irregularity of the dis-
charge.

$22. Lead.—In free air about as much evidence of the
first stage could be found as with bismuth.

* Pogg. Ann., Ixxvi, 270, 1849.

Cady and Arnold—FElectrie Are. 397

To get rid of the oxidation, an apparatus was constructed
for forming in nitrogen ares with anodes of various metals.
This consisted of a bell-jar about 27x12" cemented onto a
brass base. The substance serving as anode, in the form of a
small pellet, was laid in a cavity in the top of a massive brass
block, in order that the heat might be conducted away as rap-
idly as possible. The upper (negative) electrode was a carbon
8™™= in diameter held by a small brass rod, which passed up
into a glass tube. The latter was cemented with ceiling wax
into the neck of the bell-jar. At the upper end of the brass

VOLT
500

J 2 AMP.

Fie. 4. Points A to F, glow and are with silver anode. Points G and
H, are with lead anode.

rod was an iron button. By means of a strong electro-magnet
whose poles approached close to the tube on each side, the
weight of the carbon was sustained and it could be raised and
lowered at will.

$23. It was thought that at low pressure in nitrogen the
critical point might be observed more readily than in free air.
This we found to be the case. With a lead anode, in nitrogen
up to 12™ pressure, only the glow discharge was observed.
The cathode glow formed a mantle covering the negative ter-
minal to an extent depending on current and gas pressure.
The dark space extended from this to the positive column,

398 Cady and Arnold—Electrie Are.

which formed merely a thin coating over the surface of the
anode.

At 30 the change from second to first stage was easily
observed. This is shown in fig. 4, points Gand H. To the
left of G the arc is stable on the first stage; to the right of H,
on the second stage.

At 55° and over only the second stage could be maintained.

When on the first stage, there is a soft diffused positive glow
over the anode and the are is very dark except close to the
cathode. On the second stage the usual bright spot appears
at the anode and the whole are becomes luminous.

§24. Szlver.In free air with a carbon cathode the silver
anode always melted, though there was no troublesome oxida-
tion. Only the second stage could be observed.

Observations in nitrogen were begun at a pressure of about
a millimeter, using the apparatus described in $22. An e.m.f.
of 460 volts was applied. The chief observations are shown
in fig. 4. A and £& are on the glow discharge, C and D on
the first stage of the arc, and # and /’ on the second stage.
The length “of are was 3™™ except for the point /, where it
was somewhat greater. The change from B to C took place
spontaneously when the pressure had increased to 4-52" | and
from PD to # when it had increased to 31°". It was not found
possible to get both changes at the same pressure. The lines
AB, CD and /F are what would result if the external resist-
ance were gradually diminished and at the same time the gas
pressure increased.

The appearance in the neighborhood of the anode on the
two stages of the are was similar to that in the case of lead,
but a decided tendency to the formation of striations was
manifest.

There can be but little doubt that most metals would yield
curves similar to those in fig. 4. In other words, between a
carbon cathode and metallic anode three distinct forms of dis-
charge are possible, for each of which a separate characteristic
curve must be drawn; the glow, the are with non-volatilizing
anode, and the are with an anode that is giving off vapor.
With electrodes of iron or copper, all three stages may be
seen in free air by merely decreasing the external resistance
(fig. 3).

$25. The Striated Arc.—When the nitrogen pressure was
in the neighborhood of 6°", the are with non- volatilizing silver
anode (first stage) showed several exceedingly distinct stria-
tions, sometimes as many as six or seven. This occurred with
a length of are from 10 to 15™™ and with current ranging
from 0°3 to 15 amp. The discharge began at an incandescent
spot on the carbon cathode, from which a luminous column 2

Cady and Arnold—Klectric Are. 399

or 3™™ long extended. At this point the bluish striations
began, growing more distinct and nearer together as the anode
was appr oached.

The stratified discharges Boe oneal by Gassiot,* de la Rue
and Miiller,+ and Child,t were doubtless manifestations of the
same phenomenon. From some of the beautiful photographs
published by de la Rue and Miiller it is clear that they had to
do with an are discharge in which the anode was not heated
to boiling. In fact they express the view that the are may be
regarded. as a modified stratified discharge. This, as we have
now seen, is strictly true only of the are on the first stage.

The small size of the striations would render the ‘experi-
mental determination of the distribution of potential between
the electrodes very difficult. An approximate value of the
potential drop for each striation may be estimated on the
assumption that the cathode drop is about 10 volts, the anode.
drop, judging from the observations of Skinner,$ on the order
of 30 volts. Ina typical case, with 164 volts across the are,
current 0°42 amp., and length 15, with six striations, a simple
computation gives about 90 volts as the average potential drop
necessary for ionization in each striation. This is less than
the drop per striation usually found in the glow discharge,
and is probably to be accounted for by the high temperature
of the gas.

$26. Carbon.—We have seen that three characteristic curves
may be drawn for the discharge between metals. For carbon
terminals still another stage exists, namely the hissing are, in
which the oxidation of the anode causes a further decrease in
voltage.

We have succeeded, by suitably varying the current and
the surrounding gas, in obtaining all four types of discharge
between carbon electrodes. The experiment was begun in
nitrogen with a 460 volt supply. Under certain conditions
striations were visible on both the glow and the first stage of
the are. After the current had been increased until the sec-
ond stage had arrived, air was slowly substituted for the nitro-
gen and the current further increased up to the hissing point.
Owing to the heating of the bell-jar, observations had to be
made ‘rapidly, so that no systematic series of readings could be
taken. The change from first to second stage was not well
marked, for the volatilization of the carbon anode seemed to
set in gradually as the current increased. Child| mentions

* Pogs, Ann., cxix, 133, 1863, +Phil. Trans., elxxi, 65, 187
t Phys. Rey., xx, 364, 1905. S$ Wied. Ann., Ixviii, 752, 1899.

| Phys. Rev., xx, 373, 1908.
Am. Jour. Sct.—FourtH SERIES, Vout. XXIV, No. 143.—NovemsBer, 1907.
28

400 Cady and Arnold—Electric Are.

these forms of discharge, referring to them merely as “ares.
He also noted still amen lhere Podineanion when cored a
were used.

Theory of the Change in Voltage at the Critical Point.

$27. Among the agents that influence the anode drop may
be mentioned TL) material and temperature of the anode, (2)
nature, degree of ionization, and temperature of gas in the
neighborhood of the anode, (3) ultra-violet light from the
cathode.

In passing from first to second stage, these factors are
affected as follows:

(1) The change in temperature of the anode may be
neglected, since, according to the theory that has been advanced,
the critical point is characterized by the arrival of the material
of the anode at its boiling point.

(2) On the first stage “the gas at the anode is chiefly that of
the surrounding space, with some vapor from the cathode. On
the second stage the current passes mainly through vapor sup-
plied by the anode. The contraction of the are at the anode
on the second stage indicates a considerable rise in temperature
of this vapor over that of the gas on the first stage.

(8) Since the change does not affect the cathode, the ultra-
violet light is unchanged, unless itbecomes partly absorbed by
the vapor before reaching the anode. On the other hand,
ultra-violet light from the anode itself on the second stage may
compensate for this. ;

It is probably. that a thermal e.m.f. exists at the anode on the
first stage in a direction opposed to the current, on account of
the high temperature of the anode, and, with iron and copper
in free air, perhaps also the fact that the oxides of metals emit
negative ions with comparative ease. When vaporization of
the anode] yegins, the increased temperature and density of the
vapor in contact with the anode, and its identity with the
anode in substance, conspire to reduce this thermal e.m-f.
and thereby to lower the anode dr op.*

Great experimental difficulties would stand in the way of
measuring the counter electromotive-force at the anode by the
method used by Duddell for the carbon are,t which so far is
the only method capable of yielding reliable results. It seems
most probable to the writers that the observed decrease in volt-
age at the critical point is due partly to this change in thermal
electromotive- force, partly to the increased conductivity of
the vapor in the arc. The theory would require modification

* Of. Stark, Retschinsky and Schaposchnikoff, Ann. Phys., xviii, 213, 1905.
+ Electrician, p. 918, 1901.

Cady and Arnold—Electric Are. 401

if it were shown that at the temperature in question the anode
emits not negative but positive ions.*

$28. A test for ionization in the neighborhood of the iron
are was made after the method of Child.t No perceptible
excess of positive over negative ionization could be detected,
either on the first or second stage. In this connection it would
be of interest to measure the rate of emission of charged
particles from iron oxide at different temperatures.

As for the absolute value of the anode drop on the first stage,
this cannot be assumed to be the same as on the glow discharge,
on account of differences in temperature, current and compo-
sition of the gas in the neighborhood of the anode, as well as
the presence of ultra-violet light from the cathode. Observations
made by means of a platinum exploring electrode in an are
with a silver anode indicated that the anode drop is Jess on the
first stage of the are than on the glow discharge. This would
be expected, in the light of Skinner’s observations on the con-
ditions that control the drop at the anode.t In our experiments
it was not possible to obtain voltage readings of more than
relative value.

$29. Melting Point and Heat Conductivity.—Much weight
has been laid by Grandgquist and others on the importance of
the parts played in the mechanism of the electric are by the
heat conductivity and melting point of the electrodes. It is
of interest to examine the connection between these quantities
and the occurrence of the critical point. For purposes of com-
parison, of still greater importance than the melting points
would be the boiling points of the substances used as anode;
but we have practically no data on this subject.

The only metals showing the critical point well in free air
were iron and copper. Of these, copper has the lower melting
point and the better heat conductivity. Heat conductivity of
the metal, however, does not play the controlling part here, for
in the case of iron it makes but little difference in the occurrence
of the critical point whether the globule is formed on an iron
rod or placed in a poorly-conducting carbon holder. This is
due perhaps partly to low heat-conductivity of the oxide globule
itself, partly to its somewhat loose connection with the metal
behind it.

We have found by tests with a blowpipe that copper oxide
melts at a lower temperature than magnetic oxide of iron. In
accordance with this, the critical point for copper comes at
smaller current and less expenditure of energy than that for iron.

* Cf. Polak, Elektrot. Zeitschr., xxiv, 599, 1907 ; Gehrke and Reichenheim,
Verh. d. Deutschen Phys. Ges., v, 76, 1907.

+ Phys. Rev., xii, 137, 1901.

+t Wied. Ann., Ixviii, 752, 1899.

402 Cady and Arnold—Electric Are.

For other metals, the data are too meager to afford comparison
between any except carbon and silver, and lead and silver.
With each of these as anode, the cathode consisted of carbon.
Among the observations recorded, the following are suitable
for comparison :

TABLE IT.
Anode Are
ao (at SS ae -—-—- ——— a ae NN
Materiai Mass Gas Pres. Length Voltage Current Watts
C IL}: Coane AMteses yom 214 0°35 48
Ag 3:0 21 5 100 14 140
Ebr 30 8 75 14 105
ee BHO 31 3 82 1°6 151

In Table II, the voltage and current are those on the first
stage, just before the critical point. The last column gives the
total number of watts expended in the are. The anode was of
approximately the same size and shape in each case, and sup-
ported in the same way in the apparatus described in $22. For

each pair of observations the gas pressure is nearly the same.
In the case of silver, the bell-jar contained air ; with lead and
carbon, nitrogen. Still, the silver did not oxidize perceptibly,
and unless the oxygen changed the electrical conditions through
catalytic action, the comparison holds good.

Comparing silver with carbon and also with lead, the expend-
iture of energy is less for the substance of lower heat condue-
tivity. Lead also melts at.a lower temperature than silver.
The constants of prime importance here are, of course, the
boiling points of silver and lead. The comparatively small
current at which the critical point occurs in the case of carbon
in spite of its high boiling point can only be explained by its
low conduetivity and on the assumption that the nature of the
surface of the anode, as regards roughness and the presence of
impurities, plays a very large part. “It may be said of the are
in general that far too ‘few trustworthy data have been assem-
bled for sweeping conclusions to be drawn as to the effect of
the physical properties of the electrodes.

Ill. Zhe Characteristic Curves of the Iron Are.

§30. In fig. 5 is shown a system of curves connecting eur-
rent with voltage across the are for various lengths. The
supply e.m.f. was 240. volts, electrodes 4"" in diameter. For
the sake of distinctness, the observations for 1, 4, and 8™™ are
represented by dots, those for 2 and 6™™ by crosses.

The meaning of the lines A.B, OD was referred to in $4,
in which the overlapping of the curves on the two stages was

elt) ee

Cady and Arnold—Klectric Are. 403

pointed out. The cause of this overlapping can now easily be
understood. Beginning with point 4, if the current is
decreased, the change back to the first stage does not take
place at once along the line BA, because the are on the second
stage is very contracted at the anode, so that the high tem-
perature of the positive base of the are is maintained until a
smaller value of current, represented at C, has been reached.
Here the expenditure of energy at the anode is so small that
vaporization is no longer possible, and the first stage enters in.

Fie. 5. Characteristic curves, iron are.

For an analogous reason the change from first to second stage
oceurs along AB, not DC. ; i
It will be noticed that the change AP occurs with a larger
current for long than for short arcs, and that the change in
voltage is also greater for long ares. The probable explana-
tion is that with a long are on the first stage the anode is com-
paratively cool, so that the current has to increase farther
before evaporation sets in. If the rate of vaporization increases
with the current, the change in current, and therefore also
in voltage, must be greater the longer the arc. This is the
opposite to what occurs at the hissing point of the carbon arc.*
If the are burns with the electrodes almost in contact, the

* Mrs. Ayrton, The Electric Arc, p. 288.

404 Cady and Arnold—Electric Arc.

first stage cannot be observed, because the anode is kept at a
high temperature from its proximity to the cathode.

In the curves for 2 and 8™™ the maxima corresponding to the
“rotating stage” of the arc are evident (see $33). So unsteady
was the are duri ing this stage at lengths of 4 and 6™™ that no
readings could be ‘taken. The maxima are presumably due to
increased voltage caused by a bending or lengthening of the are.

$31. Up to the present time, no “satisfactory mathematical
expression for the mechanism of the metallic arc has been pro-
duced. Steinmetz (l.c.) has derived a formula for the potential
difference across the iron arc as a function of current and length,
which is in fair agreement with observation over the range
experimented on. This formula is open to criticism, however, in
that it takes no account of the drop at the anode, ‘and assumes
the temperature throughout the are to be that of the boiling
point of the electrodes.

As for the curves in fig. 5, their range is limited by the
comparatively low voltage of the supply. The presence of the
critical points and the irregular ities due to the rotation of the
are also contribute toward “making the observations worthless
for analysis.

832. Wide discrepancies exist between the values given by
different observers as the “constants” of the metallicare. Thus
for the value of @ in the equation

DRO A EH fs aeates Sarees ces (i)

v. Lang* gives for iron 25 volts, Lechert 20, and Child (1.c.)
25 volts. The value obtained from the experiments of Stein-
metz (I. ¢.), who observed the potential difference the instant
before the electrodes (magnetite rods) were brought together,
is from 28 to 31 volts.

These differences are inherent in the different methods
employed. When the are approaches the length zero, the anode
receives negative ions directly from the cathode and its drop is
all but eliminated. Moreover, the resistance of the iron oxide
from which the discharge takes place depends upon its temper-
ature, and this forms a very appreciable source of error for
which it is almost impossible to correct. The value of @ ob-
tained by this method cannot be compared with one derived,
for example, by extrapolation from the £7 curve, nor with one
obtained with the use of an exploring electrode, Differences
in the currents used by various observers account also fora
part of the discrepancies.

We have made a large number of observations in which the
iron terminals were slowly brought together and the last volt-

* Wied. Ann., xxxi, 384, 1887. + Ibid., xxxiii, 609, 1888.

Cady and Arnold—Electrie Are. 405

meter reading recorded before the electrodes touched. Little
reliance is to be placed on the precision of these values, on
account of the variable resistance of the globules themselves,
as stated above. Taken all together, our observations indicate
a value of from 21 to 23 volts as the critical potential differ-
ence for zero length of are. Probably the best method wouid
be to record the drop in voltage by means of an oscillograph,
which would indicate the true residual drop after the globules
had touched and before they had time to cool perceptibly.

Observations of the critical voltage for copper by the same
method gave about 22 volts at 4 amp., in fair agreement with
v. Lang’s 23:9. Child (1. c.) found as the sum of the positive and
negative drops for copper, using a carbon exploring electrode,
at 6 amp., 25 volts.

From the observations represented in fig. 5, together with
other data covering a wider range, we have constructed a
system of curves connecting voltage with length of are, for
the arc between iron terminals. These make it evident that
the linear law represented by equation (1) does not hold,
especially for small lengths. The curves are convex upward,
suggesting that an equation of the form

B=a+bl+eV1.

in which a, 6, and ¢ may be functions of the current, would be
nearer the truth. We prefer not to assign numerical values
to these constants until more complete data have been obtained.

For the iron are in nitrogen, the 4 curves have the same
general form as in air. For the same current and length of
are, the potential difference is about twelve volts less than in
air, as was found by Arons.* The critical voltage (zero length)
was perhaps two or three volts less than in air. :

IV. Electrical Oscillations in the Metallic Are.

$33. Lotations in the Iron Arc.—In the first of the prelim-
inary papers already referred to, mention was made of an effect
noted with the iron are in free air. After the current has been
increased on the second stage to a little below two amperes,
the positive end of the are begins to rotate, so that on the
anode a ring instead of a spot of light appears. This is accom-
panied by a high-pitched whistling sound, which as the current
is still further increased degenerates into a sputter, and this in
turn into a steady, strong hiss, the ring meanwhile having
disappeared. The arc is in great agitation, swaying about
from side to side, and sometimes separating into several ares.
At the beginning of whistling, there is a curious tendency for

* Ann. Phys., i, 706, 1900.

406 Cady and Arnold—Electric Are.

the are to spring back on to the first stage, so that for an instant
hissing ceases, current falls abruptly, and e.m.f. rises several
volts.

If one begins to increase the external resistance immediately
after one of these changes, the first stage can sometimes be
maintained steadily for a short time, even though the current
is far greater than that at which the critical point normally
occurs. The reason evidently is that the anode becomes so
far cooled that evaporation ceases. A maximum of the charac-
teristic curve is often found in the neighborhood of this stage
(cf. fig. 5).

The effect mentioned reminds one of Trotter’s observation
of rotations in the carbon are at the beginning of hissing, but
the cause can hardly be the same. It is more likely that it is
a case of electromagnetic rotations. Opportunity has not yet
been found to examine this more closely, but it is proposed to do
so with the aid of an oscillograph. That marked oscillations in
current are here taking place was shown by the fact that when
the are circuit included the primar y of an induction coil whose
secondary was connected to a quadrant electrometer, a large
deflection ensued.

$34. Lests for Continuity of Discharge.—The following
experiments were made with a view to testing the assertion of
Lecher,* which has found general acceptance, that the are
between terminals of silver, copper, or of carbon on the quiet
stage, is perfectly continuous, while that between terminals of
iron, platinum, or of carbon on the hissing stage is discontin-
uous or intermittent.

The nature of the fluctuations between carbon terminals has
been investigated by Duddell, who found that on the hissing
stage the are is not ‘inter mittent, but subject to irregular fluc.
tuations.

Lecher’s method was, to connect in parallel with the are a
condenser and stretched brass wire in series. Changes in
potential difference between the are terminals set up_ oscilla-
tions in the condenser circuit, thereby heating the wire and
causing it to expand. Now, recent work on the “singing”
metallic are indicates the possibility of oscillations being set up
in such a circuit as this, in spite of its small self- inductance,
even though the are of itself is not discontinuous. We decided
to investigate this, and at the same time to study the oscillations
in the neighborhood of the critical point.

$35. To measure the intensity of the oscillations, we used a
bolometer similar to Fessenden’s barretter, as this was easier
to construct and more sensitive than Lecher’s hot wire. A
platinum wire about 4™ long and 0:024"™" in diameter, having

* Wied. Ann., xxxiii, 609, 1888.

Cady and Arnold—Electric Are. 407

a resistance of about 12 ohms, was connected in series with a
mica condenser of one-third m.f. capacity, the two forming

a shunt across the are. Recent experiments of Austin* call
ee to the importance of having a low resistance in series
with the condenser. * Nevertheless, the resistance of our bolom-
eter did not prevent the production of measurable oscillations
as will be seen. The platinum wire was connected as one
arm of a Wheatstone bridge.

At the beginning of the experiment the are was struck and
the bridge balanced with battery current flowing, the condenser
circuit remaining open. When the condenser switch was
closed, the resulting deflection of the galvanometer in the
bridge was taken as a measure of the intensity of the oscilla-
tions. Current for the are was taken from the 240 volt mains.

As a preliminary test, carbon terminals were employed. No
deflection could be observed when the are was quiet, but the
light went violently off the scale when the are hissed... On the
quiet state of the carbon are oscillations are not to be expected,
on account of the small self-inductance and large resistance of
the shunt circuit.

§36. Lron terminals in free air were next used. ‘The self-
_ inductance of our shunt circuit cannot have been appreciably
greater than Lecher’s. Yet we found it possible to make the
iron are “sing” audibly, even on the first stage, with the cur-
rent a little below one ampere. The tone was a high-pitched
squeak or whistle. Inserting a self-inductance of 0-000 04
henry caused a lowering of pitch, though not as much as would
be expected, showing that under these conditions the simple
formula for period,

= Inv LC
does not hold.

When the are was singing, the heating of the bolometer
threw the light off the scale; the are current increased, while
the voltage dropped several volts, as has been observed by
Austin (I. ¢.).

Whether the are sang audibly or not, the galvanometer was
always found to be deflected, but in general much more on the
second than on the first stage. As the oscillations were gener-
ally of too high frequency to produce an audible sound, no
direct estimate of their period was made, nor could their
dependence on current and length of are be observed. The
galvanometer deflections depend on the total quantity of elec-
tricity passing per second through the shunt circuit, and this is
a function not only of frequency, but also of amplitude and
wave-form. We found the deflections to increase as the cur-

* Bull. Bur. Stand., iii, 325, 1907.

408 Cady and Arnold—Electric Are.

rent increased and also as the length diminished. This may
have been due to increasing frequency, since Simon* has shown
the frequency to vary in the same manner. Adding non-
inductive resistance to the shunt circuit always greatly reduced
the deflections.

The experiment of inserting in series with the condenser a
small coil having a self-inductance of 0-000 04 henry was also
tried. This always diminished the deflections when the cur-
rent was over 2amp., but as the current was decreased, its

To BRIDGE

le 240 VoLt
in SuPPLy

Fia. 6.

damping effect was reduced until at small currents it sometimes
even caused an increase in the deflection. This points to the
existence of a sort of resonance between the are and the shunt
circuit, and is probably related to the fact that a decrease in
the are current tends to reduce the frequency.

$37. Hxperiments with the Copper Arc, Lecher’s Method.—
With no self-induction in the shunt cireuit, closing the switch
of the latter always put the are out unless the current was over
three amperes. Galvanometer deflections were not large, and
they grew smaller as the current increased. That Lecher
found no effect with copper is probably due to the lower sensi-
tiveness of his apparatus.

With 0-000 04 henry in series with the condenser, the cur-
rent could be reduced to one ampere before the are went out.
Deflections increased with decreasing current, and before the
are was extinguished the lhght disappear ed from the scale.
This illustrates the effectiveness of the copper are at small cur-
rents for the production of high frequency oscillations.

* Phys. Zeitschr., vii, 495, 1906.

Cady and Arnold—Electric Are. 409

$38. It seemed desirable to test the continuity of the are by
some method less likely to set up disturbing oscillations.
When the condenser-bolometer system was connected in paral-
lel with a non-inductive resistance which was in series with the
are, no deflection whatever could be observed. But the con-
nections shown in fig. 6 were found to give excellent results.
A is the arc, R a regulating resistance, Z the primary of a
large induction coil whose iron core was removed, consisting of
200 turns of heavy wire, with a self-inductance of about 0-0012
henry. In parallel with Z are the capacity C of 3°6 m.f., the
bolometer wire & and the switch A.

When A’ is closed, the bolometer is not heated as long as the
are current remains constant. But any fluctuation of current,
producing a varying potential difference between the ter minals
of Z, sends an alternating eurrent through the shunt circuit
and causes a deflection of the oalvanometer in the bridge.
Correction was made for a slight deflection due to the passage
of the commutator bars under the brushes of the generator.

With the iron are in air, deflections of the order of 30 div.
were found at the beginning of the second stage, which reached
a maximum while the are was rotating at about 2 amp. and then
decreased again. There was no deflection on the first stage
after the effect of the commutator had been eliminated.
When a singing are was produced on the first stage by con-
necting a second capacity and inductance in parallel with the
are, a lar ge deflection resulted.

Sahulka* finds that a self-inductance connected in series
with a singing are oscillates with a frequency of its own, and
not simply with the are frequency. This point we have not
tested.

A simple method of comparing frequencies is to insert an
iron core in the coil Z, fig. 6. When this was done, it was
found that for a singing are on the first stage the insertion of
the core produced no perceptible effect on the galvanometer
deflection. With the are on the second stage, not singing, the
iron core increased the deflections about tenfold, while the
fluctuations due to the commutator of the generator gave a
deflection that was multiplied about thirty times by the inser-
tion of the core. Thus the oscillations due to the volatilization
of the anode on the second stage are intermediate in frequency
between those due to the commutator and those of the singing
are. The last are so rapid that the self-inductance of the coil
is no longer increased on inserting the iron core.

Tests were also made with the iron are in nitrogen at atmos-
pheric pressure. The irregular sputtering condition of the
are gave rise to a continual heating of the bolometer, which

*Kel. Electr., 1, 396, 1907.

410 Cady and Arnold—Electric Are.

was most pronounced in the unstable state just before the are
went out ($8).

$39. Oscillations in the Silver and Copper Arcs.—The eop-
per are was formed in air, the silver in nitrogen. The appara-
tus shown in fig. 6 was ‘used, with 460 volt supply. In the
case of copper, the oalvanometer deflections at currents over
1-5 amperes were small, but they increased with decreasing
current, and at a little below one ampere the hight disappeared
from the scale. This is similar to the effect found when the
bolometer circuit was in parallel with the are, and indicates
that we have to do with a peculiarity of the are itself, and not
with oscillations imposed upon it from without. No special
effect was noted on passing from the second to the first stage
of the copper are.

In the case of silver, the effect was even more pronounced,
especially at low pressures.

A silver anode and carbon cathode were used. At two
amperes and over the deflections were very small, but as the
current was decreased, the deflections grew larger, becoming
very pronounced on the first stage. With small currents at
7" pressure, the deflection was far too great to be observed.
The tests with silver indicate that the cause.of the effect lies
chiefly at the anode. There are reasons for thinking that these
oscillations may be due to an exceedingly rapid changing back
and forth between the two stages ‘of the are, or perhaps
between are and glow: this point is to be investigated further.

It is possible that the sensitive condition of the are in the
neighborhood of its critical point may be such that this region
will prove to be a particularly good one to select for the pro-
duction of high frequency oscillations.

Summary.

$40. I. There exist between iron terminals in free air two
ty a of are, distinct from each other in appearance and in
characteristic curves. To these types the terms first stage and
second stage have been applied for the sake of brevity. The
transition from one stage to the other is somewhat similar to
the hissing point of the carbon arc, but the cause is not the
same. Evidence from loss in mass of the electrodes, spectrum
of the arc, and temperature of the anode, shows that the first
stage is characterized by the absence of vaporization of the
positive terminal.

II. Experiments are described on the change from the glow
discharge to the are in free air between terminals of iron,
platinum and copper. Attempts were made at identifying
the two stages for arcs between several different metals in air.

As oe Rall

Cady and Arnold—EHlectrie Are. 411

Both stages could be maintained steadily only in the cases of
copper and of iron. In nitrogen at reduced pressures the glow
and both stages of the are were observed between a carbon
eathode and anodes of lead, silver, and carbon. There is
reason to believe that similar results ean be obtained with
most other metals. Platinum, and iron free from traces of
oxide, did not show a stable first stage under any conditions.

The diminution in potential difference between the arc ter-
minals in passing through the critical point from the first to the
second stage is attributed partly to a decrease in the thermal
e.m.f, at the anode, partly to increased conductivity of the are
vapor.

There is some evidence of a connection between the critical
point and the melting temperature of the material of the
anode, but it is not well marked, and is likely to be masked
by effects due to the physical condition of the surtace of the
anode.

III. The characteristic curves of the iron are are discussed,
together with the probable form of the equation connecting
voltage, current, and length.

IV. A series of experiments was carried out on electrical
oscillations in ares with iron, copper, and silver anodes.

At a current a little below two amperes, on the second stage,
the iron are rotates and emits a whistling sound. This is
often attended by spasmodic changes back to the first stage.

The application of Lecher’s test for continuity led to observa-
tions on the “singing” iron and copper arcs, the energy
expended in the oscillations on both stages of the arc being
observed by means of a bolometer. ‘To test the continuity of
ares between various metals by a method more free from error,
a new manner of connecting the bolometer was devised. <A
general agreement with Lecher’s results for large currents was
found. But the iron are became continuous on the first stage,
while in the case of ares with silver or copper anodes, marked
oscillations were detected as the current decreased.

Scott Laboratory of Physics, Wesleyan University,
Middletown, Conn., July, 1907.

412 Ewell—Gibbs’ Geometrical Presentation of the

Arr. XXX VI.—Gibbs’ Geometrical Presentation of the Phe-
nomena of Reflection of Light ; by Anruur W. Ewnut.

In his lectures at Yale upon the Electromagnetic Theory of
Light, Prof. Gibbs briefly outlined a ceometrical method of
representing and interpreting the phenomena of total and
metallic reflection, which, to the writer’s knowledge, has never
been published, and is not included in his collected papers.
The writer has found this method so valuable in teaching and
in experiments upon metallic reflection, that he believes others,
outside the small number who were privileeed to hear Prof.
Gibbs, will find it a clear, simple and practical method of
representing the relative amplitude and phases and optical
constants, the analytical expressions for which are complicated.

The foundations of this paper are lecture notes of the years
1898-99. In place of Prof. Gibbs’ vector notation, the more
common scalar quantities will be used, and the writer has added

extensions of this method to the simpler cases of reflection,
applications to the determination of refractive indices, coefti-
cients of absorption, and reflecting power, and numerical
ilustrations.

a le

Phenomena of Reflection of Light. 413

The method, in brief, consists in representing by lines the
reciprocals of the velocities of light in the two media between
which reflection occurs, and aiso their components parallel and
perpendicular to the surface. These lines are so chosen that
very simple geometrical quantities represent Fresnel’s Laws,
which are assumed, together with the ordinary laws of reflec-
tion and refraction. As the lines and angles vary with the
angle of incidence and the optical constants of the media, the
relative amplitudes, phases, ete., are given by simple geomet-
rical relations.

Light is reflected at the surface, separating medium (1) from
medium (2). Let C, and C, be the components, normal to the

; : 5 cer il 1 h
Surmtace, of the reciprocal velocities, —\and!~, (fie. 1). ) Let)?
v v
j i
be the angle of incidence, or reflection, and 7 the angle of refrac-
tion. ince the angles io enone and reflection are equal

and since the ratio of the sines of the ancl: of incidence

es 1
and refraction is Ui v,, the components, Ay of = and —, parallel
v v

1 2

to the surface, are equal.

For simplicity, we will take as unity the components of the
incident light vibration, parallel and perpendicular to the plane
of incidence (i. e., the ght is polarized-at 45°). Let the
amplitude of the component of the reflected vibration, parallel
to the plane of incidence, be R,, and of the perpendicular
component, R,. Fresnel’s Laws of Reflection are then:

enh EU (Cael, R, 1s G08 (i+7r) sl tan(¢—r)
: sin(?+7r)’ RK, cos (¢—71)’ " tan(@+7)
From the above figure and notatian,
%

Sind — er, Cos) 1 © w sin 7 Aw cos == Oly)

Substituting :

CLG TR MeeCG

Ro BRO,
The latter is of course the more interesting and important.
The product of the two will give R,.

2
<-------- Ge) ee se isa) Soran ae
«---G--
Gas of Gee By ing eauen, Sune Nes GH) cee SSS OES
7, / 6 5 2] 3
Sed age oper as Oral hose ap? ULL a a ener ae ema ?
< Tes --5---- >

These quantities will be represented graphically. Horizontal
distances we will consider real, and positive if to the mght:

414 Ewel— Gibbs Geometrical Presentation of the

: : ; . : ae Le ykaee 1
Vertical lines are imaginary. In fig. 2, let 13 = =, 1b Sea
1 2
aS INS Since - WAG (OF. 63 =C,’,65=C,°. Let 62 bea mean
proportional Bela een 63 and 65. Then 62 = C,C,, 64= — 62
and we have the simple formule :
= CeCe 32 tet

C74 SCC = aa: ReaD)

To apply these expressions to any case of reflection, take

2 as unity, and lay off sin “2 as (A®*, 16); cos “7 as C,’, (63);
1

2
ae nN, : , :
and 15 as —, ( where 7 is the refractive index). Locate
Ve
2 1

62 and 64, the mean proportional, and find the above ratios.
Fig. 2 is thus drawn for light, incident in air, at 48°, upon
elass of refractive index 1°53.

Bey Se aye ees etecene

Jag 12 84
Both are negative, i. e., there is a difference of phase of +
between R, and R,, and between R, and the perpendicular
component ‘in the incident light.

As the angle of incidence increases, 6 moves to the right and

therefore 14 decreases until, at the Polarizing Angle, it becomes

R
P af . . .
zero and hence also Ro: Near grazing incidence A and

as R
hence 16 is large, R is now positive and both it and Kh. are
increasing with increasing angle of incidence.
Suppose the light is passing in the opposite direction, i. e.,
from a medium where the velocity is less to one where the

2
v0

ee ESD FIR pen rele Fe (Coe ea a = =)
eee Pee - -Ct«t)- -->
3 J
3 e
7) ‘ CATR cat yh -f- = = - RIT
y ale Se en aS
CE i oo ECS re yey) a

velocity is greater. The various quantities for small angles of
incidence are represented in fig. 8. R, = ae positive,
B)

—? — ~~ js negative and as the angle of incidence is increased,

Phenomena of Reflection of Light. 415

6 moves to the right and as before, at the Polarizing Angle 14
becomes zero. Fig. 3 is drawn for light incident at 30° upon
a glass (1=1'53) — air surface.

R 14 32
2 = — 185, Re = 528
R, 12 34
As the angle of incidence is still more increased, 16 A* exceeds
Ae l ae ‘ é ie .
5 = —, and 65 = C,’ becomes negative, and hence C, is imagin-
Ae 2

2

: : iy 1 :
ary. At the point where 6 is at 5, A* = —, or the emerging
a

2
waves, travel parallel to the surface and the angle of incidence
is the Critical Angle. Beyond this angle, conditions are repre-
sented by fig. 4, which may be taken as typical of total retlec-
tion. Since 62 (= C,C,) and 64 are mean proportionals of 65
and 63 and ©, is imaginary, 2 and 4 must lie on a circle whose
diameter is 53. Evidently canprinemeal les,

R ze neds
rR =1=h,=R,, and the angle between 12 and 14 gives
Vs

the phase difference between the two components and the angle
between 32 and 34, the phase difference between R, and the

incident, perpendicular, component. The latter phase difference
is evidently zero when total reflection begins and 7 at grazing
incidence. The difference of phase between the two compo-
nents, angle 412, is seen to be zero at both extremes and a
maximum when 72 and 14 are tangents. Tor this case, by
geometry,

Qe oo! —— Pras po — Piel se Kehoe G
52:23) = 56:68 = 18:13 = Ro TER OMS Be = Oa Oe

the maximum angle 412 =2—4 angle 235 =z — 4 tan”? —

Am. Jour. Sci1.—FourtH SERIES, VoL. XXIV, No. 143.—NovemseEr, 1907.
29

416 Ewell—Gibbs’ Geometrical Presentation of the

Fig. 4 is drawn for light incident upon a glass (7 =1°53) — air
surface at 48°. The angles of phase difference are : angle

RK, ;
412 (R) =46 °°, angle 234 (R,) = 56°. The maximum
difference of phase between the two components = 7 — 4
1
tana Ae
1°58

We have hitherto considered perfectly transparent bodies.
We will now consider the general case, so called Metallic
Reflection. The general expression for the light disturbance

may be written:
mal Ww
Qarkw Qari (is )
SS ING, FN é w
where X is the instantaneous force, A = amplitude, ’ = wave
length, T = period, ¢ = time, # = coordinate for the direc-
tion in which the light waves are advancing, and &

oO

Coefficient of Absorption, i. e., the amplitude is reduced to

1 Ra Siecle. : ‘ Xr
a of its initial value in a distance -

If we write the
ork

above equation in the form:

Qi [ t w
Z <2 (-1)
X = Ae G Rie
and determine the reciprocal velocity of the light waves, we

: ih
obtain the complex expression : as ik). The real part,

Jha. : ; : 1 p
38 the ordinary reciprocal velocity, >, or proportional to

Phenomena of Reflection of Light. AIT

the refractive index, and the imaginary portion is proportional
to the product of the refractive index and the coefficient of
absorption. If we chose as unity the velocity in pure ether,
the real part is the Refractive Index, m, and the imaginary
portion is 2h.

In total reflection, O°, as we have seen, is a pure imaginary,
i. €., NO waves are propagated in the second medium perpen-
dicular to the surface. In the general case of light meeting
the surface separating two media, waves are propagated in the
second medium with some absorption, 1. €., Cis complex. Fig:
5 represents a typical case of so-called ’ Metallic Rédectiea
We will suppose that the first medium has no appreciable

é : : ; Hai i subs
alsonption., As in previous meures, 13 — —,, 16 = Av, 63 — ©.
v,

ve 7 = 1 a= .
a = On ols = Aue ©,”, (65), being complex, the mean propor-

tional between it and 63 will be 62, whose length is the arith-
metical mean proportional and such that the angle 268 is half
the angle 563. _

iw :

Since) = Bet the phase difference must decrease from 7 at

elie

perpendicular incidence (6 at 1), to zero at grazing incidence
32

(6 at 3), while the phase of R, = = varies from zero to 7. At

the Principal Incidence, I, the phase difference is + or 14 and

72 are at right angles. The numerical value of = is the

12

Principal Azimuth.
To determine the refractive index, m, and ae coefficient of

absorption, %, we extract the square root of - i Giana bhie

real part of the root is (; being unity) ane ie Imaginary
- s

zt
portion is nk.
Fig. 5 represents reflection from copper in air, when the angle

of incidence is 71° 35’, the Principal Incidence (Drude). The

velocity of light in air, v,, is taken as unity and = is repre-
1
Senedd) Dyesls a 6) — es — She Soamds 634—" C1 i—
71° 35’. The Principal Azimuth of copper is 88° 57’.. Since
a
2

at the principal incidence, I, the difference of phase is ->-

418 Ewell— Phenomena of Reflection of Light.

and 14 are drawn at right angles, in the ratio 14:12 = tan
38°57’, and such that 24 is bisected at 6. From the pre-
ceding, 62 = O,C, and 64 = — C,C,. C,’= 65 is now con-
structed. Since 62 18 a mean proportional bemveen 63 and d 63,
the angle 563 is made twice angle 263 and the length of 65 is

I Bld . a :
such that 62° = 63 X . 65. =a =A°*+C> isithen 15-5 Lhe pomine

2

being determined, the relative amplitudes = and phases

\5
(angle 412) may be determined for any angle of incidence, 2,
since sin’ 7 = 16 and 62 = — 641s a eeometric mean proportional

between 63 and 65.

Let is be the geometric square root of 15 (or the numerical
square root at an angle 816 = 4% angle 516). The real part,
19, is the refractive index, n, and the i imaginary part, 98, is nk,
where & is the coefficient of absor ption. Fig. 5. gives, for
copper, 7 = 62, nk = 2:6, hence k= 411.

It is not difficult to show also that # is the tangent of twice
angle 124, which is half the angle 263 (Drude’s Q).

R, (= a is numerically -94 and the angular difference of
phase (angle 234) is 167°

The intensity of the reflected light is:

B2\7 14\?
Ree Re ene E 5
ie a af al |
(which equals 146 for the above illustration, the incident
intensity being 2 or 73 per cent), and evidently increases rapidly
1
with increase of the imaginary portion of —z, 1. e., of #, thus

illustrating Selective Reflection.

Worcester Polytechnic Institute,
Worcester, Mass., August, 1907.

CO. Barus—Decay of Ionized Nuctes. 419

Art. XXXVII.—The Decay of Ionized Nuclet im the Fog
Chamber, in the Lapse of Time; by C. Barus.

1. Introduction. —The attempt was made in an earlier
paper to standardize the coronas by aid of the decay curves of
radium. The method is apparently very simple’and requires
the knowledge merely of the coronas appearing under given
circumstances when the radium tube is in place d on the out-
side of the fog chamber, in comparison with the coronas
observed under the same circumstances when the radium has
suddenly been removed for different lengths of time before
condensation. From electrical observations with condensers,
the equation

dn / dt = — bn? or 1/n=1/n'+b (t—-7’)

is found to be adequate if m and 2’ denote the ionizations occur-
ring at the times ¢ and ¢’, and the same would appear to be the
case with the corresponding nucleations. Moreover, if the
relative nucleations for two coronas obtained at a given
exhaustion are known (for instance by the earlier method of
geometric sequences) the absolute values of the nucleations
will follow. With aradium ionization at ¢ and ¢’ seconds after

its removal
i ae pase
p= cE 1) / 6-7’).

But the attempt to carry out this apparently straightforward
method leads to grave complications. If mn be reckoned in
thousands per cubic centimeter, the electrical value of 6 may
be taken as of the order of 6=-001; while the value of 6
which I deduce from the decay of ions in the fog chamber, is
more than two times as large as this, increasing moreover very
rapidly as the nucleation is smaller. True it is possible that
the method for finding the nucleations, absolutely, may be at
fault. If relative values seem to be trustworthy, absolute
data are not to the same degree substantiated ; but even if this
were granted, however improbable, the march in the values of
6 would be unaccounted for and seems to be a new phenome-
non.

2. Data. Exhaustion above the fog limit of air.—In the
first series of experiments the adiabatic drop of pressure 6p
was somewhat larger than the condensation limit of dust free
wet air. The initial coronas were small as the radium was
weak (10,000, 100 mg).

When the values of 6 were computed from the means of
‘successive pairs of measurements of nucleation n, at different

420 C. Barus—Decay of Lonized Nuclei.

times, ¢, a somewhat irregular increase of b, was observed as n
deer Areca When the first observation was combined with the
fourth, ete., the values were, z = °29 being the relative drop
of pressure, , 8p/p,
b = 0029
34
36
4]

or a mean value, 6 = :00338 (m reckoned in thousands), if the
last observation is ignored, since the coronas are just visible
here.

If the electrical datum, / = -0014 be correct, the present
nucleations n are to be increased on the average, 0003/" 0014=
23 times, If the last datum for were included much more.
This is quite unreasonable. One must conclude therefore that
6 for nuclei is larger than 6 for ions or that an ion, acting as a
nucleus in a saturated atmosphere, decays (dn/dt=—bn’) sey-
eral times as rapidly,as the same ion in a dry atmosphere
when tested by the electrical conduction of the medium.

If but a part, x, of all the ions are captured, m’ escaping,
we may write

—dn/dt — dn'/ di = bn? +2 bnn' + bn”

so that both da /dt and dn’/ dt are larger than bn’* and bn” ;
bi
— 2dn/dt= 4 bn’, or — dn / dt = 2 bn’.

If but 1/8 ofall the ions, 3n, are captured, — dn /dt = 9 bn’;
etc. Hence, if but 1 / mm of all the ions are captured the coeffi-
cient of decay being as found should be about m times too
large as compared with the true value.

This does not explain, however, why the coefficient }
increases when ¢ is larger and 7 is smaller; if it were addition-
ally assumed that ions decrease regularly in size as they decay
more and more, so that they withdraw more and more fully
beyond the given range of supersaturation applied, the second
part of these occurrences would also be accounted for ; ; but the
assumption is not probable.

3. Exhaustion below the fog limit of dust free air.—lt would
follow from what has just been stated that if the drop of pres-
sure is lower, the values of 6 obtained must be larger. For not
only are few of the ions caught but the diminution of bulk
(virtually) which may accompany the decay would place them
sooner out of reach of the given exhaustion as the interval of
decay increases. But in experiments of this kind, the succes-
sive values of 6 again show an outspoken march into larger
values as the time ¢ increases.

eae

C. Barus—Decay of Llonized Nucle. 421

If we combine the first observation with the fourth, etc., as
before, and
$p/p =a = 27,6 = 0038,
041,
057,
134, or a mean value of 6 = ‘0045
(when m is reckoned in thousands), if the last observation is
ignored. But to ignore this value is here quite inadmissible,
as the data for a parallel series where @ = °25, viz.,
Or 02
sided

fully show.

4. Data for weak ionization. Radium at a distance.—
In the above work the initial intensity of radiation was the
same. It was suggested that the average size of a nucleus
might decrease in the lapse of time. Thus a variety of further
questions arise; 1, whether weak radiation produces a smaller
average nucleus; 2, whether a stronger radiation does the
reverse ; 3, whether the limit of 6 decreases as the exhaustion
increases and finally approaches 6 = :001 (counting 7 in thou-
sands), etc. The experiments of the following work show that
6 varies with the number of nuclei present, no matter whether
a given nucleation is due to weak radiation, or to decay from

a stronger radiation, or finally to low exhaustion ; or that the
nuclei probably br eak to pieces as a whole.

The data, moreover, were investigated by the new method of
two diffraction sources of light, S em. apart, at a distance
from the fog chamber. The number of nuclei, 2, found in the

exhausted fog chamber, is corrected by multiplying by the
volume expansion. Finally, b was computed from pairs of
observations about 20 seconds apart. - Water nuclei were
always precipitated before each test. With the exhaustion
slightly above the condensation limit of air, the data were
constructed in comparison with cases for stronger radiation
and of weaker radiation (by decay) in the above experiments.
Together they formed a coberent series of curves, proving
that it is the number 72 present which determines the value of
6, no matter whether the small number is due to low exhaus-
tion (6p/p near the condensation limit), or to decay of ions in
the lapse of time (exhaustion ¢ seconds after removing the
radium from the fog chamber), or due to lower radiation
(radiation at some distance 40, from the fog chamber.)

The results may be otherwise summarized, by giving
—h = (dn/dt)/n* in terms of the nucleation n, from which the
decay takes place. The rapidly increasing values of 6 when 2
is smaller and their tendency towards constant values when 7

422 C. Barus—Decay of Ionized Nuclei.

is larger (remembering always that the ionization is through-
out low) are then apparent.

Exhaustions above the condensation lmit of air fails to
bring out the usual high values of 4, for the ionized nucleation
eventually emerges into the vapor nucleation of dust free air.
These high values appear if the exhaustion is low enough to
catch but few vapor nuclei while being high enough to insure
large coronas due to ions.

Two series of experiments made with this end in view con-
firm the occurrence of large values of 6 associated with small
values of », no matter how the latter are obtained.

If the true equation of the decay curve, dn/dt, were known,
it would then be worth while to reduce all the data to a com-
mon seale; but the graphs obtained oy that the values of 6
rather suddenly increase below 10-*n, = 10, so that a simple
relation is not suggested for the aneco

The question arises incidentally whether the ions may not
vanish by accretion, i. e., their number may be reduced because
individual ions cohere ; in such a case the fog limits should be
reduced for which there is no evidence. There seems to be an
independent second cause for decay entering efficiently when
the nucleation becomes smaller. We may, therefore, pertinently
inquire into its nature.

5. Case of coinbined absorption and decay of ions.—The
most promising method of accounting for the above results has
been suggested by the work done in connection with the behav-
ior of phosphorus nuclei.* There may be either generation
or destruction of ions proportional to the number 7, present
per cubic centimeter, in addition to the mutual destruction on
combination of opposite charges. In other words the equation

—dn/dt=a+ten+ br’

is now applicable, where @ is the number generated per second
by the radiation, ¢ the number independently absorbed per sec-
ond and bn? the number decaying by mutual destruction per
second. Here ¢ is negative for generation and positive for
absorption. If @ is zero

1/n ib /n,+ (1/n,+ b/c) (e c(t—t,) <3 1h).

where the nucleation 2 and nm, occurs at the times ¢ and f,
respectively. Hence, when ¢ becomes appreciable
dn | dt C
= = = — b,

Te 7

or the usual decay coefficient increases as m diminishes, becom-

*Barus: Experiments with ionized air; Smiths. Contr., No. 1309, 1901,
pp. 34 to 36.

CU. Barus—Decay of Ionized Nuclei. 493

ing infinite when n= 0. This is precisely what the above
experiments have brought out. The value of 6 does not appear
except when 7 is very large. Since 6 is of the order of 10~°,
if c is of the order of 3 < 10-* (as will presently appear), ¢ /n
will not be a predominating quantity when v is of the order of
10° or ¢/n =3 X10"; but it will rapidly become so as n

approaches the order of 10‘ or ¢/n=8 X 10-°, which again
is closely verified by the above data.

Again, if —dn / dt = — a + cn + bn’, the conditions of equi-
librium are modified and become, since dn / dt = 0,

a=cn + bn?

where @ measures the intensity of radiation. It no longer
varies as n”, for

n= = ( 144/144 ab/¢c).

6. Absorption of phosphorus nuclei in tubes.—The method
of the preceding paragraph, applied to the data obtained in
the given paper with phosphorus nuclei, leads to striking
results. It shows the possibility of computing nucleation by
passing a current of highly ionized air through tubes of known
length and section (absorption tubes) into the steam jet appa-
ratus there developed. But there is no room for these results
here.

t. Hurther data and results—Experiments with special
reference to the views just given were made at some length.
Their general character is shown in figures 1, 2, 38, where the
abscissas are the times elapsed since radiation was cut off and
the ordinates the number of nuclei caught in thousands per

424 C. Barus—Decay of Lonized Nuclei.

ceubicem. It is not possible, however, from results of the char-
acter of the present, to discriminate sharply between e and 4,
and the endeavor will have to be made to select the best values
from inspection.

In these series the constants obtainable for different inter-
vals of time separately for each series would be as follows:

Series i 1 2 2) 3 3 4
10°O out, 2°9 “82 °88 ‘61 56 Oi
10° e — 18 — 20 45 32 4] 40 39

The mean data of series 2 to 4 would then be 6 = -000000,79,
¢=°'039. There is a curious consistency in the constants
separately determined, even when the compensating values of
6 and ¢ are of different signs, as for instance in series 1. The
reason is not apparent. The constants will necessarily be
correct at three values of 7, but the computed values of 2 are
no better asa whole than will be the case if the first set of
constants of series 2 for instance were used. In fact the con-
cents 6 may be arbitrarily put at a reasonable estimate*
b = -000,001 with ¢ = -0356 and a fair reproduction of the
ae vation is obtained. This is shown in the charts where
these computed values (6 = 10°) are incorporated. * Close
inspection, however, shows that in all cases the fall of computed
curves, while not quite rapid enough at ¢—¢,< 10, is somewhat
too rapid for higher time inter -vals. Thus 6 should be less
than 10~* and ¢ greater than ‘035, to be adapted to the present
results.

The question finally arises whether any. systematie error
in the standardization of coronas and hence in the values of 7,
could have produced an effect equivalent to the occurrence of
a constant c. Suppose that ea for the true nuclea-
tion, and that V=A+ An as the result of systematic errors
of standardization. Then —dN/dt=b/NV +e N+d’, an equa-
tion broader in form than the one accepted. The constants
d’ and c¢’ both vanish with A, the former’ more rapidly.
Hence, the possible introduction of ¢ through the method of
standardization is not excluded however how i improbable, since
the equation is conditioned by the occurrence of A.

8. Conclusion.—lf the rate of decay of ionized nuclei be
written bn”, the coefticient 6 as found by the fog chamber
increases as 7 decreases and may reach tenfold the order of the
usual electrical value of the order of /=10-". The endeavor to
explain this by supposing that but 1/m of all the ions are
caught and da/dt= —mbn, is not satisfactory.

*Townsend, McClung, Langevin, find b=1°1x10—* about, using the usual
electrical method. See Rutherford’s Radioactivity, pp. 41, 42, 1905.

a

a
S
Or

C. Barus—Decay of Lonized Nuctes.

It makes no difference how the small efficient nucleation is
produced, whether by weak radiation, or by decay (time loss)
from a larger nucleation, or by small exhaustion catching but
few nuclei.

The data of the fog chamber may be explained by postulat-
ing the absorption coefficient ¢ so that if @ be the number
eenerated per second —dn/dt=—a+en+bn’*. In such a ease,
if & is 10-* the order of the corresponding decay of ions as
found by condenser, and if ¢ is of the order of 38°510~, the
results of the fog ‘chamber are closely reproduced for all
values of nucleation.

A similar theory may possibly be extended to include the
absorption of phosphorus nuclei, carried by an air current
through thin tubes of different length and section (absorption
tubes).

Finally it is improbable that the constant ¢ should be
introduced by a systematic error in the standardization of the
coronas of cloudy condensation.

Brown University, Providence, R. I.

426 Whitlock—Caleite from West Paterson, NV. J.

Arr. XXX VIII.—Some new Crystallographic Combinations
of Calcite from West Paterson, V.J.; by H. P. Wurrtocx.

Tue calcite crystals which furnish the material for this
paper were found in the trap rock quarry one mile northwest of
the village ot Haledon, N. J. They were collected by Mr.
H. H. Hindshaw in the summer of 1904 and are now in the
collection of the New York State Museum. They are essen-
tially different from any of the types described by Rogers* and
present two forms which are new to the species. The interest
attached to the remarkable crystallized datolite from the local-
ity added to the above facts seems to justify a short crystallo-
graphic description of these crystals.
The writer wishes to express his thanks
to Mr. Hindshaw for the material to be
described, as well as for facts relating to
the location and geology of the locality.

Type l. Crystals of type I occur iso-
lated or in small parallel aggregates,
; immediately associated with the light
y greenish datolite previously described
from an adjacent localityt. In several
instances they were obtained from the
casts left by the flat plates of some min-
eral of previous generation, presumably
a mica, which has been dissolved out of
the matrix. of datolite. Minute rosettes
of specular hematite accompany the erys-
tals of this type, aa well as a thin coating
of limonite.

These crystals are 8" to 8™™ in vertical’
length, transparent and colorless. They are rhombohedral in
habit, ‘the dominant form being the negative rhombohedron
x.(0° 9-5: 4). This rhombohedron is modified on the polar edges
by the positive rhombohedrons p. (1011) and m. (4041) in
small development. The positive scalenohedron U: (5491) is
present as a series of small striated faces modifying the basal
angles of x. Fig. 1 shows this combination.

Type Il. The er ee of this type occur implanted directly
on the walls of the open seams and cavities in the diabase.
They are quite uniform in size, averaging 5™™ in vertical length
and are developed with the vertical axis in every instance nor-

*A. FF. Rogers, The crystallography of the Calcites of the New Jersey
Trap Region. School of Mines Quarterly, xxiii, 336, 1902.

+ Dana, E. S., On the Datolite from Bergen Hill, New Jersey. This Jour-
nal (3), iv, 16, 1872.

Whitlock— Calcite from West Paterson, N. J. 427

mal to the bounding surface. They are semi-transparent and
contain numerous inclusions of specular hematite in minute
ageregates. Hematite im microscopic, reddish, metallic plates
in some instances fills the interstices between the crystals. The
erystals of type II are rhombohedral-scalenohedral in habit.
A well developed series of negative rhombohedrons character-
izes this type, the planes in the rhombohedral zone being sharp
and brilliant. Vie

The two positive scalenohedrons noted, M (8:4°12°5) and H:
(8695) in the zone [0001-2131], consist of smooth and some-

what dull planes but yielded fair reflections. The latter of
these is new. Two negative scalenohedrons in the zone
[0221-1220], p: (1341) and q: (2461), give sharp, brilliant reflec-
tions. A new negative scalenohedron (1:13-14-10) lies close to
the zone of negative rhombohedrons between (0553) and
(0443). The planes are small but well developed and agree
fairly well with theory as to the measured angles. The letter
C has been assigned to this form. The forms observed on
erystals of this type are :— a

0 (0001), p. (1011), X. (0887), &. (0443), v. (0553), o. (0221),
(0551), S(O-1L-11-1), p: (1341), 9: (2461), H: (8695) new,
M (8°4-12°5), and C (1:13°14:10) new. Fig. 2 shows this com-
bination.

Type III. Crystals of this type are translucent, milky white
and average 20™™ in vertical length. They occur with consid-
erable amethystine quartz implanted in irregular aggregates on
the walls of partly filled seams in the diabase. In habit these

428 Whitlock— Calcite from West Paterson, LV. J.

erystals are distinctly scalenohedral, the dominant sealenohe-
dron being K: (2131). No instance of a doubly terminated
erystal was noted. A strongly developed zone of negative
rhombohedrons connects this” type with type II, the forms
being identical in both types. Of these the negative rhombo-
hedron $.(0221) is developed to a considerable habit. Broad
but rough and irregular planes of K. (5052) replace the obtuse
polar angles of K:. The negative sealenohedron q: (2461) of
type IT is present in small development. The positive scaleno-
hedrons ¥ : (4°8°12:5) and D : (4°16-20-9) in the zone [2131- 0221]
are present as extremely narrow modifications. The termina-
tion in crystals of the type is rarely complete, the polar angle
of the dominant scalenohedron ordinarily being replaced by a
“built up” rim surrounding a shght depression, which latter is
frequently partly filled with rosettes of hematite and prochlo-
rite. The following forms were observed :

b Gee mM. (4041), K. eee 2), A.(O887 ee 443), $.(0221),
(0551), 2.(O-11-11- 1), Ke O13), a: (24a ic : (48-125) and
®: (4:16:20-9). Fig. 3 shows’ this Vaneuienae

TABLE OF ANGLES.

Meas- = Meas- :

Angle pee es Angle ned tere
0001 ; - 1010 90 25 90 0 5491 4591 1G G8 1G, 30
NOM re OA 3 10s) Sie 103 1341 0221 UG AG ON. &
LOMA 2 5052 23030 23 194 2461 4961 30033 BOM
Ol ee OS Si OS mul 93 14 6395 69385 - 6456) 66058
ONIN 0443 97 23 97 214 6395 9365 32 594 32 2
OMI ge ORS LOS NS NOS. Vis 6395 3695 BU OS, 147
On e020 107 444 107 484 6395 8°4:19°5 “ALS 6 58
ODI G09 94 1091739) 11.020 8°4°19°5 Say G2. GQ DD
0994: 9904 104 44 104 174 84:19:5 : 1248: ~ 84 44°) 34°90
Oda 05541 1QO SA OS Be eed l uo 4:83:19 5 |) Sa 2 ope O
OM Wiles Wore Te TAG. Tay MG) OEE oy «assy Tc) 0221 16 464 17 10
ONS go) TSI 75 94 [522 ANG 2.0: 9)45 8) O22 I 9 50 9 444
PSL g OYA SoS 8740) VAS AL 10 fA 38 10S Oi Gemeameol
5491: 5941 66 984 66 494 (1°13:14:10 : 14:13-1:10-~ 4,50 5 54
5491: 9451 5 Owls OR TAL eae) 2° 9 Oeule) BBS 3 ©

Mineralogical Laboratory, N. Y. State Museum, Albany.

I. K. and M. A. Phelps—Preparation of Acetamide. 429

Arr. XX XIX.—The Preparation of Acetamide by the Action
of Ammonium Hydroxide and Hihyl Acetate; by I. K.and
M. A. Puetps.

[Contributions from the Kent Chemical Laboratory of Yale Univ.—clxv. ]

Ir has been stated by Hofmann* that the action of aqueous
ammonia upon ethyl acetate at ordinary temperatures yields
atter several days’ standing considerable amounts of acetamide,
but that the amount of acetamide by no means corresponds to
the amount of ethyl acetate taken. Hofmann further states
that according to a communication from Dr. Bannow the yield
of acetamide, even when formed in large quantities, is usually
not much above 70 per cent of that theoretically demanded.
In this same communication Hofmann makes a similar state-
ment in regard to the action of ethyl formate. :

In a former paper from this laboratory+ conditions have
been shown under which very nearly the theoretical yield of
formamide by the action of ammonium hydroxide and ethyl
formate is obtained. Now it has been found that conditions
closely similar will give theoretical yield as well in the case of
ethyl acetate and aqueous ammonia.

For the work given here ethyl acetate of commerce was
treated in a separating funnel with scdium carbonate solution,
and, after separating from this, was washed with distilled
water. The ethyl! acetate thus purified from acid impurities
was separated as completely as possible from the water, dried
over fused calcium chloride and then treated again with a fresh
portion of fused calcein in chloride before fractioning. Portions
boiling between 77° and 77°-2 were taken as pure ethyl acetate.

Definite portions of the pure ethyl acetate were weighed,
and, after chilling below zero in an ice and salt mixture, were
mixed ina stoppered reagent bottle with definite volumes of
ammonium hydroxide. In some of these experiments the
ammonium hydroxide was the pure concentrated ammonium
hydroxide of commerce ; in others the ammonium hydroxide
was made more concentrated by saturating at —10° the pure
concentrated ammonium hydroxide of commerce with dry
ammonia gas obtained by heating concentrated ammonium
hydroxide in a flask connected with a return condenser and
drying further the ammonia by passing it through a lime
tower ; while in a third series of experiments the produet
obtained by mixing in the cold in a stoppered reagent bottle
the ethyl acetate and ammonium hydroxide was saturated in
the cold with dry ammonia gas obtained in the manner given

* Berichte, xv, 977. + This Journal, xxiv, 173.

430 J. kK. and M. A. Phelps—Preparation of Acetamide.

above. In every case, after the solution had stood a sufficient
time in the reagent bottle stoppered tightly so that the ammo-
nia gas should not escape, it was transferred to a 250™° distil-
ling flask, connected with a 100°* distilling flask in the usual
way for a vacuum distillation, with the use “of the least amount
of absolute alcohol to rinse the sides of the bottle. The low
boiling impurities, ammonia, alcohol, and water, were removed
by fr actioning in vacuo in the usual way, the 250° flask being
heated in a bath of hot water finally at 60° for fifteen min-
utes after the pressure on the manometer registered 15™™,
The acetamide left in the flask was distilled by heating the
flask in an acid potassium sulphate bath at 140° to 150°, and
was collected in the receiver, cooled by a stream of cold water,
and weighed.

The experiments of section A are those in which the pure
ammonium hydroxide of commerce was used with the ethyl
acetate, and after standing suitably the mixtures were distilled
in vacuo as given above.

The experiments of section B were conducted in the same
way as those in section A, except that the ammonium hydrox-
ide used was saturated at —10° with dry ammonia before
mixing in the cold with the ethyl acetate.

The experiments of section C were conducted in the same
way as those of A excepting that the entire mass of ethyl ace-
tate and ammonium hydroxide after mixing in the cold was
saturated with dry ammonia gas at —8° to —10°.

From an inspection of the results recorded in section A it is
seen that the volume of ammonium hydroxide taken for a given
weight of ethyl acetate, as well as the time allowed for the
interaction of the ethyl acetate and ammonia, influences the
yield of acetamide. The theoretical yield can be obtained
with the proportion and concentration of the reagents used
here only on long standing. Two weeks’ standing at ordinary
temperatures with so large an amount as 75° of the ammo-
nium hydroxide for 50 grm. of ethyl acetate will give the
yield required by theory, although for smaller proportions of
the ammonium hydroxide that. time is not sufficient.

It is evident from the results given in section B that a solu-
tion of saturated aqueous ammonia tends to give a larger yield
of acetamide in a given time than can be obtained by weaker
aqueous ammonia.

‘In section © the results show that in shorter time than by
the procedure in experiments given in A and B of the table
the theoretical yield of acetamide may be obtained by satura-
ting in the cold the mixture of ethyl acetate and ammonium
hydroxide and allowing it to stand either four or six days
according to the proportion of the aqueous ammonia present.

I. K. and M. A. Phelps— Preparation of Acetamide. 431

TABLE I.
A
Treatment with NH,OH. ee
cetamide
Ethyl Ammonium = ——S SS
acetate hydroxide Reaction time Theory Found

No. grm, em?® Sp. ¢. Days Hrs. grm. grm.

(1) 50 50 0°90 3 ros 30 O70 21.70

(2) 50 50 0:90 6 19 B35) 25°60

(3) 50 50 0:90 13 aye Sononk 29°00

(4) 45°4 50 0:90 126 ae 30°80 30°36

(5) 50 (® 0°90 3 23 BiB) 1 26°89

(6) 50 75 0:90 6 ae Sonor 30°15

(7) 50 75 0-90 8 us 33°57 31°18

(8) 50 7d 0°90 13 oe 33°57 34°10

B
Treatment with NH.OH saturated with NH3.

(9) 5O 50 ee 3 19 SOOM 21:03
(10) 50 50 are =1'2 ee SOE 32°80
(i 1) 50 50 ne 49 i SOnONd 3a 712
(12) 50 78 here 3 22 Sono 28°81
(13) 50 78 pats 6 22 33°57 32°53

C
Treatment with NH,OH and saturation of the mixture with NHs.
(14) 50 50 0:90 3 16 33°57 95°78
(15) 50 50 0:90 6 She Soway Souee
(16) 50 50 0°90 12 6 Boyd 33°82
(17) 50 50 0-90 20 Ne B03) 7 Socks
(18) 50 75 0:90 ae 23 Sonor Tye ge
(19). 50 75 0°90 2 te 33°75 30°08
(20) 50 75 0°90 - 3 oe Baad 31°63
(21) 50 vi) 0°90 4 at Sowa 33°62
(22) 50 75 0:90 4. 6 Bayt) 7 34°00

It is evident that the time of completion of the reaction is
dependent upon the concentration of the ammonia.

It was found that the mixture of ammonium hydroxide and
ethyl acetate became homogeneous in the experiments of sec-
tion A in about three days, in B in somewhat less time, and
twenty-four hours in C. In experiment (18) the mass became
homogeneous at the end of twenty-three hours, and in this
single instance distillation was made as soon as this phenome-
non appeared. It is evident that the formation of acetamide
progresses slowly and is not at an end as soon as the mass
becomes homogeneous.

It was found by experiment that a known weight of pure
acetamide treated with 10°™* of water and fractioned in vacuo

Am. Jour. Sct.—FourtH SERIES, Vou. XXIV, No. 143.—NovemsBer, 1907.

432 I. Ki and M. A. Phelps—Preparation of Acetamide.

could be recovered with a loss of less than 0°05 germ. The
acetamide tends to hold traces of water; and it was not found
possible to remove it by fractional distillation in vacuo without
danger of loss of very small amonnts of acetamide.

The presence and amount of ammonium salt in the aceta-
mide obtained in the procedure outlined above were tested for
by the use of a solution of sodium cobalti-nitrite. An experi-
ment showed that 0:0002 orm. of ammonium chloride could be
readily detected in the presence of 0°50 gram of acetamide. The
acetainide obtained directly by fractioning in vacuo, as given
above, was found to contain traces of ammonia and ammonium
salt. In experiment (8) the crude material was transferred after
weighing the product obtained from the vacuum distillation to
a watchelass and allowed to stand in a desiccator over sulphu-
ric acid for tw enty-four hours, and the loss sustained was 0°62
erm., presumably largely water with some ammonia. Some
of this loss must have been acetamide also, for by a separate
experiment with pure acetamide recrystallized from benzene
it was found that acetamide continually lost in weight in a
sulphuric acid desiccator. The acetamide from experiment (8)
after being dried showed the presence of not more than 0°30
erm. of ammonium salt, estimated by the amount of precipitate
produced with sodium cobalti-nitrite as compared with the
amount of precipitate obtained under similar conditions with
ammonium choride, pure acetamide and sodium cobalti-nitrite.
The material obtained in experiment (22) was redistilled with
an air condenser under ordinary atmospheric pressure and
yielded 27-5 grm. of product boiling between 221° and 222°.
This product showed on testing with sodium cobalti-nitrite no
ammonium salt. Theoretically, more ammonium salt might
be present in those cases where the standing is longest. But
it was found on redistillation, under ordinary atmospheri 1¢
conditions, of the acetamide obtained directly by the processes
given above, that ammonium salt was not present in sutticiently
large amounts in the different experiments to be noticeable.

It is clear from the work given that acetamide with only
traces of lL npurity may be obtained in quantities barely less
than quantitative for the amount of ethyl acetate taken, if
ethyl acetate and ammonium hydroxide are mixed in the
cold and allowed to stand a suitable length of time. Inereas-
ing the amount of ammonium hydroxide employed shortens
the time of standing necessary for a theoretical yield, and
increasing the concentration of ammonia by saturating the
mixture of ethyl acetate and ammonium hydroxide in these
proportions with ammonia gas further diminishes, by one-half
or more, the time of standing.

W. A. Drushel— Potassium as the Cobalti-nitrite. 433

Arr. XL.—On the Volumetric Estimation of Potassiwm as
the Cobalti-nitrite ; by W. A. DrusHet.

[Contributions from the Kent Chemical Laboratory of Yale Univ.—clxvi. |

Tue use of sodium cobalti-nitrite for the qualitative detec-
tion of potassium is well known and its use as a quantitative
reagent has been described by R. H. Adie and T. B. Wood,*
whose results are fairly accurate and favorably comparable
with results obtained by the platinic chloride gravimetric
method. In the process worked out by these investigators a
solution of a potassium salt containing the equivalent of 0:5
per cent to 1 per cent of K,O is acidified with acetic acid and
precipitated by an excess of sodium cobalti-nitrite. The mix-
ture is allowed to stand at least a few hours, preferably over

night, and is then filtered through a perforated crucible fitted
with an asbestos felt. The precipitate is washed with 10 per
cent acetic acid. According to Sutton it is important that the
precipitation should be made in a solution containing the equiv-
alent of 0°5 per cent to 1 per cent of K,O, since in solutions
of lower concentration the precipitate comes down in a condi-
tion in which it is apt to run through the filter in washing.
The precipitate is then decomposed by boiling in dilute sodium
hydroxide, and the cobalt is removed as the hydroxide by fil-
tration. The nitrites, which are a measure of the potassium
in the precipitate, are estimated by titrating with standard
potassium permanganate. Adie and Wood found by analysis
that the composition of the precipitated potassium salt is rep-
resented by the formula K,NaCo(NO,),.H,O. According to
their method a cubic centimeter of strictly > potassium per-

manganate is equivalent to 0-000785 grm. K Oo:

The object of this investigation was to determine the best
conditions for precipitating and filtering the potassium cobalti-
nitrite, and to shorten the work of estimating the potassium by
oxidizing directly with potassium permanganate without the
preliminary decomposition of the precipitate and removal of
cobalt recommended by Adie and Wood. In a series of pre-
liminary experiments the precipitated cobaltinitrite was oxi-
dized by an excess of potassium permanganate, the excess of
permanganate reduced by standard oxalic acid, and the
remaining oxalic acid titrated to color. In this tr eatment triva-
lent cobalt is reduced to the bivalent condition, and from the
formula of potassium sodium cobalti-nitrite it would appear
that the oxygen thus made available should be equivalent to
one-twelfth of that necessary to oxidize the nitrites. The
results of these experiments are given in Table I.

* Jour. Chem. Soc., lxxvii, 1076. Sutton’s Vol. Anal., 9th ed., p. 62.

434. W. A. Drushel—Potassium as the Cobalti-nitrite.

Portions of a potassium chloride solution of known strength
were treated with an excess of sodium cobalti-nitrite* and
filtered on perforated crucibles fitted with asbestos felts. The
precipitates were first washed with a 10 per cent acetic acid
solution, then once with water. In experiments 1 to 5 the
precipitate was decomposed by boiling with sodium hydroxide
and the nitrites estimated according to the method of Adie
and Wood, giving the results in the second column of Table
I. The cobaltic hy droxide filtered off on asbestos was reduced
by heating nearly to boiling in a measured amount of standard
oxalic acid containing a little sulphuric acid. The excess of
oxalic acid was estimated by titrating with standard potassium
permanganate, and from this the equivalent of the cobaltic
hydroxide in terms of permanganate was found by subtrac-
tion, giving the results in the third column of the table. In
experiments 6 to 10 the precipitated potassium s salt together
with the crucible and asbestos felt, after stirring the precipitate
and felt loose from the crucible, was placed in a beaker con-
taining # measured amount of standard permanganate, taking
care to use an excess, diluted to about ten times its volume
and heated nearly to boiling. After five to eight minutes, or
when the manganese hydroxide formed gave the solution a
ae color, it was acidified with 5% to 20° of sulphuric acid
(1:7). After a few minutes a measured excess of standard
oxalic acid was run in from a burette, the temperature beige
kept a little below the boiling point until the solution became
clear, and then titrated to color with permanganate. The
whole amount of permanganate used less the equivalent of the
oxalic acid used is the amount necessary for the oxidation of
the precipitate. The results are given in the fourth column.

TABLE I.
KMn0O, used in
titration of nitrites KMnO, KMn0O,
after removal equivalent used in
KO of cobaltic to cobaltic direct
taken hydroxide hydroxide titration
No. erm. em?, em?, em,
ie 0°0235 B24 2°5
2. 0°0235 32°25 2°5 ——
3. 0°0235 32°65 2°55 —
4. 0°0555 48°55 3°88 —
D. 0°0555 49° 3°95 —_——
6. 0°0255 —_—— —= 30°
Ue 0°0235 —— —— 29°65
8. 0°0235 —_—— —— 29-4
OF 0°0355 — —— 43°65
10. 0°0355 —— —_—— 44°4

* Prepared according to the directions given by Adie and Wood loc. cit., also
given in Sutton’s Volumetric Analysis, 9th ed., p. 62.

W. A. Drushel—Potassium as the Cobalti-nitrite. 485

From the results of Table I it appears that the oxidizing
value of the cobaltic hydroxide in terms of permanganate is
nearly one-twelfth of that required for the oxidation of the
nitrites, while the amount of permanganate necessary in the
presence of the cobalt is nearly eleven- twelfths of that required
for the oxidation of the nitrites after the removal of the
cobalt. The factor used, therefore, in calculating the results
from the direct titration should be twelve-elevenths of that
given by Adie and Wood ; that is, in titrating the precipitate
without first separating the cobalt one cubic centimeter of
strictly 7 potassium permanganate is equivalent to 0°000856
grm. K,O. (

Unless the potassium salt solution is of the proper concen-
tration the precipitate is very difficult to filter and wash, and
shows a tendency to pass through the felt. By repeated experi-
ments it was found that this difficulty as well as the necessity
for allowing the precipitate to stamd over night is avoided by
evaporating the mixture nearly to dryness on the steam bath
after adding the sodium cobalti-nitrite solution in considerable
eXcess. Upon cooling the pasty residue it becomes hard and
dry. It is then treated with cold water to dissolve the excess
of sodium cobalti-nitrite, and the insoluble portion is collected
on the filter. This precipitate may be freely washed with cold
water without showing a tendency to pass through the filter,
and is so insoluble that less than 0°5 of a milligram of the
dried precipitate will dissolve in a liter of water at the room
temperature during 24 hours standing with occasional shaking.
This mode of treatment was found to work well and was used
in all the subsequent experiments.

The method as worked out and used in all the experiments
except those of Table I is as follows: The solution of a
potassium salt, containing not more than 0-2 grm. K,O and
free from ammonium salt, was treated with a rather large
excess of sodium cobalti-nitrite solution acidified with acetic
acid, and evaporated to a pasty condition over the steam bath.
It was then cooled and treated with 50° to 100° of cold
water and stirred until the excess of sodium cobalti-nitrite was
dissolved. It was allowed to settle and decanted through a
perforated crucible fitted with an asbestos felt. The precipi-
tate was washed two or three times by decantation, after which
it was transferred to the crucible and thoroughly washed with
cold water. In the meantime a measured excess of standard
potassium permanganate was diluted to ten times its volume
and heated nearly to boiling. Into this the precipitate and
felt were transferred and stirred up, after which the crucible
was also put into the solution, since particles of the precipitate
stick persistently to the sides of the crucible. After the oxi-
dation had proceeded five or six minutes manganese hydroxide

436 W.A. Drushel—Potassium as the Cobalti-nitrite.

separated out and the color of the solution darkened. At this
point 5°™* to 25°" of sulphuric acid (1:7) were added, and
the solution, after stirring, was allowed to stand a few minutes.
Then a measured amount of standard oxalic acid, containing
50°™* strong sulphuric acid per liter, was run in from a burette,
taking care to add an excess. The temperature was maintained
a little below the boiling point until the solution became color-
less and-the manganese hydroxide had completely dissolved.
It was then titrated to color by permanganate in the usual
manner. From the whole amount of permanganate used the
permanganate equivalent of the oxalic acid used was subtracted
and the remainder multiplied by the factor calculated for the
strength of permanganate used, 0-000856 being the factor for
strictly ov potassium per manganate.

To make the =. oxalic acid solution, exactly 71066 grm. of
pure recrystallized ammonium oxalate were dissolved in about
700™ of cold distilled water contained in a liter flask. To
this solution were then added 50™* of strong sulphuric acid.
The contents of the flask were cooled to 15° C. and made up
to the mark with distilled water. The potassium permanga-
nate solution was made approximately decinormal and stand-
ardized in the usual way. ‘This standard was checked by
standardizing under conditions as nearly as possible like those
under which the solution was used. A measured portion was
diluted ten times, heated nearly to boiling, acidified with dilute
sulphuric acid and allowed to stand a few minutes. It was
then bleached with a measured amount of oxalic acid, using it
in slight excess, and titrated to color. The two ‘methods
agreed very well, the difference in permanganate seldom being

greater than one-tenth to two-tenths of a cubic centimeter in
95cem*
a °

TasueE II.
K,O taken K.O found Error in K,0O
tan aos = (ae SSS =>
As KCl Gravi- Volu- Gravi- Volu-
metrically metrically metrically metrically
No. grm. grm. erm. erm. erm.
1 Os0 237 0:°0240 0:0238 0°0008 + 0:0001 +
Deen OsO2arn 0:0243 0°0242 00006 + 00005 +
3. 0°0354 0:0359 0°0355 0°0004-+4 0°0000 +
4. 0:°0474 00478 0°0471 0:0004 +4 0:0003 —
5. 0:0048 0:0048 0°0050 0°0000 + 0-0002 +
6. 0°0024 0°0024 00025 0°0000 + 0-000L—
Bo. “OOOO SSS OOO OG a= 2 epee 0-0001 +
Se VOHO OMe res eee O° OO ter Gee ee Bee 0-0002 +
Yo OWRD eae OOS S50 Ree eee 0°0000 +

In the first six experiments of this series the precipitate
was dried at 115° C. and weighed. It was then treated with

W. A. Drushel— Potassium as the Cobalti-nitrite. 437

permanganate by the method previously described. Experi-
ments 6, 7 and 8 show that very small amounts of potassium
may be estimated with a fair degree of accuracy.

In Table III the effect of the presence of members of the
calcium group was investigated. Calcium and magnesium
apparently do not interfere, while barium and strontium tend
to give high results.

TARE
CaCl, MgCls, BaCle, Sr(NOs)s K.O K.,O
taken taken found Error
grm a grm 5 germ J grm 2 grm. grm 6 germ 6
ee) 240.0 Olees O;20O Oh piertaeas ens) (ental aie 0:0005 0:0007 0:0002+
Do OBOOO™> OFHOOO) oo ke Be ee Le 0:0237 0°6234 0:0003—
Bo OH OOOs: TOO OO ee ee ee 0:0829 0:°0824 0:0005—
AL OS OOO MOON: < we oe OLDOO OM OSOTE i OL073iie 0.0 O26EE
5. 05000 1:0000 0°5000' 0:5000 0:0474 -0:0493 0:-0019+
GB, | CHSOOO “WeOOO). “OsHOOO sain oe OLO23 0202 ai 520. 0 004
Us. OFB00O - WOOOO: oeses = Senses OO Olas, OOo 4

The method may also be used in the presence of phosphoric
acid and is therefore applicable to the estimation of K,O in fer-
tilizers. In Table IV are the results obtained in nine fertili-
zers by the platinic chloride method and the cobalti-nitrite
volumetric method. In columns one and two are the duplicate
results obtained by two analysts of the Connecticut Agricul-
tural Experiment Station, and in column three are the results by
the volumetric method. The water-soluble phosphoric acid
present in these samples is given in the fourth column.

TaBLe IV.
K.O in Mixed Fertilizers.
K,.O by platinum K.O by vol. Water-soluble
chloride method cobalti-nitrite meth. P.O; in sample
Gam ST ay
No. per cent per cent per cent per cent

A; 9°22 5°18 5°18 4°16
Dye 6°53 6°56 6°56 3°10
3. 2°23 2°24 2°24 MeS2
4, 8 68 8°64 878 0°94
ay 6°37 6°42 6°38 6°62
6. 6°08 6°13 6°13 561
ke 4°08 4°02 4°02 B31 105)
8. 4°62 4°66 4°67 2°43
9. 1°68 1°67 O97 6:03

Ten grams of the fertilizer were placed in a 500™ flask and
300" of water were added. The contents were boiled for 30
minutes and ammonia water was added to slight alkalinity.
Enough ammonium oxalate was added to precipitate all the
calcium and, after cooling, the solution was made up to the mark
on the neck of the flask and well shaken. The solution was

438 W.A. Drushel— Potassium as the Cobalti-nitrite.

then filtered through a dry filter into a dry flask. Two 50°™*
portions of the filtrate were transferred with a pipette to plati-
num dishes, one portion being used for the gravimetric estima-
tion by the platinum chloride method and the other for the
volumetric estimation by the cobalti-nitrite method. After
evaporating these portions to half their volume over the steam
bath, 1°™* sulphuric acid (1:1) was added and the evaporation
was continued as far as possible over the steam bath, and
finally over a low flame. After the danger of spattering was
over the flame was increased and the charred organic matter
was burned off, finally, over the blast lamp. The potassium
sulphate was dissolved by adding a little water and heating
over the steam bath, and the potassium was estimated as
previously described.

The volumetric method may be summed up thus: The
potassium is precipitated as potassium sodium cobalti-nitrite
by an excess of sodium cobalti-nitrite and the mixture is
evaporated on the steam bath. The precipitate is separated by
filtration through asbestos and oxidized by hot standard potas-
sium per manganate. The excess of permanganate is bleached
by an excess ‘of standard oxalic acid and the solution is then
titrated to color by permanganate. The amount of potassium
oxide is found by wultiplying the oe value of the amount
of potassium permanganate used by the factor 1:09.

This method has the advantages over the platinum chloride
method that no expensive reagents are used and that the time
required for a determination is materially reduced. The
method is considerably shorter than that of Adie and Wood
and does not require the potassium solution to be of any
definite concentration to work well.

In closing, the author desires to acknowledge his indebted-
ness to Dr. R.G. Van Name for many helpful suggestions
during the progress of the work.

Chemistry and Physics. 439

SCIENTIFIC INTELLIGENCE.

I. CHEMISTRY AND PHYSICS.

1. Atomic Weights of Silver, Nitrogen and Sulphur.—The
masterly researches conducted by Prof. Theodore W. Richards
are being continued with apparently still more wonderful preci-
sion than heretofore. RrtcHarps and Forses have studied the
quantitative synthesis of silver nitrate. Fer this purpose they
employed a bulb of fused quartz for evaporating and drying the
silver nitrate produced by dissolving pure silver in pure nitric
acid. They found that 100,000 parts of pure silver produce
157°479 parts of silver nitrate. If silver is assumed to be 107°93,
nitrogen must be 14:037, while if silver is taken as 107°880, nitro-
gen must be 14:008. The question of the exact atomic weight
of silver must be determined by further work, particularly
upon the composition of the chlorates and the ammonium salts.

Ricuarps and Jones have investigated the comparative
weights of silver sulphate and silver chloride, carrying out the
conversion in a tube of fused quartz. They found that 100-000
parts of silver sulphate gave 91-933 parts of silver chloride, and
calculated, if silver is taken as 107°93, sulphur is 317113, while if
silver is 107°88, sulphur is 32°069. The tinal decision in regard
to the atomic weight of sulphur must await further work upon
that of silver.— Carnegie Inst. Pub. 69 (1907). Hagel aWis

2. The Atomic Weight of Radiwn.—Movme. Curte made
determinations of this atomic weight in 1902, using nine centi-
grams of radium chloride. With this small quantity the value
225 was obtained. Having now at her disposal a few decigrams
of the salt, she has carefully purified it, and has obtained four
decigrams of very pure radium chloride, upon which she has car-
ried out three determinations under much more favorable condi-
tions than before. The determinations were made by comparing
the weights of anhydrous radium chloride with the weights of
silver chloride produced from them. The results led to the con-
clusion that the atomic weight of radium is 226°2, where Ag is
107°8 and Cl is 35°4, with a probable error of less than half a
unit. If the international atomic weights for silver and chlorine
are used as a basis, the value is Ra=226°45. The radium salt
was not absolutely free from barium, as shown by the spectro-
scopic test, but it is the author’s opinion that the trace of impur-
ity present had an inappreciable effect upon the results, and it is
her opinion that the difference between the results of 1902 and
1907 should be attributed to the inferior accuracy of the experi-
ments performed with only nine centigrams of radium salt, and
with less pure reagents.— Comptes Rendus, cxiv, 422. ~

H. L. W.

440 Scientifie Intelligence.

3. The Melting-Point of Pure Tungsten.—In connection with
the commercial application of tungsten filaments in electric light-
ing, it is well known that the melting point of this metal is very
high, and in the neighborhood of 3000° C. Warrrnspere has
now made some careful determinations of the melting point of
pure metallic tungsten by heating it electrically in a vacuous bulb
and employing a carefully calibrated optical pyrometer. He con-
cludes that the true melting point of the metal is at least 2800°
and probably not more than 2850°. For comparison the author
measured the temperature of the positive crater of the arc lamp,
which varies somewhat with the kind of carbon used, and deter-
mined this to be 3350°, while Reich had found 3430° with the
same kind of pyrometer.— Berichte, x1, 3287. H. L. W.

4. Solubilities of Inorganic and Organic Substances, by Atu-
ERTON SEIDELL. 8vo, pp. 867. New York, 1907 (D. Van Nos-
trand Company).—This is a useful and important book of
reference which will be appreciated by all classes of chemists.
It differs from Comey’s well known dictionary in confining itself
entirely to quantitative data, or in other words, it deals with solu-
bilities only in the sense of the composition of saturated solutions.
This limitation has made possible the introduction ef the solubili-
ties of organic compounds, as far as they have been determined
quantitatively, and this addition has still left the book much
smaller than Comey’s. The latter admirable work will still be of
use when qualitative data are sought, for instance, when we wish
to know if an inorganic compound, whose solubility has not been
studied quantitatively, is readily or sparingly soluble in water,
or when we desire to find out how readily a substance dissolves
in the ordinary acids. The new book is attractive in its mechani-
cal details, and is a very convenient one for use, as the matter is
arranged alphabetically, and besides it has a full index. The
author has not made a mere compilation, but has devoted much
labor to re-caleulations, and to the selection of the most reliable
results. 1b Up, Wo

5. Practical Chemistry for Army and Matriculation Can-
didates, and for Use in Schools, by GEOFFREY Martin. 12mo,
pp. 144. London, 1907 (Crosby Lockwood & Son).—This little
book outlines a course of laboratory work covering a wide range
of subjects. It starts with elementary chemistry, and takes up
quantitative work, solubility determinations, volumetric analysis,
blowpiping, melting- and boiling-points, ete. Necessarily, in so
small a book, the various subjects must be treated very briefly.
However, the experiments appear to be well selected and clearly
described. The book does not deal with chemical theory, except
that it gives numerous equations, and pays considerable attention
to chemical arithmetic. 1: Oa

6. The Elements of Physical Chemistry, by Harry C. JonES;
Pp. 650, 8vo. New York, 1907 (The Macmillan Company).—-The
appearance of a third edition, revised and enlarged, indicates an
extensive use of this well known, excellent text-book. Recent

Chemistry and Physics. 44]

advances in the science have been incorporated, many minor
improvements have been made, and numerous new references to
literature increase the value of the book for reference.
Hees We

7. Canal Rays.—A number of papers on this subject have
recently appeared: and the conviction appears to be gaining
that in these rays one has the best means of ascertaining the
nature of the positive atom. I. Pascnen (Ann. der Physik,
No. 7, pp. 247-260) gives some remarkable photographs of the
Doppler effect in the canal rays produced in hydrogen. In the
main he substantiates Stark’s investigation, but he differs from
him in regard to the distribution in the light of the displaced
lines or bands, in certain series of lines. In a second paper (pp.
261-206) Paschen gives an investigation of the Doppler effect in
oxygen. He employed a concave grating of 10° diameter and
3°03 metersradius. The papers contain interesting reflections upon
the relation of series lines and the Doppler effect. Srarxk (Ann.
der Physik, No. 9, pp. 798-804) answers the objections of Paschen,
and maintains that the latter’s assertion that the series lines of
oxygen do not show the Doppler effect is not correct. Paschen
maintains his position in Ann. der Physik, No. 10, pp. 997-
1000. The most suggestive paper on the subject of canal rays
is one by J. J. Tuomson (Phil. Mag., May, 1907, pp. 561-575).
The vacuum tube employed by Thomson was terminated at the
end of the tube in which the canal ravs were formed, by a screen
covered with powdered willemite. The canal rays falling on
this screen produced a fleck of light and the deflection of this
spot of light by electric and magnetic fields was studied in vari-
ous gases. A variety of rays were discovered : for one kind

é

mal has the value of 10°, that of an atom of hydrogen; for another

kind < has half this value. A paper in Phil. Mag., Sept., 1907,
7

p- 359-364) also by J. J. Thomson, shows that particles of posi-
tive electricity are shot off in all directions from the gas trav-
ersed by the canal rays. Tels

8. Propagation of Plane Electromagnetic Waves over Plane
Surfaces and their relation to Wireless Telegraphy.—J. ZENNECK
gives a mathematical discussion of this subject, and its bearing
upon the absorption of such waves by the atmosphere. Marconi
has shown that the distance one can reach by wireless telegraphy
is 24 times greater by night than by day and he attributes this
phenomenon to the increased absorption of the waves due to
ionization of the air by daylight. Zenneck’s calculation shows
that the layers of air less than 6,000 meters from the earth’s sur-
face cannot change their conductivity by daylight sutticiently to
account for the absorption of the waves, and he believes that this
absorption is due to the loss of energy from the antenne due to
daylight. It is probable also that the good effect of clouds and
fog is due to the protection of the antenne from this loss of

449 Scientific Intelligence.

energy due to light. The waves employed in wireless telegraphy
in passing from water to land and in the reverse direction must
suffer partial deflection. The amplitude, therefore, of the wave
depends not only upon the distance the waves have traversed
over land and sea but also upon the shore form, or barrier between
land and sea. On this partial reflection depends the fact that
less distorted waves are received at a distance from the sender
than at a station near the sender.—Ann. der Physik, No. 10, pp.
846—866. Bo au,

9. Influence of Magnetic Fields on the Resistances of EHlectro-
lytes.—G. Berxnpr shows that the change of resistance of metals
in magnetic fields depends greatly upon temperature conditions
and he gives a method for controlling temperature. He found
that electrolytes in fields up to 3000 Gauss units, submitted to
perpendicular lines of magnetic force and to parallel lines up to
1000 Gauss units, did not change in resistance more than s!, per
cent. Mercury showed with lines perpendicular to the layer no
change greater than y;1,, percent, and with lines parallel none
greater than s>4,7,7 per cent. A very small change in bismuth
was attributed to an electro-dynamic effect. In general: fluid
bodies suffer no change in magnetic fields.—Ann. der Physik.,
No. 10, pp. 932-950. if 1th

10. Change of Resistance in Metal Wires with Occlusion of
Oxygen.—Guipvo SzivEssy reviews the subject of occlusion of
hydrogen by palladium and gives a formula which directs his
work upon the occlusion of oxygen. He finds marked changes
due to this cause in silver wires, and in platinum. Gold wires
showed no increase in resistance. The results with palladium
were doubtful. Ann. der Physik, No. 10, pp. 963-974. 35. 7.

Atlas of Absorption Spectra; by H.S. Unter and R. W.
Woop. Pp. 59, with 26 plates. Washington, 1907 (published by
the Carnegie Institution).—This is a collection of more than one
hundred photographie maps of the absorption spectra of solutions
of various aniline dyes and also of some inorganic salts. The
maps are so arranged as to show the variation of the width of
the absorption bands with the thickness and concentration of the
solution. The spectra comprise the visible and ultra-violet
regions from about 0°6lu to nearly 0:20u. The photographs are
beautiful and accurate and reflect much credit upon the experi-
mental skill of Dr. Uhler, who has devised novel and ingenious
methods of experiment that cannot fail to be of service in future
work on absorption spectra. ee ANG 13

12. Bulletin of the Bureau of Standards ; 8. W. STRATTON,
Director.—The third number of Volume III of the Bulletin of
the Bureau: of Standards has recently appeared. One of the
papers by G. K. Burgess gives the following as the approximate
melting points of metals of the iron group; the specimens exam-
ined ranged in purity from 98 per cent to 99°95 per cent:

Tron, ae ee 1505° C. Nickel seas ee 1435°
Chromium_-*_- 1489° Manganese ._.. 1207°
Cobalt aaa es 1464°

Geology and Natural History. 443

The melting points of cobalt and nickel are regarded as correct
to within 5°, while the possible error of the others is probably
less than 10°. The method employed was based on the measure-
ment of the intensity of a particular monochromatic radiation
from platinum or other substance, as discussed in an earlier paper
noted below. An atmosphere of pure hydrogen was found appli-
cable in the case of these easily oxidized elements. For melting
points above that of platinum, it is suggested that iridium, or
perhaps tungsten, may be used.

The latter half of the same number (pp. 433-540) is occupied
by a paper by E. B. Rosa and NE. Dorsey, describing the
results of a new determination of the ratio of the electromagnetic
to the electrostatic unit of electricity. In this the method of
capacities has been employed but witha higher degree of accu-
racy than has been realized hitherto. The conclusion of the
paper with the final results is held over for another number. In
a preceding issue of the Bulletin, a paper by Waidner and
Burgess describes the radiation from, and melting points of, pal-
ladium and platinum. The final values obtained are 1546° for
the melting point of palladium and 1753° for that of platinum.
The whole paper is an important contribution to the difficult sub-
ject of pyrometry, leading to the establishment of a definite high-
temperature scale.

Il. Grotogy anp Naturat History.

1. Devonic fer of the New York Formations ; by CHARLES
R. Eastman. New York State Museum, Mem, 10, 1907, pp. 235
pls. 15.—Vhis ‘chan written and interesting treatise on “ne
Devonian fishes (Agnatha and Pisces) of New York is of far
wider scope than is indicated by the title. .The work treats, in one
form or another, of all American Devonian fishes, while the legen:
cation and Beetncien takes into account all that is known of these
Precarboniferous animals. Of Species described there are fifty-

eight and of these thirty-six are found in New York. The volume
should be owned by all paleontologists, and is one of the best ot
the New York State Survey publications.

After a short introduction there is presented a list of the Ameri-
can Ordovician, Silurian and Devonian fishes geologically arranged
with the inealaud of their known occurrence. A perusal of this
list, taking into consideration also the nature of the deposits and
ae invertebrate contents, brings out strongly the conclusion that
nearly all of these fishes are of a normal marine habitat and that
at but two localities is the evidence decidedly in favor of estuarine
waters (Campbellton, New Brunswick and Scaumenac, Quebec).
Therefore a fresh-water fish fauna is as yet unknown in hese older
Paleozoic strata. The tables further indicate that an abundant
fish fauna, remarkable for its fishes, the Arthrodires, appears

4i4 Scientific Intelligence.

with the warm water coral fauna represented by that of the
Onondaga formation; this the reviewer holds is an invasion
from the Gulf region through the Indiana Basin, thence spreading
eastward into New York while one (Macher acunthus sulcatus)
continues as far as Gaspé, Quebec. A further analysis of the
American species shows that of this warm water fauna at least
three genera (Ptyctodus, Acantholepis, and Dinichthys) spread
westward into the Dakota sea while the western province furnished
but one migrant (fleteracanthus) into the Mississippian area.

Under Geological Conclusions, in regard to the paths of migra-
tion as worked out by Se huchert and Clarke, the author remarks
that “the known distribution of the fishes is in all respects con-
sonant with, and one is tempted to add, confirmatory of the prin-
ciples that have been formulated from a study of the invertebrates.”
The American Middle Devonian fish assemblage of the Appa-
lachian basin is found to occur somewhat earlier in Bohemia,
migrating westward across the Atlantic (probably not northwest-
ward as stated by the author) and into the Mississippian sea by
way of the Indiana basin. ‘The most conspicuous elements of
the fauna are Arthrodires and Ptyctodonts, groups which began
immediately upon their introduction to attain a most remarkable
development. Throughout the Hamilton, and later Devonic, con-
ditions must have been eminently favorable in the Appalachian
sea for the furthur specialization of armor-clad Dipnoans of the
type represented by Dinichthys and its congeners. Like their
earliest predecessors, they became of greatest importance locally
in Ohio.”

Of the Agnatha, or fish-like vertebrates, the author does not at
all follow the suggestion of Professor Patten that these animals
were developed out of some Arthropod stock. He holds the class
to be an independent one, differing from the true fishes in not
having “the slightest trace of ordinary jaws, of a segmented axial
skeleton in the trunk, or of arches for the support of paired limbs.
Indeed, appendages themselves are confined to a single order, the
Antiar cha, where oarlike swimming organs appear to have de vel-
oped from an integumentary fold on either side of the body,
unsupported by rays, and in a manner fundamentally different
from the fins of the ty pical tishes.” That they are not transitional
between vertebrates and invertebrates (either Ascidia, Crustacea
or Arachnida) the author holds is disproved as follows: (1) the
dermal plates are composed of true bone ; (2) the head shield and
body armor of Asterolepids have a well developed sensory canal
system; (3) in Pterichthys, at least, there is a tail covered with
scales, a dorsal tin, and a genuinely piscine heterocercal caudal fin.”

The oldest fish remains so far discovered (Ordovician of Colo-
rado) and made known by Walcott, the author is not called upon
to treat but admits one of the three species, Astraspis desiderata,
to the Agnatha.

The Dipneusti or lung-fishes are treated in considerable detail.
There are a large number of them in the American Devonian

Geology and Natural story. 445

because the author, following Smith Woodward, refers to this
sub-class the Arthrodira. The test for relationship the author
finds in the jaw of Arthrodires, for it conforms “strictly to the
Dipnoan type, a fact of cardinal importance for their classifica-
tion.” C. 8.

2. The Paleontology of the Niagaran Limestone in the Chi-
cago Area. The Trilobita; by Stuart WeEtiErR. Nat. Uist.
Survey, Chicago Acad. Sci., Bull. iv, pt. 11, 1907, pp. 163-281,
pls. xvi-xxv.—Professor Weller gives here a complete account
of the Silurian trilobites of the Chicago area which extends
from Milwaukee, Wisconsin on the north, south to Joliet, Illinois.
There are 41 species and of these 19 are new. ‘There is also one
new genus, Jéduenoides. The detailed stratigraphy is not yet
determined but four horizons are recognized in the dolomites,
seemingly ranging from the Clinton well up into the Guelph. On
pages 181-210 a complete bibliography is given of all North
American species of trilobites, of which there are 105. The
illustrations are good, being photographic reproductions of crayon
work on stipple paper. C. §.

3. Revision der Ostbaltischen Silurischen Trilobiten ; Abb.
VI, von Fr. Scumipr. Mém. Acad. Imp. Sci. St. Pétersbourg,
xx, No. 8, 1907, pp. 104, pls. 3.—This extensive and valuable
work, treating of the Ordovician and Silurian trilobites of the
eastern Baltic region, begun thirty years ago by the author, is
now completed. In this part a general review of the work is
given, supplementing the old species with such information as
has been gleaned from new material and such changes as are due
to the interaction of the work of others during the interim. In
1858 Estland, Livland and Osel had furnished Schmidt fifty
species. Since 1876 he has made known from these and the St.
Petersburg regions 256 species or varieties, and of these he has
named about 105 forms. From the Lower Cambrian there is but
a single species, QOlenellus mickwitzi. The next trilobite zones
are to be correlated with the American Stones River and Chazy
formations of the Middle Ordovician and from these and the
Upper Ordovician horizons there are not less then 229 forms.
Not a single one of these passes into the Silurian, a fact of far
greater significance in the Baltic region than in America because
of the many prophetic species of the marine invertebrates in the
higher Ordovician beds of Estland. Of Silurian species there are
-26 (the Swedish island, Gotland, alone having 76) and 6 of these
are cosmopolitan forms, ranging as far south as Ireland and Eng-
land. Of the southern European faunas there is but one in
common with Bohemia, Deiphon forbesi.

Of the 229 western Russian Ordovician species, at least 64 are
also found in Sweden or Norway. The close proximity of these
regions leads one to look for a greater number of forms in com-
mon than is here indicated. As there was then, seemingly, no
land barrier between these regions, and as the fossils show that
all belong to one marine province, the discrepancy must be

446 Scientific Intelligence.

explained in difference of geological horizons. In Sweden
(Oeland) the Upper Cambrian passes gradually into the Ordovi-
cian while in Estland a sandstone and black-shale zone of no great
thickness represents the last of the Cambrian and reposes con-
formably upon the Lower Cambrian. Upon the former then
follow glauconitic sands and glauconite-bearing dolomites hold-
ing a fauna that can be compared in America only with the
highest members of the Lower Ordovician. The sequence in
Estland then seems to continue without any marked break to the
end of the Ordovician, but in southern Sweden these higher hori-
zons are certainly not faunally present in Oeland. They are,
however, sparingly represented to the north in the Leptaena Kalk
of Dalarne. These regions have therefore dissimilar sediments
of various transgressions of a sea from the south and west over-
lapping on the Baltic shield of Suess. With England and Ire-
land there is far lessin common, there being but 11-species, and
nearly all of these are from the uppermost Ordovician or Lyck-
holm (F) formation. With the Lower Ordovician of England
there seems to be nothing in common, but this must be ascribed to
the very backward condition of British Ordovician paleontology.
Cus:

4. The Stratigraphy of the Western Ameri ican. Trias - bye te
Sairn. Festschrift zum -siebzigsten Geburtstage von “Adolph V.
Koenen, 1907, pp. 377-434.—In this important paper Professor
Smith gives a general summary of the western American marine
Triassic formations, their faunas, and the probable waterways of
invertebrate migrations. These inter-migrations are complex and
in the main are based on abundant ammonite evidence. Having
shown that migrations take place from the Mediterranean across
the Atlantic and Mexico to California and, in the opposite diree-
tion from India around the northern shores of the Pacific, the
‘author then takes up the later migrations of Mesozoic and Ceno-
zoic time. ‘The hindrance to continuous northern Pacific migra-
tions during geological time he thinks is due to the deep channel

east of Kamehatka, , through which now courses the cold water cur-
rent from the Bering Sea. This barrier has been effective at dif-
ferent times and, at others, has been considerably shallowed
through elevation. ‘A rise of 200 meters would close Bering
Strait, and about one-half of ‘Bering Sea, giving a shoreline coin-
ciding approximately with a oreat, circle. It would then leave
the Aleutian chain as a long narrow peninsula reaching out from’
Alaska towards Siberia, separated from Kamchatka by a narrow
but deep channel ; while the mainland of Alaska and Siberia would
be united by a broad land-bridge. This change in the height of
the land would cut off ail influx of cold water from the Arctic
Sea.” Cos:

5. Remarks on and Descriptions of new Fossil Unionidee
Jrom the Laramie Clays of Montana; by R. P. WauirFriexp.
Amer. Mus. Nat. Hist., Bull. xxiii, 1907, pp. 623-628 ; pls. 38-42.
Herein are described eight new species of Unio, six other

Geology and Natural History. 447

forms from near the same locality having been defined by
the author in-an earlier volume. In regard to these the writer
remarks: ‘ Considering all the similarities between these Laramie
fossils and their representatives in the Mississippi and Ohio water-
sheds, I venture to state that these farther western waters of the
Laramie times were the original home of much of the Unio fauna
of these more eastern recent localities.” C. 8.

6. Palaeontologia Universalis, fasc. iii, ser. 11, August, 1907,
sheets 112-125.—This fasciculus treats of fourteen species, being
the work of Bézier, Boussac, Chelot, Cossmann, Lemoine and
Sacco.

7. Die Fossilen Insekten ; von ANTon Hanprirscu.—The sixth
Lieferung of this monograph, published by W. Engelmann, Leipzig,
has come to hand. It has pages 801-960 and plates 46-51 and
treats of the Tertiary insects. The remainder of the work will
appear during the coming winter. C8.

8. Illinois State Geological Survey. Bulletin No. 4. Year
Book for 1906. H. Foster Bain, Director. Pp. 260, with 4
plates and 4 figures. Urbana (University of Illinois), 1907.—
Earlier bulletins have already been noticed in the Journal (xxii,
543, xxill, 227). The present one, No. 4, contains the adminis-
trative report of the Director for 1906, with also an account of
the progress made in the topographic survey being carried on
with the codperation of the U. 8. Geological Survey and the
State Geological Commission. The State has hitherto had no
accurate and official topographic map, so that the results of the
present work are most important. Of the special subjects also
discussed may be mentioned several chapters on the study of
coal and also an account of two remarkable drill cores, both from
Hamilton county, one 920 feet deep (at Delafield) and the other
(near McLeansboro) extending to a depth of 1294 feet. The
records kept and here summarized are very complete and yield
valuable geological section's.

9. Connecticut Geological and Natural History Survey,
Bulletin No. 8. Bibliography of the Geology of Connecticut ;
by Herpert E. Gregory. Pp. 123. Hartford, 1907 (The Case,
Lockwood & Brainard Company).—The eighth bulletin of the
Connecticut Geological Survey is devoted to the bibliography of
the geology of the State. This has been prepared by Prof.
Gregory after a critical study of the literature extending over a
number of years. It presents a practically complete list of titles
of papers published up to January, 1906, with brief statements
giving the main results brought out in each. Upwards of three
hundred entries are included in the bibliography, and a list of
nearly one hundred references to maps is also added. The work
cannot fail to be of great value to all concerned with Connecti-
cut geology.

Am. Jour. Sci.—FourtH SERIES, VoL. XXIV, No. 143,—Novemser, 1907.
31

448 Scientific Intelligence.

10. Tables of Minerals including the Uses of Minerals and
Statistics of the Domestic Production ; by Samurn L. PENFIELD.
Second edition. Pp. vi, 88. New York, 1907 (John Wiley &
Sons).—The first edition of these very useful tables was prepared
by Professor Penfield in 1903 and is noticed on page 330 of vol.
xv. The second edition, now issued by Dr. W. E. Ford, has
been brought down to date, especially with respect to statistics
of mineral production ; Part III, dealing with minerals useful
in the arts, has also been re-written and enlarged.

11. New j arlosite are new
species recently discovered near the head waters of the San
Benito river in San Benito County, California, and described by
G. D. LovuprersBack. Lenitoite occurs in small hexagonal crys-
tals of pyramidal habit, referred to the trigonal division. The
hardness is 6°25-6°5 and the specific gravity 3°64-3°65. The
color varies, sometimes in the same crystal, from colorless to
deep sapphire-blue ; the latter variety has been cut as a gem and
is of particularly brilliant luster—rivalling the sapphire—because
of the high refractive index (= 1°77, e= 1°80 for sodium light).
In composition it is a titano-silicate of barium, BaTiSi,O,. An
analysis by W. C. Blasdale gave:

SiO, 43°68 TiO, 20°09 BaO 36°33 = 100-10

Carlosite is associated with benitoite as a black, or brownish
black, prismatic mineral, with perfect prismatic cleavage yielding
an angle of 80° 10’. Its hardness is 5-6 and as it is biaxial and
shows oblique extinction, it is inferred to be monoclinic. The
composition is as yet undetermined, but it fuses easily (1°5) to a
black enamel bead yielding a soda flame. These minerals occur
disseminated in narrow veins in a basic igneous rock. A more
eomplete examination is in progress— Bul. Geol. Univ. Cali-
fornia, vol. v, 149, 1907.

12, Elements of ‘Biology : A Practical Text- Book Correlating
Botany, Zoology, and Human Physiology ; by GrorcE WILLIAM
Hunter. Pp. 445; New York, Cincinnati, Chicago, 1907
(American Book Company).—The aim of the book is to present
in simple language such of the more important principles and
facts of botany, zoology, and human physiology as can be read-
ily comprehended by the pupil in the first year in the high school.
Features which are of more popular interest and practical import-
ance are emphasized and all unnecessary description is omitted.
Numerous suggestions as to laboratory and field work encourage
the pupil to carry on personal investigations outside the class-
room. At the conclusion of each chapter are lists of reference
books for the pupil and for the teacher. The illustrations are
numerous and well chosen. This book should prove of unusual
value, not only for the pupil beginning the study of biology, but
also for the teacher without wide experience, who is called upon
to give elementary courses in biology or nature study. w.R.c.

13. Elements of Physiology ; ‘by TuEropore Hoven and
Wituiam T. Sepewick. Pp. 321; Boston, New York, 1907
(Ginn & Company).—This book consists of a reprint of Part I

Geology and Natural History. 449

of The Human Mechanism (see this Journal, vol. xxii, p. 549)
and contains that portion of the larger work which treats of
Physiology. <A single chapter on drugs, alcohol and tobacco
from Part II has been added to meet the requirements of certain
State laws. W. RB. C.
14. The Young of the Crayfishes Astacus and Cambarus ;
by E. A. ANpDrEws. Smithsonian Contributions to Knowledge,
vol. xxxv, pp. 1-79, pl. 1-10. Washington, 1907.—“ The memoir
describes and illustrates the young of two kinds of crayfishes,
one from Oregon and one from Maryland, which represent the
two most diverse forms found in North America. . . . It deter-
mines the form and habits of the first, second, and third larval
stages, .... describes the hitherto unknown nature of succes-
sive mechanical attachments of the offspring to the parent, and
opens up the problem of the nature and causes of the incipient
family life in the crayfish.” New data are supplied which lead
toward the solution of the problems of the geographical distri-
bution and the origin of the species of crayfish, the evidence
furnished pointing toward the wider departure from the ances-
tral state and the more highly evolved condition on the part of
Cambarus than of Astacus. B. W. K.
15. Evolution and Animal Life. An elementary discussion
of facts, processes, laws, and theories relating to the Life and
Hvolution of Animals ; by Davin Srarr Jordan and VERNON
Lyman KELLoGe.. 8vo, 489 pp., 298 cuts, 2 colored plates.
New York: D. Appleton & Co. 1907.—This is a very complete
and comprehensive discussion of the various factors and theories
of evolution, as indicated by the title. It includes, besides the
subjects ordinarily discussed in elementary works on evolution,
useful chapters on Paleontology and Geographical Distribution,
Adaptations, Parasitism and Degeneration, Commensalism, In-
stinct and Reason, etc. Altogether it is the most useful and up
to date treatise on evolution that we have seen. AS ES Vie
16. Report on the Crustacea (Brachyura and Anomura) col-
lected by the North Pacific Exploring Hxupedition, 1853-1856 ;
by Wirtram Stimpson (edited by Miss M. J. Rarupun).
Smithsonian Miscellaneous Collections, part of vol. xlix, 1907,
8vo, 240 pp., 26 plates.—This is the full report on the Crustacea
of the two groups named, prepared by Dr. Stimpson and sent in
to the Navy Department just before the great Chicago fire in
1871. Dr. Stimpson, himself, seems to have forgotten that he
had delivered this report, for in his statement of the losses by
the destruction of the Chicago. Academy, he enumerated this as
well as his other unfinished reports, and all his drawings. This
report was, however, discovered in 1872, shortly after his death,
accompanied by a full series of carefully executed drawings,
largely made with a silver point on hard cardboard, as Dr. Stimp-
son told the writer during an interview in Chicago about six
months before the fire. At the same time he showed large num-
bers of drawings of Crustacea of the remaining groups, as well
as of mollusca, tunicates, etc., done in the same painstaking man-

450 Scientific Intelligence.

ner, in addition to a large collection of colored drawings made
from life by himself, while on the expedition. All that was
saved was this forgotten report. Shortly after its discovery it
was submitted by Prof. 8. F. Baird, Secretary of the Smithsonian
Institution, to specialists for their opinions as to its value; among
others, to Prof. 8. I. Smith, of Yale, then the leading carcinolo-
gist in this country, and to the writer. Prof. Smith and others,
including the writer, reported that it was a very important and
valuable paper, that should be published at once. It gave full
descriptions of large numbers of new genera and species of
which Dr. Stimpson had published only very brief Latin diagnoses,
without figures, and of which the types had all been lost in the
Chicago fire. The writer and others repeatedly at later times
urged its publication, but without avail. The only reason given
was the lack of funds to reproduce the figures. It had to wait
till the cheap modern methods of reproduction were invented
before any department of the U.S. Government or the Smith-
sonian Institution could afford to print it! Yetit was an official
report of an important government expedition, made by the
naturalist of the expedition, and the very great expense of mak-
ing the drawings had already been paid for, as well as the great
cost of making the report. It is an illustration of the lack of
appreciation of the value of high class, painstaking scientific
illustrations, and also of the difficulty of getting posthumous
works published, no matter how valuable. Of course, the
processes of engraving have been so reduced in cost for more
than ten years that the original excuse has not been valid for a
long time. It is, then, a matter for congratulation, that under
the present efficient officers of the Smithsonian, and with the
able editing of Miss Rathbun, this valuable report has at length
been brought out. It includes 358 species and the numerous
forms described by Stimpson as new are nearly all illustrated.
The editor has given in footnotes the modern names when they
have been changed. INGE Ns
17. Reports on the Scientific Results of the Expedition to the
Eastern Tropical Pacific, in charge of Alexander Agassiz, by
the U. S. Fish Com. Steamer Albatross, from Oct., 1904 to
March, 1905. Lieut. Com. L. M. Garret, U.S. N. command-
ing. X. The Brachyura, by Mary J. Rathbun. 4to, 54 pp., 9
plates. Memoirs Mus. Comp. Zoology, vol. xxxv, No. 2, Aug.,
1907.—This is an important contribution to our knowledge of
the Crustacean fauna of the Central and South Pacific. The
species are in large part from the deep sea, but many shore and
shallow water forms are also included. The total number of
species included is 136, of which 18 species and one genus are
new to science. ‘The numerous excellent illustrations are in large
part reproductions of photographs. A. E. V.

OBITUARY.
M. Maurice Lorwy, the eminent French astronomer, Director
of the Paris Observatory, died suddenly on Oct. 15 at the age
of seventy-five years.

THE

AMERICAN JOURNAL OF SCIENCE

[FOURTH SERIES,]

Coe

Arr. XLIL.—The Internal Temperature Gradient of Metals ;

by Scuurter B. SrErviss.

In the Physical Review for October, 1906, Dr. C. B. Thwing
reported experiments in which he found that cylindrical speci-
mens of common materials showed at points 3°" from their
surfaces an excess of temperature above their surroundings,
ranging from 0°-000015 in the case of marble to 0°-000308 for
aluminum oxide. This he ascribed to their radio-activity.
Now the work of Strutt, Eve, McLennan and Burton, Cooke,
Wood, and Campbell, shows that common materials have the
power of ionizing air to a slight extent. But Strutt and
Rutherford are of the opinion that the feeble radio- -activity of
common materials is a superficial rather than a volume effect,
so further experiments seemed advisable. Thus far the results
have been entirely negative.

In the meantime H. Greinacher has reopened the subject by
publishing in the Annalen der Physik for October, 1907, an
interesting computation of the magnitude of the effect to be
expected, showing that Dr. Thwing’s results are much too
large. It has, therefore, seemed worth while to publish these
experimental results immediately, without waiting for the com-
pletion of certain other experiments of rather a different sort,
which were to have been made.

Since the value of any work of this sort is entirely dependent
upon the accuracy of the manipulation, the reader must pardon
even a tedious fullness of detailin the description of apparatus
which follows :

Am. Jour. Sct.—FourtH Serius, Vou. XXIV, No. 144.—DrEcEMBER, 1907.
32

452 Serviss—Internal Temperature Gradient of Metals.

The Apparatus.

Dr. Thwing’s general method was followed ; viz., the sub-
stance to be tested is used in the form of two equal cylinders,
set in an air bath, one on top of the other, with a thermopile
consisting of a large number of junctions wound radially
around a ring of mica, between them to measure the mean
gradient along the radii. Certain modifications in the details
were, however, made; these will be pointed out as the appa-
ratus is described. There may be some further departures
from Dr. Thwing’s construction—his paper is unfortunately
brief, especially in the description of his apparatus.

The Specomens.—For these experiments cylinders of lead
and of iron were cast, about 12°" in diameter and 4°5 or 5:0™
thick, and turned down to 10x 2°5™. As the castings appeared
less porous on the bottom and consequently that end of the
cylinders seemed likely to be more uniform, the lower end of
each casting was turned down until a satisfactory surface was
obtained and then the top cut off until the cylinder was of the
desired thickness. It was necessary to oil the cutting tool to
obtain smooth surfaces, so after the cylinders were taken from
the chuck they were washed with naphtha to remove the oil,
and placed in a cabinet until tested. The top side of each (as
cast) was marked; when the specimens were not in use, the
unmarked base was exposed to the air, and when tested, placed
next to the thermal couples. The weights of the cylinders
after turning indicated that they were solid. The specimens
were of ordinary commercial purity. The lead was cast in
the laboratory by melting clean weights which were no longer
needed. The iron came from a local foundry and was fine and
homogeneous.

The thermopile consisted of 100 copper-constantan thermal
couples, made from wires 0:020™ in diameter (No. 32 on the
Brown and Sharpe gauge). All the wire of each kind was cut
from the same strand. Pieces of the constantan 1°5 meters
Jong were annealed by clamping the ends to wide U-shaped,
rigid, wooden frame and heating to incandescence in free air
by a current of 3°5 amperes from a direct current dynamo.
The current was of uniform strength and was maintained for
about three seconds in each case. The ends of wire were held
at the same level, so that while hot, it hung in approximately a
catenary. When the wire was clamped on the frame it was
drawn just taut, and after cooling returned to its original ten-
sion, so it was not permanently extended to any appreciable
degree.

Two lengths were selected at random and laid aside for
calibration. The remainder was cut into pieces exactly 35

rnal Temperature Gradient of Metals. 453

long and the copper into lengths of 7-6". Pairs of these
were butted and brazed with an alcohol flame. Silver solder
was used with borax as a flux. Care was taken to use as
small a quantity of silver as practicable, but if too small a
bead is used, it flows entirely to one or the other of the wires.
All junctions which suffered an accident in making, were bent,
or happened to have an undesirably large excess “of silver, or
which did not successfully flow the first time, were rejected.
After the experience of making 100 good junctions, I was able
to join these couples end to end in a similar manner, winding
them on the mica ring as completed, so that ali the junctions
of the thermopile were as uniform as they well could be.
The outer diameter of the mica ring was 12°38 and the
diameter of the hole was 2-4. Dr. Thwine placed the outer
row of junctions of his thermopile outside the cylinders and

1

the inner row between them. I so placed the constantan that
my outer row of junctions would come 0-6™ inside the con-
vex surface of the cylinders, and the inner row was, of course,
3° nearer the center. This change was made in order to
expose all the junctions in the same manner. This is a funda-
mental difference between my work and his, and it will be
discussed at length in connection with the results.
Experiments in this laboratory have shown “ that moderately
sharp bending (with a radius of 0°5°™) is not very injurious in
the ordinary thermo-electrie use of german- -silver wire ’*
but, of course, the bend about the ring of mica is necessarily
much shar per than this. Under these circumstances, it
seemed best, since the softer and more homogeneous copper
would suffer less than the harder alloy, to leave the constan-
tan straight; the junctions were, therefore, all on the same

* Proc. Amer. Acad., xli, p. 559, 1906.

454 Serviss

Internal Temperature Gradient of Metals.

side of the thermopile, and the copper was bent twice to pass
around the ring. The edges of the mica were marked at
regular intervals and very ‘slightly notched to hold the wires
in “position. The first and the last members of the thermopile
were, of course, copper, and were soldered directly to longer
pieces of the same wire. These were slipped through small
rubber tubing, coiled twice around the inside of the ther-
mostat at the level of the thermopile, and then through
85°" of copper tubing passing through the ice, and then
soldered to the galvanometer leads. Thus fully a meter and
a half of each lead wire was at zero, adequately protecting the
end: junctions of the thermopile from heating by conduction
through them. In Dr. Thwing’s apparatus, these wires came
in through a short tube between the thermostat and the ice
vat. Now it is probable that any constant error arising from
this source would tend to warm the outer junction rather than
the inner one, which is just the reverse of what Dr. Thwing
found. Therefore, if in spite of the possible masking of ‘hie
effect by conduction along his short lead wires, Dr. Thwi ing is
able to find an excess of. temperature at the inner row, then
that excess ought to be as great, if not greater, when, as here,
this source of error is certainly eliminated.

After the thermal couples were all in place on the mica
frame, they were painted with a thin coat of asphaltum var-
nish and baked in an electric oven for several hours at 120° C.
The insulation was found to be defective in a few places, but
after a second treatment stood a severe test. This varnish
proved a very satisfactory insulator when used for two years
on the main bars of the Thomson effect researches of Pro-
fessor EK. H. Hall, in which the writer had a part.* It was
accordingly adopted in preference to the beeswax and rosin
used by Dr. Thwing, which may have contained acid, since
there is no evidence to show that it was especially prepared.
Furthermore, the use of asphaltum and thinner wires enabled
me to make a thermopile only a third as thick as Dr. Thwing’s,
which is of course an advantage. And finally there was a
mechanical advantage. Although the couples with large
joints had been rejected, still the junctions were unavoid-
ably slightly larger than the wires themselves, and the weight
of the upper cylinder 1 ‘ested upon these two hundred joints.
This insulation was a thin, smooth, hard coating which was
not rubbed off or squeezed aside by this pressure.

There remains some possibility of disturbing effects due to
the circulation of air in the channels between the wires. To
prevent this, three rings of very thin mica (a, 6, ¢ in fig. 1)

*Proc. Amer. Acad., xli, 25-55, 1905, and xlii, 595-626, 1907.

Serviss—Internal Temperature Gradient of Metals. 455

were cut and rings of asphaltum varnish spread around them
as shown. The weight of the upper cylinder pressed this
between the wires, leaving the junctions in narrow annular
air spaces in contact with the upper cylinder. Possibly it may
seem preferable to have placed the rings of additional asphal-
tum directly on the thermopile, without the additional mica.
But as it was advisable to keep the thermopile flexible, this
last asphaltum was dried without baking, hence these thin
rings were needed to prevent the cylinders from sticking to
the thermopile. If the whole frame had been baked hard,
the wires would have been much more liable to accident in
changing the specimens, the channels between the wires would
not have been so thoroughly closed, and the thermopile
would have been more likely to warp, opening a wide gap or
producing a strain under the weight of the upper cylinder.
Incidentally the inner diameter of the ring in the top of
the thermostat had been made slightly too small to admit the
thermopile without bending it a little, and by leaving the
mica frame flexible it was not necessary to cut this ring. As
the distance between the cylinders was not materially increased
by the presence of the additional mica, it did no harm, but
increased the insulation, distributed the weight of the upper
cylinder over a wider area, and also protected the copper wire
projecting beyond the cylinders from possible air currents,
which might cause trouble. The thickness of the thermopile
was 0°74" and its.resistance at 0° was 48°6 ohms.

Calibration of the Couples.—bd0™ was cut from the middle
of each specimen selected for the calibration tests. These were
used to make couples as nearly like those on the thermopile as
possible, and tested between 0° and 8°. They gave 42°2 micro-
volts per degree, which agrees very well with the ordinarily
accepted value of 43 microvolts. A more careful calibration
was not necessary, as the very small differences of temperature
which are the subject of this investigation cannot be measured
within this degree of percentage accuracy. As it is difficult to
maintain small constant differences of temperature for calibra-
tion purposes, it is necessary to assume that the thermo-electric
height of constantan against copper is a right line in the neigh-
borhood of zero. A mean difference of 0°-000001 between the
inner and the outer row of junctions of the thermopile will,
then, generate 4°3x10~° volts.

The thermostat was of thin sheet copper, 17°5°" in diam-
eter and 25™ in extreme height. A composition casting
was fitted and soldered in the top. The cover was a second
casting held to this by twelve machine screws with a gasket
between. A light brass ring was suspended by double silk
threads, attached, through swivels, to screws (for leveling) in

456 Serviss—Internal Temperature Gradient of Metals.

the under side of the ring at the top. The suspended ring had
a socket which held the lower cylinder without slipping. The
threads were set vertically on a circle slightly larger than the
circumference of the thermopile. The junctions were thus
easily put in position symmetrically and could not be displaced
by any jar during the experiment. The weight of the upper
cylinder was relied upon to hold it in place. On opening the
thermostat nothing was ever found disarranged or wet.

The thermostat was clamped to a beveled ring, which was
firmly supported by three lengths of maple dowelling rod.
The ice vat was 42° in diameter and 72™ tall at the center.
A one-inch pipe was inserted in the bottom to insure com-

zi
ya
¥ A N
For convenience in draughting the three suspending threads and the
three legs are shown as two.

plete drainage. The whole was set in a barrel and packed
loosely with sawdust.

The Ice Bath.—As an ice bath is often used to produce a
constant temperature, I here give a detailed account of the
various schemes tried, hoping that my experience may be of
service to other experimenters. I may state here that this was
the most troublesome part of the whole investigation.

Dr. Thwing mentions a noticeable rise of temperature
(0°-00001 to 0°-00002) caused by the melting of the ice from
the bottom of the thermostat during the twelve-hour intervals
between which he packed the ice down. In other words,

‘

Serviss

Internal Temperature Gradient of Metals. 457

steady conditions were not realized. I attempted to obviate
this difficulty by making the bottom of both vessels conical at
an angle of 45°, hoping that the weight of the ice would cause
it to settle and replace that which melted, and so establish a
steady condition. But I found almost immediately that owing
to the comparatively narrow annular space between the two
cans and the regelation of the ice, this device alone was insnt-
ficient. Still it is an improvement over a flat-bottomed
thermostat, for the ice is kept in contact with the whole of the
conical bottom all the time, which materially aids in maintain-
ing constant conditions.

"Throughout the work, ice was added at regular intervals,
usually every four or five hours, sufficient to keep it near a fixed
level. At the same time, the drain water was weighed.
During the use of the first. arr angement, these weights - were
irregular, indicating a honeycombing of the ice and regelation.
When the i ice was crowded down, considerable changes were
noticed in the temperature, similar to those observed by Dr.
Thwing. About 12 kgms. of ice were used in 24 hours, the
amount depending somewhat on the room temperature.

Next a spudger was built out of an iron casting and six

4inch iron rods projecting vertically downward in a cirele mid-
way between the thermostat and the vat. A fan motor and
reducing gear raised and lowered this, with some freedom of
rotation, every seventeen seconds. A very good quality of
natural ice was used throughout this work. During these tests
it was ground into pieces weighing 10 grams or less. Fora
time this seemed to be successful. But the pressure of the
prongs raised the melting point sufficiently to disturb the
gradient and regelation followed. The ice from the upper
third of the vat was removed, and the remainder was found
frozen into nearly a solid mass, in which the melting point was
variable under the changing pressure of the spudger. The
external junctions showed sudden and considerable changes in
temperature within a few minutes. Consequently this scheme
was abandoned.

The third attempt was successful. Larger lumps of ice
were used, averaging 35 or 40 grams in weight. These were
cut from the 100-Ib. pieces with a sharp chisel, and the corners
rounded by washing in running tap water, which prevented
relegation, and the consequent freezing into a single, solid
mass. But the melting point still varied a little, as indicated
by the fluctuations in the gradient. It isa well known fact
that the temperature of a mixture of ice and water rises con-
siderably above zero, depending on the amount of water that
has accumulated. By stirring the ice and collecting the drip
at short intervals, it was found that water was held between
the lumps of ice by capillary action. \ As the stirring was con-

458 Serviss—Internal Temperature Gradient of Metals.

tinued, this water was released, the drip became uniform, and
the temperature gradient steady. Only slight stirring between
the observations was necessary to keep the conditions sensibly
constant. A mechanical device, consisting of two endless
chains driven in opposite directions ver tically through the ice,
has been installed to accomplish the stirring unifor my.

Connections.— All of the wiring outside of the thermostat
and the copper tube was from the same bobbin of No. 18 cop-
per annunciator wire, with double cotton insulation. The
connections throughout were hard soldered, in the same man-
ner as the thermal junctions, except at the valyanometer coils

and at the points where the shift was made from direct circuit
through the galvanometer to a potentiometer. At these places,
copper binding screws were used to clamp the ends of the
wires in contact. The connections in the galvanometer were
undisturbed; those at the observing table were cleaned daily
or oftener with infusorial ear th, in addition to the care which
was always taken to avoid contact of the bared ends with con-
taminating objects, especially the fingers. Warming one or
another of these binding screws pr oduced no perceptible deflec-
tion in the ealvanometer. The constant temperature room of
the laboratory was unfortunately occupied by another research,
but every precaution was taken to protect the apparatus from
changes of temperature. The doors and the windows were
kept closed, the galvanometer and the observing table were
further protected from air currents by screens, the copper
lead wires were wound thickly with strips of cotton batting,
and carried from one instrument to another over a wooden
frame. The resistance of the lead wires was 0°48 ohms at 20°.

Commutators—When the specimens were first placed in
position in the thermostat, the current from the thermopile
was sent directly through the galvanometer, shunted by a
standard coil, or with a _ resistance in series. The lead
wires were branched so that the current through the galvano-
meter could be reversed either by a four-point mercury com-
mutator or by a double throw copper switch. Both were
tried and the mercury commutator was found to be less dis-
turbed by local thermal effects; at least they were more uni-
form and so more perfectly allowed for. The copper switch
was, therefore, removed from the circuit, so as to have no
branch wires projecting, and the mercury commutator used
exclusively.

The mercury commutator was made from a new block of
mahogany, the mercury was freshly distilled, and the lead
wires amalgamated by cleaning thoroughly with emery cloth
and rubbing with the mercury ; no acid was used. This com-
mutator was placed in a small wooden box, and well protected
by two thicknesses of heavy asbestos paper and loose asbestos

Serviss—Internal Temperature Gradient of Metals. 459

fiber. A small wooden dowelling rod passed up through the
cover to reverse the key. ~The thermal disturbances in the
commutator and galvanometer could be found by disconnecting
the thermopile and short- circuiting the lead wires beyond the
commutator. The effects due to the galvanometer could be
measured by lifting the reversing key and closing the galvano-
meter terminals at their mercury cups by a copper strap, oper-
ated through the side of the protecting box. These deflections
were always small, never exceeding a few millimeters, and
rarely amounting to 1™™, although the entire resistance of the
circuit was about 11 ohms.

The galvanometer was a three-cased “ Panzer’ of the
duBbois and Rubens design.* This was an extremely satisfac-
tory instrument because of its high and nearly constant sensi-
tiveness, its freedom from external magnetic disturbances, and
the steadiness of the zero. The resistances of the coils were
5°5 and 5:3 ohms at 20° C., the heavier suspension (800 mgm.)
was used, hung by a very fine quartz fiber. The sensitiveness
was tested twice a day or oftener with a potentiometer, using
a cadmium cell for a standard of electromotive force and sev-
eral standard coils. This circuit was entirely independent of
the rest of the apparatus and was disconnected from the gal-
vanometer when not in use. The telescope, made by Alvan
Clark of Cambridgeport, gave a large, clear image and tenths
of millimeters could easily be read. A current of 4:1 10-"
amperes gave 1™™ deflection from zero at a scale distance of
4-2 meters. When the thermopile was connected directly
through the galvanometer, 60 ohms were in cireuit and a dif-
ference of 0°-000001 between the inner and the outer row of
junctions would give a deflection of 1-7" from zero. so that a
mean gradient of 0°-0000001 per cm. could certainly have
been detected.

The Potentiometer.—The electromotive force of the thermo-
pile was usually measured by a potentiometer, kindly placed
at my disposal by Professor Bb. O. Peirce. The slide wire,
which could be easily changed, rested upon a glass scale. The
keys were provided with springs, insuring uniform pressure on
the wire. The contacts and the cable were copper, and like
all the other connections were covered with cotton. A stand-
ard cadmium cell was used, and the resistance (R) in the main
circuit was adjusted so that the fall of potential was 1 volt
over 10,000 ohms in 7. A D’Arsonval galvanometer (D), sen-
sitive to a change of 1 ohm in the box R, was used for this
adjustment. Unless the glass scale of the potentiometer was
kept very clean, troublesome leakages and short circuits oc-
curred. Washing the glass with alcohol and with ether was
found to be effective.

* Zeitschr. fiir Instr., xx, pp. 65-78, 1900.

460 Serviss—Internal Temperature Gradient of Metals.

When the cylinders were first placed in the thermostat, the
outside cooled down rapidly and the gradient was large. It
was more convenient to follow this down by sending the eur-
rent directly through the galvanometer. But after some hours
the difference of temperature was small, and it was then meas-
ured on the potentiometer.

Three copper slide wires were used, as follows :

Size. Diam. Ohms/meter. Cm./0°:000001.
18 1:0247™, 0:01979 0°22
12 2°053 6°004922 0°86

3 5°827 0:0006106 (lil

Summary.—A brief enumeration of the changes made in
Dr. Thwing’s apparatus may now facilitate comparison. These
changes were made, not so much because they were necessary
as because they seemed to add to the certainty of the results.

(1) The junctions were all placed inside the cylinders, in
order to expose them in the same manner.

(2) In the final observations, the circuit was all of copper to
avoid disturbing thermo-electric forces.

(8) An armored astatic galvanometer of unusual qualities
was available, instead of a D’Arsonval, which usually has
internal electromotive forces.

(4) A potentiometer was used to measure the e.m.f. of the
thermal couples instead of sending the current produced by
them through the galvanometer.

(5) Constant conditions were maintained by slowly stirring
the ice to drain it dry and using conical bottoms on the ther-
mostat and ice vat to prevent the ice from melting away from
the thermostat. To this same end the external parts were pro-
tected from large or sudden changes of temperature.

Serviss—Internal Temperature Gradient of Metals. 461

(6) The specimens were suspended in the air bath so as to
expose them at all surfaces equally, instead of piling them on
a block of stearie and a paper box.

(7) Finer wires and thinner insulation were used in the ther-
mopile, bringing the cylinders closer together.

(8) The lead wires from the thermopile to the outside con-
nections were much longer, insuring against conduction of
heat to the end junctions.

Observations and Results.

I shall now give a brief history of the experiments. As
Dr. Thwing found the largest effect in lead, it was selected ;
iron was taken for a second metal, chiefly because of the ease
with which it can be worked. Four cylinders of each metal
125™ and two 129°", were cast early in January; all of
each metal were of the same pouring. Two of the smaller size
were also cast from a different sample of lead. One pair of
the smaller cylinders of each metal was turned at once and
laid aside while the rest of the apparatus was being con-
structed; owing to delays and trouble with the ice bath, the
final observations were not made until July.

When I outlined this investigation, [ provided for certain
variations and check experiments. I planned to use cylinders
from the same pouring with fresh surfaces and with aged sur-
faces, hence the extra small cylinders and the considerable
allowance for turning; also as a check experiment, to use
cylinders twice as thick; but for this, I should have made my
thermostat 5 to 8°" shorter. As my results have been entirely
negative, these other forms of the experiment have not been
made. The work on the radio-activity of metals shows that
it may be much lessened by scraping the surfaces, and if a
specimen long exposed to the air does not give a positive
result, a fresh one could not be expected to do so. The varia-
tion with thicker cylinders was also given up for the same
reason.

Throughout the work, all the ice and drip-water was weighed
to detect honeycombing of the ice. The ice was kept at nearly
the same level by the addition of a sufficient quantity every
four or five hours to fill the vat to a certain mark. The ther-
mopile was tested before and after each pair of specimens.
There was a very slight e.m.f. detected, due probably to uncon-
trolled external conditions. This was less than the slight
continuous movement of the zero.

The first run (April 5-15) was made on lead. In this set of
observations the current from the junctions was sent directly
through the galvanometer, which was only about 1/20 as sen-

462 Serviss—Internal Temperature Gradient of Metals.

sitive as in the other runs. The ice was kept covered with a
blanket, and crowded down from time to time. A steady con-
dition was not reached, owing to the manner of managing the
ice-bath. The eradient was “variable, even reversing at times,
after the ice had melted out, at times rising to values fully as
large as those observed by Dr. Thwi ing ; but « even this imperfect
series gave no indication of a steady oradient, i. e. there was
no positive effect toward which it returned after being dis-
turbed.

The second run (May 8-20). The spudger had meanwhile
been installed, the sensitiveness of the galvanometer increased
to its final value (41<10-" °™?/__, “from zero at a scale
distance of 4:2"). Owing to trouble with the spudger and a
final breakdown, this run gave no additional evidence on the
preblem.

The third run (June 29-August 1). The use of the spudger
was at first continued, but after a ie days was replaced by
stirring the ice at frequent intervals. In this run the work
was continued day and night, much of the observing being
done at night when the room temperature was less variable
and the building was free from mechanical jars. The inter-
missions were brief, and during them the apparatus was left
under the care of some other person, who stirred the ice occa-
sionally te prevent too great an accumulation of water and
who added more ice if neces sary. Twosamples of lead and one
of iron were very carefully tested and sufficiently steady con-
ditions were maintained for three or four hours at a time to
show unmistakably that no mean gradient as great as 0°-0000001
existed.

It would be interesting to compare the cooling curves with
those computed, but as the labor of maintaining steady condi-
tions was so oreat I did not attempt to make observations until
I knew the cylinders had cooled practically to zero throughout.
The specimens were placed in the thermostat at about 20°,
and it seemed to require fully 24 hours to bring the junctions
of the thermopile to sensibly the same temperature. This is
considerably longer than is required for the cylinders to cool
to the same gradient, the difference indicating a lag between
the junctions and the cylinders, due to the insulation of the
asphaltum, but this is no evidence that they did not ultimately
assume the temperature of the adjacent metal. The pressure
at the bottom of the thermostat, due to the weight of 25° of
ice, is about 1/50 of an atmosphere, which will cause a lower-
ing of the melting temperature of 0°-00015. Possibly, when
the vat is jarred by the stirring, convection currents may be
set up, but they will not affect the temperature of the copper
wire projecting beyond the cylinders, for the additional mica

Serviss—Internal Temperature Gradient of Metals. 463

covers them (see fig. 1), and may be relied upon to protect
them from so small an effect as that due to the passing of the
small quantity of air transferred by convection.

In the early part of this paper (p. 453), I called attention to the
fundamental change I made by placing all my junctions be-
tween the cylinders in order to expose both rows equally. I
wish now to explain why this change was made.

Let us compute the gradient in a sphere of radius ¢, in
which 6 calories are generated per gram per second. Take
p = density, / = conductivity, A = surface conductivity. Then
when a steady state has been set up,

4 if
Se
3

8k dt

He po dy

; Te 3k
Integrating, Fas t+C0
2 p

To determine the constant of integration,

A
when 7 = ¢, = Tr pb = 4nc°hkt,
tS)
pes epb
ier Ue
Ci c
Os So =
2 a h

from which it appears that the gradient is a parabola.

Strutt* in adetermination of the ionizing power of common
materials, states “radium is 100,000 times more active than
uranium, and uranium 3000 times more active than the most
active material I have experimented with. So that one part of
uranium in 300,000,000 would suffice to account for the
observed effects.” His results show as wide a range for differ-
ent specimens of the same metal as for different metals. So
although he did not try iron, we may reasonably suppose that
it would give an effect similar to that set for the maximum ;
the order of magnitude is all that can be shown by such a com-
putation as this. Curie and Labordet found that radium

* Phil. Mag. (5), v, pp. 680-685, 1903.
+ Comptes Rendus, exxxvi, p. 673, 1904.

464 Serviss—Internal Temperature Gradient of Metals.

radiated heat at the constant rate of about 100 gram ealories
per hour per gram. Applying this, Strutt’s most active material
would radiate 9X10~-" gram calories per second per gram.
Substituting this and the other constants for iron, we have,
with ¢ = 5™,

15 X 10% = 8025—7°,

from which we get,

r 15¢ t
(Ose 8:025 x 10-° HS ppl Ome
1:0 8°024
hot 8:°023
2 8021
4°4 8:006
5:0 8:000 5533

The junctions of my thermopile came at the third and the
fifth of these points, and it will be seen that the difference
between these could certainly have been detected with my
apparatus. Dr. Thwing gives the temperature of iron at 7 = 2-1
as 3°210~*, 1. e. his iron was about six times as radio-active as
Strutt’s most active substance, if we grant that the outer junc-
tions of the ther mopile 1 in the air bath were at zero, and that
the entire leak in Strutt’s electrometer was due to the radio-
activity of the specimens tested and not partially due to other
causes, as the emanations from the earth.

It is interesting to compare, in this connection, the gradients
in the same sphere of iron, heated to 20° thr oughout and
allowed to cool in the air bath at 0° sa We Geb:

r 5 min. 15 min. 1 hour 2 hours 3 hours 6 hours
On elO: S99 SalaiG Sie lod sO 0°1010 0:00717 Dos 7p lO
1°4 17°688

Pri 17-679

4°4 17°645

SEO FLO: 1611 17°632 12°410 0°1007 0:00715 2516 xl Ome
Diff. 0°060 0°055 0°040 0:0003 0:00002(2) 0°08x10~*
% O51 0°31 0°52 0°32 0°31 0°31

Similar computations have not been made for lead, because,
so far as I can find, experiments have failed to measure the
radio-activity of lead.

These tables show the decided discontinuity at the surface ; in
fact, in. both cases, only about 0°3 of 1 per cent of the total differ-
ence of temperature between the center and the air bath oc-
curs in the sphere itself; and moreover, this fraction remains
very nearly constant throughout the range computed. The
poorer the surface conductivity, other things remaining equal,

*W. E. Byerly: Fourier’s Series, etc., pp. 116-122. Boston, 1893.

Serviss—Internal Temperature Gradient of Metals. 465

the greater will be the fraction of the total difference occur-
ring at the surface. The computations were made for spheres
rather than for a cylinder, because of the greater simplicity
of the formulas. The gradients in the plane through the
center and parallel to the bases of a cylinder 10% in diameter
and 5 tall will be similar to those in the spheres but the
part of the drop which is in the metal will be smaller, because
of the greater surface in proportion to the volume and the fact
that the center is nearer the surface along the axis than before ;
hence the relative discontinuity at the surface will be even
greater for the cylinder.

Suppose, then, that the cylinders are cooling in air, either
with or without a generation of internal heat; the gradient is
CA-BD. My inner junction at S, will come to the tem-
perature of the adjacent metal, unless there is some small
gradient in the thin coating of asphaltum, and this is certainly
slight, as the quantity of heat transferred to maintain equality

4

A

i
!
!
'
!
!
1
|
!
i)
|
!
!
!
|
|
!
{
!

B
0 5

is small. So long as the sphere is cooling, 1. e. so long as the
gradient CA exists, heat will flow into this junction and down
the constantan wire, as well as through the cylinders them-
selves. Again, the ‘outer junction comes very nearly to the
temperature of the adjacent metal; there may be a gradient
through the insulation here as at the inner junction, and this
eradient may be from wire to cylinder or from cylinder to
wire, depending on the relative ease of conduction and radia-
tion from the cylinders and from the thermopile to the sur-
rounding air. So the difference of temperature between the

466 Serviss—Internal Temperature Gradient of Metals.

junctions is (except for the possible case of opposite gradients
in the insulation, which are probably very small compared with
that between the junctions) equal to that in the adjacent metal.
This method measures the gradient in the metal itself, while
Dr. Thwing’s does not, although he states his results in that
way. But it is to be noted that any method working on gra-
dients in the metal is insensitive, because the large discontinuity
at the surface makes the gradient in the metal itself small,

Suppose, on the other hand, that the outer junction be
placed in the air bath, as in Dr. Thwing’s arrangement. It
would seem, at first sioht, that this would be a much more
sensitive method for detecting internal generation of heat,
because while the inner junction is at the ‘temperature of the
adjacent metal as before, the outer junction is supposedly at
the temperature of the air bath. This reasoning I believe is
dubious. Heat is still flowing down the wire, and must
now be radiated into the air from the junction itself through
its coating of asphaltum, and while we may well believe that
the inner junction is at, or very near, the temperature of the
metal at that point, and that the temperature of the wire as it
passes along the radius of the cylinder is not far from that of
the adjacent metal; still the outer junction, cooling as it does,
is above the temperature of the air bath, due to “the greater
discontinuity when cooling takes place with the poorer surface
conductivity through the insulation. This well known fact
must be borne in mind when, as here, a thermojunction is
exposed in air, and proper precautions must be taken to insure
that the true temperature of the sur roundings is attained.

In conclusion, I desire to express my sincere thanks to Pro-
fessor. EK. H. Hall, who first suggested the work to me and who
made valuable suggestions at the beginning, and to Professor
B. O. Peirce for his kind interest and encouragement and the
generous loan of apparatus.

Jefferson Physical Laboratcry,
Harvard University,
Noy. 11, 1907.

C. R. Keyes—Aggraded Terraces of the Rio Grande. 46%

Art. XLIU.—Aggraded Terraces of the Rio Grande; by

Cuarues R. Keyes.

Tu Rio Grande is probably the most remarkably terraced
drainage-way in the world. Bordering the stream is an exten-
sive succession of high-level mesas that constitute the most
striking feature of the great valley’s surface relief.

These escarpmented plains, or mesas, inclining strongly
towards the river, are abruptly cut off as they near the banks
of the stream. The different series of terraces are at levels 20
to 250 or 300 feet above the flood-plain. The areal shape of
the plains is roughly that of a parallelogram, very long in pro-
rs to the width, the longer axis disposed at right angles
o the stream. Viewed from the side, the relationships of sey-
eral mesas is as represented in diagram below (fig. 1) :

1

Relationship of Rio Grande mesas.

In their wide extent, in their marked inclination towards
the river, in their isolated character, and in their smooth and
even surface, these high-lying plains present notable accentu-
ations of physiographic expression that are rarely recognizable
outside of the semi-arid regions. These elevated plains of the
broad river valley have long attracted the attention of trav-
elers. Many have been the explanations of their origin. No
solution of their genesis has yet appeared that is at all satisfac-
tory, or that takes into account the unusual climatic conditions
of the region, and the novel geologic agencies at work.

The Rio Grande is one of the great rivers of the American
continent. It is as long as the Mississippi. Unlike the case of
the latter stream, it has, for a long river, a very high gradient.
For the first thousand miles, from its headwaters, the average
fall is over five feet to the mile. In times of flood, when the
snows are melting in the mountains, the waters are almost of
torrential nature. At other times, asin the months of July
and August, the stream is often very nearly dry, although
there is ‘always a strong underflow beneath the sandy bottom.

After leaving the headwaters region, a distance of perhaps
Am. Jour. Sct.—FourrH Serres, Vou. XXIV, No. 144.—Decemper, 1907.

468 OC. R. Keyes—Aggraded Terraces of the Rio Grande.

300 miles, the Rio Grande receives no lateral augmentation
from perennial tributaries in a distance of uearly one thousand
miles, or nearly to the mouth of the Rio Pecos in Texas.
Through most ‘of this part of its course the river-bed is 1 ,000
to 2 500 feet below the level of the bolson plains lying on

either side of the great valley.

Phy siographically the Rio Grande has had a complex origin,
That portion of its course which lies in New Mexico is largely
antecedent m character. Other parts may be consequent.
Climatic conditions with which most of us are not very famil-
iar are chiefly responsible for many of the bold and unusual
relief effects that characterize the region through which the
Rio Grande flows. The valley mesas, or table- -plains, are
among the most instructive of all “of these topographic features.

Being a stream of such high gradient, with an average fall,
as already stated, for its 300 miles in New Mexico of over five
feet to the mile, its corrading powers are little short of mar-
velous and its carrying capacity amazing. During the dry
season of the year the waters of this stream are greatly over-
burdened with sediments and the channels rapidly silt up: In
late years, since most of the water has been taken out for irri-
gation purposes, the silting-up process has become a very seri-
ous matter. Many localities show clearly that within the last
dozen years the deposition has filled up the valley to heights

of 20 to 30 feet.

As already mentioned, no lateral drainage of a perennial na-
ture is received by the Rio Grande within the borders of New
Mexico. All increase of the grand stream from the sides takes
place only during brief periods of very heavy rainfall. The
side waters are then torrential. At other seasons of the year
these tributaries are true arroyos, or dry creeks, as their Span-
ish title signifies. These arroyos have very steep gradients,
often two to four and even five per cent.

In the mountain ranges on either side of the valley of the
great stream the arroyos occupy deep canyons. In this part
‘of their courses the channels are being rapidly cut deeper and
deeper into the indurated bed- rocks. After emer ging from
the mountains these lateral drainage-ways are as pronouncedly
constructive in character as they were destructive before.

Into the broad Rio Grande valley the side arroyos pour vast
quantities of coarse mountain waste. The alluvial fans which
are formed become confluent. On either side of the river
broad plains several miles wide are built up; and these are
inclined strongly towards the channel of the master stream
Were the river free from meandering, these plains on each
side would doubtless become continuous and even.

In the course of its wide meandering the stream cuts rapidly

C. R. Keyes—Aggraded Terraces of the Rio Grande. 469

into its banks on the convex side of its broad swings. In a
distance of two or three miles lateral change of channel a cliff
100 to 200 feet in height may be formed. Arroyos entering
the river from the bowed side quickly shorten their paths,
assume new and higher gradients, and scoop out canyons in the
old fans. Thus, between the adjacent arroyo-courses there is
left a high-level terrace, or mesa, bordering the main water-
way. On the opposite side of the bow the arroyos lengthen their
courses, lower their gr adients, and build out new extensions of
their fan-plains to the water’s edge. This phenomenon is best
shown in diagram, by an actual section across the Rio Grande
at Socorro (fig. 2).

Augh-lerel Plavm

i)
a
2
<
Bias
16)
2
fai

High-level plains of the Rio Grande, at Socorro.

The processes described are repeated again and again, along
the entire course of the river. As a final result there are
found sloping terraces at many different levels. The effects
are apparently as unique as they are striking. Closer inquiry
clearly shows that the same relief effects exist along the streams
in more humid regions: only in the last named cases the char-
acteristic phenomena are all but totally obscured because here
destructive processes are far more active than the constructive.
In the semi-arid districts the gradation conditions are reversed.

The typical characters of the high-level plains along the Rio
Grande are well displayed at inany points. At El Paso, at
Rincon, at Socorro, at Albuquerque and at San Filipe they are
finely shown. Near the last named place another interesting
factor comes into play. At different times during the latest
geologic epoch great basalt- flows have moved down the plains
towards the river. These have preserved the surfaces of the
old mesas at different levels. In some instances the river has
quite recently cut through the lava-cap, as represented below
(fig. 3).

The best description of these high-level terraces is given by
Herrick.* Although they were thought by this writer to be
striking enough to deserve the distinguishing title of clino-
plains, ‘he did not even hint at their real origin.

* American Geologist, vol. xxxiii, p. 376, 1904.

470 ©: R. Keyes—Aggraded Terraces of the Rio Grande.

A more recent description of what appears to be a very
similar phenomenon, but upon a miniature scale and in the
humid region of Vermont, is that of Fisher.* This account
is important in this connection in showing that the general
features are also recognizable beyond the arid districts.

The original surface of the line of bolson plains, through
which the Rio Grande now treads its way, appears to be from
700 to 1000 feet above the present bed of the stream. At the

= Zatec 2 Cravds E

=

Basalt flows cut by the Rio Grande at San Filipé.

present time there are remaining only few traces of these old
bolson surfaces. Most of these remnants have been preserved
only on account of being covered by extensive lava sheets.
The Albuquerque volcanoes and basalt sheets, six miles west
of the city, on the divide between the Rio Gr ande and the Rio
Puerco, seem to rest on the old bolson surface. The basalt
fields near Santa Fe have similar relations. The elevated
plain, southwest of Socorro, also appears to belong to the same
class. Farther south at Paraje, at old Fort Selden and at
El Paso, the same remnantal levels are noted. All of these
‘mesas are to be carefully distinguished from the so-called
clinoplains of the Rio Grande valley proper, that are seldom
found rising more than 200 to 800 feet above the water-level
in that stream.

There is still another class of plains which might be easily
taken for terraces bordering the Rio Grande, that should be
clearly separated from all the others. These may be in a
sense regarded as lake-benches. They are due to the damming
of the river for a brief period. Evidences are abundant
indicating that the Rio Grande has been in very recent times
and at various points dammed by the crossing of voleanic
flows. At old Fort Selden, at San Acacia above Socorro, at
Algodones above Albuquerque, at the White-Rock canyon
west of Santa Fe and at the Black canyon south of San
Antonio, basalt-flows have covered the course of the river and
produced a temporary choking up of its waters. These bar-
riers thrown across the river were from 200 to 500 feet in
height.

*Proc. Boston Soc. Nat. Hist., vol. xxxili, pp. 942, 1906.

C. R. Keyes—Aggraded Terraces of the Rio Grande. 471

Under the conditions now existing it would be expected that
such dams would leave conspicuous traces of the existence of
the lakes. Such, however, does not appear to have been the
ease. At least, the lakes were so ephemeral in character that
they left no evidences of distinct beach-levels, or elevated lines
of delta-fans. It may be that the fans were confluent in such
cases, and that the entire areas of the high plains, as those
existing above the White Rock canyon, at Rincon, and above
the mountain barrier at El Paso, represent these temporar y lake
deposits.

Notwithstanding the turbid character of the river waters
and the prodigious quantities of coarse materials carried by the
swiftly flowing stream, any lake that may have been formed
must have quickly disappear ed through the cutting of the dam
or holding barrier. Even with some extraneous accumulations
of sediments the local gradient was soon lowered, and the great
influx of coarse sands and gravels from the lateral arroyos
must have rapidly destroyed all traces of the lake beds and
moulded them into new aggradation surfaces. In any ease it
would be exceedingly difficult to differentiate these plains of
old lake-basins from ancient arroyo fillings. Moreover, the
plains aggraded by the arroyos extend, at the distal and more
elevated side, very much higher than would be possible for
lake beds to do. If any deposits of the last mentioned class
were formed, they would be quickly destroyed, or so disguised
as to be almost unrecognizable.

When the barriers finally disappeared, enabling the river to .
recover its original lower level and to renew its old meander-
ings, the gradients of the lower courses of the arroyos were
greatly increased, and the high aggradation-plains were cut by
canyons more or less parallel to one another. This may be the
explanation of certain local series of close-set, parallel canyons
found at such places as in the Pajarito Park district west of
Santa Ie.

Many erroneous notions are entertained regarding the his-
tory of the Rio Grande. Travelers unaccustomed to the novel
effects of the workings of the geologic agencies in the arid and
semi-arid regions are prone to explain everything in terms of
their native humid heaths. As a result much that is fanciful
has been indulged in. With the exception of W. J. McGee
even the trained geologists, who have traveled through the
region, seem to have entirely overlooked the two most vigorous
geologic processes of the arid regions—the eolian and the
planorosive, the latter more widely known among the dwellers
of those regions as sheet-flood erosion. Both are practically
unknown to people living in those parts of the world where
there are abundant rains. They give rise to phenomena the

472. CO. R. Keyes—Agyraded Terraces of the Rio Grande.

explanation of which present great difficulties by the other and
more familiar agencies.

There appears to be much misconception regarding the com-
position of the deposits which go to make up the agoraded
plains of the Rio Grande and other large rivers of the South-
west. The substructure is almost invariably spoken of as
gravel. This seems hardly correct. To be sure, the entire
surface of the plains might easily be mistaken for exposures of
gravel-beds. Closer inspection clearly shows that by far the
oveater part of such deposits is made up of clay, called by the
the Mexicans adobe. In this adobe there are commonly a few
small pebbles. For the most part the most gravelly surface
when turned by the plow gives a loamy soil, : such as is pos-
sessed by the prairies ot the Mississippi valley, except the
color. There are some gravel-beds, but on careful examination
they are found to be very limited and to represent gravel-
trains of former arroyo-courses. They are sharply defined
gravel streams traversing the old fans.

Over most of the arid regions the winds blow away the fine
soil at the surface until there is left a layer of closely set
pebbles which act as a protection to the further action of this
kind. The fact that the plains deposits are largely adobe
instead of coarser materials, as one would naturally expect
from surface appearances and from long experience in humid
regions, is at first very surprising. That the deposits in ques-
tion are composed chiefly of fine clays is due mainly to the
effects of planorosion, or flood-sheet erosion as it is termed by
McGee. The deposits of this vigorous geologic process are
principally clays rather than sands and gravels, This is probably
an explanation for most of the extensive so-called Ter tiary and
Quaternary “lake-deposits” of many portions of the West.
Planorosion is peculiar to arid regions. It’ is neither fluviatile,
nor lacustrine, nor maritime in nature, but sheet-flood erosion
in its strictest sense, modified by eolian influences during dry
times. .

J. M. Ordway— Waterglass. t

Arr. XLIII.— Waterglass ; Part V1; by Joun M. Orpway.
[Continued from vol. xl, p. 190, 1865. ]

Tue urgent duties of a busy life have interrupted this inves-
tigation for many years, but latterly changed circumstances
have allowed the work to be resuined, and now something may
be added to what has already been set forth.

It has been shown that alcohol throws down the silicates of
potassium and sodium from their solutions, retaining at the
same time a portion of their base. Acetone and methylic alco-
hol are found to answer equally well as precipitants. Some-
times neutral saline solutions may be advantageously used for
the same purpose. And most of such deposits, as well as pre-
cipitated silicates of the other metals and hydrated silicic acid
itself, can be nearly freed from the mother liquor by folding
the drained curd in closely woven cotton cloth and subjecting
to strong pressure applied gradually. The cloth can be peeled
off from the cake with very little loss.

Of course the water in the cake is partly water of hydration
and partly absorbed mother liquor; and how much there is
of each cannot always be determined. But when a chloride,
incapable of forming a basic salt, has been used in the precipi-
tation, the amount of retained ‘mother liquor may be ascer-
tained by comparing the percentage of chlorine in the pr essed
curd with that in the remaining liquid.

Experiments 3a, 6 given below will serve to show what dif-
ferences there may be between the gross analyses and the net
composition obtained by deducting the small quantities of solids
due to retained mother liquor. But the differences are slight,
and for our present purpose perfect exactness has been deemed
unnecessary. Therefore, except when otherwise specified, the
gross analysis is given ; ‘and for the sake of convenience and
brevity, the results are expressed approximately in empirical,
dualistic formulas, the water of combination being left out of
account.

Silicates of Lithiwm.—These are sometimes spoken of as
being insoluble in water. But this really applies only to the
anhydrous compounds such as are formed by fusing the car
bonate with silica. In fact when we mix weak solutions of
sodium waterglass and lithium chloride there is no precipita-
tion. Hence either no exchange of bases takes place or the
lithium waterglass remains dissolved. But when strong solu-
tions are put together an abundant deposit is formed, the nature
of which is shown by the following experiments :

1. 13 grams of a 31 p. c. sbbatoul of lithium chloride added to
100 cubic centimeters of a solution containing 10 g. of Na,O-3Si0,

474 J. M. Ordway— Waterglass.

gave no precipitate, but alcohol threw down a curd which when
hard pressed weighed 21 g. and contained 65 p. c. of 5(Li,O°3Si0,)
+6(Na,O-3Si0,). This dissolved readily in a little more than
twice its weight of water.

2. 40 g. of the 31 p. ¢. lithium chloride solution with 100°
of the same sodium trisilicate gave 17 g. of cake showing 55 p. ¢.
of 3(Li,0°3810,) + Na,O'3810,, and soluble in 1°6 times its weight
of water.

a. 86 ¢. of a 20 p. ¢. lithium chloride solution were put with

5 of a 20 p. c. solution of Na,O-48i0,. The cake weighed
48g. It contained gross 56 p. ¢. of 7 (Li, O-48i0. »)+6(Na,O° 48i0 By:
Net 54°4 p. c. of 7(Li, 0'5S10,) + 6(Na, ,0-4Si0,).

3,6. The cake of a was sore in twice its weight of water.
133 g. of this liquid with 75 g. of the lithium chloride solution
gave 46 g. of cake canuainine gross 51°7 p. c. of 7(Li,0-4810,)
+3(Na, O-3810,) ; net 50°7 p.c. of 9(Li,O-48i0,) + 5(Na,O-5Si0,).

The cake of 6 treated with 102 g. of water left a very
Teste, sediment; 120 g. of the filtered solution with 60 g. of a 25
p. ¢ lithium chloride solution gave 49 o. of cake not entirely
soluble in 150 g. of water. The dissolved portion contained net
42 p. c. of 2(Li, O:4810,) + (Na,O-48i0.,).

4, 110 g. of a 12 p.c¢. solution of Na,O-3Si0, with 40 g. of a
31 p. ¢. lithium chloride solution yielded a curd containing 54
p. ¢ of 3(Li,0°38i0,) +Na,O°38i0,. This dissolved readily in
twice its weight of water.

5. 50g. of a 20 p. ¢. solution o K,O:2810, with 30 g. of an 18
pac solution of LiCl gave 7 g. of ‘cake containing 40 Pp. Cs OF
Li,O-28i0, + K,0-28i0,..

6. 30 g. of a 20 p.c. solution of K,O'3Si0, with 30 g. of an
18 p. c. solution of LiCl gave 10 g. of cake mostly soluble in
4 parts of water. The clear solution contained 7 p. c. of
3(Li,O-3Si0,) + 2(K,O-2S8i0.,).

So we may make an endless variety of mixed or double sili-
cates of lithium and sodium or lithium and potassium. Start-
ing with a given potassium or sodium waterglass, the relative
proportion of lithium may be gradually increased by dissolving
the successive cakes and reprecipitating with lithium chloride.
And until we get beyond a tetrasilicate the deposit is com-
pletely soluble in a small amount of water. Were the precip-
itations continued, it would be necessary to add hthium
hydroxide to the later solutions to keep them from being too
silicious. But an attempt to eliminate the sodium entirely
would probably develop into an infinite series of trials.

A saturated solution of the lithium hydroxide contains but -
a small percentage of the alkali and to concentrate a siliceous
liquor which has been much diluted with such a solution we
cannot resort to evaporation by heat. For it has been found
that when any sodium lithium waterglass is heated a precipi-

b)

J. M. Ordway— Waterglass. 475

tate is formed which, if the heating is not long continued, is
mostly or wholly redissolved on cooling. This precipitation
must be due to the lithium constituent, since sodium water-
glass alone is not so affected. Could the deposit then be a
simple lithium silicate ?

In making quantitative trials with respect to this matter it is,
of course, necessary to keep up the heat till the curd is
removed from the solvent action of the mother liquor. There-
fore the dish holding the liquid, the stirring stick, the press
cloth and frame, the draining dish and the press boards were
heated together in an oven kept at the desired temperature.
While still in the hot space the curd was transferred to the
cloth and drained and the folded cloth was placed between the
boards. These were then quickly withdrawn from the oven
and subjected to strong pressure.

7. 37 g. of a solution containing 19 p. c. of 2(Li,0°3Si0,) +
Na,O- 28i0, were heated to 80° C. The cake was transparent
and weighed a little over one gram and had 49 p. ec. of
Li,O-38i0, + Na,O-3Si0,, It dissolved readily in cold water.

8. 85 g. of the same solution heated to 90° C. yielded 6°5 g
of transparent cake containing 58 p. ec. of 5(Li,O-3SiO,) +
2(Na,O°38i0,). This dissolved completely in three times its
weight of water. 24 p.c. of the solid matter had been thrown
down.

So this lithium constituent imparts its thermic character to
its associate and the two refuse to part company.

It seemed barely possible that mixed monosilicates with the
lithium element largely predominant might be made to yield a
pure lithium silicate by fractional crystallization.

9a. 56 g. of a 13 p. c. solution of Na,O-4S8i0, with 60 g. of a 7
p. ¢ solution of LiHO produced a coagulum which soon redis-
solved. A dish holding the clear solution was put under a bell
glass with two dishes of calcium chloride in lumps. In four
weeks, the temperature ranging sometimes as high as 40° C., the
weight was reduced to 21 g. In this residue were some bits of
an apparently crystalline crust amounting to nearly two grams.
These were perfectly soluble and consisted of 7(Li,O'SiO, 4H ,O) +
Na,O'SiO,-4H,O. The rest was a somewhat pasty mass of opaque
amorphous matter.

From what has been detailed so far it may fairly be inferred
that the clean dissociation of double silicates cannot be brought
about by any mode of treatment.

After the discovery of the peculiar effect of heat on lithium-
bearing silicates, the way was clear to attempt the formation of
a simple hydrated lithium: silicate by direct synthesis. The
union must be effected without the use of heat.

476 J. M. Ordway— Waterglass.

In preparing silicic acid hydrate by mixing weak solutions
of sodium waterglass and ammonium chloride, it was found
advantageous to add some methyl alcohol to aid in the disinte-
gration ‘of the coagulum. The precipitate was hard pressed,
crushed and washed with water and alcohol. After a second
washing the pressed cake was dried in the air. The hydrate
seems to have no fixed and definite composition, but after
becoming apparently dry it may continue to part with water,
like an eftlorescing ‘salt. Thus when some which appeared at
first to be a sexhydrate was kept in dry air three weeks it was
reduced to 28i0,+H,0.

Lithium sulphate being decomposed by barium hydroxide
the very dilute solution was boiled down in a nickel dish til! it
let fall some of the oxide. This saturated solution contaimed
about seven per cent of LiH-O. When it was digested, at the
ordinary temperature of the air, with enough of the silica to
form a monosilicate, very nearly all of the silica was taken up
and the filtered solution showed 8°5 p. ¢. of Li,SiO, besides a
slight excess of LiHO.

9b. 128 g. of this clear liquid, in a glass dish which would allow
the slightest change to be seen, were confined over calcium chlor-
ide, the temperature at no time exceeding 22° C. In five days a
faint deposit began to appear and the loss amounted to 32 g.
The point of saturation being thus reached, it appears that lith-
ium monosilicate is perfectly soluble in water and the solution
may contain as much as eleven per cent of Li,SiQ,.

10. 30° of dilute chlorhydric acid (3°65 g. of HCl in 100°)
mixed with 50 g. of a5 p. c. solution of Li,SiO, caused no appa-
rent change. The addition of 40° of alcohol produced a
soluble GELS weighing 4°7 g. and containing 35 p. c. of Li,O-2Si0,.

11. 304 9. of a solution containing 6 p. ¢. of Li,SiO, and 08
p. ¢. of Li, “0 treated with 149 g. of dilute nitric acid (5: 4 p.
N,O,) gave with 370° of methy! alcohol 36 g. of cake eee
soluble in less than three parts of water. It showed 36° Pach OF
10Li,0-23Si0,. The residue consisted chiefly of 0°5 g. of SiO,.

Some of the cake solution being evaporated began to show
a deposit when it contained 24 p. ¢. of the silicate. The evap-
oration being continued till the percentage amounted to 30 it
became a very stiff, opaque, amorphous mass. This dissolved
completely in its own weight of water. This latter solution
with half its bulk of alcohol gave a deposit holding 39 p. c. of
10Li,O-27S8i0, and soluble in twice its weight of water.

12. 105 g. of a solution containing 7 p. ¢. of Li »10, and 0-2
p. ¢. of Li,O with 43 g. of dilute nitric acid (10 p. ¢. N ,O,) and
i) GeO MS i LiCl yielded 9 g. of cake not entirely soluble.

The soluble part amounted to 21 p. ¢. of MOLTO SiO;
The mother liquor with 100° of methyl alcohol gave 7-9 g.

J. M. Ordway— Waterglass. ATT

of cake almost entirely soluble and containing 38 p. c. of
10Li,0°238i0,,.

13. 50 g. of a7 p. e. Li,SiO, solution heated almost to the boil-
ing point g ‘eave an abundant precipitate of which about five-sixths

redissolved on standing two days.
14. 47 g. of a liquor holding 21 p. ¢. of 10Li1,0°23Si0, heated

to 90° Gs made a hard mass that could not be got out of the
beaker, On cooling it liquefied again and the next day showed
very little sediment.

Rubidium Silicates.

From the chemical similarity of the light metals we might
suppose that rubidium would afford a series of silicates like
those of potassium. But positive proof is better than analogi-
cal conjecture. And verification was possible because happily
the lepidolite that furnished material for the lthium experi-
ments contained rubidium also. Some rubidium hydroxide
was made by decomposing the sulphate with baryta, and some
by the better method of burning the bitartrate and boiling the
resulting carbonate with lime. A clear solution was evapor-
ated in a nickel crucible till it contained 21 p. c. of Rb,O.
Nearly enough silicic acid hydrate being added to make a
monosilicate, “the boiling down was continued till the Rb, SiO,
amounted ‘0 Sap uc: The thick liquid on standing showed no
disposition to crystallize. Diluted and heated with more silica
it took up sufficient to form a bisilicate. Precipitation with
alcohol gave more silicious compounds.

15. 100 g. of a solution containing 16°5 p. c. of 3Rb,O°10Si0
with 35°° a alcohol gave a cake weighing 16 g. and contain-
ing 71 p.c. of Rb,O- 48i0,,. Thjs dissolved in twice its weight
on water.

The mother liquor with 40° of alcohol let fall a curd
weighing 2°5 g. which showed 71 p. c. of 2Rb,O-7Si0,,.

16. 160 g. of 21 p. ¢. 4Rb,0°7810, with 150°° of alcohol
gave a clear liquid precipitate weighing 43 g. It contained 45
p. ¢. of 10Rb,0-238i0,,

17. 92 g. of liquor containing 15 p. ¢. of 10Rb,0°29S810, with
22°° of alcohol produced 22 g. of cake not entirely soluble. The
soluble part amounted to 58 p. c. of 10Rb,0°468i0,. The resi-
due consisted chiefly of 1:5 g. of SiO,.

18. 45 g. of 15 p.c. 4Rb,0-98i0, with 30 g. of 11 p. ¢. LiCl
let fall 4 g. of cake, not completely soluble, The soluble portion
amounted to 29 p. ¢. of 17(Rb,O°2S8i0,) + 14Li1,0°Si0,,.

Action of Ammonia on Waterglass.

When caustic soda is poured into a strong waterglass solu-
tion it produces a momentary cloud ; lithium: hydroxide gives

478 J. M. Ordway— Waterglass.

a temporary precipitate; but ammonia has a more lasting
effect.

The ammonia water used in the following trials was of
sp. gr. 0°900.

19. 40°° of ammonia water mixed with 50 @. of a 29 p. «.
solution of 3Na,O°11SiO, gave a curd weighing 16°8 g. and show-
ing 43 p. c. of 5Na,0717Si0,,.

20. 50 g. of a 21 p. ¢. solution of 5Na,0-19SiO, with 40°
a ammonia water threw down a cake containing 40 Pp. CANOE

Na,O-4810.,.

21. 20 g. of a 31 p.¢. solution of 10K,0:33Si0, gave 6°2 g. of
cake containing 54 p.¢. of 10K,0:388i0,.

22. 40 g. of K,O-38i0, with 50° of ammonia water afforded
a 3g. cake containing 55 ?. c. of K,O-48i0,,.

23. 50 g. of a 9 p. ©. solution of Li SiO, with 30° of ammo-
nia water made a 4°3 g. cake with 44 p. ¢. of 3Li, ,0'4810..

24. 50 g. of a 33 yp. c. solution of 10Rb,0° s78i0, with 30°
of ammonia water produced a cake weighing 10 g. and contain-
ing 23 p. c. of 1ORb,O'448i0%.

25, 15 g. of a 45 Denke: Rb,O-9Si0, and 15° of ammonia
water gave 9°7 g. of thin liquid precipitate. Evaporated to 7:2 g.
it became very thick and almost solid. It then contained 70
p. ¢. of 5Rb,0-128i0,,.

In summing up results it appears that while the solubility of
the silicates of sodium, potassium and rubidium is unlimited
the lithium silicates have each its definite point of saturation,
and the monosilicate is less soluble than the more silicious ones.

Simple lithium silicate solutions give precipitates when
heated which are mostly redissolved on Y cooling.

Hence to concentrate weak solutions we must resort to
spontaneous evaporation in dr¥ air or in a vacuum.

So, too, we cannot expect to obtain the hydrates by boiling
the anhydrous silicates in water.

Combinations of lithium silicates with those of the other
alkaline metals when precipitated by heat or by other means
are not dissociated, though the composition is changed some-
what.

The composition of soluble rubidium silicates, like that of
potassium waterglass, may vary infinitely between the propor-
tions of one equivalent of silicate, one equivalent of rubidium
oxide and nine of silicic acid to two of base.

With lithium silicates ihe range is not so great.

Ammonia may give precipitates with strong solutions of all
the more silicious of all silicates.

Sometimes when hthium-bearing Precipitates are dissolved
there is a residue which makes much show but really weighs
very little.

Phelps, Weed, and Housum—Action of Dry Ammonia. 479

Arr. XLIV.— The Action of Dry Ammonia upon Ethyl
Oxalate; by I. K.3 Puerpes, L. H. Weep, and C. R.

Hovsvum.

[Contributions from the Kent Chemical Laboratory of Yale Univ.—clxvii. |

Liepie* states that dry ammonia acts upon ethyl oxalate only
with difficulty to form ethyl oxamate and a little oxamide ;
and further that the greater part of the ethyl oxalate remains
undecomposed. Within our knowledge this is the only state-
ment in the literature concerning the action of dry ammonia
upon anhydrous ethyl oxalate. This paper records the results of
experiments made to show the action both in the cold and at a
higher temperature of dry ammonia upon dry ethyl oxalate.

For this work ethyl oxalate was purified by repeated frac-
tional distillations until a product was obtained that boiled
within the limits of two-tenths of a degree. For each experi-
ment portions of this sample were freshly distilled into a dry
flask. Dry ammonia was obtained by boiling concentrated
ammonium hydroxide ina flask connected with a return con-
denser and leading the ammonia gas obtained in this way
through a lime tower.

To show what effect ammonia might have on ethyl oxalate
in absence of water, a qualitative fest was made by passing
ammonia, prepared as stated above, through a glass tube, into
pure ethyl oxalate chilled in a bath of ice and salt. A white
precipitate appeared after some minutes in the liquid still
chilled in the mixture, but wherever the interior of the tube
had been wet with the ethyl oxalate a white precipitate
appeared almost immediately. This precipitate proved to be
a mixture of ethyl oxamate and oxamide. The crude material
obtained was freed from ethyl oxalate and the large mass of
oxamide by crystallization from alcohol, and from the remain-
ing traces of oxamide by distillation under diminished pressure.
The ethyl oxamate separated by distillation in this way was
recognized as pure by its melting point—114:5°. It boiled at

~C. under 15™™ pressure. It was found that the ethyl oxam-
ate may be quantitatively distilled 7m vacuo as is shown by the
following experiment. A portion of the ester weighing 7°6496
grm. was introduced into a Claissen flask of 100°™° capacity
connected through a rubber stopper for a vacuum distillation
with a 100% flask used as a receiver. The side tube of the
Claissen flask was short—5‘5°"—and was connected with
another tube of 6™° bore of sufficient length to reach well into
the bulb of the receiver, the rubber stopper closing the receiver
serving also as the connector between the sideneck of the
Claissen flask and its extension.

* Ann. Chem., ix, 129.

480 . Phelps, Weed, and Housum—Action of Dry

TABLE I.
Ethyl

Unchanged oxalate Loss of

Ethyl Ethyl Oxamic Oxamide accounted Reaction Ethyl
oxalate oxalate ester for time oxalate
No. grm. erm. erm. erm erm, min. grm.
(1) 22°319 0545 12°714 3:46] 22°144 20 0°175
(2) 13°667 3°243 SIs). Wei) 13°580 27 0.087
(3) 9°627 9°240 4°375 2°300 9°510 25 OSTA 7/
(4) 10:944 0°828 4408 2°713 10°825 23 OPIN)

After the flasks had been evacuated to a pressure of 15™™, the
mass of the ester was distilled by plunging the flask into a bath
of acid potassium sulphate heated to 200°. By flaming the
Claissen flask and increasing slightly the current of air through
the tlasks, the remaining traces of ester are driven either into
the flask used as a receiver or into the extension tube leading
into it. The increase in weight of this flask with its rubber
stopper and the extension tube was taken as the weight of the
ethyl oxamate distilled. This amounted to 7-6361 grm., show-
ing a loss of 0°0135 grm. in distillation under these conditions.
Tt was found further that by the use of this apparatus a mix-
ture of ethyl oxalate and ethyl oxamate could be separated and
each determined. The ethyl oxalate was first distilled off by
heating the Claissen flask after evacuating the flasks to a
pressure of 15™™ in a deep bath of acid potassium sulphate, the
flask being plunged into the bath as deeply as possible and the
receiver being kept cool by a stream of cold water striking it
constantly during distillation. This operation was continued
until the Claissen flask when heated for half an hour under
these conditions showed a loss in weight as small as 0°010 grm.
between successive weighings. A mixture composed of
10:7749 orm. of ethyl oxalate and 3:°3184 germ. of ethyl oxam-
ate gave by heating with the arrangement of apparatus as
described above e, W ith the bath at 115°, a distillate of 10°7151
erm., showing a loss consequently of only 0°0598 orm., of ethyl
oxalate; and by heating further with the bath at 200°, a distill-
ate of 3°3015 grm., showing here a loss of only 0:0169 orm. of
ethyl oxamate. Since oxamide undergoes no change when
heated to 200° under diminished pressure, it is possible to sepa-
rate and estimate mixtures of ethyl oxalate, ethyl oxamate, and
oxamide. Table I records the results of experiments obtained
by treating the recorded weights of the pure ethyl oxalate with
the dry ammonia gas prepared as already given. The ethyl
oxalate was weighed out in the Claissen flask of the apparatus
described above after the apparatus had been dried by leading
through it dry air for half an hour. In place of a capillary
tube usually used in a vacuum distillation, a tube not con-

Ammonia upon Ethyl Oxalate. 481

stricted in bore was used, since in the operation to follow
ammonia gas was led in ‘through this tube. The weighed
ethyl oxalate was thoroughly chilled in a mixture of ice
and salt by immersion of the Claissen flask connected with
its receiver ready for the vacuum distillation to follow, and
was kept as cool as possible with this mixture during the
period of treatment with ammonia. The flask was shaken
violently at frequent intervals to facilitate the cooling of
the mixture, but more particularly to break up the solid mass
ot precipitate and allow as complete contact as possible
between the ammonia gas and the liquid ethyl oxalate. After
the ammonia had been passed in for the time stated, the
flask was removed from the freezing mixture and the system
evacuated by the water pump. The receiver being kept
cool by a stream of cold water, the Claissen flask was heated
first with a water bath, brought slowly to the boiling point,
and finally at 115° with a bath of acid potassium sulphate to
drive off completely the ethyl oxalate, which was collected as
already described and weighed; similarly the ethyl oxamate
was distilled and weighed. The weight of the residue left in
the Claissen flask was taken as the weight of the oxamide pro-
duced. The first period of heating at 115° to drive off the
ethyl oxalate was uniformly made ‘of a half hour’s duration.
It was presumed that during this heating all aleohol together
with any excess of ammonia would be driven out. Incidentally
there must have been a slight loss of ethyl oxalate carried off
in vapor form to the pump. In subsequent heatings to com-
plete the separation of ethyl oxalate, the loss in weight of the
Claissen flask was taken to represent the weight of ethyl
oxalate separated, rather than the weight of ethyl oxalate con-
densed in the receiver. The weight was, however, taken as a
check. The loss of ethyl oxalate in vapor form was naturally
largest when perceptible amounts of ethyl oxalate were distil-
ling during the longest periods of time. In the experiments
of Table I the largest amount of this loss observed was 0:048
grm. At the higher temperatures at which the oxamate
distilled the rubber stoppers lost somewhat in weight, but no
correction was made for the error thus introduced. :

In experiment (1) of Table I the ratio between the ethyl
oxamate and the oxamide formed is 3:09:13; in (2), 3°67:1
although in (2) there is a much larger amount of unchanged
ethyl oxalate. We observed nothing which could be definitely
assigned as the cause of this fact. ~ There was a falling off in
the amount of oxamate found in relation to the amount of
oxamide in experiments (3) and (4) when compared with that
in experiments (1) and (2). The solid mass was better broken
up by shaking in the case of (8) and (4), thus indicating that

482 Phelps, Weed, and Housum—Action of Dry Ammonia.

under these conditions the production of the end produet,
oxamide, is only a question of contact between the active

materials. To further test this conclusion the experiments of
Table II were tried.
TABLE IT.
Oxamide

Ethyl Absolute Tore Reaction - ——_---—— _

oxalate alcohol ‘) time Theory Found Error
No. grm. em’, em’, min. erm. grm. grm.
(1) 8-231 20 — 20 4°965 4°968 0°003 +
(2) 6°554 20 — 20 3°953 3°956 0°003 +
(8) 6°088 = 10 20 3°639 3°638 0'001—

Dry ethyl oxalate was weighed out in the dried apparatus
and treated with the dry ammonia in the cold, in presence of
absolute alcohol in experiments (1) and (2), and in presence of
a purified low-boiling ligroin in experiment (8). The time of
action of the ammonia was twenty minutes in each case before
the distillation which was made in (1) and (3) as previously
described, while in (2) all volatile material was removed with-
out raising the temperature. No evidence of unchanged
ethyl oxalate or of any ethyl oxamate appeared in any of these
experiments, although the flask was heated at 115° for the
removal of ethyl oxalate before weighing, and at 200° for the
removal of ethyl oxamate. The residue left in the flask was
taken, as in the experiments given in Table I, as the weight of
the oxamide produced.

Hence it appears that dry ammonia in excess acts upon
anhydrous ethyl oxalate in the cold to form a mixture of ethyl
oxamate and oxamide as far as the ethyl oxalate is acted
upon; and, if with the same proportion of reagents contact
between the dry ammonia and ethyl oxalate is facilitated
by a suitable medium such as ethyl alcohol or ligroin, the final
product produced in the same time appears to be oxamide
exclusively.

C. Barus— Note on Volcanic Activity. 483

Arr. XLV.—Wote on Volcanic Activity ; by C. Barus.

1. Lasr year* I published a brief discussion on the origin
of voleanic heat containing, as I think, certain features of
sufficient interest to deserve a more definite statement. The
remarks were based on the experiments which I described some
time ago,t showing that water glasses may be made in the
laboratory which melt (aqueous fusion) even below 200° C.
In their general physical properties like density, hardness,
refraction, ete., these glasses need not differ remarkably from
ordinary igneous glasses, in spite of the content of water in the
first case. The solution takes place under a contraction of the
system (liquid water and glass) of from 20 to 30 per cent ; and
it occurs more rapidly as the temperature of the system is
higher, the observations extending from between 180° and
300° C.

That water may be kept liquid near the shores of an ocean,
~ even if it should, by some accident, penetrate as far as an iso-
therm of several hundred degrees C., will at once be granted.
Similarly other properties of this voleanie mechanism fit the
actual occurrences fairly well, as was pointed out in the articles
cited. Thus the contraction of the system is favorable to the
evolution of heat ; a region of activity comparable in depth with
the depth of the ocean may be inferred ; ete.

2. My purpose in the present note, however, is to dwell more
specifically on the phenomenon of diffusion. I have argued
that since water is more rapidly soluble in glass at 300° than at
200°, diffusion of liquid water would take place from the lower
to the higher temperatures, and that the larger vapor pressures
associated with the higher temperatures would have no appre-
ciable bearing on the phenomenon. However we may approach
the order of values of osmotic pressure, they are usually to be
estimated in thousands of atmospheres.

In the case of dilute solutions (for which there is an adequate
theory) the increase of a diffusivity D, with temperature, 0,
is sufficiently well substantiated. If we take Nernst’s expres-

2U, U,

N+,’

sion, ) increases both with the absolute temperature 6 and witha
mobility, UY, and WY, of the ions. True the case of solution of

* Science, xxiv, p. 400, 1906.
{ This Journal, xlvi, p. 110, 1891: vii, p. 1, 1899: ix, p. 161, 1900; Phil.
Mag., xlvii, pp. 104, 461, 1899.

Am. Jour. Sc1.—FourtuH SERIES, VoL. XXIV, No. 144.—DrcemBeEr, 1907.
d4

JO) z= Jefe)

484 C. Bar Uae ote on Volcanic Activity.

glass in water will not be amenable to these simplified con-
ditions at once, but there can be no doubt that the general
bearing of the effect of temperature on /) is definitely fore-
shadowed by Nernst’s s equation, evenif the case above is ‘rather
one of the solution of water in glass. If it were known that
this liquid is more soluble in hot glass than in colder glass,
under the requisite temperatures, similar conclusions as to the
diffusion of cold to hot might be drawn; but all that has been
proved by experiments is the occurrence of a larger rate of
solution for the hot glass.

Apart from theoretical considerations , however, the question
as to the distribution of a solvent in a solution, the parts of
which are kept at different temperatures, is one of considerable
interest. It is well worthy of extended experimental investiga-
tion, particularly in those cases in which a high degree of
concentration obtains. On this subject I hope to offer some
contributions myself, at another opportunity.

3. To return to the subject of vulcanism; as I take it, the
diffusion of water takes place from colder to hotter strata under
the relatively small pressure required to keep it liquid. Thus
it is conceivable in a favorable case, that the lower strata will
in the course of time be more fluid than the upper strata both
from heat and from solution, so that heat may pass from lower
to higher levels by convection. In such regions, in other words,
convection would bring an excessively high temperature
abnormally near the surface; or there would be an abnormal
local rise of the isotherms. Hence, recalling that the solution
of water in glass is accompanied by a remarkable contraction
of the system of glass and water, and is therefore in itself
favorable to the localized production of heat, while the identical
phenomenon (by reason of the diffusion of water from cold to
hot) is at the same time tending to raise the deeper and hotter
isothermal levels nearer to the earth’s surface, one may reason-
ably conclude that two phenomena both conducive to volcanic
activity are acting in concert. Furthermore they are liable to
be evoked along ‘the shore line of the ocean, and particularly
in those places where this line is unstable and undergoing
displacement. For it is here that water will most probably
have access accidentally to unusual depths.

Brown University, Providence, R. I.

C. E Munroe—Artificial Hematite Orystats. 485

Arr. XLVI.—Artificial Hematite Crystals; by Cuar.es E.

MunROok.

Iy Deacon’s process for the isolation of chlorine hydrogen
chloride is decomposed by passing it, in the state of a gas, over
a “contact” substance which practically acts as a conveyor of
oxygen from the air to the hydrogen chloride, so that the
reaction appears to result only in the oxidation of the HCl as
follows: (HCl),+0,=(H,O), +(Cl,),. The most satisfactory
“ contact ” substance for use in this art is cupric chloride, for
when heated to 400° C., it dissociates into cuprous chloride
and chlorine (CuCl,),=Cu,Cl,+Cl,. On exposing the cuprous
chioride to oxygen, cupric oxide is formed and the remainder
of the chlorine is liberated in accordance with the following
equation: Ou,Cl,+O,=(CuO),+Cl,. If now the cupric oxide
be exposed to hydrogen chloride gas it reacts with it, forming
cupric chloride and water as follows (CuO), +(HCl),= (CuCl,),+
(H,O),, so that the “contact” substance is regenerated and
the cycle of changes begins anew. ‘Theoretically it is unneces-
sary to employ an external source of heat in the process after it
is once begun, because while 32 calories are absorbed in the
dissociation of the cupric chloride 60-4 calories are evolved in
the two subsequent reactions, showing a gain of 28-4 calories
for the cycle, but in practice it is found advantageous to heat
the air and hydrogen chloride up to 400° or 500° C. before they
are brought to the contact substance.

Although there are differences in the details of the different
‘apparatus in which this manufacture is carried on commer-
cially, yet they in general consist of a salt-cake pan or retort, in
which the hydrogen chloride is liberated; a set of cooling pipes
and drying tower, in which the moisture is condensed; super-
heaters, in which the gases are heated in iron pipes, set in masonry,
to the desired temperature; the ‘‘decomposer”’ or “ digester, ”
containing the contact substance absorbed in porous material,
where the above reactions take place; and the condensing
apparatus and coke towers for drying and purifying the chlorine.
As the catalytic substance becomes inactive after a time, the
process is interrupted every two or three months and air is
passed through the heated pipes to the contact mass to regen-
erate it.

A Belgian apparatus of this general description had been
employed in the generation of chlorine at the Ammonia Com-
pany’s works in Philadelphia for about seven years when the
iron heater pipes about the digestor became burned through.
This was evidently due to the action of the heated air and sul-

486 C. EB. Munroe—Artificial Hematite Crystals.

phur dioxide on the ontside, for the inside of the pipes was not
pitted. When the pipes were removed they were found
incrusted on their interior surfaces with finely developed
and brilliant crystals of hematite, which had evidently been
formed by the action of the hydrochloric acid on the iron,
producing iron chloride, and the subsequent conversion of this
compound to the oxide by the air passed through the pipes.
These crystals were attached to the surfaces of the pipes by
their edges and many of them were grouped in rosettes like the
famous iron roses of the Alps. A noticeable fact in connection
with their occurrence was that the size of the crystals varied with
the size of the pipe in which they were produced. A six-inch
pipe, for instance, furnished crystals whose longest diameter
was about one centimeter, while a twenty-inch pipe yielded
erystals whose longest diameter was slightly over three
centimeters.

The George Washington University, Washington, D. C.

Bascom and Goldschinidt—Anhydrite Twin. 487

Arr. XLVII.—Anhydrite Twin from Aussee ; by F. Bascom

and V. GoLDSCHMIDT.

Tue subject of this paper is a twin crystal, obtained from
F. Krantz in Bonn by Mr. R. Schroeder.

Description of Crystal.—The erystal is colorless and trans-
parent. Its dimensions are 15™™:13™™:5™™. A twinning plane
divides the crystal into two nearly equal parts. Figure 1
shows both a top-view of the crystal, projected upon the plane
6 = 0x (010) of Goldschmidt’s Winkeltabelle (WT),* and also
the crystal in perspective; figure 2 is a gnomonic projection of
the same with a similar orientation.

Combination.—The forms present on the crystal are the
following:

a b CAE) ORM Ip ae
0 Oxcop econ) ze Obma lent imal satel (WT)
DOP enOm Olco- “col IS ae (projection on: })
Of these forms 7 appears only as the twinning plane and

* WT means here and in the following,—V. Goldschmidt, Krystallo-
graphische Winkeltabelle, 1897.

488 Bascom and Goldschmidt—Anhydrite Twin.

constitutes the plane of contact for the two individuals.
p = 15(151) is a new form for anhydrite. It appears on the
smaller individual with a well defined face and occupies the
place of the plane 7, which, though present on the larger
individual, is missing on the smaller.
The face gives a fair signal and the form was determined by
the following measurements :
Measured: ¢p = 41° 57’ : 16° 23’
Calculated’) 10= 412945) = lene4 2)
Not only is the agreement between the measured and ecaleu-
lated angles assuring, but the form fits perfectly into the series,
as is shown by the following discussion :
Posy 40 f Foe to)

/XG = NOs TN ale} lis Oe
a= AO; ela See Oi te Oe OL Rmco
4 (G21 c=) Ones Se eo eA ye tton

The series of is a normal one and the form p is accordingly
established.

General Discussion.—Anhy drite twins, with twinning plane
parallel to 7 (10), were first described by Hessenberg.+ These
erystals came from Berchtesgaden and are an inch in diameter.

OS Cryst.1. o ---Cryst.2.

* Senkenb. Abh. 1872, vol. viii, p. 12, pl. 1, fig. 14.

Bascom and Goldschmidt—Anhydrite Twin. 489

They are bounded only by the three pinacoids and show twin-
ning striations parallel to the twinning plane. Later Preis-
work* described some more highly developed crystals twinned
according to the same law.

The erystal in question is still richer in forms: The parallel
orientation of the faces (0 »)and 7 (10), which is characteristic
and genetic, is the decisive factor for the twinning law of
anhydrite. They are the chief planes of coincidence, and
moreover all planes of the zone br coincide. We therefore
eall this a zone of absolute coincidence. This zone is next in
importance to the dome-zones ac and ab, which are the most
richly developed zones in anhydrite. It includes the forms
b.p.jn.or. The gnomonic projection (figure 2) makes quite
apparent these relations.

It should be observed that the 4 faces of both individuals
fall into a single plane, as is also the case with several
of the twins fieured by Preiswork,+ and further that the
7 face is a plane of contact (gr enzfliche) for both individuals.
The transformation of the elements and symbols of the
Winkeltabelle (WT) to those of the projection on 6 = 0m. (B)
is made according to the formule.

. 70
pa(WT) = a (B)

1 Pp
p yD ss So ee (183
PIA WT) neg (B)

The elements and table of angles on p. 490 belong to the pro-
jection on 6 and may sometimes prove convenient for use.

The following forms, which are distinguished thus (¢) in
the table, are uncertain :

a © be p o
Orientation =i OF 02m OS 02) 03! == Hessenberg
Orientation, (B) = 20 20 20 40 30

The forms a and p are given by W. H. Miller} but desig-
nated by him as uncertain. The forms 7, w, and o are figured
by A. Schrauf§ without further description. They also cannot
be considered assured forms. All these forms should be omitted
from the list until they have received additional confirmation.

Heidelberg, July, 1907.

t Jahrbuch Min., 1905, vol. i, p. 39, pl. 3, fig. 5-8.
+ Jahrbuch Min., "1905, vol. i, pl. 3, ‘figs. Deity (Sh

+ Phil. Mag. 1874, vol. xlvii, p. 124.

+ Atlas 1871, pl. 15, fig. 4, 5

Bascom and Goldschmidt—Anhydrite Twin.

490

00 (WT).

Anhydrite Elements (B) Projection on b

are 0008-0
So 0008-0
Boe 082-0
ll || GOOG: T
QS ”
fo a) ”
= aout
aoe) ”
eal tS)
II \| ”
oS ”

64
Sos
DS)
Il T ”
ao ”
ep ”

?
Ho
pcien ”
aan 2
ai 18e8-0
|
oo 9667-0
3 Z
ener = 8608:0
sical 6666-0
DD sH 0646-1
oS =
aS S86F-T
Se V866-
ll Il oO
a)
op ON 2)
mr D
ae (J33)
ee p
else 7]
do
ll |]
8 oO

6866-0
6&LE-0

-L6GG-0
COLL T

3)

??
?
)

”)
?
?

a)
Le)

9)

”?

?

8661-0
1886.0

966F-0
6666-0
OLCE-0

GOFF: 0
ccEc-0
F699-0

{668-0
9CTT- 1
OO6E- TE

L8&6-T
CL8P-T
C909-T

OS8L- 1
GIEG:6
CLLO-G

OOLG-§
CoOr-P
I&&6-0

9669-0
C66¢-0
6666-0
O6PG:T
886P- 1
F866: 1

”?

0

ea)
(Ae)

(stust1g)
n&

?

os)
?
)

”
9

00
00

eG
G

0
06

00 .0

(3)
i

60 IT
06 LT
& &@
68 &E&
68 61

60 VG
Ol 86
SP &

cy IP
LQ) 8F
“4G 6P
‘FL &¢
GQ 9G
90 8&

‘Tr 09
“TG 99
Ig 69
Io FL
66 LL
96 8I
6& 96
“99 O&
‘8G TP
61 Tg
LT 9G
Ge &9

”

00 0

00 206

(d—(6)
S

un

Le ot
86 06
‘PL 66
‘ET 8h

3

”?)

”

9
pe)

”

%)

7
00 0
00 06
00 0
00 .06

(9)
you

81
Me
ee

86

)

2)
2
9

2?
9
)

pe)
”
”?

”
?

?

00
96
&&
“9G
“8g
6T
LI
G6
00
00

IT
8I
96
VP

06
gI
98
0g
tP
Ig
g¢
89
06
0

00 .06

(°4— 06)
Par

or OT
VE 96

6 9&
61 9G

00
96
&&
9G
8G
61
iT
C6

00

2?

2)
”
)

9
”

?

?
)
”?

)
”?
”?)

9?

06
8I
96
0S
bv

Tg
9¢
69

06

00 0
00 006

(°4— (6)

v7]

(q) SHIONV JO WIG

d

>
”

1)

cy IP
68 61

60
OL 8@
8h &&

‘cv IP
‘LO 89
LG 67

‘PL &¢
‘G0 9G
90 8g
‘bP 09
-T¢ G9
& 69
Io bh
CG Ld
)

”?
2)

7

”

00 06
00 0

00 .06

(°3—06)
é

7]

CTT ¢ a 6%
SIT 4 f 8
aa ¢ wu LS
et 0 96
ogg =x y cB
OGL gx v ¥
Ogg= = a2 A ee
Ore 72 o ra
OL a m7 1@
Org we 1 06
08h 2% a 61
088 OE b ST
06G Hy B LT
O88 Oy a gt
01 2g n CT
03G w~§ a iat
ole og a eT
Olp FP 7 ra
O1¢ og 7) IL
1) oi OF
Col | OF di 6
COS 08 Hi 8
10T Or s L
FOS OF Ti 9
BOE 0% Di g
10608 ° 7
010 20 0 g
100 0 q 3
0OL 02 » if

ITIL “quikg 1044077 “ON

C. Palache—Oceurrence of Olivine. 491

Arr. XLVIII.—Occurrence of Olivine in the Serpentine of
Ohester and Middlefield, Mass.; by Cuaries Panacne.

Some years since, while examining the private mineral collec-
tion of Mr. E. L. Cowles of Chester, the writer’s attention was
drawn to certain specimens which were identified as olivine,
a mineral which had long been sought for in the region but
without success. Mr. Cowles was so kind as to supply speci-
mens for study and, on a later excursion to Chester, conducted
the writer and several students to the locality where he had
found the olivine, giving us opportunity to collect abundant
material and to see the nature of the occurrence. At the same
time specimens of the serpentine, which occurs in a large mass
at the locality, were obtained and olivine was found in the rock
in subsequent study of thin sections.

As considerable interest attaches to these occurrences of
olivine, publication of the observed facts seems desirable.

Both occurrences of olivine are in a lense-shaped mass of ser-
pentine, about a mile and a half long and nearly a half mile
wide, that extends from the town of Middlefield into the town
of Chester. According to Professor Emerson* this serpentine
contains chromite locally, and also supplied the specimens of
serpentine pseudomorphs after olivine known as “hampshirite,”
to which reference will again be made in these pages. But not-
withstanding these suggestions of the derivation of the serpen-
tine from a peridotite, Emerson was unable to definitely deter-
mine olivine in any of his slides, which were made from the
western half of the bed, and came to the conclusion that the
serpentine was in large part at least derived from the associated
amphibolite and not from olivine. Professor Emerson found
much olivine in the continuation of this bed to the south and
much coarse enstatite rock.

In thin sections of a massive dark green serpentine collected
near the eastern boundary of this serpentine area where the
Chester-Middlefield road crosses it, olivine was found in abun-
dance, in complete anhedraand as centers of a network of platy
serpentine developed in characteristic fashion by the alteration
of the olivine. Much of the serpentine in the slides bore marks
of the same derivation ; other smaller areas had a different charac-
ter, suggesting rather the alteration of a pyroxene, but no fresh
pre was seen. Grains of magnetite are sparsely present.

n every detail the specimen is a typical peridotite and seems
conclusive evidence to the writer of the igneous origin of this
serpentine mass.

*U.S. G. S., Monograph xxix, p. 81, and pp. 99-101.

492 C. Palache

Occurrence of Olivine.

The specimens of olivine discovered by Mr. Cowles occurred
in the railroad cut where it passes through this same Middle-
field serpentine mass, probably on the Chester side of the town
line. He noticed the mineral, which he took for apatite, in a
narrow vein on the side of the cut, exposed during the widening
of the road bed; collecting specimens at the time, he also noted
the spot beyond the cut where the rock was being dumped by
the workmen, and it was from blocks thus located, several hun-
dred yards from the place of occurrence, that we were able to
collect material. The olivine forms a narrow vein, two inches
or less in width, cutting massive serpentine like that described
above. The olivine is dark to light green in color, vitreous in
appearance and hard; parts of the vein are completely filled by
granular olivine ; other parts of the vein show the olivine in rude,
rounded crystals, up to an inch in length, embedded in a matrix
consisting of greenish white serpentine with the structure of
picrolite, densely felted white chrysotile and _ occasionally
broad plates of clear cleavable brucite. Large anhedra of mag-
netite up to an inch across occur rarely in the vein and, like the
olivine crystals, these are wrapped around by the matrix in in-
timate fashion. Although some of the olivine erystals have
undergone partial serpentinization and show glistening scales
of brucite, it is easy to find those which appear perfectly fresh
and show under the microscope the characteristic appearance
and optical characters of the mineral.

Such material, carefully selected as free from visible impurities
as possible, was analyzed in the laboratory of the U. S. Geolog-
ical Sur vey by Mr. Schaller and the analysis is published here by
permission of the Director and through the kindness of Profes-
sor Emerson, for whose studies it was ; made.

Penk esas a artes, 39°43% Note by Mr. Schaller. “Some
Re @ Nes wether = 7°83 impurity from the grinding ma-
MoO 449296 chine was accidentally introduced
Mn@ 0 =" Seen oD into the sample. The value for
CAD EE Sse weeks MONS ferrous iron represents total iron.”
ee ne aee tal none
RO eels Sen
Le AOE ASS ae 1:49
COMBe aes 2 77

100°10

The analysis shows that the material was less fresh than it
appeared. ‘But if CO, be regarded as present in form of mag-
nesite and H,O as equally divided between brucite and serpentine,
both known to be present in the sample, we have

CO. Palache—Occurrence of Olivine. 493

Mialomesitelaayss ees ae 147%

IBrUCIte ny: Mame ole een 4°34

Serpentimewy Has. nese 10°35
16°16

Deducting this 16°16 per cent alteration products from the
analysis and recaleulating to 100 per cent, we obtain the fol-
lowing figures, which give a ratio almost exactly that of oli-
vine.

Theory for
olivine with
Mol. ratio Ratio Mg:Fe=9:1
SLO Me ee 41°58% 688 1 41°00%
HeOu esr 9°33 °150 9°8
Iu lex Os Sea 48°94 1223 1°355 1°97 49°2
Mn@r a0 "14 002
100°00 100°00

In view of the discussion on a later page (p. 495) the absence
of fluorine is important.

The vein containing the olivine is sharply defined in the en-
closing serpentine and is evidently a younger formation. We
have here apparently another case of the regeneration of olivine
in a rock mass which has undergone a general serpentiniza-
tion—a process previously recorded by Weinschenk* in serpen-
tine in the Tyrol and observed by the writer} in the peridotite
of Mine Hill, Cumberland, R. I.

The curious appearance of these specimens of olivine embed-
ded in serpentine recalled the description of the above mentioned
hampshirite pseudomorphs as given by Emerson}; it seems
to the writer, and his conclusion is wholly confirmed by Profes-
sor Emerson after seeing the specimens, that the close accord-
ance in general locality of Mr. Cowles’ specimens and the orig-
inal hampshirite and the general similarity of the minerals,
save that in the single specimen of hampshirite studied by Pro-
fessor Emerson the olivine was wholly altered to serpentine and
brucite was not developed with it, point to the conclusion that
we have here a rediscovery of the long-lost locality of the pseu-
domorphs and final proof of their derivation from olivine.

On the latter point it is necessary to refer to a recent paper
by Mr. A. D. Roe and Mr. A. L Parsons,§ in which the history
and nature of these pseudomorphs is discussed.

* Beitriige zur Petrographie der dstlichen Centralalpen speciel des Gross-
venedigersstockes. Abh. Kgl. bayer. Akad. Wiss. II cl. 1894, xviii, 651.

+An occurrence soon to be deseribed by Dr. C. H. Warren in a paper on
this interesting locality.

{ Emerson, B. K., Mineralogical Lexicon, Bull. 126, U.S. G.S., pp. 92, 146.

$A Mineral Resembling Meerschaum from the Serpentine Range of Hamp-

den County, Mass., with Descriptions of Interesting Included Crystals, Bull.
Minnesota Acad. Sci. IV, No. 2, 1906, pp. 268, 276.

494 C. Palache—Occurrence of Olivine.

First discovered by Dr. E. Emmons and described by Dewey*
as crystals of steatite, they came later to be regarded as steatite
pseudomor phs after quartz. Emerson first assioned olivine as
the original mineral, basing the determination on measurements
of the crystals, and. comparison with serpentine pseudomorphs
after olivine from Snarum, of similar size and color.

In this paper Mr. Roe describes the locality and the finding
by himself of all extant specimens of hampshirite so far as
known, and this locality agrees exactly with the one from which
our material comes. ie gives analyses of the psendomorphs
and of the meerschaum-like serpentine matrix made by E. E.
Nicholson : both correspond fairly well with ordinary analyses
of serpentine although somewhat low in water. To the ma-
trix serpentine is given the name hampdenite, hampshirite being
retained for the serpentine of the pseudomorphs; both names
seem to the writer superfluous since no varietal distinction from
serpentine is established, and the name picrolite embraces
varieties of serpentine with the characters of the so-called
hampdenite. Large magnetite crystals showing dodecahedral
and octahedral planes were associated with the serpentine pseu-
domorphs.

Mr. Parsons decribes the crystals, giving contact measure-
ments and sketches of a number of them, and pointing out the
close resemblance to humite which thev_ present in form and
angles. He regards as strongly confirmatory of the derivation
of the crystals from humite the facts : (1) that minerals of the
humite group are abundant in other Massachusetts localitiest
and at Tilly Foster Mine, N. Y., in the last case in somewhat
similar paragenesis: (2) that er ystallized olivine in good-sized
crystals has never been found in the region: (3) that the size of
these pseudomorphs is altogether exceptional for olivine.

In view of the discovery of olivine crystals close at hand as
described above, quite comparable i in size with the pseudomorphs,
although not so perfect in form, the confirmatory facts given
by Parcons of course lose all weight. The agreement in
crystal measurements is, it is true, less satisfactory for olivine
than for humite, as the following table, taken from Parsons’
paper with the addition of the tiewr es for olivine, shows:

Humite Pseudomorphs Olivine
(measured)

210 to 210 49° 40’4 49°-50° I1O%ton 110" 495 oie
001 to014 45 324 464-47 KO ON Ae $33

OOM RCONOMN SG 74 OOL to 041 66 55
001 to 103 55 44 55 OO1l to 10L 51 33
001 to 216 58 16 58 O01 to lil 54 15

* See Dana, System, 1892, p. 675
+Mr Emerson informs us that the localities cited are many miles distant
from Chester, and of different geological age and association.

C. Palache—Occurrence of Oliwine. 495

The diticulty of securing accurate contact measurements on
material of this sort with more or less curved faces is, however,
so great that too much weight should not be attached to the
discrepancies shown on the olivine side of the table. And it is
further to be noted that, of the angles given by Parsons as
“measured,” only the first between the dominant prism planes
could actually have been measured on these crystals since the
basal pinacoid is absent, and this first angle agrees equally well
with humite and olivine ; the other angles must have been
derived from the actual measurements, presumably by halving
angles measured over the summit of the crystal between small
faces; such measurements are liable to much greater error than
those on larger faces at obtuse angles. On the whole then, the
crystallographic evidence alone seems too weak to establish the
derivation of the hampshirite pseudomorphs from humite.

The possibility that the mineral described above as olivine
might be humite was carefully considered, especially when the
fact was noted that the optical characters of the two minerals
are so similar that in granular form they are practically
indistinguishable under the microscope. The result of the
analysis and the proved absence of fluorine seem to settle this
point conclusively.

Harvard University, June, 1907.

496 Scientific Intelligence.

SCIENTIFIC INTELLIGENCE

I. CHEMISTRY AND PHysIcs.

1. Speculations in Regard to Atomic Weight mM umbers.—
Starting with the old assumption that the atomic weights are
obvious whole numbers and the old deduction that the elements
are multiples of the element hydrogen, H. Cottrys has made
some observations which appear to be of more interest than is
usual in the case of such speculations. He observes that the
‘“artiads ” below 63 are generally odd numbers, while the “ peris-
sads” are usually even. The antiquated terms just used refer to
odd and even valency, respectively. He states that nitrogen is
the single exception, but it seems to the reviewer that beryllium
and scandium are absolute exceptions, and that helium and argon,
haying no known valency, are not strictly ‘“perissads.” It
appears further that this application is of little importance with
Ti, V, Cr, Mn, Fe, Co, and Ni, which show both odd and even
valencies. However, there are enough cases conforming with
this rule to make the observation an interesting one. Another
observation is that these atomic weights, below 63, based on oxy-
gen as 16, are either whole numbers or greater than whole num-
bers, with the exceptions of argon, iron and nickel. The author
makes the deduction that the increase in the atomic weights
above whole numbers is due to the condensation of a “‘protyle”
within each atom, but he does not explain why the oxygen atom
may not also possess a “ protyle.” He cbserves further that in
nine instances a difference of 4 in the list of “obvious whole
numbers” (below 63) corresponds to a difference of two valencies ;
for instance,

Li Monad 7 a Na Monad 23 ) 4
B Triad 11 ce Al Triad 27 §
Meg Dyad 24 La Al Triad 27 LA
Si Tetrad 28 ee P Pentad 31 |

He observes also that several changes of a single valency involve
changes in weight of 1 and 3 ; for example,

Bie einiad: 11) 1 Na Monad 23) 1
C Tetrad 12: Mg Dyad 24

Me Dyad 24), Si Tetrad 28),
Al Triad OT Ns P Pentad Sa

He makes the deduction that the acidity or non-metallic nature of
an element is always due to a pair or pairs of electro-positive
forces, each pair emanating from a portion of the element, of
which the mass is 4, taking the mass of an atom of hydrogen as
unity. He believes that the constitution and structure of each

Chemistry and Phystes. 497

element can be deduced in a rational and uniform manner by
taking into consideration all of his observations and deductions
(only a part of which have been touched upon here), and the por-
tion of each element which is metallic is thereby made evident.
It is not clear that the author has any good reason for using
metallic elements as a basis, for it might be argued that the
opposite change—non-metallic to metallic—was brought about
by increases in mass. It is not probable that many chemists will
accept these views as possessing any significance.— Chem. News,
ENT, INAS H. L. W.

2, The Vapor-tension of Sulphur at Low Temperatures.—By
passing gases over sulphur in such a manner that saturation took
place, condensing the sulphur in a cold tube and weighing it, Dr.
H. Grurner of Adelheid College, Cleveland, O., has succeeded
in determining its vapor tension at temperatures between about
50 and 120°C. At the two temperatures just mentioned the pres-
sure of the vapor was found to be -00008 and :0339™™ respectively.
At 100° C, it was found to be -007™™, and the same result was
obtained by boiling water in which finely divided sulphur was
suspended, passing the vapor through a column of solid sulphur,
condensing the water, and determining the sulphur in it. It is of
interest to know that as much as -013 g. of sulphur may be carried
off with the steam of 100 g. of water. Another interesting point
brought out in this investigation is the fact that sulphur is
scarcely oxidized at all by air below 100°, but that a trace of SO,:
is formed at that temperature. The author states that the odor
perceived when sulphur is heated to 100° is due to volatilized
sulphur and not to SO,.—Zeitschr. anorgan. Chem., \xvi, 145.

H. L. W.

3. Heliumin Natural Gas.—The announcement of the discovery
of considerable quantities of helium in the natural gas from a
well in Kansas was made some time ago by D. F. McFarland.
Copy and McFaruanp have now examined for helium some 47
samples of natural gas, most of them from the Kansas region,
but including samples from Ohio, Indiana, West Virginia, Cali-
fornia and Louisiana. It appears that the gas originally exam-
ined contains more helium, 1°84 per cent, than any other of the
numerous samples, although there are two or three other gases
that contain nearly as much. However, it is noteworthy that in
only one case was no helium found, and that in most of the
gases its quantity was over ‘10 per cent. The authors observe
that the helium content of a gas tends to increase with an increase
in nitrogen, although no direct relation between the two was
observed. The gases richest in helium contained over 82 per
cent of nitrogen. A further observation was made that in the
Kansas region the amount of helium present in the gases varied
according to the geological strata.—Jour. Amer. Chem. Soc.,
290g. WE AB} 8 iby A

498 Scientific Intelligence.

4. Electro-Analysis ; by Epgar F. Suir. 12 mo, pp. 336.
Philadelphia; 1907 (D. ‘Blakiston’s Son & Co. )—This is the fourth
edition, revised and enlarged, with 42 illustrations, of a well
known, valuable text-book. The changes in the new edition are
far more important than is usual in new revisions, for here is in-
cluded an account of recent developments in the subject from
the author’s laboratory, which are of the greatest importance and
interest to analytical chemists. Electrolytic analysis heretofore
has dealt chiefly with the determination of a few of the heavy
metals, but the range of its applications has been gradually en-
larged by the discovery of new applications. An instance of this
kind is the determination of the halogens by collecting them upon

a silver-plated anode. Dr. Smith’s new work has greatly im-
foie and extended this application of electrolysis, and he has
shown that rapid and accurate determinations of not only the
halogens, but of such anions as the carbonate, ferrocyanide, ferri-
cyanide, phosphate and sulphide radicals, can be made readily.
These remarkable results have been made possible by an ingenious
device whereby the alkali metals, as well as barium and stron-
tium, are passed through a layer of mercury into an outer com-
partment of the electroly tic cell. This same device has enabled
the author to make various separations of metals in an exceed-
ingly simple and accurate manner; for example, an alkali metal,
or barium or strontium, from calcium, magnesium, etc. These
new methods devised by Professor Smith will certainly find
extensive practical applications, and they mark an important
epoch in electro-analysis. H. L. W.

5. Kizeaws Research on the Change of the Azimuth of
Polarization due to Movement of the Karth.—TYhe endeavors to
discover an effect on optical phenomena due to movements of
the earth have led to negative results. Fizeau, however, be-
lieved that he had noticed such an effect. He passed a polarized
light ray through a series of inclined glass plates and noticed a
change in the polarizing angle, as the ray passed in the direction
of the earth’s movement or in the opposite. Fizeau found a dif-
ference in the polarizing angle of 0°024°. Bruce repeated
Fizeau’s experiments with a different order of apparatus and
found an angle change of only 0:008°, which is within the limits
of error, and Bruce concludes that Fizeau’s result must have
been due to some other cause than that of the movement of the
earth. 5B. SrrassEer’s has taken up the subject and shows that
Fizeau’s use of a reflecting mirror to change the direction of the
ray of light was faulty. It is important that the source of light
should not be changed, but should move with the apparatus.
Strasser gives a diagram of his apparatus which shows how this
result is accomplished, .and his tabulated results show that the
Fizeau effect does not exist, and that no experiment shows any
effect upon optical phenomena due to movement of the earth.—
Ann. der Physik, No. 11, 1907, pp. 137-144. ap tt

Chemistry and Physies. 499

6. Secondary Cathode Rays Emitted by Substances Haposed
to y-Rays.—R. D. Kuxeman concludes that the y-rays from
radium consist principally of two groups of rays, the constituent
rays of each group differing not much from one ‘another in their
properties. The rays of one of the groups are more efficient in
producing secondary cathode radiation from aluminium, sulphur,
iron, nickel, zinc, and tin, than from lead, and are all more or
less easily absorbed by each of these substances excepting lead.
the absorption by lead being much less. The rays of the other
group are more efficient in producing secondary cathode radiation
from lead than from the other substances, and are more easily
absorbed by lead, mercury, and bismuth, than by any of the
other substances. There is also a third, apparently weak group
of rays which is most efficient in producing secondary radiation
from carbon. This group of rays is less easily absorbed by the
above mentioned substances than either of the other groups.—
Phil. Mag., Nov. 1907, pp. 618-644. Jeu

7. Secondary Rontgen Radiators from Gases and Vapors.—
Careful investigation of this subject was made by BarKkua
(Phil. Mag. v, 1903, vii, 1904), who concluded that :

(1) All gases, when subjécted to X-rays, are a source of sec-
ondary radiation.

(2) The absorbability of the secondary radiation is, within the
limits of experimental error, the same as that of the primary pro-
ducing it.

(3) For a given primary radiation, the intensity of the second-
ary radiation is proportional to the density of the gas from which
it proceeds.

(4) The ratio of the intensities of the primary and secondary
beam is independent of the hardness and intensity of the primary
rays.

eecla: experiments were performed on few gases and not of
a varied type. Mr. J. A. CRowrneEer has repeated the experi- °
ments with a large number of gases—of varied atomic weights—
and confirms Barkla’s results.— Pid, Mag., Nov. ee ae
653-675.

8. Abrupt Limit of Distance in the Power of the Peta ie
to Produce Phosphorescence.—It has been found by the various
observers that the a-rays from polonium and radium lose their
power abruptly of producing phosphorescence and of affecting
the photographic plate. Dr. Jacos Kunz, reflecting that the
positive rays are similar in nature to the a-rays of the radio-
active elements, was led to believe that the positive rays would
also show an abrupt falling off in a manner similar to the a-rays.
The tube used for the production of the canal or positive rays
was enclosed in a larger tube, exhausted to the same degree :
and the inner tube by means of spiral springs connected to the
electrodes could be moved to varying distances from a_willemite-

Am. Jour. Sc1.—FourtH Series, Vou. XXIV, No. 144.—DeEcrempBer, 1907.
3d

500 Scientific Intelligence.

screen. It was found that there was, as expected, an abrupt
limit of distance in the power of the positive rays to produce
phosphorescence.— Phil. Mag., Nov. 1907, pp. 614-617. 5. 7.

9. The Vacuum Bolometer.—It has often been observed that
a bolometer strip placed in a vacuum is more sensitive than in air.
EK. Warsure, G. Leiraiuser and Ep. JoHANSEN have investi-
gated this phenomenon and give a tabulated résumé of their
results. The conditions of sensitiveness vary with the strength
of currents employed and the breadth of the bolometer strips ; in
general the sensitiveness in a vacuum is from three to four times
that in air.— Ann. der Physik, No. 11, 1907, pp. 25-42. Sait

10. Ratio of the Electrical Units. ~The paper by E. B. Rosa
and N. E. DorsEy on a new determination of the ratio of the
electromagnetic to the electrostatic unit of electricity, alluded to
in the last number (p. 442), is completed in the current issue of
the Bulletin of the Bureau of Standards (pp. 541-604). The
final value obtained reduced to vacuo (assuming the dielectric
constant of air at 20°C and 760™™ as 1:00055) is

Up = 29 Oral allO ns

Accepting the velocity of light as 2°9986 10", this value of »,
shows a difference of 5 parts in 10,000 with a possible uncertainty
of 2 parts in 10,000. The explanation of the resulting difference
(1 in 3000) is as yet uncertain. A supplementary paper by the
same authors (pp. 605-622) gives a critical comparison of the
various methods of determining the above ratio.

IJ. GEoLoGy AND MINERALOGY.

1. The Geology of North Central Wisconsin ; by SAMUEL W EIb-
MAN. Wisconsin Geol. and Nat. Hist. Surv., Bull. xvi. Madison,
1907, Svo, pp. 697, maps in cover.—The area whose geology is
described in this memoir contains about 7200 square mies, about
one-eighth of the state, and is situated as described in the title.
The city of Wausau (15,000 pop.) near the center is the largest
place in the district. It is without especially characteristic topo-
graphic features and is chiefly an agricultural country.

The geological problems of the district are mainly those relating
to the pre-Cambrian rocks, which consist to a relatively small ex-
tent of metamorphosed sedimentaries and very largely of igneous
intrusives, and those relating to the latest deposits, which are Pleis-
tocene, or glacial. One-half of the volume is, therefore, devoted
to the working out of the petrographical problems presented, the
other to glacial geology and physiography. in the time-interval
between these two, the Paleozoic alone is represented by the Pots-
dam sandstone. Of the igneous rocks, the most interesting are

Geology and Mineralogy. 501

those found in a series of intrusions in the neighborhood of
Wausau, which consist of syenites and nephelite-syenites. One
of the latter is noteworthy in that its pyroxene is a hedenbergite-
and it contains fayalite. As is so often the case, these rocks in
their pegmatitic developments contain many interesting minerals.
These have been carefully investigated and many of them have
been analyzed, as well as the rocks and rock-minerals. The whole
makes a very thorough and excellent piece of petrographic inves-
tigation. It isan interesting fact that the intrusion of the alkalic
rocks has no accompanying retinue of differentiated dikes and satel-
lite masses which are so common a feature in such occurrences else-
where.

Following the petrography, the matter of chief interest is the
result of the study of the glacial geology. The various characters
of the ice invasion and of the deposits it left are described in
detail for all parts of the area. The writer finds evidence of
four distinct glacial formations, each believed to have been formed
by a separate ice invasion. One part of the region is driftless
and nonglaciated, and the author, in explanation of this, adopts the
view of Chamberlain for the larger areas to the southwest, that is
the diversion of the ice currents by the highlands of northern
Wisconsin and Michigan.

The work concludes with a description of the topographic fea-
tures of the region and a discussion of its physiographic develop-
ment. ‘The volume is well printed and embellished by many fine
half-tones, and as a whole, both in the results obtained and in
the manner in which they are presented, it is an excellent work,
of general as well as of local interest, reflecting credit on the
author and on the State survey. Lavine

2. Research in China (in 3 vols, and atlas): Vol. I, Part IT;
Petrography and Zoology; by E1z10tT BLACK WELDER. 4°, pp. 357— »
528, plates 12. Washington, 1907 (published by the Carnegie
Institution). —Some years since, as is well known, the Carnegie
Institution despatched to China an expedition under the leadership
of Mr. Bailey Willis of the United States Geological Survey.
One of its chief purposes was the study there of the earliest strati-
fied rocks, in the hopes of throwing light on important questions
concerning Cambrian and pre-Cambrian geology and paleontology.
While the success of the expedition in this particular direction
was perhaps not greater than that which has attended the study
of these strata in other places, a considerable amount of material,
valuable in several branches of science, and of interesting infor-
mation concerning the regions traversed, was obtained and, under
the auspices of the Carnegie Institution, these have been studied
and collated and the results are now being published.

The present volume by Mr. Blackwelder, who was Mr. Willis’s
chief assistant, describes the petrography of the rock specimens
collected along the route of travel. They represent a large var-
iety of types, igneous, sedimentary and metamorphic, which have

502 Scientific Intelligence.

been studied under the microscope and classified. Among the
igneous rocks, granites, diorites, gabbros and several por phyries,
with effusives of rhyolite, andesite and basalt,are the r uling types,
and it is interesting to note that no rocks of pronounced alkalic
nature were encountered. The work is of a purely descriptive
character, and as no summation or generalizations of the observed
facts are given, it is inferred that these are reserved for that part
of the work dealing with the general geology of the region visited.

Of the zoological material gathered by Mr. Blackwelder, the
description of the reptiles and Dirds is given by himself with the
assistance of Drs. Stejneger and C. W. Richmond. Of the Reptilia
only a few specimens were obtained, consisting of several lizards
and one snake, of already described species. The Birds are rep-
resented, however, by a larger collection of 64 specimens contain-
ing 49 species, Most of these are naturally of small kinds, larks,
finches, sparrows, wrens, thrushes, ete. While none of the speci-
mens proved to be of an absolutely new species, an interesting
new variety of the Chi-li winter wren was obtained— Olbiorchilus
Sumigatus ideus Richmond. This part of the text is embellished
by several fine colored plates of the birds collected.

The volume concludes with a syllabary for the transcription of
Chinese sounds in the dialect of Peking, modified for literary
purposes by Dr. Friepricn Hrrru of Columbia University.

LisVee:

3. Miscellanea Paleontclogica; von Prof. Dr. ANTton Frirscu.

Paleozoica. Pp. 23, plts. 12. Published by the author, Prag,
Bohemia, 1907.—In this quarto paper the author describes and
illustrates seventeen invertebrate Paleozoic animal remains.
Among them are discussions of several American species, as fol-
lows: Proscorpia osborni, Paleocampa anthrax (gives also a
restoration of this Polechet worm; states that it is not a Myria-
pod), Propolynoé laccoei (a new genus and species of annelid
from Mazon Creek, Indiana), Hestonites bioculata (Cheetopod
from Mazon Creek), Latzelia primordialis (the oldest Chilopod
and related to the recent genus Scutigera).

There is also described a new Camerocrinus quarzitarum, ex-
tending the range of these extraordinary crinoid floats to the
Ordovician (etage Dd2). The balance of the work is devoted
to a description of the Permian Coprolites of Bohemia (26 species!).

C8:

4. EKvidences of a Coblenzian invasion in the Devonic of Hastern
America; by Joun M. Crarxe. Festschrift zum siebzigsten
Geburtstage von Adolf v. Koenen, pp. 359-368, Stuttgart, 1907.
—In this short paper the author gives his views in regard to the
relation of the Helderbergian and Oriskanian and the probable
waterways of migration of the faunas about Gaspé, Quebec, Dal-
housie, New Brunswick, and localities in Maine. It is in the
latter region that marked Coblenzian affinities are shown in the
faunas. C. S.

Geology and Mineralogy. 503

5. The Geology of Islay ; by 8. B. Witkrinson ; with notes
by J. J. H. Teatt and B. N. Peacu. Memoirs of the Geol. Surv.,
Glasgow, 1907; 8°, pp. 82, pls. 8.—This Memoir describes the
Geology of Islay and Oronsay with portions of Colonsay and Jura,
islands on the west coast of Scotland. In the introduction a brief
reference is made to the physical features ofthe islands. A short
sketch is given of the progress of geological research in Islay.
The special feature of the Memoir is the detailed account of the
metamorphic rocks which enter into the structure of Islay, the
correlation of the gneisses of the Rhinns of Islay (the western hills)
with the Lewisian gneisses of the counties of Sutherlandand Ross
in Scotland and the description of the sediments overlying the
gneisses of the Rhinns which resemble subdivisions of the Torri-
don Sandstone in the Northwest Highlands. A detailed account
is given of the rock groups in the southeast of Islay which have
been linked with Eastern Highland types. The glacial and recent
deposits are also described.

6. Geology and Wuter Resources of the Bighorn Basin,
Wyoming ; by Cassius A. Fisuer. Professional Paper No. 53
U.S. Geol. Survey.—As stated in the introduction, this paper is
designed mainly to furnish information regarding geologic
structure and the prospects for underground water. A general
account of the surface waters is given, including a statement of
their present and proposed uses for irrigation, and the economic
products of a geologic nature are also described. The region
considered comprises the Bighorn basin, a part of the Clark Fork
basin, and the slopes of the adjoining mountain ridges, the entire
area comprising 8,500 square miles. The basin is floored by
Eocene strata overlying Laramie strata of great thickness. The
surface is now dissected by streams which flow in deep but broad
sloping valleys, bordered by terraces rising to adjoining highlands.
Bad-land structure is rather common in these Eocene strata.

J.B.

7. The Geology of the Guaynopita District, Chihuahua. A
contribution to the knowledge of the structure of the Western
Sierra Madre of Mexico; by Epmunp Ottis Hovey, Pu.D.—
This paper occupies pages 78 to 95 of the volume written by the
students of Professor Rosenbusch in celebration of his seventieth
birthday. The first pages sketch the general structural features
of Mexico, pointing out that the limits of the plateau have been
determined by profound faulting, and calling attention to that
feature of the surface consisting in the “ Bolsons ” or basin deserts.
Following this introduction the details of the Guaynopita district,
lying in the heart of the Western Sierra Madre, are given. ‘The
fundamental rocks are gneiss and schist overlain by limestone, the
whole now folded. These are capped by a series of eruptives and
are cut by granite which Hovey regards as probably of inter-
mediate age in the eruptive series. ey

504 Scientific Intelligence.

8. Tertiary Mammal Horizons of North America; by Henry

F. Osporn.

Bull. Amer. Mus. Nat. Hist., xxii, 1907, pp.

237-253.—This is one of the most valuable geological summaries
of our present knowledge of North American mammal horizons.

The thickness of these Tertiary strata is over 18

major facts may be tabulated as follows:

,000 feet. The

Seventh mammal

Provisional

Fourth

Up to 1,000 feet.

phase. Increas- correlations
ing cold, moist- with other
ure, and forests. countries.
Sueur Rare bower Pleistocene or Preglacial.
Slew eal Sheridan or Hquus beds:
lan intermigra-)
tion. Much ex- |
tinction of pre-
vious native
stocks. |
Sixth mammal) Upper Phocene development as Yetlas chien
phase. Inter- not recognized in America. ;
migrations with Middle Pliocene. Blanco or Glypto- I sor
South America. therium beds. Up to 3,000 feet. 3
Fifth mammal|Lower Pliocene. Republican or|/Messinien-
phase. Inter-| Peraceras beds 100 feet. Plaisancien.
migration with|Upper Miocene. Loup Fork or|
Eurasia. In-| Protohippus beds. Up to 120)Tortonien.
crease of sum-| feet.
mers droughts.|Lower Miocene. Ticholeptus beds. Langhien-

Helvétien.

mammal
phase. First
Great Plains!
forms. Second,
or Lower Oligo-
cene intermi-)
gration with)
western Eu-
rope.

Lower Oligocene.

Arikaree

Up-

Transitional to Miocene.
or Promerycocheerus beds.
per John Day 400 feet.

Upper Oligocene. John Day or
Diceratherium beds. Mountain
fauna. Up to 1,000 feet.

Middle Oligocene. Upper Brule
or Leptauchenia beds. Lower
John Day. Up to 300 feet.

Middle Oligocene. Lower Brule
or Oreodon beds.

Titanotherium

beds. 200 feet. This and the

Lower Brule horizons have de-

cided European affinities in the

Perissodactyla.

Langhien-
A quitanien.

A quitanien.

About
Stampien.

?Ludien in
part.
Sannoisien.

Third mammal

phase. Con-

tinued physical

conditions
without new in-
termigrations.
Eocene deposits
probably main-
ly of volcanic
origin.

|
|

Geology and Mineralogy.

Transition to Oligocene and Upper}
Eocene. Uinta or Uintatherium
beds. Up to 1,750 feet. Ter-
tiary genera 87 per cent; Creta-
ceous genera 13 per cent.

Middle Eocene. |Washakie
Eobasileus beds, 500 feet.

Middle Eocene. Bridger or Oro-
hippus beds. Up to 1,800 feet.
Modernized mammal genera 81
per cent; Cretaceous genera 19

or

per cent.
Lower Eocene. Wind River or
Lambdotherium beds. Up to

1,400 feet. Modernized mammal
genera 63 per cent; Cretaceous
genera 37 per cent.

Second mammal

phase. First
decided migra-
tion of modern-

Wasatch or Cory-
Up to 2,500 feet.

Lower Eocene.
phodon beds.
Modernized mammal genera 42

505

Absolute
dissimilar-
ity between
America
and Europe.

Bartonien.

Bartonien.
Upper Lute-
tien in part.

Lower Luté-
tien in part.
Yprésien.

Yprésien.

ized forms be-|) per cent; Cretaceous genera 58 Spar nael a:
tween America| per cent.
and Europe.
Basal Eocene. 'Torrejon or Panto-
lambda beds, 300 feet. Huropean
connections seen in the similar|/Thanétien
stages of development in America or
and France. Modernized mam-|Cernaysien.
mal genera 4 per cent; Mesozoic
First mammall 2¢™er? 96 per cent. f
phase. Archeeic elinniyy
P ; F with Nosto-
or Mesozoic
fae stylops beds
: Basal Eocene. Puerco or Poly-jof Patago-
mastodon beds, 500 feet. Meso-|nia. Creta-
zoic mammal genera 100 per|ceous or Ba-
cent. sal Kocene.
| Intermigra-
tion late
Cretaceous.

Cs ast
9. Gold Nuggets from New Guinea.—Professor A. Liver-
sipGE of Sydney describes two small gold nuggets from New
Guinea, which, after being polished and etched with aqua regia,
showed near the edges a clearly marked concentric structure.
This is regarded as probably indicating successive deposition in
the walls of the cavity analogous to that observed in agates.—

Roy. Soc. NV. S. W., xl, 161.

506 Scientific Intelligence.

Ill. Miscertanerovus Screntiric INTELLIGENCE.

1. Annual Report of the Board of Regents of the Smithsonian
Institution, showing the Operations, Expenditures, and Condi-
tions of the Institution for the year ending June 30, 1906. Pp.
hi, 546, with 41 plates. Washington, 1907.—The annual vol-
ume now issued contains the report of the Acting Secretary, Mr.
Richard Rathbun, which (as separately published) was noticed in
an earlier number of this Journal (vol. xxiii, p. 242). The gen-
eral appendix (pp. 91-546) contains as usual a series of illus-
trated papers, selected for republication here as giving concise
accounts of recent scientific discovery in different fields. The
subjects included range from radio-activity and wireless teleg-
raphy to ethnography and geography. The volume closes with a
biographical notice of the late Professor Langley by Cyrus
Adler.

This report by the Acting Secretary shows that the Institution
was maintained at its full efficiency during the time following
the death of the former Secretary. The recent accession of Dr.
Charles D. Walcott, who was elected Secretary 1 in January last,
gives promise of new activity and expansion for the varied
interests involved.

Some of the prominent publications recently issued under the
auspices of the Smithsonian Institution are given in the follow-
ing list :

Twenty-fifth Annual Report of the Bureau of American Eth-
nology for 1903-04. Pp. xxix, 296, with 129 plates and 70
figures.—This volume, besides the report of the Chief of the
Bureau, Mr. W. H. Holmes, contains two memoirs by Jesse
Walter Fewkes; one of these is devoted to the Aborigines of
Porto Rico and Neighboring Islands, the other to Certain
Antiquities of Eastern Mexico.

Smithsonian Contributions to Knowledge: Part of Volume
xxxv. The Young of the Crayfishes Astacus and Cambarus ; by
K. A. AnpRews. Pp. 79, with 10 plates and 93 tigures. See
0. 449.

Butretins.—No. 50. The Birds of North and Middle America.
Part IV ; by Ropert Ripveway. Pp. vi, 973, with 34 plates.

No. 53. Part II. Catalogue of the Type and Figured Speci-
mens of Fossils, Minerals, Rocks and Ores in the Department of
Geology, United States National Museum. Prepared under the
Direction of Georer P. Merrity. Part II. Fossil Vertebrates ;
Fossil Plants ; Minerals, Rocks, and Ores. Pp. v, 370.

No.57. The Families and Genera of Bats; by GrErrir S.
Miter, Jr. Pp. v, 282, with 14 plates.

No. 58. Herpetology of Japan and Adjacent Territory; by
LEONHARD STEJNEGER. Pp. xx, 577, with 35 plates and 409
figures.

Miscellaneous Intelligence. 507

Contributions from the United States National Herbarium,
Vol. x, Pt. 5. Report on the Diatoms of the Albatross Voyages
in the Pacific Ocean, 1888-1894 ; by ALBERT Mann. [Assisted
in the bibliography and citations by P. L. Ricksr.| Pp. v,
221-419, with plates xliv—liv.

Smithsonian Miscellaneous Collections ; Quarterly Issue, Vol.
iv, Pts. 1, 2. Among the papers here published may be men-
tioned one by Merrill and Tassin (pp. 208-214) describing the
remarkable shale balls found among the Canyon Diablo meteor-
ites ; also (from vol. ii1) acatalogue of earthquakes on the Pacific
coast, 1897-1906, by A. G. McAdie.

. National Academy of Sciences.—The autumn meeting of
the National Academy was held in New York City on Nov.
19-20, President Remsen presiding. About forty members
were in attendance. The following is a list of papers presented :

M. I. Pupin: A new application of dynamics to electrical circuits.

LrercHton B. Morse: The selective reflection characteristic of carbonates ;
wave length of displacement a function of the atomic weight of the base.
Oxygen the active atom in the characteristic reflection of carbonates, nitrates,
sulphates and silicates.

A. P. Witus: A modification of the Bjerkness hydrodynamics analogy.

A. G. WexBsTER: On Rayleigh’s disc as an absolute measure of sound.

Gro. E. Stepsins: On the minimum audible sound.

J. F. Kemp: Buried river channels of the Hudson Valley.

W. M. Davis: Glacial erosion in Wales.

Cuas. D. Watcorr: Summary of studies of Cambrian brachiopods.

Cuas. S. Minor: On certain changes of nuclei in relation to age.

J. McK. Carretnt: Researches from the Psychological Laboratory of
Columbia University.

H. F. Osporn: Additions to the Collections of Extinct Vertebrates in the
American Museum of Natural History.

W. K. Brooks: A biographical memoir of Alpheus Hyatt.

RESTON STEVENSON and J. Livinaston Morean: Drop weight and the law
of Tate; the’ determination of the molecular weight in the liquid state by
the aid of drop weights.

E. C. Pickerine: The relation of the spectra, magnitudes, and colors of
stars.

Stuon Newcoms: Tables of Minor Planets discovered by James C. Watson,
prepared by A. O. Leuschner under the direction of the Watson Trustees of
the National Academy of Sciences.

It was announced that General Cyrus B. Comstock had given
to the National Academy of Sciences $10,000 as a fund, the in-
come to be used for the advancement of knowledge in magnetism,
engineering and radiant energy.

3. American Association for the Advancement of Science.—
The fifty-eighth meeting of the American Association will be held
at Chicago, in the buildings of Chicago University, from Decem-
ber 30, 1907, to January 4, 1908, with Professor E. L. Nichols
as the president. The meetings ‘of the usual affiliated societies
will also take place at the same time. A preliminary announce-
ment relating to this, the sixth of the Convocation week meet-
ings, has recently been issued by the Permanent Secretary, Mr.
L. O. Howard of the Smithsonian Institution, Washington.

508 Scientific Intelligence.

4. Annual Report of the Board of Scientific Advice for
India for the year 1905-1906. Pp. 171. Calcutta, 1907
(Government Central Press).—The Board of Scientific Advice
for India was established in 1902 as a central authority for the
coérdination of official scientific inquiry, with the design of mak-
ing research work as effective as possible and also of aiding the
Government of India in connection with the investigation of
questions of economic and applied science. The subjects included
are economic and agricultural chemistry, astronomy and meteor-
ology, geology, eeodesy and geography, botany and zoology.
The volume now issued contains concise reports by different
authors in these different departments and presents many inter-
esting facts. In geodesy, Lieut.-Col. Burrard gives the results of
determination of the altitude of some of the peaks of the Hima-
layas as influenced by atmospheric refraction; the height of Mt.
Everest, for example, as observed from the plains of India, is
some 29,002°3 feet, while from the Darjeeling Hills it is 29,141.
The same author in connection with the Seistan (Afghanistan)
geography, discusses desiccation in Central Asia, expressing the
opinion that the theory of a permanent climatic change within
the human period calls for meteorological proof before it can be
accepted. The constant tendency of the sand to inerease while
the amount of water is constant is shown to have an important
bearing on the problem.

5. Mendelism ; by R. C. Punnett. Second edition, pp. vil
+ 85. Cambridge, 1907 (MacMillan & Bowes).—The appear-
ance of a new edition of this well-written essay on Mendel’s
principles of heredity within two years after the first printing
of the work indicates the cordial reception which the little book
has received. The stories of Mendel’s discoveries and their
applications in the breeding of plants and animals is presented in
popular language by one who has had wide experience in testing
the applicability of the so-called law. Even in the short time
that has elapsed since the printing of the first edition, discoveries
have been made which have necessitated considerable revision of
the original text, and it is obvious that the end is not yet.

W. B. C:

6. Les Prix Nobel in 1904. Stockholm, 1907 (P. A. Norstedt
& Séner).—This annual volume describes the distribution of the
Nobel prizes in 1904. It contains sketches and portraits of the
recipients of the prizes—in science, Lord Rayleigh, Sir William
Ramsay and Dr. I. P. Pawlow,—and also reproductions of the
Nobel medals and diplomas. The Nobel lectures, delivered at
Stockholm by the three gentlemen named, are also reproduced.

7. Memorials of Linneus. British Museum of Natural His-
tory, Special Guide No. 3. Pp. 16. London, 1907.—This pamph-
let contains a list of the collection of portraits, manuscripts,
specimens, and books brought together at the British Museum to
commemorate the bicentenary of the birth of Linnezus.

Scientific Intelligence. 509

8. Astronomical Observatory of Harvard College, Epwarp
C. Pickerrne, Director.—Recent publications from the Harvard
College Observatory are included in the following list (continued
from vol. xxiii, 328) :

Annats.—Volume XLVII, Part I. A photographic Study of
Variable Stars, forming a part of The Henry Draper Memorial ;
prepared by Wittramina V. Fiemine under the direction of
Epwarp C. Pickering. Pp. 113, with 5 tables. Volume LVII,
Part I. Observations of Seventy-five Variable Stars of Long
Period during the years 1902-1905 ; prepared for publication by
Leon CampseE_y under the direction of Epwarp C. PIcKERING.
Pp. iv, 210, with 2 plates and 13 tables. Vol. LX, No. IV.
1777 Variables in the Magellanic Clouds; -by Henrierta S.
Leavitt. Pp. 87-108, with 2 plates and 6 tables. No. V. Ten
Variable Stars of the Algol Type; by Henrierra S. Leavirrt.
Pp. 109-146, with 3 plates and 15 figures.

Volume LXIIJ, Part I. Determination of Constants for the
Reduction of Zones Observed with the Meridian Circle during the
years 1888 1898; by ArrHuR SEARLE. Pp. 145, with 9 tables.

Crrcutars—No. 125. Stellar Magnitudes; by Epwarp C.
PickERiING. Pp. 3, with one table.

No. 126. Two Variables discovered by M. Baillaud; by
Epwarp C. Pickrerine. Pp. 3, with two tables.

No. 127. New Variable Stars in Harvard Map, Nos. 3 and 6;
by Epwarp C. Pickrertne. Pp. 4, with three tables.

No. 128. Missing Durchmusterung Stars; by Epwarp C.,
Pickering. Pp. 4, with one table.

No. 129. 15 New Variable Stars in Harvard Maps Nos. 31 and
62 ; by Epwarp C, Pickrerine. Pp. 4, with two tables.

No. 130. 71 New Variable Stars in Harvard Maps, Nos. 9, 12,
21, 48, and 51; by Epwarp C. Pickrrine. Pp. 4, with two
tables.

9. New York State Museum, Albany, N. 'Y. Joun M. Cr. ARKE,
Director.—The following publications have recently been issued :

Third Report of the Director of the Science Division, 1906,
including the 60th Report of the State Museum, the 26th Report
of the State Geologist, and the Report of the State Paleontol-
ogist for 1906. Pp. 182.

Bulletin 111, Geology 13. Drumlins of Central Western New
York; by H. L. Faircuitp. Pp. 391-443.

Bulletin 112, Economic Geology 16. The Mining and Quarry
Industry of New York State: Report of Operations and Produe-
tion during 1906; by D. H. Newtanp. Pp. 80.

10. Dew-ponds ; by Epwarp A. Martin, F.G.S. Reprinted
from “ Knowledge and Scientific News,” Maryland, June 1907.—
These ponds are shallow artificial hollows, without inlet or outlet,
made on the English downs. ‘The writer discusses their construc-
tion and the theories accounting for their filling. He regards it
as certain that mists contribute largely to these ponds, and points
out that if dews contribute also it may be necessary to revise
somewhat the theory that dew is really formed from moisture
which rises out of the soil rather than from moisture condensed
from the air. iy 1B

INDEX,” ZO” VOUGIMIn) OxXexeies

A

Academy, National, meeting at New |
York, 507.

Agassiz’s Expedition to the Tropical |

Pacific, reports, 450,
Alabama, underground water
sources, Smith, 84.
Arnold, H. D., electric arc, 383.

re-

Association, American, meeting at |

Chicago, 507.
Astronomical papers, Lehigh Uni-
versity, Ogburn, 283.
—Observatory, Cambridge, 509.

B

Bacterial Infections of Digestive |
Tract, Herter, 91.

Barker, H. C., thermoelectromo-
tive forces of. potassium and so- |
dium, 159.

Barrell, Votes

geology of Marysville |
mining district, Montana, 85. |

Barus, ice method for observ ation |
of coronas, 277, 3876; cycles of |
coronas, 309 ; decay of nuclei, 419 ;
volcanic activity, 483.

Bascom, F., anhydrite twin from
Aussee, 487.

Becker, G. F., current theories of
slaty cleavage, iL.

Benton, J. R., Senet and elasticity |
of spider thread,

Bermuda Islands, Verrill, 179, 180.

Biology, Elements, Hunter, 448,

Birds of the Chicago area, Wood-
ruff, 92.

Blackwelder, E., Research in China,
501.

Bolometer, vacuum, Warburg, Leit-
hiuser and Johansen, 500.

Boltwood, B. B., radio-activity of

thorium salts, 93; new radio-ac-
tive element, 370.
BOTANY.
Anemonella thalictroides, Holm,
245.

Plant Chemistry, studies in, and |

literary Papers, Michael, 90.

Textbook of Botany, Campbell, 91. |
Brogger, W.C., minerals of South- |

ern Norway, 282.

LS

Cady, W. G., electric arc, 383.

Campbell, D. H., Textbook of Bot-

any, 91.

| Canal rays, Paschen, 441.

eepe Colony, plains in, Schwarz,

OF

Carnegie Institution, publications,
87, 882.

Catalonia, volcanoes and _ rocks,
Washington, 217; Calderon, Ca-
ZUrTtO and Fernandez- Navarro, 282.

Cathode rays, secondary, Kleeman,
| 499.

CHEMISTRY.
Acetamide, preparation, Phelps,
429.

Ammonia, action upon ethyl oxa-
late, Phelps, Weed and Housum,
479.

Atomic weights, speculations in re-
gard to, Collins, 496.

Chromium, new variety, Jasson-
neix, 81.

Copper, determination, Gooch and
Heath, 65.

Fluorides, interference with preci-
pitations of alumina, Hinrichsen,
49).

Formamide, preparation,
and Deming, 173.

Helium in natural gas, Cody and
McFarland, 497.

Hydrocarbons, decomposition of
gaseous, Kusnetzow, 374.

Jonium, Boltwood, 370.

Lanthanum, estimation,
197.

Manganese, magnetic compounds
with boron, ete., Wadekind, 80.

Molybdic acid, behavior, Randall,
313.

Potassium as the cobalti-nitrite,
Drushel, 435.

— — aluminium sulphate,
and Osborne, 167.

Radium, atomic weight, Curie, 439.

Silicon and silicon carbide, com-
bustion, Mixter, 130.

Silver, nitrogen, etc., atomic weight,
Richards and Forbes, 439.

Phelps

Drushel,

Gooch

* This Index contains the general heads, BoTANY, CHEMISTRY (incl. chem. physics), GEOLOGY,

MINERALS, OBITUARY, ROCKS, ZOOLOGY, and und
mentioned.

er each the titles of Articles referring thereto are

INDEX.

CHEMISTRY—continued.

Sulphur, vapor-tension, Gruener,
497.

Tellurium, separation, Brauner and
Kuzma, 373.

Thorium, new intermediate pro-

duct, Hahn, uo}

Zine chloride, use of, Phelps, 194. |

Chemistry, Physical, Jones, 440.
— Practical, Martin, 440.
Chicago area, birds, Woodruff, 92.
China, Research in, Blackwelder,
501.
Connecticut geol. survey, 447.

Cooksey, C. D., corpuscular rays |
_ produced in metals by Rontgen |

rays, 285.
Coronas, observation of, Barus, 277,
376 ; cycles of, Barus, 309.
Crandall, R., Cretaceous of Santa |
Clara Valley, California, 33.

D

Deming, C. D., preparation of form- |
amide, 173.

Dew-ponds, Martin, 509.

Dominica Island, Hercules beetles,
A. Hyatt Verrill, 305. |

Drew, G, A., Invertebrate Zoology,
o82.

Drushell, W. A., volumetric estima- |
tion of lanthanum, 197; potassium |
as the cobalti-nitrite, 433.

E

Earthquake Investigation
mittee, Imperial, bulletin, 90.
Eastman, C. R., Devonic fishes of
the New York formations, 448.
Eiszeit und Urgeschichte der Men-
schen, Pohlig, 381.
Electric arc between metallic elec-
trodes, Cady and Arnold, 383.
Electrical units, ratio of, Rosa and

Com-

Dorsey, 443, 500.
Electro- ~Analysis, Smith, E. F., 498. |
Electrolytes, infiuence of nae |
fields on, Berndt, 442.
Electromagnetic waves over plane
surfaces, Zenneck, 441. |
Ethnology, Bureau of American
24th annual report, 89, 91.
Evolution and Animal Life, Jordan |
and Kellogg, 449.

335)|

Ewell, A. W., Gibbs’ Theory of re- |

flection of light, 412.

511

F

Hemphotozsaphic, Elektrische,
Korn, 82.

Field Museum of Natural History,
publications, 88.

|Fizeau on change of azimuth of
polarization, 498.

Fossils, see GEOLOGY.

G
GEOLOGICAL REPORTS.

Alabama, 84.
Connecticut, bulletin No. 8, 447.
Illinois, bulletin No. 4, 447.

India, 181.

Maryland, 180, 181.

United States, folios, 376; topo-
graphic atlas, 82; Mineral re-
sources, 82; professional papers,
82, 376; bulletins, 82, 376 ; water
supply papers, 83, 377.

Western ‘Australia, bulletin,
24, 84.

Wisconsin, 83, 500.

Geological Society of London, Cen-

tenary, 92.

No.

2

GEOLOGY.

Cambrian transition fauna of Brain-
tree, Mass., Shimer, 176.

Cleavage, slaty, current theories,
Becker, 1.

Cretaceous of Santa Clara region,
California, Crandall, 33.

Devonian of eastern America,
Coblenzian invasion, Clarke, 502.

— fishes of the New York forma-
tions, Eastman, 443.

Fauna, lower Miocene from So.
Dakota, Matthew, 379.

Fossil Insects, Handlirsch, 447.

Guaynopita district, Mexico, geol-
ogy, Hovey, 503.

Laramie, meaning of term, Veatch,

8.

Mammal horizons, Tertiary, of No.

America, Osborn, 504.

Marysville mining district, Mon-
tana, Barrell, 35.

Miocene, Lower, fauna from So.
Dakota, Matthew, 379.

Mississippian formations of Rio
Grande Valley, N. M., Gordon,
58.

Niagaran limestone in the Chicago

area, Weller, 4495.
Paleontologica, Miscellanea,
Fritsch, 502.

553 BY. INDEX.

GEOLOGY—continued. | Hough, T., Physiology, 448.
-Housu G R., acti fd -
Paleontologia Universalis, 447. mania 470 pam
pene in Cape Colony, Schwarz, | Hunter, G. w., Biology, 448.
Stromatoporoids in Ontario, Parks, I
86. ERS 2)
Terraces, aggraded, of the Rio} Illinois geol. survey, 447. f
Grande, Keyes, 467. |India, Board of Scientific Advice,
Tertiary formations of the John|_ report, 508 ; geol. survey, 181.
Day region, Merriam, Sele Indians, Handbook of American,
— mammal horizons of No. Amer- | Hodge, 91. ‘
ica, Osborn, 504. Inorganic and Organic Substances,
—peneplain in Arizona and New _ Solubilities, Seidell, 440.
Mexico, Robinson, 109. Islay, geology, Wilkinson, Teall and
Trias, stratigraphy of the Western Peach, 003.
American, Smith, 446. |
Trilobites in the Chicago area, | J
Weller, 445. Jahrbuch fiir Mineralogie, etc., 92.
— Silurian, of the East Baltic, Jones, H. C., Physical Chemistry,
Schmidt, 445. 440,
Unionide, from Montana Laramie Jordan, D. S., Evolution and Animal
clays, Whitfield, 446. | Life, 449,
Upper Permian red beds of Okla- |
homa and Texas, Beede, 86. K
Voleanic activity, Barus, 483. Kell V. L., Evoluti d Ani-
Voleanoes, active, of the earth, Sees 449.” yee eae
Merealli, 282. Keyes, C. R., aggraded terraces of

— of Catalonia, Washington, 217;) the Rio Grande, 467.
Calderon, Cazurro and Fernan- |Kip, H. Z., determination ae he

dez-Navarro, 282. | hardness of minerals, 23.
Wyoming, Big Horn basin, geology. | Korn, A., Wlolniceehe Fernphoto-
Fisher, 503. | graphie, 82.

Gibbs’ ceometrical theory of reflec- | }zristall Hie Phvaileaech
tion of light, Ewell, 412. Sea a yes a

Goldschmidt, V., anhydrite twin Kunz, G. F., forms Of Atieancactdiee

from Aussee, 487. monds, O75.
Gooch, F. A., determination of cop-
per, 65; potassium aluminium sul- | fe
phate, 167. |Larmar, J., Memoir and Correspond-

Gordon, C. H., Mississippian forma-| ance of Sir G. G. Stokes, 81.
tions of Rio Grande valley, New | Lawton, E. E., bands ati spec-

Mexico, 58. | trum of nitro 101
eee gen ;
Gregory, H. E., Bibliography of | ehigh University, Astronomical
Connecticut Geology, 447. papers, Ogburn, 283.
| Light, Gibbs’ ceometrical theory of
H reflection of, Ewell, 412,

ag | Lifneus, Memorials of, 508.
Pervate College Observatory, 509. otha Al J., Mode ‘of growth of
eath, F. H., determination of cop- Le Ala cent net UOMSRG
oe, material aggregates, ; :
Herter, C. A., Bacterial Infections | M
of the Digestive Tract, 91.
Hillebrand, W. F., vanadium sul- Magnetic compounds of manganese

phide patronite, etc., from Peru, | and boron, 80.

141; Texas, mercury minerals, 259.|— fields on the resistance of elec-
Hofmeister, F., Beitriige zur chem-_ trolytes, Berndt, 442.

ischen Phy siologie, 91. | Martin, G., Practical Chemistry, 440.
Holm, T., Anemonella thalictroides, | Maryland ¢ geol. survey, 180, 181.

243. Material aggregates, mode of growth,

Horse, skeleton of Arab, Osborn, 380.| Lotka, 199, 375.

INDEX. 513

Matthew, W. D., Lower Miocene

fauna from So. Dakota, 379.
Mendelism, Punnett, 508.

Mercalli, G., Active Volcanoes of the |

Earth. 282.
Merriam, J. C., Tertiary formations
of the John Day region, 377.
Metals, internal temperature gra-
dient, Serviss, 451.

Michael, H. A., Studies in Plant)
OBITUARY.

Chemistry, etc., 90.

MINERALS.

Albite, 255. Anhydrite twin crys-
tal, 487.

Benitoite, California, 448. Bra-|
voite, Peru, 151.

Calcite, New Jersey, 426. Calo- |

mel, Texas, 273. Carlosite, Cali-
fornia, 448. Chalmersite, Brazil,
250. Chiastolite, So. Australia,
183. Chlorite, 255.

Diamonds, Arkansas, 275.

Eglestonite, Texas, 271. Evansite, |

Idaho and Alabama, 155.
Gold nuggets from New Guinea,

Liversidge, 505. Gorceixite, Bra. |

zil, 182.

Harttite, Brazil, 182. Hellandite,
Norway, 182. Hematite, 255;
artificial crystals, 485.

Kleinite, Texas, 261.

Manganotantalite, Maine, 154. Mer-
cury, native, Texas, 274. Mon-
troydite, Texas, 269.

Nepouite, New Caledonia, 182.

Olivine in serpentine of Chester, |

Mass., 491.
Patronite, Peru, 141. Phenacite,

Gloucester, Mass.. 252. Purpur- | ie
fate f Pd | Phosphorescenice, power of positive

ite, So. Dakota, 152. Pyrite, 254.
Quisqueite, Peru, 141.
Rutherfordine, East Africa, 181.
Siderite, 255. Sphalerite, 204.
Terlinguaite, Texas, 270. TVour-

maline, Elba, 157.

Zinuwaldite, Alaska, 158. Zoisite |

crystals, Chester, Mass., 249.

Minerals, determination of the hard- |

ness, Kip, 25.

— measurement of the optic axial
angle of, Wright, 317.

— mercury, from Texas, Hillebrand
and Schaller, 259.

— of Southern Norway, Brogger, 282.

— Tables of, Penfield, 448.

Mixter, W. G., combustion of silicon
and silicon carbide, 130.

Munroe, Chas. E., artificial hematite
crystals, 485.

N

|New Mexico, Mississippian forma-

tions, Gordon, 58.
Nitrogen spectrum, Lawton, 101.
Nobel prizes in 1904, 508.

| Nuclei, decay of ionized, Barus, 419.

O

Atwater, W. A., 382.
Heilprin, A., 184, 284.
Loewy, M., 450.
Safford, J. 'M. , 284,

Occlusion of oxygen, Szivessy, 442.
Optic axial angle of minerals, meas-

urement, Wright, 317.

Ordway, J. M., waterglass, 473.
Osborn, H. Ee skeleton of Arab

horse, 380; Tertiary mammal
horizons of "America, 504.

Osborne, R. W., potassium alu-

minium sulphate, 167.

/Ozone, action on metallic silver,

Manchot and Kampschulte, 373.

P

Palache C., mineralogical notes, 249 ;

occurrence of olivine, 491.

| Penfield, S. L., Tables of Minerals,

448.

|Phelps, I. K., preparation of for-

mamide, 173 ; action of dry
ammonia, 479.

|—and M. A., use of zinechloride,

194 ; preparation of acetamide,
429,

rays to produce, Kunz, 499.

| Physiologie, Beitrige zur chemi-

schen, Hofmeister, 91.

Physiology, Hough and Sedgwick,

448.
Polarization, Fizeau’s research on
the change of azimuth, 498.
|Positive rays, power to produce
phosphorescence, Kunz, 499.
Punnett, R. C., Mendelism, 508.

R

Radio-active element, new, Bolt-
wood, 370.

Radio-activity of thorium salts,
Boltwood, 93.

Randall, D. L., behavior of molyb-
dic acid, 513.

~

514

467.
— Mississippian of, Gordon, 58.

Robinson, H. H., Tertiary peneplain |

of Plateau district, 109.

ROCKS.

Volcanoes, Catalan, and their rocks,
Washington, 217.
—— Calderon, Cazurro and Fer-
nandez-Navarro, 282.
Rontgen radiators, secondary, Barkla,
Crowther, 499.
—rays, production of corpuscular
rays by, Cooksey, 285.

S)

San Domingo Solenodon, Verrill, 55,

Schaller, W. T., mineralogical notes,
152; mercury minerals from Texas,
259.

Schwarz, E. H.L., plains in Cape |

Colony, 185.
Schwingungserzeugung,
der, Barkhausen, 283.
Sedgwick, W. T., Physiology, 448.

Seidell, A., Solubilities of Inorganic |

and Organic Substances, 440.
Serviss, S. B., internal temperature
gradient of metals, 401.

Shimer, H. W., Cambrian transition |

fauna of Braintree, Mass., 176.

Smith, E. F., Electro-Analysis, 498. |

Smithsonian Institution, annual
report, 506; other publications, 506.

Spectra, absorption, Uhler and
Wood, 442.

Spectrum of nitrogen, Lawton, 101.

Spider thread, strength, Benton, 75. |

Standards, Bureau of, bulletin, 87,
442.

Stokes, Sir G. G., Memoir and Cor-
respondence, Larmor, 81.

40

Thermoelectro-motive forces of po-
tassium and sodium, Barker, 159.
Thorium products, rays from, Hahn,

374,
— salts, radio-activity, Boltwood, 93.
Tungsten, melting point of pure,
Wartenberg, 440.

Rio Grande, aggraded terraces, Keyes, |
|
|

Problem |

INDEX.

U

United States geol. survey, see
GEOL. REPORTS AND SUR-
VEYS.

V

| Veatch, A. C., meaning of term

Laramie, 18.

| Verrill, A. E., Bermuda Islands,

| 179; 180:

| Verrill, A. H., Solenodon of San
Domingo, 55; Hercules beetles
from Dominica Island, 305.

Volcanic activity, Barus, 483.

Volcanoes active, Mercalli, 282; of
Catalonia, 217, 282.

WwW

| Washington, H. S., Catalan vol-
canoes and their rocks, 217; forms
of Arkansas diamonds, 275.

Waterglass, Ordway, 473.

Weed, L. H., action of dry ammo-
nia, 479.

| Western Australia geol. survey, 84.

| Whitlock, H. P., calcite from West
Paterson, N. J., 426.

Wireless telegraphy, relation of elec-
tromagnetic waves to, Zenneck,
441.

| Wisconsin geol. survey, 838.

— geology of north central, Weid- _
man, 900.

Wright, F. E., measurement of the
optic axial angle of minerals, 317.

Z

| Zoological Congress, seventh inter-
national, meeting at Boston, 92.

| ZOOLOGY.

Birds of Chicago, Woodruff 92.

Brachyura of the Eastern Tropical
Pacific Expedition, Rathbun, 450.

Crayfishes, young of, Andrews, 449.

Crustacea of the North Pacific Ex-
ploring Expedition, Stimpson,
449.

Hercules beetles from Dominica
Island, A. Hyatt Verrill, 305.
Solenodon of San Domingo, Verrill,

Zoology, Invertebrate, Drew, 382.

Dr. Cyrus Adler,.

Librarian U. S. Nat. Museum.

%

VOL. XXIV. JULY, 1907.
SR os ER I TI SSIS ES FE DE I ET

Established by BENJAMIN SILLIMAN in 1818.

agetterts Wikis. ,

SMW nR le AN
JOURNAL OF SCIENCE,

Epirorn: EDWARD S. DANA.

ASSOCIATE EDITORS
Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE,
W. G. FARLOW anp WM. M. DAVIS, or CamBrinceE,

Proressors ADDISON E. VERRILL, HORACE L. WELLS,
L. V. PIRSSON anp H. E. GREGORY, or New Haven,

Proressor GEORGE F. BARKER, or PHILADELPHIA,
Proressor HENRY S. WILLIAMS, or Irwaca,
Proressor JOSEPH S. AMES, or Battiwore,

Mr. J. S. DILLER, or Wasuinerton.

FOURTH SERIES
VOL. XXIV—[WHOLE NUMBER, CLXXIV.]
Nos 139 JULY, 1907.

NEW HAVEN, CONNECTICUT.
TORO G

THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET.

Published monthly. Six dollars per year, in advance. $6.40 to countries in the
Postal Union ; $6.25 to Canada. Remittances should be made etre orders,

aes

JUL 2 1907

registered letters, or bank checks (preferably on New York banks).

st

IMPORTANT NOTICE.

If you wish to secure choice rare or ordinary minerals and cut gems, now
is the time to buy them, as we have inaugurated our yearly summer reduc-
tions of 10 per cent. on rare, showy minerals and cut gems, and 20 per cent.
on ordinary minerals. This reduction is for the months of July and August
only. Our stock is the largest and finest that has ever been. Especial
attention is called to the immense size of some of our preciovs and semi-
precious cut gems, running as high as 30 carats, all of which will make a
brilliant addition to your collections :—Siberian Amethysts, deep purple
color, at night reddish color, like Arizona Garnets, from 1 to 30 carats;
Topaz, Bahia, Brazil, from 1 to 40 carats, deep golden color; Red, pink,
green and blue Tourmalines, from Minas Geraes, Brazil, from 1 to 10 carats;
Red Tourmalines, from Ural Mts., from 1 to 8 carats; Opals, Whitecliff,
New South Wales, Australia, from 1 to 8 carats, milky with deep fire. Fine
Opal Shells, perfect ; are considered very rare when perfect.

RARE NORWAY MINERALS

Leucophane crystals, in the matrix, finest in the world; Polyerase xls in
matrix; Thorite xls; Gadolinite xls; Broggerite xls; Hellandite, new min-
eral, xls in matrix; Monazite xls in matrix, and loose xls; Rutile xls;
Malakon xls in the matrix; Apatite xls; Xenotime xls; Euxenite xls in
matrix ; Katapleit xls in matrix.

NEW CRYSTALLIZED NATIVE COPPER

which was advertised and illustrated in the American Journal of Science and
Mineral Collector. We secured the whole output of one pocket. Five of
the finest specimens still remain, price from $10 to $70. (Write for further
particulars and illustrated circular.)
SCIENTIFIC RUBIES

We have an extra fine lot of these Scientific Rubies, from 14 to 4-carats,

at 3) per carat.
RARE MINERALS :

Anatase, Binnenthal, and St. Gothard, Switz., $4-$10. Dioptase, Siberia,
$7.50-S20. Phosgenite, Eng., $2-$10. The new mineral Zeophyllite, Rad-
zein, Bohemia, $3-$7.50. Bismuthinite, $4-$6. Eulytite, Saxony, $0-S6.
Alexandrite, Ural Mts., xls from $3-$5, matrix specimens from $20-$25.
Zeinerite, Joachimsthal, Bohemia, $3. Terlinguaite, Terlingua, Texas, $3.
Graftonite, Grafton, N. H., from $2-$5.

CALIFORNIA MINERALS
Pink Beryls, Pala, in matrix, and loose xls, $8-S35. Blue and white
Topaz, Romana Co., $8-$10. Colemanite, San Bernardino Co.,$2-$5. Cali-
fornite, Pala, polished slabs, $1-$5. Kunzite, from 50c. to $50. Cinnabar,
from Sonoma Co., and New Almaden, fine x]s in matrix, $2-$5. Tourmalines,
Mesa Grande, and Pala, in matrix and xls, different colors, 50c.—$100.
HUNGARIAN MINERALS
Stibnite, from 25c¢ to $7.50. Barite, different colors, $1-S5. Realger, $4—
$5. Orpiment, $1.50-83. Cinnabar, $2-S5. Bournonite, $1-S3. Sphalerite
and Quartz, 50c-34. Blue Chaleedony Pseudomorph, 50c-S2.
Crystallized Goid, Silver, Calaverite, and Copper
We have a fine lot of crystallized Gold, Silver, Calaverite, and Copper,
from the different localities; also Calcite enclosing Copper.

s

Write for further particulars.

ALBERT H. PETEREIT,
81—83 Fulton Street, New York City.

Dr. Cyrus Adler,

Librarian U. S. Nat. Museum. ©

VOE XXIV. AUGUST, 1907

Established by BENJAMIN SILLIMAN in 1818.

THE

AMERICAN
JOURNAL OF SCIENCE.

Eprror: EDWARD S. DANA..

ASSOCIATE EDITORS

Proressors GEORGE L. sOODALE, JOHN TROWBRIDGE,
W. G. FARLOW anp WM. M. DAVIS, or Camsrincs,

Proressorss ADDISON E. VERRILL, HORACE L. WELLS,
L. V. PIRSSON anp H. E. GREGORY, or New Haven,

Proressor GEORGE F. BARKER, or PHILADELPHIA,
A Proressor HENRY S. WILLIAMS, or Itwaca,
Proressor JOSEPH S. AMES, or Batriore,

- Mr. J. 8S. DILLER, of Wasurncton.

FOURTH SERIES
| VOL. XXIV—[WHOLE NUMBER, CLXXIV.]

No. 140—AUGUST, 1907.

We, eae ee

NEW HAVEN, CONNECTICUT.
een OeG

THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET.

Published monthly. Six dollars per year, in advance.
Postal Union ; $6.25 to Canada. Remittances should be m
registered letters, or bank checks (preferably on New Yor

IMPORTANT NOTICE.

If you wish to secure choice, rare or ordinary minerals and cut gems, now
is the time to buy them, as we have inaugurated our yearly summer reduc-
tions of 10 per cent. on rare, showy minerals and cut gems, and 20 per cent.
on ordinary minerals, This reduction is for the months of July and August
only. Our stock is the largest and finest that has ever been. Especial
attention is called to the immense size of some of our precious and semi-
precious cut gems, running as high as 30 carats, all of which will make a
brilliant addition to your collections :—Siberian Amethysts, deep purple
color, at night reddish color, like Arizona Garnets, from 1 to 40 carats ;
Topaz, Bahia, Brazil, from 1 to 40 carats, deep golden color; Red, pink,
green and blue Tourmalines, from Minas Geraes, Brazil, from 1 to 14 carats;
Red Tourmalines, from Ural Mts., from 1 to 8 carats; Opals, Whitecliff,
Australia, milky with deep fire. Fine Opal Shells, perfect ;;are considered
very rare when perfect.

RECENT ARRIVALS

Scheelite, massive xls, and yellowish Powellite, Oak Springs, Nevada, 32.50
to $5; Tetradymite and Green Bismutite, Nevada, $1 to $8; Rose Chalce-
dony, Aurora, Nevada, $1 to $5; Zinkenite, Mosey, Nevada, $1.50 to $3;
Copalite on Coal, Castle Gate, Utah, 50c. to $1; Calaverite, Cripple Creek,
Colorado, 50c. to $5; Onegite, El Paso Co., Colorado, $1 to $8; Cobaltite,
Cobalt, Canada, $3 to $7.50; loose crystals, 10c. to 25c.; Niccolite, Cobalt,
Ontario, Canada, 75c. to $4; Native Silver and Niccolite, Cobalt, Ontario,
Canada, $2 to $10; Erythrite, in native silver, Cobalt, Ontario, $1 to $0;
Rhodonite, Franklin Furnace, N. J., $2.50 to $10; Petrified Wood, Chalee-
dony Park, Wyoming, $2.50 to $5; Opal, Barcoo River, Queensland, Austra-
lia, 50c. to $5; Tourmaline, Mesa Grande, California, different colors, in
matrix, and loose xls, from $2 to $50; Hubnerite, 75c. to $0; Bismuth and
Cassiterite, Bohemia, $2 to $5; Native Antimony, Prince William, New
Brunswick, $1 to $2; Tellurium, Boulder, Colorado, $1 to $5; Native Plati-
num, Columbia, S. America and Oregon, from $2 to $9.

RARE NORWAY MINERALS

Leucophane crystals, in the matrix, finest in the world ; Polykrase, xls in
matrix; Thorite xls; Gadolinite xls; Bréggerite xls; Hellandite, new min-
eral, xls in matrix; Monazite xls in matrix, and loose xls; Rutile xls;
Malakon xls in the matrix; Apatite xls; Xenotime xls; Huxenite xls in
matrix.

SCIENTIFIC RUBIES

We have an extra fine lot of these Scientific Rubies, from 14 to 4 carats,
at $d per carat.

RARE MINERALS

Anatase, Binnenthal, and St. Gothard, Switz., $4-$10. Dioptase, Siberia,
$7.50-$20. Phosgenite, Eng., $2-$10. The new mineral Zeophyllite, Rad-
zein, Bohemia, $3-$7.50. Bismuthinite, $4-$6. Eulytite, Saxony, $5-$6.
Alexandrite, Ural Mts., xls from $3-$5, matrix specimens from $20-$20.
Graftonite, Grafton, N. H., from $2-S5.

CALIFORNIA MINERALS

Pink Beryls, Pala, in matrix, and loose xls, $8-$35. Blue and white
Topaz, Romana Co., $8-$10. Colemanite, San Bernardino Co., $2-$d. Cali-
fornite, Pala, polished slabs, $1-$5. Kunzite, from 50c. to $50. Cinnabar,
from Sonoma Co., and New Almaden, fine xls in matrix, $2-$). Tourmalines,
Mesa Grande, and Pala, in matrix and xls, different colors, 50c¢.—$100.

Write for further particulars.

ALBERT H. PETEREIT,
81—83 Fulton Street, New York City. -

‘Dr. Cyrus Adler, : es

Librarian U. S. Nat. Museum.

VUbS AAT: : SEPTEMBER, 1907.

Established by BENJAMIN SILLIMAN in 1818.

THE

AMERICAN
JOURNAL OF SCIENCE.

- Epirorn: EDWARD S. DANA.

ASSOCIATE EDITORS

Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE,
W. G. FARLOW ann WM. M. DAVIS, or Camprincz,

Proressors ADDISON E. VERRILL, HORACE L. WELLS,
LL. V. PIRSSON anp H. E. GREGORY, or New Haven,

Proressor GEORGE F. BARKER, or PHILADELPHIA,
Proressor HENRY S. WILLIAMS, or ItwHaca,
Proressor JOSEPH S. AMES, or Battimore,
Mr. J. S. DILLER, or Wasuineton.

FOURTH SERIES
VOL. XXIV—[WHOLE NUMBER, CLXXIV.]

No. 141—SEPTEMBER, 1907.

NEW HAVEN, CONNECTICUT.
1907

THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET.

Published monthly. Six dollars per year, in advance. $6.40 to countries in the
Postal Union ; $6.25 to Canada. Remittances should be made either by money orders,
registered letters, or bank checks (preferably on New York banks). ;

IMPORTANT NEWS.

We have secured one of the finest collections ever gotten together by a
Brooklynite, and we think all will know its owner without mentioning his
name. It is rich in the much-sought-after old finds and rare minerals which
were plentiful 25 or 30 years ago, but which cannot be secured at any price
to-day, except out of very old collections. It is valued at over $10,000.
Some of these specimens are so rare and fine that the owner has been impor-
tuned many times to part with them by universities, museums and the better
class of collectors, but in every case he has refused their offers, not wishing
to break the collection. It will now be offered on and after September 1st,
in lots to suit purchasers, at a very reasonable figure. First come first
served will be the rule. Don’t delay and then blame us, as we know there
will be a rush to secure the finest of these specimens. If you cannot call on
us send us your wishes and they will be filled to best advantage. This col-
lection has been correctly labelled according to the Dana System by its
owner, so that all purchasers are assured of the accuracy of these specimens.
Further particulars on application.

“OUR NEW CIRCULAR.”

We have just issued a new 8-page Circular, covering almost all our stock
at present date, except the Brooklyn Collection, the East Indian and Russian
Gem Collections, and two other Kuropean collections that are on their way.
Special Circular will be issued of these as soon as they arrive and are
arranged.

RECENT ARRIVALS

Tetradymite and Green Bismutite, Nevada, $1 to $3; Rose Chalcedony,
Aurora, Nevada, $1 to $0; Zinkenite, Mosey, Nevada, $1.50 to $8; Copal-
ite on Coal, Castle Gate, Utah, 50c. to $1; Calaverite, Cripple Creek,
Colorado, 50c. to $5; Onegite, Kl Paso Co., Colorado, $1 to $3; Cobaltite,
Cobalt, Canada, $3 to $7.50; loose crystals, 10c. to 25c,; Niecolite, Cobalt,
Canada, 7dc. to $4; Native Silver and Niccolite, Cobalt, Canada, $2 to $10;
Hrythrite, in native silver, Cobalt, Canada, $1 to $5; Rhodonite, Franklin
Furnace, N. J., $2.50 to $10; Petrified Wood, Chalcedony Park, Wyoming,
$2.50 to $5 ; Opal, Barcoo River, Queensland, ’ Australia, 50c. to $5; Tour-
maline, Mesa Grande, California, different colors in matrix and loose xls,
from $2 to $50; Hubnerite, "5c. to $0; Native Antimony, Prince William.
New Brunswick, $1 to $2; Tellurium, Boulder, Colorado, $1 to $5; Native
Platinum, Columbia, S. America and Oregon, from $2 to $5.

RARE NORWAY MINERALS
Leucophane crystals, in the matrix, finest in the world; Polykrase, xls in
matrix; Thorite xls; Gadolinite xls; Bréggerite xls; Hellandite, new min-
eral, xls in matrix; Apatite xls; Xenotime xls; Huxenite xls in matrix.

RARE MINERALS
Anatase, Binnenthal, and St. Gothard, Switz., $4-$10. Dioptase, Siberia,
$7.50-$20. Phosgenite, Eng., $2-$10. The new mineral Zeophyllite, Rad-
zein, Bohemia, $3-$7.50. Bismuthinite, $4-$6. Kulytite, Saxony, $5-$6.
Alexandrite, Ural Mts., xls from $3- $5, matrix specimens from $20-$25,
Graftonite, Grafton, N. EL from $2-$5.

ENGLISH MINERALS

Violet Thane. Durham ; Gray Fluorite; Fluorite and Galena; Yellow
Fluorite, a great variety of different colors; Emerald green Fluorites, very _
rare ; Quartz, studded with fine Fluorites ; ’ Barytes, different varieties and-*
colors, double termination ; Cockscomb Baryte, Frizington ; Fleam Calcite ;
Calcites, different varieties and colors; Calcite, twin forms with Dolomite,
Park House ; Specular Iron with pearl calcite; Hexagonal Calcite, showing
moss structure ; Pyrites, Frizington ; Iridescents, Woodend ; Hausmannite,
Cumberland—this is quite a new local mineral.

SCIENTIFIC RUBIES
We have an extra fine lot of these Scientific Rubies, from 14 to 4 carats
at $5 per carat. Write for further particulars.

ALBERT H. PETEREIT,
81—83 Fulton Street, New York City.

ir. Cyrus Adler,

Librarian U. S. Nat. Museum.

;

i 4

» 43

OCTOBER, 1907.

Mer XXIV:

Established by BENJAMIN SILLIMAN in 1818.

THE

AMERICAN
JOURNAL OF SCIENCE.

Epitorn: EDWARD S. DANA.

ASSOCIATE EDITORS

Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE,
W. G. FARLOW anp WM. M. DAVIS, or Camprivcz,

Proressors ADDISON E. VERRILL, HORACE L. WELLS,
L. VY. PIRSSON ann H. E. GREGORY, or New Haven,

Proressor GEORGE F. BARKER, or PutvapEeLpPHia,
Proressor HENRY S. WILLIAMS, or Iruaca,
Proressor JOSEPH S. AMES, or Baurtiore,
Mr. J. S. DILLER, or Wasutneron.

FOURTH SERIES
VOL. XXIV—[WHOLE NUMBER, CLXXIV.]

No. 142—OCTOBER, 1907.

WITH PLATES I, II.

NEW HAVEN, CONNECTICUT.
1950.7,

THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET.

Published monthly, Six dollars per year, in advance. $6.40 to eouméxies 4a
oe 3 : >) . dU. 20 quixilés Say
Postal Union ; $6.25 to Canada. Remittances should be made eithe /GSsoney ord
registered letters, or bank checks (preferably on New York banks). / ae +
\

que
fe Geese |

SSS SSS SSS

‘the,
BEB

Gh?

IMPORTANT NOTICE.

We have just issued a new 8-page circular, covering almost all our stock
at present date. In addition to this we also issued a complete catalogue of
the Brooklyn Collection which was mentioned in the September number.
In addition to these we have printed a new Cut-Gem Circular, which super.
sedes the other Gem Circular we got out, and gives weight and price of each.
You will be astonished at the quality and low prices of these cut gems.
Write for circular.

NEW ARRIVALS.

Azurite Balls, Bisbee, Arizona, 25c. to $5. Azurite, Copper Queen, Ari-
zona, 50c. to $10. Psilomelane, Inwood and Montreal, Mich., 25c. to $3.50.
Herderite, Poland, Me., 50c. to $20. Malachites, Russia and Arizona, 50c.
to $5. Aragonite, Chessy, sixlings, doubly terminated, 25e: to $1. Stalac-
tites, Arizona, 20c. to $15. Diaspore, crystallized, Chester, Mass., $2.50 to
$5. Wulfenite, Organ Mts., Arizona, $1 to $7.50. Chaleotrichite, Morenci,
7dc. to $d. Calamine, Montana and New Jersey, 50c. to $5. Chalcanthite,
Arizona, 50c. to $4. Topaz, Utah, matrix, $1 to $1.50. Native Antimony,
Prince William, N. B., $1 to $2. Native Copper, Lake Superior, Mich., 50c.
to $5. Native Silver and Copper, 50c. to $12. Native Gold, crystallized,
Hungary, $15, and Nova Scotia, in quartz, $8 to $15. Tourmaline, Mesa
Grande and Pala, Calif., Hamburg, N. J., Haddam, Conn., Paris, Me., 25ce.
to $1.50. Chrysocolla, Arizona and Chili, 50c. to $8. Azurite and Malachite,
Arizona, 50c. to $5. Freibergite, Cobalt, 50c. to $1. Cobaltite, $1 to $5;
loose crystals, 10c. to 25c. Smaltite and Silver, 75c. to $10. Native Silver
on Cobalt, $2 to $10. Erythrite on Native Silver, $1 to $5. Agates, Brazil,
Michigan, Germany, etc., 20c. to $d. Petrified Wood, Arizona, $1.50 to $5.
Opal, Australia, 50c. to $25. Phenacite, Chatham, N. H., 50c. to $2. Topaz,
Chatham, N. H., $1 to $8. Amethyst, 50c. to $5. Amazonstone, 35c. to $5.

ENCLISH MINERALS.

Calcites, nailhead, flame, red, white, pink, many forms and colors, 25e. to
$3.50. Fluorites, green, purple, brown, yellow, 35c. to $7.50. Hausmanite,
new find, 35c. to $2.50. Chalcosiderite, $1 to $5). Hematite, specular with
Pearl Spar, 35c. to $2.50. Dolomite with Calcite Twins, 35c. to $1.50.
Pyrite, iridescent, 50c. to $1.50; rare form, 7dc. to $1.50. Barite, 2dc. 40 $9.

NORWAY MINERALS.
Gadolinite, $8. Hellandite, $5. Monazite crystals, $1 to $5. Rutile, $3.
Malakon, $1. Xenotime, $8. Thorite, $2.50. Apatite, 25c. to 7dc.

OTHER LOCALITIES.

Zinkenite, Nevada, $1.50 to $38.50. Calaverite, $1.50 to $3. Mixite, 50c.
to $1.50. Linarite, 50c. to $1.50. Bixbyite, $5 to $7. Brochantite, 50c.
to $1.50. Zeunerite, Utah, 75c. to $2. Hessite, Hungary, $10 to $25. Ana-
tase, Binnenthal, $4 to $10. Crocoite, Brazil and Dundas, Tasmania, $3 to
$15. Columbite, N. C., $1 to $5. Graftonite, N. H., $1 to $4. Microlite
in Albite, and loose crystals, Amelia, Va., $4 to $7. Phosgenite, England,
$2 to $15. Samarskite, N. C., 50c. to $5. Euclase, Urals and Brazil, -$5 to
$40. Argentite, $5 to $12. Zeophyllite, Radzein, $1 to $5. Enargite, $2.50
to $5. Bournonite, $2 to $5. Tetrahedrite, $1.50 to $10. Cinnabar, $3 to
$10. Pink Beryl, Pala, Calif., $10 to $35.

We still have three of those rare native coppers from Bisbee, Arizona,
described in this Journal of March. Write for further particulars.

CGEM MATRIX SPECIMENS.

Diamond in the matrix, Old Mine, Kimberley, South Africa, $25 to $40.
Ruby in limestone, Burmah, $25 to $35. Alexandrite in matrix, fine crystals,
$20 to $25. Hiddenite in matrix, $10 to $25. Ruby Spinel in limestone,
Burmah, $5 to $10. Ruby Spinel Crystals from Ceylon, 35c. to $2.50.
- Emerald in matrix, Bogota, $10 to $50; Ural Mts., $10 to $50; Habacktel-
thal, Austria, $2 to $10; Alexander Co., N. C., $2 to $10. Rubellite in matrix,
Alabashka. $5 to $15. Lot of Topaz, Ural Mts., crystals and matrix specimens,
$7.50 to $20; lot of crystals, Minas Geraes, Brazil, 75c. to $5; crystals from
Mexico, 25c. to $5; matrix specimens, 50c. to $10; from Siberia, $5 to $10.
Beryls, a variety from different localities, 25c. to $10. Kunzite, $1 to $20.

A. H. PETERETIT,
81-83 Fulton St.; New York City.

Dr. Cyrus Adler,

Librarian U. S. Nat. Museum.

VOLP XXIV. NOVEMBER, 1907.

Established by BENJAMIN SILLIMAN in 1818.

THE

AMERICAN
JOURNAL OF SCIENCE.

Epitork: EDWARD S. DANA.

ASSOCIATE EDITORS

Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE,
W. G. FARLOW anp WM. M. DAVIS, or Camrrincs,

Proressors ADDISON E. VERRILL, HORACE L. WELLS,
L. V. PIRSSON ann H. E. GREGORY, or New Haven,

Proressor GEORGE F. BARKER, “OF PHILADELPHIA,
ProressoR HENRY S. WILLIAMS, or Iruaca,
Proressor JOSEPH S. AMES, or Battrmore,
Me. J. S. DILLER, or Wasuineton.

FOURTH SERIES

VOL. XXIV—[WHOLE NUMBER, CLXXIV.]

No. 143—NOVEMBER, 1907.

107,

NEW HAVEN, CONNECTICUT. |

| THE TUTTLE, MOREHOUSE & TAYLOR ©O., PRINTERS, 123 TEMPLE STREET.

EE TES TE AE ETD DOT DA ESP DIE TTA TREE BET HOE BNE ESTE BES RE TI EON

Published monthly. Six dollars per year, in advance. $6.40 to countries i= the
Postal Union ; $6.20 to Canada. Remittances should be made either by money orders,
registered letters, or bank checks (preferably on New York banks). fa

WHAT WE HAVE DONE, AND WHAT WE INTEND 10 00.

OUR PAST ACHIEVEMENTS.

Another Fall and Winter Campaign in the Mineral Business is before us.
A little review of our work in the past and prospects for the future is not,
therefore, at this time out of place. We are going to ‘‘ blow our horn,” but
will simply state facts.

Two years ago we were almost unknown in the Mineral World; to-day we
are the leading emporium of minerals in America, and the leading firm in
the distribution of foreign minerals. We have in the past two years bought
out and sold wholly or in part thirty-five collections, ranging in value from
$500 to $5,000. We have had exhibitions at the Miners’ Exposition, the
American Museum of Natural History, and the Brooklyn Institute of Arts
and Sciences. We have sold a number of notable and rare minerals, among
them a remarkable Tourmaline Crystal from California, remarkable Kunzite
Crystals, extraordinary Cut Gems, some extremely rare and new German
minerals, etc., ete.

OUR FUTURE PROSPECTS.

The prospects for this Fall were never so promising in the number of
remarkable collections, large or small, we have secured for our patrons.
First and most notable is the consignment of Russian and Hast Indian Gem
Minerals, which will soon be here. This consignment is so large and so
remarkable that no dealer in gems in this country can equal it. Some of
the individual specimens are so fine that it will be almost impossible to
satisfy all our patrons without auctioning them off. But of this more in the
future number.

Next in importance is the “‘ Brooklyn” Collection, so called because the
owner does not wish his name known. This is full of what are called unat-
tainable specimens, that is, specimens from exhausted localities, in so fine
examples that other localities producing the same minerals have never
equalled them in crystallization or beauty. In other words, just those
specimens you have ‘‘hankered after” but could not get, even though you
carried ‘‘a wad of the long green.” This collection is now on sale.

Then next in importance comes the English consignment, comprising a
large shipment from the well-known English localities. It is not necessary
to speak of the great beauty of these specimens, as you are all familiar with
them, except to say they are fresh specimens, never before offered for sale,
and contain several new shades on color in the Fluorite, Calcites, etc.

There are still three other Huropean consignments on the way that we will
describe later.

Of small lots of minerals there is an innumerable number either here, on
the way, or under negotiation. So that we are not exaggerating when we
say that never has there been so many and so fine and rare specimens offered
at one time; and while we would be justified in feeling proud of our success,
we take createst pride in our file of letters from well pleased and satisfied
patrons, both private and public.

Send for our three new circulars describing our stock, our gems, and the
Brooklyn Collection.

A. PETEREI
81—83 Fulton Street, New York City.

Dr. Cyrus Adler,
ee a DECEMBER, 1907.

tl Established by BENJAMIN SILLIMAN in 1818.

THE

AMERICAN
JOURNAL OF SCIENCE.

Epirorn: EDWARD S. DANA.

ASSOCIATE EDITORS

Proressors GEORGE L. GOODALE, JOHN TROWBRIDGE,
W. G. FARLOW anp WM. M. DAVIS, or CampBrwnce,

Proressorns ADDISON E. VERRILL, HORACE L. WELLS,
L. V. PIRSSON anp H. E. GREGORY, or New Haven,

Proressor GEORGE F. BARKER, or PHILADELPHIA,
Proressor HENRY S. WILLIAMS, oF Iruaca,
Proressor JOSEPH S. AMES, or Battimors,

| Me. J. S. DILLER, or Wasuineron.

FOURTH SERIES

VOL. XXIV—[WHOLE NUMBER, CLXXIV.]

id lid a te tut i ee A ven tN

No. 144—-DECEMBER, 1907.

~

NEW HAVEN, CONNECTICUT.
HO! Oee

| THE TUTTLE, MOREHOUSE & TAYLOR CO., PRINTERS, 123 TEMPLE STREET.

Published monthly. Six dollars per year, in advance. $6.40 to countries in the
Postal Union ; $6.25 to Canada. Remittances should be made either by money orders,
registered letters, or bank checks (preferably on New York banks). i, <

Ceaahieraa

NEW ARRIVALS.

These minerals come direct from the mine, not from any dealer, and is
the finest consignment ever received from Europe. There are large and
small specimens, some excellent for Museums, and all at very reasonable
prices. Those from Hungary are the most important. On account of the
large number of specimens and in order to introduce them to you, we will
have Bargain Sales every day this month, so do not fail to come and see
them. If not convenient for you to call, we would be pleased to send the

goods on approval. A few of them we name below: E

Rhodochrosite, Kapnik, 50¢ to $5; with Quartz and Calcite, Kapnik
d0¢ to $5; with pink quartz, 50¢ to $5; Barite, different-colors, Felsébanya,
with stibnite, realgar, etc., 7o¢ to $5; Stibnite, Felsébanya, star groups,
fine sarge groups, with stout crystals on smoky quartz, 75¢ to $10; Plumos-
ite, Felsdbanya 75¢ to $4; Calcite, pink, from Borpasak, Kapnik, and
Feketebanya, 75¢ to $2; Pyrite, Felsobanya, and Kapnik, with bournonite and’
braunspar, 25¢ to $3; Fluorite, Kapnik, lilac color, 50¢ to $8; Bournonite,
Kapnik and Felsébanya, with tetrahedrite, pyrite, sphalerite, etc., 50¢ to $d;
Cinnabar, Kapnik, $1 to $4; Chalcopyrite, Felsobanya, and Kapnik, with
wurtzite, sphalerite, braunspar, etc., 75¢ to $38; Pearlspar, Kapnik, 70¢ to
$2.50; Libethenite, Libethen, $2.50 to $5; Sphalerite, Kapnik, and Felso-
banya different colors, with chalcopyrite, galena, tetrahedrite, ete., 50¢ to $4 ;
Galenite, Felsbbanya, and Kapnik, with tetrahedrite, calcite and braunspar,
50¢ to $4; Galena, twin, Rodna, $1 to $1.50; Tetrahedrite, Kapnik with
bournonite, sphalerite and chalcopyrite, $1 to $5; Marcasite, Felsébanya,
50¢ to $2; Braunspar, different tints, Felsdbanya, $1 to $2.50; Quartz,
different colors, Felsébanya, with citrine, braunspar, and chalcopyrite,
50¢ to $2.50; Sphereosiderites with stibnite, Felsdbanya, $5 to $7.00,
Amethyst, Nagybanya, with Marcasite, 50¢ to $3; Helvite, Kapnik, $1 to
$2.50; Realgar, Felsébanya, beautiful crystals, $1 to $5; Chalcedony,
Trestia, different colors, 50¢ to $7.50; Semseyite, very rare, with Galenite,
Felsébanya, $5 to $12; Gypsum, crystals beautiful, Diembrava, 75¢ to $2.50 ;
Grossularite, Vasco, $1 to $4; Cerussite, Rodna, 75¢ to $2.50 ; Greenockite,
Dognacska, $3 to $5; Sylvanite, Nagyag, $5 to $7.50; Hessite, Botes, $20
to $35; Gold, beautifully crystallized and in leaves, some in matrix, $2 to
$15 ; Topaz, Schneckenstein, Saxony, in matrix and loose crystals, 20¢ to $3.

We have a very fine lot of all the known gems on hand, which will be
suitable for a Christmas Gift; write for our gem circular.

There are a number of other important consignments on the way, one
of which is now in the Custom: House, and will be on exhibition and for
sale at the same time as above. Some of these are so extremely rare that
they will repay a long trip to see. Further particulars cheerfully furnished.

A. H. PETEREIT,
81—83 Fulton Street, New York City.

63 Le

oes,

Ad.
baa tt

oer

SMITHSONIAN INSTITUTION LIBRARIES
“AACA
3 9088 01298 5750