h^(rJt>/t'^<^^

PROCEEDINGS

AMERICAN ACADEMY OF ARTS AND SCIENCES.

PROCEEDINGS

AMERICAN ACADEMY

ARTS AND SCIENCES.

NEW SERIES.

Vol. XXIII.

WHOLE SERIES.
Vol. XXXI.

FROM MAY, 1895, TO MAY, 1896.

SELECTED FROM THE RECORDS.

BOSTON:

UNIVERSITY PRESS: JOHN WILSON AND SON.

1896.

<(i'*

2>

li'l 0

CONTENTS.

Page
I. On the Composition of the Ohio and Canadian Sulphur

Petroleums. By Charles F. Mabery 1

II. On the Occlusion of Baric Chloride by Baric Sulphate. By
Theodore William Richards and Harry George
Parker 67

III. On the C upriammonium Double Salts. Third Paper. By

Theodore William Richards and George Oenslager 78

IV. ' On the Cuprianiline Acetobromides. By Theodore William

Richards and Frederic Charles Moulton ... 87

"V. The Chemical Potential of the Metals. By Wilder D.

Bancroft 96

VI. On the Behavior of Certain Derivatives of Benzol containing

Halogens. By C. Loring Jackson and Sidney Calvert 123

VII. Bromine Derivatives of Metaphenylene Diamine. By C. Loring

Jackson and Sidney Calvert 136

VIII. A Revision of the Atomic Weight of Zinc. First Paper: The
Analysis of Zincic Bromide. By Theodore William
Richards and Elliot Folger Rogers 158

IX. Note on the Automorphic Linear Transformation of a Bilinear

Form. By Henry Taber 181

X. Thermo-electric Interpolation Formulce. By Silas W. Holman 193

XI. Melting Points of A luminum, Silver, Gold, Copper, and Platinum.

By S. W. Holman, with R. R. Lawrence and L. Barr 218

VI CONTENTS.

Page
XII. Pyrometry : Calibration of the Le Chatelier Thermo-electric

Pyrometer. By Silas W. Holman 234

XIII. Calorimetry : Methods of Cooling Correction. By Silas W.

IIOLMAN 245

XIV. On some Points in the Development of JEcidia. By Herbert

Maule Richards 255

XV. On the Thermal Conductivity of Mild Steel. By Edwin H.

Hall 271

XVI. The Outline of Cape Cod. By William Morris Davis . 303

XVII. Preliminary Notes on the Embryology of the Starfish {Aste-

rias pallida'). By Seitaro Goto 333

XVIII. On the Group of Real Linear Transformations ichose Invariant

is an Alternate Bilinear Form. By Henry Taber . . 336

Proceedings 339

Biographical Notices: —

Richard Manning Hodges 355

Harold Whiting 356

Edward Samuel Ritchie 359

Martin Brimmer 360

Henry Wheatland 363

James Edward Oliver 367

Viscount Ferdinand de Lesseps 370

List of the Fellows and Foreign Honorary Members . . 385

Statutes and Standing Votes . 393

Index *^'^

PROCEEDINGS

OF THE

AMEHICAN ACADEMY

OF

ARTS AND SCIENCES.

VOL. XXXI.
PAPERS READ BEFORE THE ACADEMY,

Aid in the work desoribed in this Paper was given by the Academy from the
C. M. Warren Fond for Chemical Research.

I.

CONTRIBUTION FROM THE CHEMICAL LABORATORY OF
CASE SCHOOL OF APPLIED SCIENCE.

XX. — ON THE COMPOSITION OF THE OHIO AND
CANADIAN SULPHUR PETROLEUMS.

By Charles F. Mabery.

Presented October 10, 1894.

Notwithstanding the great number of investigations that have
been undertaken on petroleums from different countries, our knowl-
edge of their composition, especially of the portions with high boiling
points, is still incomplete. What is known concerning American
petroleums is based chiefly on the results of investigations which were
carried on thirty years ago, before the discovery of several series of
organic compounds, since shown to be contained in certain petroleums.
With these facts in view, and having found in the study of the sulphur
compounds from Ohio and Canadian petroleums that these oils are in
certain respects peculiar in their composition, I have undertaken a
comprehensive examination of them with reference to the different
series of products which they may be found to contain. The study of
Ohio petroleum seemed especially inviting on account of its origin in
the Trenton limestone, a new horizon for a prolific oil supply, and
its associations with animal remains in the oldest geological forraa-

VOL. XXXI. (n. s. xxiii.) 1

2 PROCEEDINGS OF THE AMERICAN ACADEMY.

rions of the contiuent. A similar interest attaches to the Cana-
dian petroleum from the Coruiferous limestone, which has yielded
since 18G2 enormous quantities of oil within the limited areas, less
than thirty square miles, at Petrolia and Oil Springs. From the in-
formation already acquired concerning the nature of the sulphur petro-
leums, they seem to possess, beside their distinctive characteristics due
to sulphur constituents, qualities of other petroleums which differ
essentially in their composition.

In the earlier attempts to ascertain the constituents of petroleum,
the methods then employed for fractional distillation were so inade-
quate that very little was accomplished. In 1862 the first systematic
examination of American petroleum was undertaken by Pelouze and
Cahours, * who showed the presence of the series of hydrocarbons
C„H2„_i_2? beginning with butane. On account of a want of suitable
apparatus for fractional distillation, their results lacked precision, and
the questionable assumption was based upon them that jDetroleura is
composed principally, including the heavier oils and paraffine, of the
homologues of marsh gas. In accordance with the suggestion of
Watts, these bodies were called the paraffine hydrocarbons, and as
such they have since been known in chemical literature. Having
obtained from coal oil a series of hydrocarbons corresponding to cer-
tain members of the series discovered by Pelouze and Cahours, Schor-
lemmer f submitted the more volatile portions of American petroleum
to distillation, and succeeded in separating hydrocarbons that had not
been recognized by Pelouze and Cahours. In a more thorough and
carefully conducted examination of Pennsylvania petroleum, carried on
contemporaneously with the investigations of the chemists mentioned
above, by means of an efficient fractional condenser devised especially
for this and other similar investigations, C. M. Warren t avoided the
errors of other experimenters and established beyond question the
presence in Pennsylvania oil of two series of hydrocarbons, each with
an homologous difference in boiling points for CHo of 30°, and each
member of one series differing in boiling point from the isomeric
member of the other series by a little less than 8°. One of the series
C„H2„ + 2 identified by Warrren terminates at 127°. 6, the other at
150°, the fractions of higher boiling points containing members of the
series C„H2„. The assumption of Pelouze and Cahours that the frac-

* Comptes Rendus, LIV. 1241, LVI. 505, LVII. 62.

t Journ. Chem. Soc, 1862, p. 419.

{ Mem. Amer. Acad. (N. S.), IX. 135; Proc. Amer. Acad., XXVII. 56.

MABERY. — SULPHUR PETROLEUMS. 3

tions with higher boiling points have the composition represented
by the general formula C„H2„ + 2 was shown to be erroneous by
the results of Warren, which excluded members of this series
above 151°.

The presence of aromatic hydrocai'bons in American petroleum was
first recognized by Schorlemraer in 1865. Pelouze and Cahours had
previously stated that American petroleum contained no aromatic
hydrocarbons, but Schorleramer collected a distillate from Canadian
petroleum below 150°, and upon treating it with nitric acid and reduc-
ing with tin and hydrochloric acid, after distillation he obtained an oil
with an odor of aniline that gave, with bleaching powder, the rosani-
line reaction. The portion distilling between 150° and 170° gave a
mixture of solid and liquid nitro-products, and the solid portion proved
to be trinitrocumol. Benzol and its homologues were also found by
Schorlemmer in Pennsylvania petroleum, in Galician petroleum by
Freund* and others, and in Hanover petroleum by Bussenius and
Einstuck.f In Galician petroleum Pawlewski$ found two per cent of
aromatic hydrocarbons, chiefly benzol and paraxylol, the latter never
having previously been recognized in any petroleum. In the fraction
170°-190° from American kerosene, Engler § discovered pseudocumol
and mesitylene by the formation of nitro-com pounds, and calculating
the weight of crude oil corresponding to the weight of kerosene taken,
it was estimated that these constituents are contained in crude Penn-
sylvania oil to the extent of 0. 2 per cent. These hydrocarbons were
also found in German, Galician, Italian, and Russian petroleum, in the
latter to the extent of 0.1 per cent. According to the results of Beil-
stein and KurbatoflT, || the petroleum of the Central Caucasus has an
essentially different composition from that of the deposits on the coast
of the Caspian Sea. Oil obtained from the region of Zarskige
Kolodzy, in the precinct of Tiflis in the Central Caucasus, proved to
contain small amounts of benzol and toluol, but to consist jirincipally,
like the Pennsylvania petroleum, of the series C^,H2„+2. Pentane,
hexane, and heptane were identified. The oil from Baku on the
coast contains the series C„H2„ + 2 in smaller quantity, and Beilstein
and KurbatofF found no trace of the aromatic hydrocarbons C„H2„_s,
but the principal constituents are members of the series C„H2„.

* Ann Chem. Pharm., CXV. 19. t Ibid., CXIIL 151.

t Ber. der deutsch. chem. Gesellsch., 1885, p. 1915.

§ Ibid., 1885, p. 2234.

II Ibid., 1880, p. 1818 ; 1881, p. 1620.

4 PROCEEDINGS OF THE AMERICAN ACADEMY.

In Pennsylvania petroleum, Beilstein and Kuibatoff* recognized
hexahydroisoxylol. In the fraction 95°-100° the same chemists de-
tected the presence of a body containing less hydrogen than is required
for the series C„H2„+2. Crude American heptane, when treated with
nitric acid, gave a nitro-product corresponding to the formula C7H15NO2.
Nearly coincident with the researches of Beilstein and Kurbatoff,
Schutzenberger and Jonine f identified, in the petroleum of Baku,
hexahydrobenzol and hexahydrotoluol.

In their classic researches on the composition of the Caucasus
petroleum, Markownikoif and Ogloblin f showed the presence, in the
oil from Baku, of benzol, toluol, isoxylol, pseudocumol, mesitylene,
isodurol, durol, and higher hydrocarbons of the composition C11H14,
C12II14, C11H12, and C13H14, and others, possibly homologues of
styrol and phenylacetylene. Contrary to the experience of others,
Markownikoff and Ogloblin found the naphtha from the Balachani
plain on the Apscheron peninsula very rich in aromatic hydrocarbons ;
in the oil from the Central Caucasus, naphtenes were found to the
extent of 80 per cent, and the aromatic hydrocarbons C„H2„_6 to
the extent of 10 per cent. At first Markownikoff looked upon
the naphtenes as isomeric with the hexahydro compounds, but later
Markownikoff and Spady § appear to accept the identity of octonaph-
tene and hexahydroisoxylol. || As members of the naphtene series
C„H2„, Markownikoff and Ogloblin identified hexahydromesitylene
and the higher homologues between CgHig and C15H30. In petro-
leum from Boryslaw in Galicia, beside benzol, toluol, isoxylol, and
mesitylene, Lachowicz 1[ found of the hexahydro series only hexa-
hydroisoxylol.

The presence of unsaturated hydrocarbons C,,!!,,, in American
petroleum was not mentioned in the early publications of Pelouze
and Cahours, nor were they referred to by Schorlemmer. Warren **
separated from Pennsylvania petroleum, rutylene, C10H20, boiling
point 174°. 9, margarylene, C11H22, boiling point 195°. 8, and laurylene,

* Eer. der deutsch. cliem. Gesellsch., 1880, p. 2028.

t Comptes Rendus, XCI. 823.

t Ann. Chim. Phys., [6 ], II. 372.

§ Ber. der deutstch. chem. Gesellsch., 1887, p. 1850.

II In the Berichte, No. 7, April, 1895, recently received, the synthesis of 1-3-
dimetliyl hexametliylen, CeHjo(CH3), is described by Zellnsky, and it is shown
to be identical with the octonaphtene of Markownikoff, or he.xahydroisoxylol.

1 Ann. Chem Pharm., CCXX. 187.
** These Proceedings, XXVII. 56. Communicated May 12, 1868.

MABKEIY. — SULPHUR PETROLEUMS. 5

C12H24, boiling point 216'^. 2. In an examination of a hydrocarbon
naphtlia obtained as a product of the destructive distillation of a lime
soap prepared from menhaden oil, Warren and Storer* discovered the
series C„H2„ + o, beginning with C5H12, members of the aromatic
series including benzol, toluol, xylol, and isocumol, and a series of
the general formula C„H2„ as follows : rutylene, CiqH^o, margarylene,
C11H20, laurylene, C12II04, identical with the hydrocarbons previously
separated by Warren from Pennsylvania petroleum. Warren and
Storer also submitted Rangoon petroleum to prolonged distillation,
and they succeeded in proving the presence in this oil of rutvlene,
margarylene, laurylene, cocinylene, and naphthalene.

As early as 1842, Pelletier and Walther separated a hydrocarbon
from the " steinol " of Amiano boiling at 80°-88°, to which they
assigned the formula C7H14. A similar product with the properties of
heptylene was obtained by Mabery and Smith t as one of the products
in the sulphuric acid extract from the refining of Ohio burning oil
distillate.

More recently Engler J subjected menhaden oil to distillation under
a pressure of ten atmospheres, and from the distillates thus obtained
hydrocarbons were separated which proved to be identical with those
contained in natural petroleum. These products included the light
constituents of gasoline, the hydrocarbons of which burning oil is
composed, the heavy oils, and paraffine. Engler, therefore, believes in
the origin of petroleum from the decomposition of animal remains.
The hydrocarbons distilling above 160°, referred by Pelouze and
Cahours to the series C,jH2,j + 2 D^ay really be naphtenes or similar
bodies. In most petroleums more recently examined, it is believed
that the unsaturated hydrocarbons are not present in the crude oil, but
when found in the products of distillation have resulted from decom-
position. Beilstein and Kurbatoff stated that the se<"ies C^Ha^i ^^ the
Caucasus petroleum does not consist of the homologues of ethylene.
In the lower fractions of Galician oil, Lachowicz obtained no reaction
with bromine even after long standing. Above 200° the ready absorp-
tion of bromine indicated the presence of unsaturated hydrocarbons ;
but it was attributed to decomposition. On the other hand, Engler §
found that petroleums from Alsace (Pechelbronn), Oelheim (Hanover),

* Mem. Amer. Acad. (N. S.), IX. 177.

t These Proceedings, XXV. 222.

I Ber. der deutsch. chem. Gesellsch;, 1888, p. 1816.

§ Zeit. Ang. Chem., 1888, p. 73.

6 PROCEEDINGS OF THE AMERICAN ACADEMY.

Tegernsee, Pennsylvania, Galicia, and Baku contain members of the
series C„H2„4.2 and C„H2,„ both unsaturated hydrocarbons and naph-
tenes. Engler asserts that all petroleums have the same composition,
differing only in the proportions of their constituents. Markownikoff
and Oglobliu * suggest the presence of unsaturated aromatic hydrocar-
bons in the oil from Baku.

Most analyses of crude petroleum have fallen short of 100 per cent,
and the deficiency has been attributed to the presence of oxygen. In
ascertaining the composition of oils from different districts, St. Claire
Devillef showed that the percentage of oxygen varies between 2.1 per
cent in the Canadian petroleum and 5.6 per cent in the oil from Zante.
In 1874 Hell and Medinger t attempted to separate acids from crude
petroleum by agitating it with a solution of sodic hydrate and precipi-
tating with sulphuric acid. The oil which separated was distilled and
converted into a methyl ether. By saponification of this ether an acid
was obtained to which was assigned provisionally the formula
C11H20O2. From the sodic hydrate solution used in the refining of
Baku oil, by the addition of sulphuric acid, Aschan § separated a mix-
ture of acids, and, taking advantage of a difference in solubility of
their salts, he obtained one acid, C8H14O0, distilling at 237°-239°, and
another, CgHieOg, that distilled at 251° -253°. The first acid, called
by Aschan octonaphtenecarboxylic acid, by distillation with hydri-
odic acid, was converted into hexahydroisoxylol.

The oxygen compounds in petroleum were considered by Zalo-
ziecki || to be lacto-alcohols which are oxidized to acids by contact
with air. By the same method which was followed by Heil and
Medinger, and Aschan, Zaloziecki obtained an acid, CioHigOa, to which
he gave the formula,

CHo. /CHo.

CH-(CHo)4-CH ^CHOH,

CHs-^ \ O /

and to the hydrocarbon obtained by distillation with hydriodic acid he
gave the formula,

C[I-(CH.,)4-CH II

* Ber. der deutsdi. cliem. Gesellsch., 1883, p. 1873.
t Comptes Rerulus. LXVI. 442 ; LXVIII. 485.
} Ber. der deutsch. chem. Gesellsch., 1874, p. 1216.
§ Ibid., 1890, p. 867 ; 1891, pp. 1864,2710; 1892, p. 3661.
II Ibid., 1891, pp. 796, 1808.

MABERY. — SULPHUR PETROLEUMS. 7

Engler looks upon these acids as formed by oxidation of other con-
stituents of the crude oil. According to the results of Markownikoff
and Ogloblin,* the fraction 7o°-8o° from Caucasus oil contains 0.76
per cent of oxj^gen com[)Ounds, and the fraction 220°— 230°, 5.21 per
cent. These oxygen compounds are in part acid, in part neutral, and
in part phenol. The acids CioHiaCOOH and CnHjiCOOH were ob-
tained as colorless oils ; f Markownikoff and Ogloblin regarded these
substances as naphtene carboxylic acids.

Most petroleums have been shown by analysis to contain nitrogen,
usually in minute quantities. In distillates from Pennsylvania oil,
Beilby t found 0.U8 per cent in the residuum or tar, and 0.375 per cent
in the coke. Since the tar was one tenth of the crude oil, the latter
contained 0.008 per cent of nitrogen. In crude Russian oil Beilby
found 0.05 per cent. Peckham § found in West Virginia oil 0.54 per
cent, in Mecca oil, 0.23 per cent, and in California oil, 0.56-1.1 per
cent of nitrogen. In Egyptian oil, Kast and Kunkler || reported 0.3
per cent of nitrogen, 1.21 per cent of sulphur, and 0.92 per cent of
oxygen. Weller IT detected certain alkaloid bases in paraffine oil, and
Bandrowski** described a thick, transparent liquid solidifying at 20°,
■which he obtained by agitating Galician oil during several weeks with
sulphuric acid. This substance gave a platinum salt containing 19.7
per cent of platinum. Upon neutralizing a sulphuric extract obtained
in refining with calcic hydrate and distilling with steam, Zalozieckift
obtained an oil containing nitrogen that formed a platosochloride
whose percentage composition corresponded with that of tetrahydro-
corridine (CloHigNCOaPtCIa, or to the formula (doHaiNCOoPtClg ;
another insoluble platoso compound was shown by analysis to have
the composition represented either by the formula (CioHi5NCl)2PtCl2
or (CioHi.NCO^PtCl^.

Several hydrocarbons have been described as occurring in the less
volatile portions of American petroleum. Hemilian :i:| obtained from
the high boiling fractions "petrocene," a crystalline body melting

* Ann. Cliim. Phys., [6.], II. 372.

t Ber. der deutsch. chem. Gesellsch., 1883, p. 1873.

t Journ. Soc. Chem. Ind., 1891, p. 120.

§ Geological Survey of California, Appendix to Vol. XL., p. 89.

II Chem. Centralb., 1890, p. 932.
1 Ber. der deutsch. chem. Gesellsch , 1887, p. 2097.
** Monatsheft fur Chem., VIII. 224.
tt Ibid., XIII. 498.
Jt Ber. der deutsch. chem. Gesellsch., 1876, p. 1604.

8 PROCEEDINGS OF THE AMERICAN ACADEMY.

above 300°. To this substance was assigned the formula C32TI22-
Prunier and David * stated that they had obtained evidence of the
I^resence in refinery residues of anthracene, chrysene, pyrene, phenan-
tiirene, chrysogene, retene, benzerythrene, fluoanthrene, parachrysene,
paraanthracene, an isomeric acenaphtylene, besides different parafilnes
and stilbene. In the " petrocene " and " carbopetrocene " prepared
from the tarry residue of petroleum, Prunier and David identified
compounds, the former melting at 1G0°-190°, the latter at 200°-238°,
and also thallene melting at 110°. They attributed the green fluo-
rescence in petroleum to the presence of chinones. Inasmuch as these
products were obtained from the distillation of tar in coking, evidently
no inference is permissible concerning their presence in the crude oil.
From " petrocene," Sadtler and McCarter f separated two hydrocar-
bons, one of which melted at 280° and the other at 178°. From these
hydrocarbons the chinones were prepared, and from one an alizarine.

In undertaking an examination of the Ohio and Canadian sulphur
oils with the advantage of former experience in studying the sulphur
constituents, it was evidently necessary to conduct the distillations
with all possible precautions to avoid decomposition. The tendency
of the sulphur oils to decomposition by heat is due as much at least
to the action of air on the hot oil as to the increased temperature.
Markownikoff and Ogloblin attributed the decomposition in the dis-
tillation of the Russian oil to polymerization of the unsaturated com-
pounds, and perhaps also to the polymerization of certain aromatic
compounds, such as phenylacetylerie, and they found that coloration of
distillates on standing was less marked when the oxygen compounds
had been removed. This tendency towards polymerization in unsatu-
rated hydrocarbons separated from sulphuric acid solutions was ob-
served by me, t and it will receive further attention. It was there-
fore deemed advisable to carry on all distillations from the crude oils,
instead of relying upon refinery products, except only the most vola-
tile distillates, and the advantage gained has been apparent in subsid-
iary distillations of refinery oils which had been subjected to the
decomposition incident to refinery distillation. Longer time is then
necessary for separations, and the odor of decomposition is retained
indefinitely during subsequent distillations. Certain constituents of
the sulphur petroleums are even more unstable than the suljjhides, as

* Bull. Soc. Chim., [2.], XXXI. 158.
t Amer. Chem. Jouru., 1879, p. 30.
t Ibid., 1893, p. 92.

MABERY. — SULPHUR PETROLEUMS. 9

shown by the rapid coloration of the oils when distilled, even after
removal of the sulphides by mercuric chloride. Nevertheless, in the
separation of coi^stituents requiring larger quantities of distillates than
can conveniently be collected in vacuo, since only porcelain or earthen
ware stills are admissible on account of the decomposition in metallic
stills, it may be necessary to depend to a limited extent upon refinery
distillates.

After the first distillation of the crude oils in vacuo, distillation of
the portion collected below 150° was continued under atmospheric
pressure, since it occasioned no appreciable decomposition. Under the
diminished pressure some loss of the constituents with low boiling
points could not be avoided ; but this was not important, since refinery
distillates could be utilized for the separation of the volatile hydrocar-
bons. Distillation of considerable quantities of oil in vacuo presented
certain difficulties. Neither glass nor metallic stills were suitable, and
no American earthen ware stills could be procured that would support
a vacuum on account of porosity from imperfect moulding and glazing.
Some of the English earthen ware has supported a vacuum, but the
ideal stills for such work, or for any distillations in large quantities,
when metals cannot be employed, are those manufactured in the
Royal Berlin Porcelain Factory. We have had a three-gallon porce-
lain still and several others of smaller capacity in operation almost
continuously during several months with apparently no deterioration.
When these stills are enclosed within a brick chamber, the tempera-
ture of distillation may, without difficulty, be carried above 350° by
means of large laboratory burners. Another serious obstacle imme-
diately presented itself in the porosity of common corks, which alone
could be used for connections. After much labor it was found that
tight joints could be secured by means of a rubber lute made by dis-
solving gum rubber in very light gasoline. Thin films of this lute
drawn by the inward pressure into the joints and imperfections in the
corks, after several applications, formed sufficiently close connections.
Any other than the lightest gasoline as a solvent leaves a sticky film
that is unpleasant to manipulate.

An important feature in prolonged fractional distillations in vacuo is
a simple and convenient means for maintaining, without too close
attention, a constant tension within the still. Air must not come in
contact with the hot oil or vapor, and it would require too large quan-
tities of an inert gas. An expedient suggested itself in the fact that
occasionally small leaks held the manometer column stationary at what-
ever height it happened to stand. It therefore seemed possible to

10

PROCEEDINGS OF THE AMERICAN ACADEMY.

graduate leaks apart from the still in such a manner that the tension
could be held constant for some time at any desired point.

The accompanying figure represents the form of regulator that has
been used in all our distillations of large as well as small quantities of
oils, together with the complete apparatus in the form for use. The
regulator consists simply of a glass stopcock, A, better of considerable
size, attached to the manometer by means of a side tube. To regu-
late closely the inflow of air an arm three to five feet in length,

according to the working of the stopcock, is attached firmly to the
head of the cock, and supported in a manner easily movable within
very small divisions on the arc of a circle of which it is the radius.
To enable the operator to make adjustments while standing in front of
the manometer, a piece of lead is attached as a weight to the upper
end of the lever, and a cord is carried over a pulley, B, and terminates
in a ring in front of transverse rows of pins a few millimeters apart.
For economy of space the pulley is placed lower in the figure than its
actual position. The upper part of the lever consists of two strips of
wood, with a space between, through which passes a rigid copper wire
as a support and guide. With the lever in a vertical position, the

MABERY. — SULPHUR PETROLEUMS. 11

stopcock is fully open ; any adjustment is easily obtained, and the
manometer may be held stationary witliin one millimeter at any
desired point during several hours.

In the distillation of small quantities of liquids requiring constant
attention, we have used a piece of glass tube forty-five to sixty milli-
meters long, attached to the side stopcock, with a sliding support near
the end. In vacuum distillations on a large scale it is more convenient
to refer to the entire length of the manometer column, since at any
time leaks may occur that are indicated only on the lower portion of
the graduated scale. Wiih a short manometer column alone, much
time may be lost in waiting for an exhaustion that is interfered with
by leaks.

In heating the still, direct contact with the flame was prevented by
maintaining an air space above the burner by means of a sheet of
asbestos. With such application of heat equally on the sides and bot-
tom, there was less danger of decomposition at high temperatures.
All but the highest distillates were collected in a Warren hot con-
denser containing a glass worm. With this condenser vacuum distilla-
tions are easily controlled, and, as in distillations under atmospheric
pressure, with a great economy in time. Continuous distillation is
possible without losing the vacuum, by drawing in consecutive frac-
tions through the rear tubulure of the retort. The two receivers
shown in the figure are convenient, and some additional advantage
would be gained by means of an independent vacuum connection with
the lower receiver. Several supports are not represented in the
figure. Much time and tedious labor were expended on this apparatus
before all the difficulties were overcome ; but the compensation was
ample, since by means of it we have been able to separate in con-
siderable quantities constituents of high boiling points without decom-
position, which otherwise would have been impossible. As an evidence
of its usefulness, during several months continuous distillations were
in progress, in charge of assistants, with highly satisfactory results.
Distillates were collected at intervals of 10°, 5°, and 2°. The depres-
sion in boiling points by the diminished pressure in vacuum distillation
varies between 60° and 65° for the lower constituents, and 125° or
more for those collected between 300° and 350°. The residue above
350° (450°-500° atmospheric pressure) in both Ohio and Canada oils
had apparently undergone but little decomposition ; in appearance, it
was a thick, ready flowing oil, with scarcely any odor.

A portion of the residual oil above 350° in vacuo was redistilled in
an ordinary boiling flask, and the temperatures of the vapor and of

12 PROCEEDINGS OP THE AMERICAN ACADEMY.

the liquid were read on a Mahlke 550° thermometer. The oil
began to distil with much decomposition at 385° in the vapor, and
415° in the ]i(|uid. One half distilled with the temperature in the
liquid below 430'^. Doubtless the temperatures of distillation were
much reduced by the decomposition. Since we liave found that dis-
tillations below 250° may be carried on successfully without serious
decomposition in an atmosphere of carbonic dioxide, when we return
to the distillation of the higher constituents this method may be
serviceable.

Ohio Petroleum.

In the study of Ohio petroleum I have been aided by Mr. E. J.
Hudson.* Besides the publications from this Laboratory on the sulphur
compounds f and those of Orton, t I am aware of no published state-
ments concerning the composition of Ohio sulphur petroleum. The
crude oil which formed the basis of our work was received from
the Peerless Refining Company, Findlay, which controls a large sec-
tion of oil territory. When received, it was of a somewhat thicker
consistency than ordinary Pennsylvania oil, with a slight odor of
hydric sulphide that is usually observed in crude sulphur oils. It con-
tained a small quantity of water, which was removed completely only
after long standing with fused calcic chloride. In determining the
specific gravity of Ohio sulphur petroleums, oils were collected from
wells at different points in the Findlay and Lima fields. The deter-
minations were made at 20°.

Findlay Field.

Specific Grayity.

(1) Barnsville 0.8272

(2) Heilstone Oil Co., Well No. 2, Hancock County 0.8296

(3) Ohio Oil Co., Wood County 0.8194

(4) Langmade Well, No. 4, Portage, Hancock County 0.8149

(5) Peerless Refining Co., Well No. 2, Liberty, Han-

cock County 0.8278

(6) Peerless Refining Co., Well No. 5, Baltimore,

Wood County, 0.8239

* A part of this work formed the subject of a tliesis by Mr. Hudson for the
degree of Bachelor of Science

t These Proceedings, XXV. '218; Amer. Chem. .Journ., XVI. 83.

t Ohio State Geological Reports, 1880, 1888, 1890 ; U. S. Geological Report,
1886-87.

MABERY. — SULPHUR PETROLEUMS.

13

Lima Field.

The following specimens of crude oil,
tions of the Lima and Findlay fields, were
Company, with a history of each well : —

representing different por-
received from the Ohio Oil

Township.

County.

Drilled.

Sand.

Oil.

Depth.

Production
after shot.

Production
at present.

(1)

St. Mary's

Auglaize

May, 1892

ft.
1132

ft.
1147

ft.
1166

Daily.
100 bbls.

Daily.
5 bbls.

(2)

Liberty

Hancock

Aug., 1893

1250

1273

1286

75 "

10 "

(3)

Montgomery

Wood

Oct., 1892

1192

1197

1242

25 "

5 "

w

Woodville

Sandusky

July, 1892

1180

1195

1205

125 "

10 "

(5)

Nottingham

Wells, Ind.

April, 1895

990

1004

1044

165 "

75 "

(6)

Liberty

Wood

Sept., 1894

1152

1172

1227

125 "

30 "

(7)

Lima

Allen

(1) (2) (3) (4)

Sp. gr. 0.8288 0.8345 0.8265 0.8254

(5) (6) (7)

0.8244 0.8428 0.851

There is evidently an appreciable variation in the composition and
properties of the oils from different points in the Ohio field. Marked
differences occur in the specific gravity as well as in the percentages
of carbon, hydrogen, nitrogen, and sulphur.

The oil that was employed in this examination was somewhat
heavier, with a specific gravity 0.8380. These results show an ap-
preciable variation even in different portions of the same field. A
similar variation in specific gravity has been observed in other fields : —

Alsace (Pechelbronn), depth 146 feet (Engler)

u (' ii 213 " "

Oelheim (Hanover) "

Tagernsee "

Pennsylvania "

u a

Galicia ♦'

Baku *'

Ohio

(Mabery)

Specific Gravity.

0.906

0.885

0.889

0.815

0.8185

0.801

0.8235

0.859

0.810

0.838

14 PROCEEDINGS OF THE AMERICAN ACADEMY.

Specific Gravity.

Ohio (MarkowiiikofF and Ogloblin) 0.887

Baku (Apscheron) " " " 0.855-0.885

Galicia " « " 0.835-0.895

American Petroleum (Petrolitz) 0.830

(Weil) 0.827

Canada (Markownikoff and Ogloblin) 0.828

" " "■ " 0.844

Alsace (Pechelbronn) « " " 0,668

The value 0.887 assigned by Markownikoff to Ohio oil is much
higher than has elsewhere been given. It must have been obtained
in an oil from another Ohio field, perhaps from the Mecca district.
The values given by Markownikoff for Canada oil must refer to a
product from the Oil Springs district, although the number 0.828 is
lower than is usually found even in that oil. The numbers given by
Redwood are 0.844-0.854 for Oil Springs oil, and 0.859-0.877 for
Petrolia oil. In Oil Springs oil we found 0.8427-0.8389 (gas oil),
0.8442, and in Petrolia oil, 0.8553, 0.8621, 0.8800.

In ascertaining the quantity of sulphur by combustion in air in the
crude oil from which distillates were prepared for examination, the
following results were obtained: (1) 0.73 per cent, (2) 0.72 per cent,
(3) 0.72 per cent. In oils previously examined the percentage of sul-
phur has not been above 0.60 per cent. Sulphur was also determined
in the oils collected in the Findlay and Lima districts in the order of
the numbers given above: —

Findlay.

Lima and Findlay.

Findlay

Lima and Findlay.

(1) 0.33

0.61

(5) 0.68

0.56

(2) 0.63

0.71

(6) 0.61

0.76

(3) 0.56

0.37

iV

0.81

(4) 0.68

0.49

In Apscheron oil, Markownikoff and Ogloblin* obtained 0.064 per
cent, and 0.16 per cent in Trans-Caspian oil.

The percentages of carbon and hydrogen in the oil from which dis-
tillates were obtained were found by combustion in air with a layer
of plumbic peroxide in front to retain the sulphur f: carbon, 84.57;
hydrogen, 13.62. In other samples from Findlay and Lima carbon
and hydrogen were also determined : —

* Ann. Cliem. Phys., [6.], II. 393. t Warren, These Proceedings, VI. 472.

c.

II.

(1)

85.76

13.56

(2)

85.82

13.80

(3)

84.33

13.46

(4)

84.35

13.36

(5)

84.20

13.41

(6)

84.18

14.60

(7)

MABERY. — SULPHUR PETROLEUMS. 15

Findlay. Lima and Findlay.

C. II.

84.73 13.48

84.03 13.05

83.89 13.18

84.55 13.55

83.41 13.13

85.07 13.33

85.00 13.05

The percentages of carbon and hydrogen that have been found in
analyses of oils from other deposits are given in the following table : —

0. II. 0.

Apscheron (MarkownikofF and Ogloblin) 86.65 13.35

" " " " 87.01 13.22

" " " « « 86.89 13.18

Trans-Caspian (Markownikoff and Ogloblin) 86.75 12.19

Egyptian (Kast and Kunkler) * 85.85 11.72

Pechelbronn (Sainte-Claire Deville) t 85.7 12.00 2.3

Galicia (Sainte-Claire Deville) t 82.2 12.10 5.7

Rangoon (Sainte-Claire Deville) t 83.8 12.7 3.5

The following values represent the percentages of carbon and hydro-
gen in crude petroleum from other American oil fields : —

W. Virginia, Scioto Well (Peckhara) J

"VV. Virginia Cumberland Well (Peckham) J

California oil (Peckham) J

Mecca oil (Peckhara) J

Canada, Manitoulin (Deville) f

Canada, Petrolia (Deville) t

Canada West (Deville) f

Ohio oil (Deville) t

Pennsylvania oil (Deville) f

Pennsylvania oil (Deville) f

* Chem. Centralb., 1890, p. 932.

t Comptes Rendus, LXVIII. 485.

t Geological Survey of California, Vol. XI., Appendix, p. 39.

0.

H.

0.

86.62

12.93
13.38
11.82

86.32

13.07

83.00

14.6

2.4

84.5

13.5

2.0

79.4

14.1

6.5

84.2

13.1

2.7

83.4

14.7

1.9

84.19

13.7

1.4

16 PROCEEDINGS OP THE AMERICAN ACADEMY.

In the crude oil used in this examination, and in the other specimens
described above, nitrogen was determined by the Kjeldahl method,
and several closely concordant results were obtained by combustion
vrith soda lime; the former gave 0.11 per cent, and the other oils the
following percentages : —

Findlay. Lima and Findlay.

(1) 0.26 per cent 0.068

(2) (a) 0.023, {b) 0.023 per cent 0.047

(3) 0.21 per cent 0.054

(4) 0.13 " 0.049

(5) 0.35 " 0.060

(6) 0.08 " 0.056

(7) 0.024

The presence of nitrogen in Ohio and Canadian petroleum will re-
ceive further attention at the close of this paper in some observations
on the origin of petroleum.

The bromine absorption in the crude oils was determined by the
method described in Allen's Commercial Organic Analysis. A
weighed quantity of the oil was allowed to stand in the dark with
a slight excess of bromine dissolved in dry carbonic disulphide, and
the portion not absorbed was titrated with standard solutions of
sodic hyposulphite and iodine. The strength of the bromine solution
was ascertained by parallel titrations. Approximately one per cent
of hydrobromic acid is evolved in these determinations in crude
oils.

Findlay. Lima and Findlay. Findlay. Lima and Findlay.

(1) 11.29 8.74 (5) 13.07 10.93

(2) 14.62 9.31 (6) 11.32 12.31

(3) 10.55 11.49 (7) 12.06

(4) 14.89 12.30

A comparison of the bromine absorption of the sulphur oils with
that of oils from other sources indicates that bromine absorption is in-
dependent of sulphur compounds, and a distinctive property of petro-
leums in general. The following determinations were made: —

Chinese petroleum 10.90 per cent
Italian petroleum 7.10 '•

Macksburg, O., petroleum 9.74 "

Berea Grit, 0., petroleum 10.71 "
California petroleum 9.88 "

MABERY. — SULPHUR PETROLEUMS. 17

The quantities of bromine absorbed by distillates from the crude
sulphur oil were also determined : —

Fraction. Percentage of Bromine absorbed.

110°-150° 0.73

150°-220° 1.74

220°-257° 4.84

257°-300° 5.04

300°-330° 12.10

Residue 19.50

Throughout this investigation some reliance has been placed on the
absorptive capacity for bromine of crude oils and distillates obtained
from them as indicating a certain unsaturated condition. While it
should be borne in mind that a considerable proportion of the bromine
absorption is due to the sulphur constituents, there is besides a large
absorption in the crude oils and in the residues of distillation above
350^ by other constituents. There is much yet to be learned concern-
ing the decompositions in distillations at high temperatures, which are
indicated by the greatly increased bromine absorption, and the study
of the higher boiling portions will be greatly facilitated by the aid of
the Mahlke thermometers for observing temperatures below 550°.

The characteristic qualities of Ohio oil appear also in the propor-
tions that distil at different temperatures ; 800 grams of the crude oil
collected in the following proportions beginning at 110° : —

llOMoO^

150^-220^

220°-257°

257°-3003

300^-350^

Residue.

Grams

76

133

86

76

69

348

Per cent

9.75

16.63

10.75

9.75

8.63

43.5

Sp. Gr. at 20°

0.7282

0.7669

0.7940

0.8138

0.8242

0.8976

Per cent sulphur

0.10

0.38

0.41

0.37

0.37

0.54

The distillates below 225° were colorless, and no odors resulting
from decomposition were observed. Above this point color appeared
in the distillates, with the odor of decomposition, which became more
marked with increasing temperatures. Above 275° the heavier
paraffine oils began to distil. In refinery distillation of Ohio petro-
leum it is therefore evident that cracking begins in the vicinity of
250°. No doubt crude sulphur petroleums undergo decomposition
spontaneously to some extent, since upon standing they always con-
tain hydric sulphide. We find that certain unstable constituents
separated from the crude oils gradually become darker in color, with
other indications of chemical change. At the beginning of the distil-

VOL. XXXI. (n. S. XXIII.) 2

18 PROCEEDINGS OP THE AMERICAN ACADEMY.

lation hydric sulphide came oflf in considerable quantities, but after the
first fraction very little appeared in the succeeding distillates below the
point where decomposition began.

On account of the viscous character of the Ohio and Canadian petro-
leums, and the large proportion of heavy oils, temperatures indicated
by the thermometer in the vapor of the distillates should he higher
than the corresponding temperatures of the oil. To ascertain this
difference, crude Findlay oil was distilled, and the temperatures of the
distillates and of the oil were noted with the following results : —

Thermometer in the Oil. Thermometer in the Vapor. Difference,

o o o

167 120 47

180 140 40

203 160 43

221 180 41

238 200 38

259 220 39

282 240 42

301 260 41

318 280 38

341 300 41

Except in the first reading the average difference in temperature is
about 40°.

Determinations of sulphur in the crude oil and in the distillates
obtained from it, showed that considerable sulphur was lost during
distillation. In order to obtain definite information concerning the
quantity lost, 100 grams of the crude oil were distilled under atmos-
pheric pressure, and attached to the receiver were flasks containing
a solution of sodic hydrate for the purpose of absorbing any hydric
sulphide that escaped. In front of the flask containing hydric sul-
phide there was connected another flask, which contained alcohol, with
a delivery tube in front to absorb volatile products that might result
from decomposition. The oil was fractioned to 300°, collecting be-
tween 115° and 250°, and between 250° and 300°, and the hydric
sulphide was determined after oxidation with bromine by precipita-
tion with baric chloride. The percentage of sulphur in the several
fractions was also determined. As usual in distillation of the sulphur
petroleums, a slight sublimate of sulphur was observed in the neck of
the condenser. Upon diluting the alcohol it became slightly turbid,
which indicated probably some volatilization of sulphur constituents.

MABERY. — SULPHUR PETROLEUMS. 19

The alkaline solution of the sulphur from the distillate 115°-250'^
gave 0.1135 gram of baric sulphate, corresponding to 0.015G gram of
sulphur. From the alkaline solution of the sulphur absorbed from the
fraction 250°-300°, 0.694G gram of baric sulphate was obtained, cor-
responding to 0.0958 gram of sulphur. A determination of sulphur in
the distillate 115°-250° gave 0.55 percent; in the distillate 250*^-
300°, 0.51 per cent; and in the residue above 300°, 0.60 per cent.
Since the weight of the distillate collected at 115°-250° was 20.55
grams, the weight at 250°-300°, 5.1 grams, and the weight of the
residue above 300°, 74.35 grams, the total weight of sulphur accounted
for in these determinations was 0.7166 gram, leaving 0.27 gram
which must have escaped in ways not determined.

In comparing the quantities of the distillates from Ohio oil, and
their specific gravities, with those obtained by MarkownikofFand Oglo-
blin * in the Apcheron oil with a specific gravity at 17° of 0.882, and
those given by Bolley from Pennsylvania petroleum with a specific
gravity of 0.816, it is evident that the properties of Ohio petroleum
place it between the Caucasus and Pennsylvania oils.

The Caucasus oil began to distil at 120° in the vajDor, and 180° in
the liquid.

Apscheron.

Pennsylvania.

Onio.

o

Parts in 100.

. Sp. Gr.

Parts in 100

. Sp. Gr.

o

Parts in 100

. Sp. Gr.

120-150

0.5

0.786

19.70

110-150

9.75

0.7282

150-200

10.9

0.824

8.85

0.757

150-220

16.63

0.7669

200-250

12.8

0.861

15.23

0.788

220-257

10.75

0.7940

250-320

24.7

20.7

0.809

257-300

9.75

0.8138

47.9

300-350

8.63
55,51

0.8242

Total

64.48

Residue

53.1

35.52

43.00

While the temperatures at which the Ohio oil was collected are
slightly different from the others, they are sufficiently close for com-
parison.

Kramer f has compared the quantities of distillates obtained from
crude petroleum of other fields : —

Sp. Gr.

-150^

150^-250°

250°-300'^

Residue.

Tagernsee

0.812

20.04

26.12

14.00

35.91

Pennsylvania

0.814

14.34

25.35

13.75

40.99

Baku

0.880

0.63

12.73

15.55

37.95

Oelheim

0.885

0.74

11.05

9.75

75.71

Alsace

0.888

1.3

16.37

17.07

47.88

* Ber. flor deutscli. cliem. Gesellsch , 1883, p. 1873.
t Chem. Centralb., 1887, p. 290.

20 PROCEEDINGS OF THE AMERICAN ACADEMY.

Taubes Barludu* distilled 1115 c.c. of crude Roumanian petroleum
from the deposits on the south slope of the Carpathians, which are
probably connected with the Galician oil zone, with the following
results: —

Q c.c. Per cent by Volume.

30-125 150 13.5

125-225 385 35.5

225-280 IfiO 14.3

280-315 98 8.1

In an examination of Burmese petroleum, Romanis f obtained an
oil from Yay-nan-Chaung with a specific gravity of 0.8590, which
solidified at 24°. An oil from Arracan with a specific gravity of
0.825 at 32° contained considerable benzol and other aromatic hydro-
carbons; upon distillation, the following results were obtained: —

Per cent.

Distilled with Steam.

70-90

3.1

o Per cent.

90-100

7.6

—100 23.3

100-130

10.6

100-110 33.0

130-200

18.7

110-130 29.3

200-300

18.7

Residue 13.3 (heavy oil)

+ 300

12.5

98.9

Oil in residue

8.0

Paraffine

3.1

Loss

17.7

100.00

Markownikoff and Ogloblin examined the ash of Caucasus petro-
leum by igniting the residue of distillation. They found 0.09 per cent
of ash calculated for the original quantity of crude oil, and it consisted
chiefly of the oxides of calcium, iron, and aluminum. Traces of cop-
per and silver were also found. We have determined by combus-
tion in oxygen the percentages of carbon and hydrogen in coke from
the refinery residue of the distillation of Ohio petroleum ; the per cent
of carbon was 95.06, and that of hydrogen 4.85. A determination of
nitrogen by the Kjeldahl method gave 0.31 per cent. The quantity of
ash in the coke was determined by burning off the carbon, and the
weight of ash thus obtained corresponded to 0.11 per cent of the coke
burned. In attempting to estimate the percentage of ash in the crude

* Zeit. Ang. Chem., 1889, p. 606. t Cliem. News, LIX. 292.

MABERY. — SULPHUR PETROLEUMS. 21

oil from the amount found in the coke, there is some uncertainty as to
the quantity of oil corresponding to the coke. In some oils the pro-
portion of residue is estimated as ten per cent of the crude oil. It
depends also upon the method followed by the refiner. Sometimes
the first distillation of the crude oil is pushed to the point of com-
plete decomposition, and the tar distillate is again distilled until it is
coked. It is well known that earthy matter frequently remains for
some time in suspension in the crude oil after it is taken from the
wells. On this account, if the oil was distilled before the suspended
material had subsided, the ash would not represent what had been in
solution in the oil. But the oil is usually allowed to stand some time
before distillation, and that the coke we examined was practically free
from suspended matter is evident from the low percentage of ash, and
corresponding results with Canadian oil where the ash was determined
in a tar distillate and in coke from crude oil ; the percentages of ash
from the two sources were not very different.*

In Findlay oil, in which our determination was made, the proportion
of still residue is doubtless somewhat smaller than that mentioned
above, probably between ten and five per cent. The corresponding
percentage of ash iu the crude oil would, therefore, be not far from
0.005 per cent, an amount considerably less than MarkownikofF and
Oglobliu found iu Russian oil.

An analysis of the ash showed that it was composed chiefly of
lime and magnesia, and the ({uantity of magnesia is at least equal to
that of the lime. Traces of iron and aluminum were found, the iron
possibly having been dissolved from the £,till. It is therefore evident
that the crude oil has exerted an appreciable solvent action on the
limestone reservoir, dissolving both constituents of the dolomitic rock.

It is maintained by some chemists that all petroleums contain the
same series of compounds iu different proportions, and that the differ-
ence in properties depends upon a variation in the quantities of the
constituents. In a general sense, with respect to the principal series
of hydrocarbons this is, doubtless true; yet there is such a wide
difference in the properties of oils like those from Pennsylvania
and the Caucasus that they are characteristic of substances quite
unlike. The Caucasus petroleum is wholly, or nearly, wanting in
the series C„H2,i^_2. and the Pennsylvania oil evidently contains the
series C„H2„ in much smaller proportion than the Russian oil. The
presence of the higher members of the latter series in the Pennsyl-

* Determination of Ash in Canadian Petroleum, page 51.

22 PROCEEDINGS OF THE AMERICAN ACADEMY.

vania oil has yet to be ascertained. It is conceivable that the differ-
ence in the composition of petroleums is due to the different influences
to which they have been exposed. Perhaps greater porosity of the
reservoir or cover where oils exist under pressure has permitted an
escape from certain oils of the more volatile constituents, especially of
the series C^Ho^ + o. If this should be demonstrated by more
extended observations, it would be reasonable to ex[)ect the same
bodies in the Pennsylvania as in the Russian oil, only in smaller
quantities of the higher constituents. Referring the sul{)hur in Ohio
petroleum to the average composition of the compounds containing it,
the crude oil should contain at least five per cent of the sulphur deriv-
atives. Evidently such a large proportion of sulphur compounds iu
petroleum must exert an important influence on its properties, and we
should therefore expect a marked difference between the sulphur
petroleums and those which contain only traces of sulphur.

From a general similarity of Ohio petroleum to the oil from
Pennsylvania, so far as it relates to hydrocarbons of the series
C„H2„ + 25 it should perhaps be expected that the composition of the
latter oil, which has been established beyond question, at least so far
as the portions of low boiling points are concerned, should represent
also similar portions of the Ohio product. As mentioned above, even
a casual examination of the sulphur oils affords abundant evidence
that their peculiar properties depend upon other constituents than the
hydrocarbons C„H2„ + 2' While these unique constituents may detract
from, rather than enhance, the value of the sulj^hur oils for commercial
purposes, it is as important for the intelligent guidance of the refiner
as it is interesting from a scientific point of view that they be well
understood.

While occupied with the sul|>hur compounds in Ohio petroleum, I
was impressed with the complexity in composition manifested by the
products of distillation, and with the importance of a thorough exami-
nation for all constituents. We therefore began with an endeavor to
separate and identify the individual homologues of methane which
have been found in Pennsylvania petroleum, including an approximate
quantitative determination of all but the more volatile members.

Hydrocarbons QJT2«-!-2-

To separate the hydrocarbons of lower boiling points we obtained
twenty-five litres of the very first distillate from a three-hundred
barrel still. This distillate contained 0.10 per cent of sulphur.

MABERY. — SULPHUR PETROLEUMS. 23

ronsiflerable gas always escapes in refinery distillation before a liquid
distillate appears, but we have not yet undertaken an examiuatioa of
its composition. Probably this is essentially the same as that of the
gis given off' in beginning a distillation of Pennsylvania oil together
with hydric sulpiiide, which is always evolved to a greater or less
extent in the distillation of sulphur petroleums. Sadtler* found that
(he gas from Pennsylvania wells consisted principally of methane, with
some ethane, nitrogen, and hydrogen. In the gas from Canada wells, at
Enniskillen, Fouque f found marsh gas, ethane, and small quantities of
carbonic dioxide. According to Ronalds, J the gas from Pennsylvania
petroleum contains 1.27 per cent of carbonic dioxide, 6.58 per cent of
oxygen, 54.00 per cent of nitrogen, and 38.15 per cent of ethane and
propane ; by exposing the escaping gas to a freezing mixture, butane
was condensed to a liquid.

In an exhaustive study of natural gas from Pennsylvania wells,
Phillips § found that the principal constituents were hydrocarbons
C,jH2„_|_2i ^^'itb nitrogen in variable proportions and carbonic dioxide
in small quantity, but no hydrogen nor carbonic oxide.

The crude distillate was subjected to fractional distillation in a
porcelain still, to which was attached a Warren condenser filled with
a mixture of salt and ice, with ice alone, or with water, according
to the fraction collected. Another ordinary condensing worm sur-
rounded with a freezing mixture was placed in front. Subsequent
distillations were conducted in glass stills, and the fractions rapidly
accumulated within limits of temperature which distinguish the hydro-
carbons C„H2„4.2' -A.t the beginning of the first and second distilla-
tions, a delivery tube was attached to the bottle receiving the distillate,
and extended so as to collect in a receiver inverted over water any
volatile constituents that might have escaped condensation. At first a
very little gas collected, which burned with a smoky flame, but none
afterwards. The following quantities were collected during the first
distillation : —

—25° 25^-30= SOO-SS" 35°-i0=>

Grams 525 400 450 400

After the fourth distillation thirty-five grams collected between 0°
and 1°, distilling for the most part at 0°, barometer 740 mm. This was
evidently butane, boiling point 0°. Inasmuch as the boiling point of

* Amer. Chemist, 1876, p. 97. t Journ. Chem. Soc, XVIII. 54.

t Comptes Rendus, LXVII. 1045. § Amer. Chem. Journ., 1894, p. 406.

24 PROCEEDINGS OF THE AMERICAN ACADEMY.

this hydrocarbon has been carefully determiued, * further exam-
ination was not deemed necessary. The temperature rose rapidly
to 5°, and between 7° and 9° 20 grams of a distillate collected, mostly
between 7° and 8°, the boiling point of a hydrocarbon which was
separated by Warren from Pennsylvania petroleum, and which was
regarded by him as one of the butanes. Since the atmospheric tem-
perature was in the vicinity of 30° when these distillations were in
progress, special care was necessary to preserve the distillates, and the
ice accidentally becoming exhausted in the ice-chest, the distillate col-
lected at 0° burst the bottle, and the one at 8° forced out the stopper
and volatilized. Of the two possible butanes, the boiling point of one
is without question 0°. The other seems to have been obtained by
Butlerow t from isobutyl alcohol, and the boiling point assigned to it
was — 17°. 5, Under more favorable conditions, we shall collect a
larger quantity of this distillate, to determine by its chemical behavior,
as well as by its constancy in boiling point, whether it be a definite
compound, t

Considerable quantities of distillates collected below 30°, but by
continued distillation they were mostly separated into higher and
lower constituents, indicating the absence of individual products. In
the vicinity of 30° the fractions were large, amounting to 300 grams
between 28° and 32°. After the ninth distillation, 75 grams collected
between 29° and 30°, with the barometer at 747 mm. A vapor
density determination gave 2.52, required for pentane 2.49. This
product was, therefore, isopentane. boiling point 30°. Between 36°
and 37°, 75 grams distilled, and this distillate was shown by its vapor
density to have the composition required for pentane ; a vapor density
determination gave 2.49, required for pentane 2.49. This sub-
stance therefore corresponded to normal pentane, boiling point 37°

* Ronalds, Journ. Cliem. Soc, XVIII. 54 ; C. M. Warren, Mem. Amer. Acad.
(N. S.). IX. 156.

t Ann. Chem. Pliarm.. CXLTV. 10.

t We have since obtained .50 .crams of an oil that collected at 8° to 9° with
very small amounts above and below these limits. A vapor density determina-
tion by the Hofmann method gave the following result: —

0.0717 gram of the oil gave 45.5 c.c. of vapor at 16°, and under a tension of
48.1 cm. of mercury.

Required for 0411,5. Found.

2.01 2.04

For further proof as to the composition of this distillate, an examination of its
halogen and other derivatives is now in progress.

MABERY. — SULPHUR PETROLEUMS. 2o

(Warren). The weij^hts collected evidently afford no information
concerning the proportions in which these hydrocarbons are contained
iu the crude oil. Other results show that they are present iu smaller
quantities than iu Pennsylvania oil.

At higlier temperatures to 60°, the weights of the distillates were
very small and irregular, which indicated the absence of definite com-
pounds. Between 60° and 62°, 150 grams collected at the end of the
fifteenth distillation, and this was still further reduced to 50 grams be-
tween 60° and 61°, with a vapor density corresponding to that of
hexane; found, 2.94; required for hexane, 2.'J8; boiling point of
isohexane, 61°. 27 (Warren). After the fourteenth distillation, with
the barometer at 749 mm., between 67° and ()S°, 75 grams collected
that distilled tolerably constant within this limit. A vapor density
determination gave 3.00 ; required for normal hexane, 2.98 ; boihng
point, 68°. 5 (Warren).

For the separation of the less volatile hydrocarbons, the fraction
— 150°, obtained from the crude oil by distillation in vacuo, was sub-
jected to further distillation under atmospheric pressure. 41.5 kilos
of crude Findlay oil were distilled under a tension of 50 mm., and
collected in the first distillation at —100°, 100°-150°, 150°-200°,
200°-250°, and 250°-350°. The decomposition was comparatively
slight, and the fractions, especially the less volatile, were free from the
disagreeable pungent odors characteristic of refinery distillates. Even
the residue above 350° had apparently undergone very little decom-
position. On account of the reduced boiling points, it was not
expected tliat the more volatile constituents could be collected, and it
was subsequently found that scarcely any distillate boiling below 30°
was condensed. The weights of the first fractions were as follows : —

-100^

100M50=

150^-200^

200^-250^

250^-350^

Resiilue.

Grams

8000

8520

6480

7700

2670

9000

Percentages

18.6

19.8

15.1

18.0

6.2

20.9

The specific gravity of the individual fractions was determined with
the following results : —

—100° 100^-150= 150^-200° 200°-250° 250°-350° Re>'iclue.

0.7445 0.7941 u.8245 0.8455 0.907 0.9139

The fraction — 100° contained no hydric sulphide; the higher frac-
tions contained it in small quantities. The percentages of sulphur in
the same distillates were obtained by combustion in air : —

26 PROCEEDINGS OF THE AMERICAN ACADEMY.

—100'^ 100^-150^ 150^-200^ 200^-250° Residue.

Sulphur 0.054 0.25 0.42 0.61 0.67

In comparing the percentages of sulphur in the vacuum distillates
with those under atmosphei-ic pressure, it is evident that the main
body of the sul[)hur compounds collects in the higher portions, leav-
ing the fractions below 150° nearly free from sulphur.

Percentages of sulphur under atmospheric pressure : —

110°-150°

150°-220<^

220^-257°

257^-300^

soo-^-ssoo

Residue,

0.10

0.38

0.41

0.37

0.37

0.54

It is also apparent that distillation in vacuo protects, to a certain
extent, the sulphur compounds from decomposition. In ordinary dis-
tillation of the crude oil, as has been mentioned, sulphur is invariably
observed in the condensing tube. Even in vacuum distillation at high
temperatures decomposition cannot be wholly avoided, although the
separation of sulphur has never been observed. It is probable that
chemical reactions occur within the oil from the action of the various
constituents upon one another, such, for example, as the action of the
oxygen compounds upon the sulphur compounds. These observations
are valuable evidence in favor of vacuum distillation for the sulphur
petroleums.

The percentage of bromine absorbed by the fractions collected in
vacuo was also determined : —

-100° 100°-150^ 150°-200° 200^-350' Residue. Crude Oil.

Bromine 0.0 4.57 6.60 7.08 24.38 10.19

It is interesting to compare the bromine absorption in the vacuum
distillates with the quantities absorbed in the distillates collected
under atmospheric pressure : —

100^-150° 150°-220° 220^-257° 257^-300'' 300''-330° +330°

Bromine 0.73 1.74 4.84 5.04 12.10 19.50

Doubtless the larger absorption of bromine in the vacuum distillates
at lower temperatures depends, at least in part, upon the fact that the
sulphur compounds are protected from decomposition during distilla-
tion, and also upon the extent to which the constituents with higher
boiling points are distilled at lower temperatures under the diminished
pressure. It would be expected that the residue in the distillation
under atmospheric pressure should sliow a liigher absorptive power
than that collected in vacuo. The behavior of the higher fractions

MABERY. — SULPHUR PETEOLEUMS. 27

toward bromine and the nature of the decompositions by cracking
will receive furtlier attention.

The portions distilling below 150° were next submitted to pro-
longed fractional separations under atmospheric pressure with the aid
of Hempel columns and Warren condensers. We were led to appre-
ciate the exhaustive labors of our predecessors in their investigations
on petroleum, and our indebtedness to them for the efficient means at
present available for conducting such distillations. For the separation
of complex mixtures, especially in considerable quantities, in point of
efficiency the Warren condenser leaves nothing to be desired. It
appears to effect a more rapid separation than the Hempel column,
although the latter is of great service. The Hempel method has the
advantage that it requires less attention, with no loss of time in heat-
ing a bath nor in maintaining a constant temperature in the bath.
From the description of Warren's hot condenser given in the
treatise on Chemistry, Vol. III. Part 1, by Roscoe and Schorlem-
mer, an erroneous im[)ression must have been received concerning this
apparatus. On pp. 149, 150, the following words appear: " An appa-
ratus has been employed by AVarren in the fractional distillation of
tar oils and petroKums. This permits a complete control over the
temperature of the vapor, accomplished by an air bath round which a
spiral tube is placed, connected with the boiling flask. The tempera-
ture of this air bath is regulated by a lamp. The licjuid used for
heating the air bath may be either water, oil, or fusible metal, and into
this the thermometer is placed. ... In distilling petroleum the differ-
ence in temperature between the boiling liquid and the air bath was,
to begin with, about 35°, or even more." The idea conveyed here is
that the constant temperature is maintained by means of an air bath,
although it is evident from the following description, taken from the
original memoir,* that there is not the remotest allusion to an air
bath : " In the new process, perfect control of the temperature of the
vapors is secured by simply conducting these vapors upwards throtigh
a worm contained in a bath the temperature of which is regulated by
means of a lamp, or by a safety furnace. The bath may be of oil or
water or metal for very high temperatures, as the case may require,
and it is furnished with a thermometer. That this bath may be
equally adapted for the separation of liqt;ids boiling below the com-
mon temperature, an empty vessel is permanently secured in the
interior of the bath by means of straps of metal across the top, to

* Mem. Amer. Acad. (N. S.), IX. 125.

28 PROCEEDINGS OF THE AMERICAN ACADEMY.

serve as a convenient receptacle for ice or ice vi^ater, by means of
which a low temperature may be steadily maintained. This interior
vessel also serves a good purpose in economizing time and fuel in
heating the bath, as it diminishes the cpnntity of oil required to cover
the worm. It is made to extend to within about three inches of tiie
bottom of the bath, and large enough to fill a greater part of the
space in the centre of the coil." It will therefore be seen that it was
uot Warren's intention to use this apparatus in any sense as an air
bath. It is to be used solely as a liquid bath.

Since the principal object was to identify the individual constituents
and to determine their approximate quantities, it was only necessary
to collect our products within such close limits of temperature com-
parable with boiling points already accurately determined that they
should yield satisfactory analytical data. In successive distillations,
collecting at first within o°, then within 2°, and finally within 1°,
after the fifth distillation the fractions collected rapidly, wiih increasing
quantities at temperatures near boiling points of well known hydro-
carbons C„H2„ + 2) ^"^^ ^t certain other points at which an equilibrium
in boiling points seemed to be established by mixtures. It was only
with much ditficulty that some of these mixtures could be se[)arated
into their constituents. We had occasion to recall the remark of War-
ren concerning the greater amount of labor involved in determining
the absence of definite compounds in such mixtures than in proving
the presence of well defined hydrocarbons. The fractions containing
the aromatic hydrocarbons will be considered together. The products
collected for vapor density determinations were purified as completely
as possible by the removal of unsaturated hydrocarbons, sulphur com-
pounds, and the aromatic hydrocarbons. For the removal of sulphur
compounds, each fraction was thoroughly agitated with alcoholic mer-
curic chloride. After washing with water, there remained in solution
not more than 0.02 or 0.03 per cent of sulphur when the mercuric
chloride gave a crystalline precipitate, which was the case in distillates
below 150°, provided they were collected at first in vacuo. In a former
paper * it was stated that alcoholic mercuric chloride removed two thirds
of the sulphur. Those experiments were made with refinery distillates,
which do not behave the same towards mercuric chloride as vacuum
distillates. In higher fractions somewhat more sulphur is retained,
and with increasing boiling points even the mercury itself in consider-
able quantity may be held in clear solution, either in the form of

* Amer. Chem. Journ., 1894, p. 88.

MABERY, — SULPHUR PETROLEUMS. 29

HgClsR^Sj or in some other combination. This peculiarity has
occasioned us much trouble in purifying distillates above 200°. In
some of these products, hydric sulphide in the cold will not precipitate
the mercury; frequently it is only after prolonged action with the aid
of heat that the mercury can be completely removed. These sulphur
oils seem to possess the property of dissolving metals, metallic oxides,
and other compounds, which has been observed in other petroleums.
The action of paraffine oils on metals has been examined by
Macadam,* who finds that lead, solder, and zinc are quite readily,
tin and iron but slightly, affected. Some oils have a greater solvent
action than others, and Macadam attributes it to the hydrocarbons.
Engler f repeated the experiments of Macadam, and observed that
metals were not affected when air is excluded. It was therefore
inferred that acid compounds are formed in the oil by exposure to air,
and also metallic oxides, which are dissolved by the acids. Engler
does not attribute the solvent action to ozone. It is probable that the
purification of refinery distillates from the sulphur petroleums, by
agitation with an alkaline solution of plumbic oxide, depends, at least
in part, upon the solvent action of certain constituents of the oil.
Oxygen compounds, which are doubtless present in these oils, may
assist the action, as has been observed in other oils. This subject will
receive further attention when we reach the higher distillates.

After removal of the sulphur compounds, each distillate for vapor
density determination was thoroughly agitated, first with concentrated
nitric, then with concentrated sulphuric acid, washed, and dried.
Finally, under a return condenser, it was heated for some time with
metallic sodium.

From 41.5 kilos distilled in vacuo, at the end of the eighth dis-
tillation, the last seven under atmospheric pressure, the following
weights were obtained with much smaller quantities outside these
limits : —

—55° 58 ^-62° 65^-68'^ 77^-83='

Grams 15 120 310 85

As already mentioned, on account of loss from the diminished
pressure, as well as the unavoidable lo^s in any distillation, the
weights of the lower fractions collected in vacuo evidently cannot be
accepted as representing even approximately the quantities present ia
the crude oil.

* Journ. Chem. Soc, 1878, p. 355.

t Ber. der deutsch. chem. Gesellsch., 1878, p. 2186.

30 PROCEEDINGS OF THE AMERICAN ACADEMY.

Between 87° and 93°, 80 grams collected after the sixth distillation,
and in the portion which distilled constant at 89°-90°, bar. 754 mm.,
a vapor density determination gave 3.43 ; required for isoheptane,
3.46 ; boiling point, 90°. 4 (Warren).

At the end of the sixth distillation, 175 grams collected at 96°-99°,
and 50 grams distilled constant at 9G°-97°, bar. 744 mm. ; a vapor
density determination of this product gave 3.42 ; required for heptane,
3.6 ; boiling point of heptane, 98°. 7 (Warren). Above this point the
distillates were small in amount to 109°, where other products began
to appear; between this limit and 120° the single degree fractions
were considerable in quantity.

At 118°-120°, 120 grams were collected, of which 50 grams dis-
tilled constant at 119°-120°, bar. 749 mm. This product gave a
value in a determination of its vapor density required for octane :
found, 3.98 ; required, 3.94. Since some doubt has been expressed
concerning the existence of an octane with this boiling point, this
fraction was carefully purified for analysis with alcoholic mercuric
chloride, nitric acid, and sulphuric acid, and it was finally submitted to
prolonged boiling with sodium. Determinations of carbon and hydro-
gen then gave the following results : —

I. 0.1707 gram of the oil gave 0.5282 gram COo, 0. 1707 gram H2O.
11. 0.2017 gram of the oil gave 0.6237 gram CO., 0.2737 gram H^O.

Calculated for

Found.

CgH,g.

I. II.

c

84.20

84.42 84.28

H

15.79

15.19 15.08

The low percentage of hydrogen evidently indicated that the
octane was still contaminated by a hydrocarbon containing less hydro-
gen. For further purification the oil was heated to boiling during
several hours with a mixture of nitric and sulphuric acids, and boiled
several times with sodium. It then gave the percentages of carbon
and hydrogen rquired for octane: —

Required for CgHu.

Found.

c

84.20

84.20

H

15.79

16.10

There seems to be no question that the fraction 118°-119° contains
a hydrocarbon with the composition required for octane, and this
observation is apparently confirmed by the results of others. War-

MABERY. — SULPHUR PETROLEUMS. 31

ren* separated a constituent of Pennsylvania petroleum distilling con-
stant at 11 9°. 5 (cor.), which gave a vapor density corresponding to
that of octane. From the extraordinary care with which the determi-
nations of Warren were made, there can be no doubt as to the
existence of a hydrocarbon with this boiling point in Pennsylvania
petroleum. Although hexahydroisoxylol has been recognized by
Beilstein and KurbatofF,t it is probably not the principal constituent
with tliis boiling point of the Pennsylvania oil ; it is certainly not of
the Ohio oil.

From coal oil, Schorlemmer $ separated an octane boiling at 119°-
120°, and subsequently he identified the same body in petroleum
boiling at 119°. The followin<x data are taken from his original

publication : —
C

Required for Cgllig.

84.20

Found.

84.10

H

15.79

16.10

Vapor density

3.98

3.95

In the treatise on Chemistry by Roscoe and Schorlemmer, New
York, 1886, it is stated that the three octanes known are normal
octane, boiling point 125°. 46, found in petroleum; tetraraethylbutane,
boiling at 108°. 5 ; and hexamethylbutane, melting at 96°-97° and
boiling at 105°-106°. This enumeration does not recognize an
octane boiling at 119° in petroleum. It would seem that our fraction
from the Ohio oil was prepared with sufficient care to preclude the
possibility of a mixture of higher and lower constituents. We shall
endeavor to obtain independent evidence as to the identity of this
fraction by a study of its chemical reactions, and the preparation of its
derivatives.

The following suggestion, which I have taken from one of the pri-
vate papers, dated 1868, of Mr. Warren, through the kindness of
Professor Storer, indicates that Warren recognized the possibility of
a series containing less hydrogen : "The samples analyzed may have
contained traces of more highly carbonized substances, and that it
would be worth while to treat with HOSOj and HONO5, and remove
these." Beilstein and KurbatofF met with the same difficulty in the
fractions 95°-100° and 118°-120° in their attempts to ascertain the

* These Proceerlings, XXVII. 78.

t Ber. der deutsch. chem. Gesellsch., 1880, p. 2028.

t Journ. Chem. Soc, XV. 419.

32 PROCEEDINGS OF THE AMERICAN ACADEMY.

presence in the Pennsylvania oil of the hexahydro series. The series
with less hydrogen seems to be removed only very slowly, even by
vigorous treatment with a mixture of nitric and sulphuric acids.

In continuing the separations above 120°, after the eleventh distilla-
tion the oils collected iu considerable quantities within the limits of one
degree : —

120^-121^ 121^-122' 122^-123° 123^-124^ 124°-125^ 125^-126' 126^-127^ 127°-128='

Grams 35 40 80 70 75 75 60 40

Notwithstanding our endeavors to separate the fraction 122°-125°
into higher and lower constituents, they may still be mixtures; aro-
matic derivatives are here present in small quantity. If in more pro-
longed separations distillates still persist within the same limits, they
will be submitted to careful study in their behavior toward chemical
reagents.

The greater portion of the fraction 125°-lo0° was unfortunately
lost by an accident after the eleventh distilUition. Sufficient however
was collected at 127°-128° for a vapor density determination; it
gave, by the method of Dumas, the value required for the formula
CgHjg ; found 3.90, required 3.94. To the liquid collected by War-
ren at this point was assigned the boiling point 127°. 6; the vapor
density found by him was 3.99, and those observations seem to be
confirmed by our results with the Ohio product. The portions dis-
tilling between 115° and 130° evidently need to be carefully examined
in larger quantities with particular reference to the octanes. The
distillates collected between 130°-145° will be considered in connec-
tion with the aromatic compounds.

Between 144° and 148°, after many distillations, the fractions
were small in quantity. At the end of the sixth distillation, 65 grams
collected at 149°-15l°, of which 40 grams distilled constant at 149°-
150°, bar. 756 mm. After prolonged heating with concentrated nitric
acid a nitro-compound was formed that remained chiefly in solution in
the oil. Upon dilution of the acid an oily liquid separated in small
quantities. But when the oil was shaken with sodic hydrate, it
became intensely red in color, and by repeated washing with the alka-
line solution the nitro-product was in part removed, and it separated
again as an oil by acidifying the solution. After washing, the remain-
ing oil w.TS dried and heated to boiling during several hours with
sodium. It was then distilled from the large quantity of solid that
had separated, and the greenish yellow distillate again boiled with
sodium. Still more solid separated, and after distillation the oil was

MABEBY. SULPHUR PETROLEUMS. 33

again treated with sodium uutil there was no further action. The
boiling point was not appreciably changed by this treatment, although
it was evident that some constituent had been removed capable of
forming a nitro-derivative soluble in alkalies. The residual oil was
shown by vapor density determinations and analysis to have the
composition required for uouane. The boiling point of nouane is
150.8° (Warren).

I. 0.1799 gram of the oil gave 0.3987 gram CO2, and 0.1749 gram

11. 0.1751 gram of the oil gave 0.5424 gram CO.,, and 0.2488 gram
HoO.

Calculated Found,

for CgHjg. I. II.

C 84.37 84.96 84.48

H 15.63 15.16 15.88

Analysis I. was made of the oil after treatment with alcoholic mer-
curic chloride, but before heating with the acid mixture ; analysis 11.
of the oil after treating with a mixture of nitric and sulphuric acids
and boiling with sodium.

The determinations of vapor density of this product in the apparatus
of Victor Meyer, with a bath of ethyl benzoatCg gave the following
results: I. 4.73; II. 4.74; required for C9H0O, 4.43.

This fraction therefore contains nonane, although it consists to a
very considerable extent of another hydrocarbon, which seems, by the
ready formation of nitro compounds and its behavior toward alkaline
solvents, to be a member of the series C„H2„. All the fractions of
Ohio oil that we have examined below 150° are similarly contami-
nated by, or rather contain in appreciable quantities, hydrocarbons
which form nitro products with the characteristic qnalities of the un-
saturated or parafRne nitro compounds.

Thus far Ohio petroleum has proved to contain members of the
series C„Ho„ + 2' corresponding to those which have been identified in
Pennsylvania oil, but in much smaller quantities. They form one
fifth of crude Pennsylvania oil and less than one tenth of Ohio oil.

In identifying these constituents it has not seemed necessary to
accumulate data further than would appear essential in showing the
similarity of our products to those whose composition has been so
carefully demonstrated in Pennsylvania petroleum. And while this
portion of our labor may be of less interest than the study of the
constituents with higher boiling points, it has seemed of sufficient
VOL. XXXI. (n. s. xxiii.) 3

34 PROCEEDINGS OF THE AMERICAN ACADEMY.

importance to justify the necessary expenditure of time and effort.
The higher fractions have been quite thoroughly distilled entirely in
vacuo to avoid decomposition so far as possible, and these products,
as well as the residue above 350° of the first vacuum distillate, are
reserved for later study.

Aromatic Hydrocarbons.

Series C„H2„_s.

So far as I am aware no attempts have hitherto been made to sepa-
rate the aromatic hydrocarbons from the Ohio sulphur oil. In Ameri-
can petroleum Pelouze and Cahours found, as already mentioned, no
appreciable quantities, but Schorlemmer recognized the presence of
benzol and toluol. Beilstein and Kurbatoff were the first to discover
the aromatic hydrocarbons in the oil from Baku, and hexahydroisoxy-
lol in Pennsylvania petroleum.

Benzol.

After the sixth distillation of the products from the original quan-
tity of crude oil, twenty-five grams collected at 77°-79°, 35 grams at
79°-8r, and 20 grams at 81°-83°. In each of these fractions the
quantity of benzol was determined by heating carefully for some time
a weighed quantity of the fraction with a mixture of nitric and sul-
phuric acids, distilling off the portion not affected by the acid mix-
ture, and weighing it and the residual nitrobenzol. The fraction 77°-
79° gave by this method 3 per cent of benzol ; the fraction 79°-81°,
15 per cent; and the fraction 81°-83°, 5.8 per cent. The fraction
75°-76° and 85°-86°, when treated in the same manner, left scarcely
any residual product after distillation, and after reduction with tin
and hydrochloric acid not a trace of color was visible in applying the
exceedingly delicate furfurol reaction for aniline. Calculating from
these numbers, the quantity of benzol in the 41.5 kilos of crude oil
taken, it amounts to 7.16 grams or 0.017 per cent, which represents
approximately the quantity of benzol in the crude oil.

Toluol.

The fractions collected between 107° and 113°, after the sixth distil-
lation under a Hempel column, were examined for toluol by treating a
weighed quantity of the oil with nitric and sulphuric acids, keeping
the solution cold. After some time crystals of dinitrotoluol sepa-
rated, and when the hydrocarbons not affected by the acid were

MABERY. — SULPHUR PETROLEUMS. 35

removed by distillation, there remained a heavy oil which consisted of
the liquid orthouitrotoluol and crystalline dinitrotoluol. The latter
soon solidified, and after crystallization from alcohol it was identified
by its melting point, 71°. The loss in weight by the removal of toluol
in the fraction 107°-109° amounted to 1.14 per cent, in the fraction
109°-lll''to 13.07 per cent, and in the fraction 111°-113° to 2.8
per cent. The total weight of the first fraction was 50 grams, of
the second 80 grams, and of the third 65 grams. These fractions,
therefore, contained altogether 12.85 grams, which corresponds to 0.03
per cent of toluol in the 41.5 kilos of crude oil taken. As in the
case of benzol, evidently the percentage of toluol is expressed only
approximately by these results. After longer distillation the fraction
114°-115° was treated in a similar manner, but after treatment with
nitric and sulphuric acids, and with tin and hydrochloric acid, the pro-
duct which remained in considerable quantity after distillation did not
dissolve in hydrochloric acid, and it gave no trace of color when heated
wich ferric chloride or with mercuric chloride. The insolubility in
acids excludes in this fraction any appreciable quantity of hexahydro-
isoxylol which is contained in the higher portions. The oily product
of the nitration must therefore be derived from another series, perhaps
from an unsaturated hydrocarbon C„H2„ ; it will receive further
attention.

Xylols.

In the first allusion to the presence of the aromatic hydrocarbons in
petroleum, by De La Rue and Miiller,* who found in the Rangoon oil
benzol, toluol, metaxylol, and cnmol, there is no evidence that the
isomeric forms of xylol were discovered in that petroleum. Except
paraxylol, which Pawlewski found in Galician petroleum and Engler
in Pennsylvania petroleum, and metaxylol, which is generally found,
it does not appear that the isomeric xylols have elsewhere been recog-
nized in petroleum.

Between 136° and 142° in our distillates, at intervals of one degree
we collected 205 grams, of which the greater portion distilled at
137°-138°, 139°-140°, and 141°-143°. After prolonged distillation
these fractions collected in increased quantities at these points ; they
were readily acted upon by nitric acid, forming nitre compounds or
oxidation products, according to the form of the reaction.

To prove the presence of paraxylol in fraction 137°-138°, a portion
was treated with a mixture of nitric and sulphuric acids, at first in the

* Proc. Roy. Soc, 1856, p. 221.

36 PROCEEDINGS OF THE AMERICAN ACADEMY.

cold, then with the aid of a gentle heat. Upon distillation of the
hydrocarbon not affected, there remained a brown oil that deposited a
crystalline product on standing. After crystallization from alcohol,
this substance appeared in the form of glistening white needles which
melted at 139°-140°. It was therefore trinitroparaxylol, melting
point lo9°-140°.* By treating carefully another portion of the same
fraction in the cold with fuming nitric acid, long yellow needles spar-
ingly soluble in alcohol were obtained, melting at 145°, corresponding
to dinitroparaxylol, melting point 147°-148°. | For further con-
firmation, another portion of the fraction 137°-138° was submitted to
oxidation with chromic acid, and the solution extracted with a con-
siderable quantity of ether. Evaporation of the ether left a solid
residue in the form of minute prisms insoluble in water, but readily
soluble in sodic hydrate. This substance sublimed witliout melting,
and in general its properties corresponded to those of terephthalic
acid. After heating for thirty hours a quantity of the same fraction
with dilute nitric acid and distilling off the hydrocarbon not affected,
there remained an oily product which contained no toluic acid. In
repeating this experiment with longer heating, on cooling an oily layer
separated above the acid, which was neutralized with sodic hydrate,
evaporated to dryness, the salt decomposed with hydrochloric acid,
and the solution extracted with ether. Upon evaporation a crystalline
solid was left, but not in sufficient quantity for further examination ; it
was probably para toluic acid.

In the fraction 139°-140° metaxylol was recognized by the forma-
tion of the trinitro-compound. A portion of the oil was heated with a
mixture of nitric and sulphuric acids during forty-eight hours, the oil
together with the crystals which formed was separated from the acid
solution, the oil distilled, and the solid remaining with the first crys-
tals was purified by crystallization from hot alcohol ; on cooling, long
slender colorless needles separated, melting at 175°-176°. This
nitro compound was therefore trinitrometaxylol, melting point 176°.
Several different melting points have been assigned to triuitroxylol.
Luhman t gave 177°, and Tilden § 182°. To explain the latter result
Tilden asserted that the lower value of Luhman was due to contam-
ination of the trinitroxylol by the isomeric nitro compounds derived

* Nolting and Geissman, Ber. der deutsch. chem. Gesellsch., XIX. 145.
t Lellmann, Ann. Chem. Pharm., CCXXVIII. 250.
} Ann. Chera. Pharm., CXLIV. 274.
§ Journ. Chem. Soc, XLV. 416.

MABERY. — SULPHUR PETROLEUMS. 37

from coal tar, the source of the metaxylol from which the trinitro com-
pound was i^repared. But in the preparation of this trinitro deriva-
tive from octonaphtene, Markovvuikoff and Spadi * gave 179°-180°
as its melting point.

An approximate quantitative determination of metaxylol was made
in the fraction 139°-140° by heating a portion during one hour with
dikite nitric acid, which should oxidize the para- and ortho-xylols to
the corresponding toluic acids, and the remaining oil was washed with
water, dried, and distilled with steam. The decrease in volume was
noted, and it corresponded to 6.3 per cent of the quantity taken. Tlie
distillate was next shaken with concentrated sulphuric acid, which dis-
solves metaxylol, and the decrease in volume corresponded to 16.6
per cent of metaxylol. In determining the quantity of paraxylol, a
weighed amount of the fraction 137°-138° was shaken thirty minutes
with concentrated sulphuric acid to dissolve the ortho- and meta-xylols,
with a loss in volume equivalent to 4 per cent of the quantity taken.
The residual oil was then agitated with fuming sulphuric acid, and the
diminished volume represented 12.2 per cent of paraxylol.

In attempting to ascertain the presence of orthoxylol, dependence
was placed upon the observation of Jacobsen, that a single drop of
orthoxylol may be distinguished in a mixture of the three isomers by
treating them in the cold with a mixture of nitric and sulphuric acids.
The fraction 142°-143° gave immediately an oily layer inconsiderable
quantity between the acid and the lighter oil, which should be the
liquid orthonitroxylol. The fraction lo9°-140° gave no trace of an
oily layer even after long standing. By further treatment with the
acid mixture with the aid of heat, the oily mononitro product from
distillate 142°-143° was converted into a crystalline dinitro com-
pound melting at 91°, the melting point of dinitrometaxylol. By long
continued action of the acid mixture, hot, a trinitroxylol was formed,
nearly insoluble in alcohol, and melting at 178°, which might be
a derivative either of ortho- or meta-xylol. Another portion of the
distillate 142°-143° was heate 1 with dilute nitric acid, which should
form orthotoluic acid without affecting metaxylol. Upon extracting
the acid solution with ether, an oily residue remained after evaporation
of the ether with the characteristic odor of toluic acid, but no crystals
separated even after long standing. An attempt was made to separate
the less soluble sodic orthoxylol sulphonate, but upon evaporation of

* Ber. der deutsch. chem. Gesellsch., 1887, p. 1850.

38 PROCEEDINGS OF THE AMERICAN ACADEMY.

the solution no crystalline product appeared until the solution was so
far concentrated that the sodic metaxylol sulphonate was deposited.
Orthoxylol cannot therefore be present in any appreciable quantity.

Determinations of the constituents capable of forming nitro products
were also made in the fractions containing the xylols by treatment
with a mixture of nitric and suphuric acids according to the method
employed for benzol and toluol. The quantity of xylol corresponding
to the nitro product represented by the portion not affected by nitric
acid in the fraction 137°-138° was 54.3 per cent of the weight of oil
taken; in the fraction 139°-140°, 72.1 per cent; and in fraction
141°-143°, 31.5 per cent. Calculating the percentages of xylols from
the results of the first determination in the 41.5 kilos of crude oil first
distilled, the quantity of metaxylol is 0.016 per cent, and of paraxylol,
0.013 per cent. Evidently the xylols were not wholly collected in
their respective fractions, although the quantity outside of the limits,
137°-143° was doubtless small. Probably at least one third should be
added to these results. But if the percentages were increased to the
results obtained by the action of concentrated nitric acid, the quantities
of the xylols in the crude oil would still be small. In the fractions 132°-
136° we observed the presence of aromatic hydrocarbons, although in
too small amounts for identification ; nitro products were formed by the
action of nitric acid. Attempts were made to ascertain the presence of
ethylbenzol and hexahydromesitylene, but vv'itliout success, on account
of the limited quantity of the distillates. After treatment of the frac-
tion 135°-136° with concentrated nitric acid to convert the hydrocar-
bons C„H2„_g into nitro products, and distilling off the hydrocarbons
not affected by the acid, the distillate was heated for some time with a
mixture of nitric and sulphuric acids. According to Baever,* fuming
nitric acid readily converts hexahydromesitylene into trinitromesity-
lene; but KonowalotFt found that a mixture of nitric and sulphuric
acids, or fuming nitric acid, forms the trinitro compound only very
slowly. After prolonged heating we obtained a crystalline nitro com-
pound, but in such small quantity that it was impossible further to
identify it. A much larger supply of these distillates will be necessary
to determine the presence of these bodies.

* Ann. Cliem. Pharra., CLV. 275.

t Ber. der deutsch. chem. Gesellsch., 1887, p. 1850.

MABERY. — SULPHUR PETROLEUMS. 39

Series C„H2„.

Hexaliydro compounds (Reilsteia and Kurbatoff) ; napliteues (Mar-
kownikott" aud Ogloblin).

The lower members of this series include : —

Hexahydrobenzol, C^Hja, boiling point, 69°.
Hexahydrotoluol, C-H14, boiling point, 97°.

Hexahydrocumol, C9H18, boiling point, 147°-150°.
Hexahydrocymol, C10H20, boiling point, 171°-173°.
Hexahydroisoxylol, CsIIig, boiling point, 118°.
Hexahydromesitylene, CaMis, boiling point, 135°-138°.

In examining Ohio petroleum for the lower hydrocarbons of this
series, the fraction 69°-70°, after the fourteenth distillation, was
shaken with concentrated nitric acid, then with concentrated sulphuric
acid, and finally heated with metallic sodium. Determinations of car-
bon and hydrogen in the purified oil gave results corresponding to the
composition of hexane : —

I. 0.1618 gram of the oil gave 0.4934 gram CO2, and 0.2299
gram HjO.
II. 0.1891 gram of the oil gave 0.5790 gram CO2, and 0.2663
gram HgO.

Required for

Found.

CeH,!. CfHij.

I. II.

c

83.72 85.71

83.18 83.51

H

16.28 14.28

15.79 15.65

Hexahydrobenzol is therefore not present in appreciable quantities
in Ohio petroleum.*

The fraction 96°-97°, bar. 750 mm., after the fourteenth distilla-
tion, was purified as before, and the composition determined by
analysis : —

* Recently Markownikoff (Ber. der deutsch. cliem Gesellsch., 1895, p. 577)
has ascertained that hexanapliten (hexahydrobenzol) which he formerly re-
ported as a constituent of Caucasus naphtha, boiling at 69°-71°, is not a single
body. Since the discovery by Baeyer of hexamethylen, boiling point 79°,
Markownikoff has discovered that Caucasus naphtha contains rather this same
hexamethylen which he has identified by the formation of its nitro products.
It will, therefore, be necessary to examine more carefully the corresponding
distillates from the oils we have under examination for this body.

40 PROCEEDINGS OF THE AMERICAN ACADEMY.

I. 0.2083 gram of the oil gave 0.6423 gram CO^, and 0.2849
gram II2O.
II. 0.2010 gram of the oil gave 0.6200 gram CO2, and 0.2744
gram H^O.

Required for - Fouud.

C7H1G. CjHi^. I. II.

C 84.00 85.71 84.09 84.72

H 16.00 14.28 15.27 15.17

This oil was evidently heptane contaminated, as shown by the low
percentage of hydrogen, by a hydrocarbon containing less hydrogen.
To remove any doubt as to the presence of heptane, the oil was heated
during fifteen hours with a mixture of nitric and sulphuric acids, the
oil separated from the acid, washed, dried, and boiled for some time
with sodium. It was then distilled from the large amount of colored
residue, again boiled with sodium, and distilled. The last distillation
left very little residue, and the distillate was nearly odorless. The
residue from the treatment with sodium crystallized well from alcohol.
Upon diluting the acid a heavy nitro product separated in consider-
able quantities. A combustion of the purified oil gave the following
percentages of carbon and hydrogen : —

0.1410 gram of the oil gave 0.4224 gram of CO2, and 0.2065 gram
H2O.

Required for CjHjg. Found,

C 84.00 83.75

H 16.00 16.28

In a distillate 95°-100° from American ligroine, Beilstein and
Kurbatoif * found 84.8 per cent of carbon and 15.4 per cent of hydro-
gen. After prolonged heating with nitric acid, the oil distilled at
98°. 5 — 99°. 5, and gave on analysis 84.2 per cent of carbon and 15.9
per cent of hydrogen, from which it was inferred that hydrocarbons
poorer in hydrogen were contained in the crude ligroine. A nitro
derivative was separated from this fraction with the composition of
nitropropane, C7Hi5N02. The fraction from Ohio petroleum gave no
nitro compound when treated with nitric acid, and the acid was diluted.

Scarcely any residue remained when the oil was distilled after the
treatment with acid. With a mixture of nitric and sulphuric acids, as
shown above, much nitro compound separated upon dilution, which was

* Ber. der deutsch. chem. Gesellsch., 1880, p. 2028.

MABERY. — SULPHUR PETROLEUMS. 41

soluble in potassic liydrate like the nitro derivatives of the series
C„H2„^2» ^^*-^ ^y boiling with sodium the solid residue vpas large.

In distillates* 85°-l 05° from Bibi-Eibat and from Balakhani, after
thirty fractional separations, Milkowsky * collected a body that distilled
constant at 100°-101°, and by the formation of its halogen derivatives
proved it to be heptanaphtene. The oil boiling at 101°-103°, which
Beilatein and Kurbatoff separated from Baku oil, was doubtless the
same product, but Beilsteiu and Kurbatoff considered it to be com-
posed for the most part of hexahydrotoluol, boiling point 97°.

Hexahydroisoxylolwas found by Beilsteiu and Kurbatoff in Cauca-
sus petroleum,! and in American ligroine ; t since the American
source was not mentioned, it is to be inferred that the ligroiue was
prepared from Pennsylvania oil. In testing the fractions 118°-119°,
fourteenth distillation in a Hempel column, for hexaisoxylol, a por-
tion of the oil was heated forty hours with a mixture of nitric and
sulphuric acids. The nitro product separated from the oil when cold
in long, flat plates, nearly insoluble in cold, more readily in hot alco-
hol. When purified by crystallization, this substance melted at 177°,
and was, therefore, trinitroisoxylol. A similar crystalline nitro pro-
duct with the same melting point was found in the fraction 123°—
124°. Since the prolonged distillation precluded the possibility that
this fraction contained metaxylol, there can be no doubt that the
trinitroxylol obtained from fraction 118°-119° indicated the presence
of hexahydroisoxylol, although it evidently formed only a small pro-
portion of this product. After a portion of the same distillate was
agitated thoroughly with a mixture of nitric and sulphuric acids, and
distilled over sodium, analysis gave the percentages of carbon and
hydrogen required for octane, as has been shown (page 30) : —

c

Required for

84.20 85.71

I.

84.42

Found.
11.

84.28

III.
84.20

H

15.79 14.28

15.19

15.08

16.10

As already explained, Analysis III. was made after prolonged treat-
ment with the mixture of acids by which the hexahydro compound is
removed very slowly, and the final results show that the fraction is
composed chiefly of the more highly hydrogenized hydrocarbon,
although it contains a considerable quantity of the other constituent,
as shown by the abundant formation of nitro product.

* Ber. der deutsch. chem. Gesellsch., 1885, c. 187.

t Ibid., 1880, p. 1818. t Ibid., p. 2088.

42 PROCEEDINGS OF THE AMERICAN ACADEMY.

In the iutermediary product from the separation of octonapliteue
and nononaphtene, by prolonged boiling with sodium and treatment
with fuming sulphuric acid, Putochin * obtained a hydrocarbon distil-
ling at 122°-] 24°, principally at 122°. 5, which corresponded in its
composition to the formula CgHis. It gave a chloride different from
the corresponding derivative of octonaphtene, and it was therefore
accepted as isooctonaphteue. By treatment of the chloride with alco-
holic potassic hydrate, isooctonaphtylene was formed with a higher
boiling point than naphtylene from octonaphtene.

The principal features of the Ohio sulphur petroleum which have
appeared in the course of this examination are the following.

1. The crude oil is heavier than the Pennsylvania, and lighter than
the Russian oil. In the quantities of the higher distillates, and in its
general properties it resembles more nearly the latter.

2. It differs from other petroleums in the large amounts of sul-
phur compounds which exert an influence on the general properties of
the oil.

3. It resembles the Pennsylvania oil in containing below 150°
members of the series C„H2,i + 2' although in much smaller quantities.
The presence in the Ohio oil of the two isomeric series C„H2„4.2 con-
firm the observations of Warren on the Pennsylvania petroleum.

4. The aromatic hydrocarbons are here present in minute quan-
tities, apparently much smaller than in other petroleums. Benzol,
toluol, meta- and para-xylol have been identified. The hexahydro
series C„H2„ is represented by hexahydroisoxjdol, and very probably
by higher members, although this has yet to be determined. Hexa-
hydrobenzol and hexahydrotoluol are not contained in this petroleum.

5. By the formation of characteristic nitro products, and the results
of bromine absorption, the presence in the crude oil of unsaturated
hydrocarbons C„H2„ seems to be indicated.

Portions of Ohio petroleum have been thoroughly fractioned, and
are now under examination to establish the identity of the octanes, to
ascertain whether isooctonaphtene and hexahydromesitylene are pres-
ent, and to ascertain whether this petroleum contains a-nonane, boiling
point 135°-137°, and /3-nonane, boiling point 129°.5- 131°.5, which
Lemoine f asserts is contained in Pennsylvania petroleum. Examina-
tion of distillates collected above 150° is now in progress.

* Ber. der deutsch. chem. Gesellsch., 1885, p. 186.
t Bull. Soc. Chim., 188i, XLI. 164.

MABERY. — SULPHUR PETROLEUMS. 43

Canadian Petroleum.

Closely connected with the chemistry of Canadian petroleum are
cei'tain features relating to its occurrence, and the associated geologi-
cal formations, which have not been fully investigated. The de-
posits of petroleum in Canada have been longer known than those
in the Lima and Findlay fields in Oiiio. As early as 1857 the exist-
ence of oil in considerable (quantity in the township of Enniskillen was
ascertained, and in 1862, the first flowing well was started. It was
estimated by Dr. Winchell* that during the summer of 18G2 not less
than 5,000,000 barrels of oil flowed off on the waters of Black Creek.
The flow of these early wells was very large. At a depth of 188
feet as much as 6,000 barrels of oil daily escaped from a single well,
and at 237 feet 7,500 barrels daily, nearly equal to the flow of the
great wells at Baku.

It is now known that these deposits of oil were mainly in the form
of "pockets," and they formed no part of the main fields which are
still productive. This oil territory is situated on two parallel anti-
clinals, about ten miles apart, with the corresponding synclinal
between, from which no oil is obtained. In various reports of these
oil fields, the larger areq, of oil-bearing strata at Petrolia is given as
twenty-seven scpiare miles, but the really productive field is actually
contained within an area of less than eight square miles. The pro-
ductive field at Oil Springs is included within an area of less than two
square miles. The oil deposits are here found in the Corniferous
limestone underlying the Hamilton group of shales and limestones.
Unlike the Trenton oil rock in Ohio, the oil-bearing limestone is quite
near the surface ; the usual depths of wells at present in the Petrolia
field is 465 feet.

I am indebted to the experience and extended observations of
Messrs. M. G. Woodward of Petrolia and F. J. Carman for valu-
able information concerning the geological features and technology of
the Canadian oil, of which a more detailed account will elsewhere be
given. f

Under the general title of American petroleum, with occasional
reference to Pennsylvania and to Canada as the particular sources,
several partial examinations of crude Canadian oil were early made
by French, and English chemists. The first examination of Canadian

* Geological Keport of Canada, 1888-89.

t A paper soon to be presented at the Franklin Institute.

44: PROCEEDINGS OF THE AMERICAN ACADEMY.

oil was undertaken by Pelouze and Caliours * for the purpose of sepa-
rating and identifying the hydrocarbons therein contained. Their
attention was confined to the series C„H2„ + 2, and published accounts
of their woik contain no allusion to any otiier constituents than this
series of hydrocarbons. They failed to observe the jiresence of aro-
matic hydrocarbons, although, as already mentioned, Schorlemmer
discovered the presence of the benzol hydrocarbons in " real Cana-
dian rock oil, a thick black liquid of a very unpleasant odor."

My attention was first attracted to Canadian petroleum in 1890,
when I procured some of the crude oil and also a quantity of the
"sludge" from the refining of burning oil, for the purpose of examin-
ing the sulphur compounds. The peculiar features of the distillates in
a preliminary examination f invited further attention, and I determined
to undertake, with the aid of the refiners, as complete an examination
as was possible with the appliances at my command. It may seem
somewhat surprising that such an examination of the sulphur petro-
leums in general has not previously been undertaken ; but it should
be borne in mind that it requires the facilities of a well equipped
organic and technological laboratory for the manipulation of con-
siderable quantities of material, and a corps of efficient chemists with
aid from the refinery of crude oils, partially refined products, and resi-
dues. Even with all necessary accessories, aside fi-om the tedious
routine labor, there are certain features of decomposition and slow
separation of constituents that render this work extremely difficult.
In view of possible changes in the composition of petroleum, in the
course of time, as wells become exhausted or the oil is taken from
different depths, or indeed with the possibility of future exhaustion of
oil fields which at present appear to be in the zenith of their produc-
tion", it would seem that a comprehensive study of these oils should not
be too long delayed.

From the peculiar nature of petroleum and its numerous constitu-
ents, all of which, so far as they have been examined aside from the
lighter hydrocarbons of the series C„H2„ + o, or the series C„H2n5 are
present in small quantities, any attempt toward a separation of these
constituents involves the manipulation of large volumes of the oils in
such a manner as to prevent so far as possible decompositions, which
cannot be entirely avoided even with the greatest care. On this
account satisfactory results can be hoped for only in products that have

* Bull. Soc. Chim., 1863, p. 228.
t Amer. Cliem. Journ., 1894, p. 89.

MABERY. — SULPHUR PETROLEUMS. 45

been obtained, at least in part, in experiments with several hundred
barrels of oil, which must be performed with the aid of the appliances
in a refinery. On the other hand, an examination of the unstable sul-
phur oils for certain constituents can be carried on satisfactorily only
on a smaller scale, with laboratory appliances.

The products which I obtained through the aid of Messrs. Samuel
Rogers & Co., of Toronto, and Mr. J. H. Fairbanks, of Petrolia, for
the study of Canadian petroleum, included a barrel of crude oil, con-
siderable quantities of the first distillate, naphtha distillate, and burn-
ing oil distillate, none of which had been further refined, besides 200
litres of tlioroughly washed sulphur oil from " sludge." The crude
oil was thick and nearly black in color ; it contained hydric sulphide
in small quantity, and some water, which was removed only after long
standing in intimate contact with fused calcic chloride, and even then
a small quantity appeared in the first distillate. A determination of
its specific gravity at 20° gave 0.8621. In a former determination
in another quantity of the crude oil we reported 0.8600.* Determi-
nations in other specimens gave the following results : —

Oil Springs 0.8442

0.8427

" (gas oil) 0.8389

Petrolia 0.8553

These numbers are not essentially different from those reported in
other examinations of these oils. H. P. Brummel f gave as the
specific gravity of the Canadian oils 0.804 and 0.808, values which
must be accepted as only approximate, rather than as results of accu-
rate determinations. MarkownikofF and Ogloblin j referred to results
of Sainte Claire Deville, which gave 0.844 as the specific gravity of
Canadian oil, and 0.887 for Ohio oil. Reference to the original pub-
lication of Deville § shows that he obtained 0.870 as the specific
gravity of Petrolia oil, and 0.844 for oil from " Canada West " (Oil
Springs?). The numbers assigned by Redwood || to the oils at Petrolia
were 0.859-0.877, and to the oils from Oil Springs 0.844-0.854. As
Engler observed in the Alsace oils, it is possible that the specific
gravity diminishes with the depth of the well.

* Amer. Chem. Journ., 1S94, p. 90.

t Canadian Geological Report, 1888-89.

t Ann. Chim. Pliys., [6.], II. 372.

§ Comptes Rendus, LXVIH. 485.

II Journ. Soc. Chem. Ind., 1887, p. 405.

46 PROCEEDINGS OP THE AMERICAN ACADEMY.

Determinations of sulphur in the crude oil gave the following
percentages: —

I. II. III. Oil Springs.

Sulphur 0.98 0.99 1.06 0.60

Canadian petroleum contains somewhat less carbon and hydro-
gen than Ohio oils : —

I. 0.1908 gram of Petrolia oil gave 0.5874 gram of CO^ and 0.2295
gram HgO.
II. 0.1914 gram of Oil Springs oil gave 0.5868 gram CO^ and 0.2305
gram HgO.

I. II.

C 83.94 83.62

H ■ 13.37 13.39

In the Canadian Geological Report above mentioned Brummel gave
85 per cent for carbon and 15 per cent for hydrogen. These percent-
tages are evidently only approximate. They do not agree with results
earlier obtained by Pelouze and Cahours, who reported a considerable
percentage of oxygen.*

c. H. o.

84.2 13.4 3.0

The variation in composition in the Canadian, Ohio, Pennsylvania, and
Russian petroleums is shown in the following table : —

Canadian.

Ohio.

Pennsylvania.

Russian.

c

83.94

84.57

84.19

86.89

H

13.37

13.62

13.70

13.18

After accounting for one per cent of sul|»hur in the Canadian oil and
0.70 per cent in the Ohio oil, the remaining percentage may reason-
ably be assigned to oxygen, yet in the Russian oils the composition is
fully accounted for by the carbon and hydrogen, although the presence
of oxygen compounds in considerable quantity has been demonstrated
in several independent investigations.

Having an opportunity to collect fragments of the oil rock at
Petrolia soon after it had been removed in drilling a well, I have
ascertained its composition. Like most wells in this territory, the oil-
bearing stratum was reached at a depth of 465 feet. The driller distin-
guishes two varieties of rock, one stratified and offering less resistance

* Loc. cit.

MABERY. — SULPHUR PETROLEUMS. 47

to the drill than the other, which is finer grained and for the most
part loose and granular like sand. For convenience these specimens
may be designated as I. and II. consecutively, and their composition
is shown by the following results of analysis, to which is appended for
comparison the composition of the Trenton limestone at Findlay, III.,
at a depth of 1,096 feet, and of the same oil rock at Lima, IV., at a
depth of 1,247 feet.*

I.

11.

III.

IV.

Calcic carbonate

49.80

49.75

53.50

52.66

]\Iagnesic carbonate

44.35

45.44

43.05

37.53

Silicious residue

0.52

0.80

1.70|

4.15

Alumina and iron

0.46

1.00

1.25]

It therefore appears that the dolomitic condition is not wanting in
the oil-bearing Corniferous limestone ; in fact, magnesic carbonate
seems to be somewhat in excess of the quantity present in the Ohio
Trenton limestone. The specimens were impregnated with oil and
were thoroughly washed with gasoline to remove so far as possible
the carbonaceous portion.

In a distillation of Petrol ia oil under atmospheric pressure, the first
distillate appeared at 115°, and the following weights in grams were
obtained from 800 grams : —

• 115^-150^ 1503-200'^ 200°-250° 250°-300° 300°-350° Residue. Loss.

Weights

22

62.5

72

43

27

561

12

Per cent

2.75

7.8

9.5

5.1

3.1

70.1

1.75

Sp. Gr.

0.767

0.8026

0.8228

0.8345

0.9037

In the proportions that distil at different temperatures and in the
specific gravity of the distillates, the Oil Springs resembles more
nearly the Findlay oil : —

— 100° 100'-150= 150^-200° 200^-250^ 250^-500° 300^-350=' Residue. Loss.

Weights 8 41 90 62 88 47 357 7

Percent 1.14 5.86 12.85 8.86 12.6 6.71 51.0 0.98
Sp. Gr. 0.7335 0.7675 0.7984 0.8222 0.8386 0.9032

A comparison of the distillates at increasing temperatures from oils
of different localities has been included in considering the proper-
ties of the Ohio oils. A clearer idea of the peculiar character of
Canadian oil may be gained by comparing the distillates from it with

* Orton, Geological Survey of Ohio, 1890, p. 13.

48

PROCEEDINGS OP THE AMERICAN ACADEMY.

those from oils of other fields. It is evident that the high specific grav-
ity of the Canadian and Ohio crude oils depends upon constituents
that do not distil below 350°. In this respect these oils differ essen-
tially from the Russian oil, in which all the distillates show a high
specific gravity. The distillates below 150° from Canadian oil are
somewhat lighter than the corresponding products from Ohio crude
oil. In considering later the properties of the vacuum distillates it
will be seen that these proportions are very materially changed by
distillation in vacuo.

Apscheron.

Pennsylvania.

Per cent.

Specific Gravity.

Per cent.

Specific Gravity.

120°-150°

0.5

19.70

150°-200°

10.9

0.786

8.85

0.757

200°-250°

12.8

0.824

15.23

0.788

250°-320°

24.7
48.9

0.861

20.70
64.48

0.809

Residue

51.1

Canada.

35.52

Petroiia.

Oil Springs.

Per cent.

Specific Gravity.

Per cent.

Specific Gravity.

— 100°

1.14

100°-150°

2.75

0.7670

5.86

0.7335

150°-200°

7.80

0.8026

12.85

0.7675

200°-250°

9.50

0.8228

8.86

0.7984

250°-300°

5.10

0.8345

12.60

0.8222

300°-350°

3.10

28.25

0.9037

6.71

48.02

0.8386

Residue

70.10

Ohio.
Per cent.

51.00

Specific Gravity.

110°

-150°

9.75

0.7282

150°

-220°

16.63

0.7669

220°

-257°

10.75

0.7940

257°

-300°

9.75

0.8138

300°

-350°

8.63
55.51

0.8242

Residue

43.00

In the percentages of the lower fractions it will be seen that the
Canadian oil resembles more nearly that from the Caucasus, but the

■ MABERY. — SULPHUR PETROLEUMS. 49

residue above 350° is mucli larger than in the oils from other fields.
As will appear later, this difference is much less in distillations con-
ducted 171 vacuo.

The percentage of sulphur was determined in each distillate by a
combustion in air : —

115^50='

150^-200°

200°-250°

250^-300^

300^-350

Residue.

0.28

0.42

0.50

0.51

0.86

0.70

Sulphur

Determinations of the quantity of bromine absorbed indicated a
greater capacity for absorption in the higher fractions, but less in the
residue than in corresponding fractions from the Ohio oil : —

Canada. Ohio.

Per cent of

Per cent of

Fraction.

Bromine.

Fraction.

Bromine.

115°-150°

0.67

110°-150°

0.73

150°-200°

1.12

150°-220°

1.74

200°-250°

3.49

220°-257°

4.84

250°-300°

8.39

257°-300°

5.04

300°-350°

14.4

300°-330°

12.10

+ 350° 17.8 +330° 19.50

The Oil Springs oil differs essentially in many respects from Petro-
lia oil, especially in its lower specific gravity, lower percentage of sul-
phur, and in the proportions in which it distils at different temperatures.
In certain peculiarities it approaches Ohio oil. The lower fractions
show a higher bromine absorption than either the Petrolia or the Ohio
oil: —

Per cent of
Bromine absorbed.

—100° 0.0

100°-150° 2.31

150°-200° 4.09
200°-250° 8.98

250°-300° 8.41

300°-350° 12.00

+ 350° 33.78

The percentage of bromine absorbed by the crude oils was also
determined : —

Petrolia.

Oil Springs.

Ohio.

15.11

17.69

10.19

VOL. XXXI. (n. S. XXIII.)

4

50 PROCEEDINGS OF THE AMERICAN ACADEMY.

Hydric sulphide escaped in small quautities during the distillation,
but below 2(30° the decomposition was slight, and the distillates were
colorless. Above this point the products were somewhat colored,
with the disagreeable odor of decomposition. It is probal)le that
cracking begins near this temperature, affecting the unsaturated hydro-
carbons if they are present, and perhaps other series as well as the
sulphur compounds. Certain constituents of the Canada oil seemed
to be more unstable than those of Ohio petroleum. The tendency
toward polymerization of unsaturated hydrocarbons separated from dis-
tillates corresponding to burning oil has been observed by me.* An oil
that had been distilled many times in vacuo and allowed to stand two
years, when again heated suddenly polymerised into a higher product
that could not be distilled at any temperature on account of complete
decomposition. The conversion of Canadian petroleum into asphalt,
upon long standing exposed to the weather, is well known. Large
masses of this material may be seen in a pitchy form in the vicinity of
Oil Springs.

Determinations of carbon and hydrogen were made in the coke from
Petrolia oil, in one sample from the crude oil, and in another from a
tar distillate with the followino: results : —

Crude Oil.

Tar Distillate.

c

94.04

94.34

H

4.19

4.34

A specimen of " surface " oil was collected at Oil Springs for exam-
ination. It was very thick, with the consistency of ordinary molasses.
A determination of its specific gravity at 20° gave 0.9059. It con-
tained 0.05 per cent of nitrogen, and 0.95 per cent of sulphur. The
weight of bromine absorbed was equivalent to 25.46 per cent. This
oil is evidently an intermediary product in the formation of the deposits
of pitch resembling asphalt, which have long been known at Oil
Springs, The pitch is evidently formed by evaporation from the oil of
the more volatile constituents, and its formation is doubtless due, in
part at least, to polymerization of lower constituents of less stability.
A combustion of this pitch gave 64.86 per cent of carbon and 8.13
per cent of hydrogen. In a determination of nitrogen 0.40 per cent
was obtained, and the percentage of sulphur was found to be 0.65,
The pitch contained 10.13 per cent of ash, and a qualitative examina-
tion showed that it was composed chiefly of calcic oxide, with smaller

* Amer. Chem. Journ., 1894, p. 92.

MABERY. — SULPHUR PETROLEUMS. 51

amounts of alumiuic and ferric oxides, besides a trace of magnesic
oxide. A determiuation of its bromine absorption gave 37.70 ^ler
cent.

Nitrogen was determined in each variety of coke, and in several
samples of crude oil by the Kjeldahl method : —

Coke from crude oil 0.38

Coke from tar distillate 0.31

Petrolia crude 0.16

Oil Springs 0.18

Oil Springs gas oil 0.21

Sulphur was determined in the two varieties of coke, and in a speci-
men of crude paraffine wax : —

Coke from crude oil 0.76

Coke from tar distillate 0.76

Crude paraffine wax 0.97

When it is remembered that this coke in the process of carboniza-
tion had been exposed to a temperature closely approaching a red
heat, it is not easy to understand how it could retain so large a pro-
portion of nitrogen and sulphur. If the ash was sufficient in quantity,
the presence of metallic cyanides and sulphides might be assumed.
But 0.07 per cent of ash is not sufficient to account for such a high
percentage of sulp.hur. Moreover, careful tests foiled to show the
presence of either suljjhides or cyanides. In determinations of sulphur
by combustion in air, we have frequently found that the carbon must
be completely burned, otherwise the percentage of sulphur is too low.

With the purpose of ascertaining the mineral constituents of Cana-
dian petroleum I procured a specimen of coke from the distillation of
the crude oil, and another from the coking of a tar distillate. Since
finely divided mineral matter is frequently held in suspension in the
crude oil long after it is taken from the well, it might be inferred that
the ash from the crude oil should consist in part of suspended material.
The specimen selected was a part of a large fragment, one side of
which had evidently been carbonized in contact with the bottom of the
still. The portion for analysis was taken from the opposite side, which
had evidently been carbonized several inches above the bottom. The
residue from combustion in oxygen corresponded in the crude oil coke
to 0.17 per cent, and in the coke from the tar distillate to 0.07 per
cent. These quantities of ash correspond to 0.012 per cent in the
crude oil. An examination of the ash showed that it was composed of

52 PROCEEDINGS OF THE AMERICAN ACADEMY.

the oxides of magnesium, calcium, iron, and aluminum. It consisted
chiefly of the oxides of calcium and magnesium, doubtless derived
from the dolomitic reservoir.

A quantity of brine was collected for examination at Petrolia from
a well recently drilled. Analysis showed that it contained calcic sul-
phate, and calcic, magnesic, and sodic chlorides, in the following pro-
portions in 1,000 parts : —

NaCl 10.71

MgClg 2.90

CaCla 1-20

CaSO, 3.20

Iron and alumina traces.

18.00

The specific gravity of this brine at 20° was 1.0165.

The composition of this brine is quite different from that given in
the early history of this oil territory by Dr. T. Sterry Hunt. We
made no examination for potassic chloride. 1,000 parts of the brine
gave in Hunt's analysis : —

NaCl 4.800

KCl 0.792

CaCl2 12.420

MgClg 3.650

21.662

It is practically impossible, at least in glass, to distil the Canadian
oil on a small scale unless it is free from water, and the water can be
removed only by long standing with large quantities of calcic chloride.

After the first distillation there is less difficulty in removing water
except in the least volatile distillates. The necessity of vacuum dis-
tillation to avoid decomposition was even more evident in Canadian
than in Ohio oil. In quantities of 12 litres each, 64.5 kilos were
distilled in a porcelain still under a tension of 50 millimeters, and
the following quantities of the distillates were collected at different
temperatures : —

— 100"

100°-150°

150°-200^

200°-250°

250^-300'"

300° -350°

Residue

Grams

3870

7288

7159

8578

7869

6698

22059

Per cent

6.00

11.3

111

13.3

12.2

10.4

84.2

Per cent in

250"-3o0°

Ohio oil

18.6

19.8

1-5.1

18.0

6.2

20.9

Sp. Gr.

0.7549

0.7852

0.8161

0.8887

0.8647

0.8759

0.9189

Sp. Gr. of

Ohio oil

0.7445

0.7941

0.8245

0.8455

0.9070

0.913

MABERY.

SULPHUR PETROLEUMS. 53

It is peculiar that in distillation under atmospheric pressure as
well as in vacuo the lowest fraction from the Canadian oil is heavier
than that from Ohio oil, while the next following distillates are re-
versed in the order of their specific gravity. Under the influence of
vacuum distillation a large portion of the heavier constituents of the
residue above 350°, under atmospheric pressure, are reduced in boil-
ing points to such an extent that the specific gravity of the lower frac-
tions is very considerably increased. Referring to the specific gravity
of the fractions from the Apscheron oil, page 48, it will be seen that
the corresponding fractions in vacuo from the Canadian oil are much
heavier, and the residue is much smaller. The ditFerences between the
weights collected at different temperatures in the Canadian and Ohio
petroleum also confirm the marked variation in composition already
referred to, and an explanation must evidently be sought in the larger
quantities of the series C2H2 „+ o in the fractions below 150° from
Ohio oil, and the greater quantity of the heavier oils of the series
C„H2« and similar series in distillates from Canadian oil above this
point.

The percentages of sulphur in these distillates were also determined :

—100= 100^-150° 150^-200= 200-^-250= 250^-300= 300^-350^ Residue.

Sulphur 0.25 0.45 0.47 0.75 0.78 0.81 0.83

When distilled without much decomposition the sulphur compounds
in Canadian oil collect in smaller quantities iu the lower distillates
than is the case in Ohio oil.

Determinations were made of the per cent of bromine absorbed by
the vacuum distillates : —

100° 100^-150= 150°-200= 200^-250= 250'=-300= 300^-350= Residue.

Bromine abs. 0 3.25 4.59 6.2 8.2 15.8 25.82

Bromine abs.

atm. pressure 0.67 1.12 3.49 8.39 14.4 17.8

Br. abs. Ohio 250°-360°

oil, vac. dist. 0 4.57 6.60 7.08 24.38

There is a marked difference in bromine absorption between the
distillates collected in vacuo and those collected under atmospheric
pressure. The capacity for absorbing bromine in the distillates from
Canadian oil is greatly increased in the lower fractions by vacuum
distillation, and this difference is even more noticeable in Ohio oil.
It is probably due to the smaller amounts of Canadian oil distilling at
lower temperatures under atmospheric pressure. The larger quanti-

54 PROCEEDINGS OF THE AMERICAN ACADEMY.

ties of bromine absorbed in the fractions collected in vacuo is suggest-
ive. It cannot be caused by decomposition, but it seems to be due
rather to the reduction in boiling points by which certain compounds
capable of absorbing bromine are carried over at lower temperatures,
and doubtless too with less decomposition. The increased absorp-
tion above 250° may indicate cracking, or the presence of normal con-
stituents that absorb bromine. Since the difference in tlie amounts
absorbed in the Canadian oil in vacuo and under atmospheric pressure
is not large except in the residues, and the conditions much less favor-
able for cracking in the vacuum distillates, it would seem that the
absorptive capacity is due to normal constituents of the crude oil.

The fraction 150°-200°, containing 0.47 per cent of sulphur, ab-
sorbed 4.59 per cent of bromine. A portion was treated with alco-
holic mercuric chloride, washed, dried, and it was then found to contain
0.063 per cent of sulphur. A determination of the amount of bromine
it absorbed gave 2.8 per cent. Tliis result is evidently independent of
the sulphur compounds. It must indicate either normal constituents
of the oil with an affinity for bromine, or the presence of decomposi-
tion products due to cracking.

Another portion of the same fraction was agitated with concentrated
sulphuric acid, and the quantity of bromine then absorbed was equiv-
alent to 1.15. Since mercuric chloride has been shown to remove
nearly all the sulphur compounds from the lower fractions, and sul-
phuric acid only partially, it is still further evident that there are pres-
ent in this fraction other bodies capable of absorbing bromine, perhaps
unsaturated hydrocarbons, either contained in the crude oil or result-
ing from decomposition during distillation.

The distillates collected in vacuo showed but slight indications of
decomposition ; they were only slightly discolored, except the residue
above 350°. In prolonged distillation the higher fractions gradually
become colored by polymerization or other decomposition. Even the
residue above 350° showed scarcely any odor, and it had apparently
undergone but little decomposition. The distillates above 150° and
the residue above 350° are reserved for further study.

As in the study of Ohio petroleum, the lower members of the series
C„H2„ + 2 were sought for in the most volatile refinery distillate.
Twenty litres of the veiy first distillate from Petrolia crude oil was
submitted to distillation, and the vapors collected in a Warren condenser
filled with a freezing mixture or with water, according to the boiling
points of the distillates, and with a condensing worm in front filled with
the freezing mixture. Distillates were collected below 45°, and the dis-

MABERY. — SULPHUR PETROLEUMS. 55

dilation of them repeated until they collected for the most part withia
well deliiied limits. The follovviug weights were collected as repre-
senting the quantities of these products in the twenty litres first
distilled : —

-10^

10^-20°

20 '-25°

28°-30°

30=-35°

36^-38°

100

40

30

175

30

80

The small quantities distilling at temperatures outside of the boiling
points of the well known hydrocarbons were not further examined,
since it was assumed that they were merely mixtures. The portion
collecting at 29°-30° in a vapor density determination gave 2.54 ; re-
quired for is opentane 2.49.

A vapor density determination of the fraction 36°— 37° gave 2.66,
which corresponds to the composition of pentane, CgHig.

Distillation of the portion — 10° was continued until 20 grams col-
lected between 7° and 8°. In a vapor density determination by the
method of Hofmann, the value 2.01 was obtained ; calculated for butane,
C^H^,, 2.01. A considerable portion of this oil collected below 5° ;
which consisted chiefly doubtless of the butane that boils at 0°. But
since this body has been identified no further attempts were made
to separate it more completely. Concerning the butane bulling at 8°,
what has been said of the same product separated from Ohio petro-
leum applies also here. It is evident from the weights collected of
these volatile hydrocarbons that they are present in much smaller
quantities than in Ohio petroleum.

The distillate below 150° in vacuo was fractioned twelve times, col-
lecting at first within 5° limits, then within 2°, and finally within 1°,
with the aid of Warren condensers containing glass coils and Hempel
columns. At the end of the eighth distillation, the last seven under
atmospheric pressure, the lower distillates collected as follows, with
smaller quantities between these limits : —

—55° 55°-60° 65°-69° 77°-83°

Grams 8 25 40 65

As was mentioned when considering the lower fractions of the Ohio
oil, these weights evidently represent only approximately the quan-
tities in the crude oil.

After the eighth distillation, the fractions 55°-60° were further
purified, until 15 grams distilled constant at 60°-61°, bar. 749 mm.,
and a vapor density determination of this product gave 2.96 ; required
for isohexane, 2.98.

66 PROCEEDINGS OF THE AMERICAN ACADEMY.

At 67°-G8°, after the twelfth distillation, 10 grams of oil collected,
which gave as its vapor density 3.01 ; required for hexaue, 2.98.
The distillates 75°-85° will be considered with the aromatic series.

At 87°-93°, after the eighth distillation, the distillates amounted to
115 grams, and after the fifteenth, 20 grams distilled at 90°-91°, bar.
745 mm. A vapor density determination of this product gave 3.51 ;
isoheptane, C7H14, requires 3.46.

At the end of the seventeenth distillation, 80 grams collected at
96°.5-97°.5, bar. 740 mm., which distilled constant within these lim-
its. A determination of its vapor density gave 3.63 ; required for
heptane, 3.46. The composition of this oil was further established by
analysis : —

0.1870 gram of the oil gave 0.5781 gram CO2, and 0.2514 gram H2O.

Required for C,Hi8.

Found.

c

84.00

84,31

H

16.00

15.77

Outside of the limits of temperature within which the hydrocarbons
Qi^2h + 2 liave been found, the distillates below 105° have been sub-
jected to prolonged distillation, but the quantities collected were so
irregular, gradually separating into higher and lower constituents, that
it excluded the presence in any considerable quantity of other bodies.
The fractions in the vicinity of 111° will be described later, in the
examination for toluol.

Concerning the distillates collected at 118°-119°, the observations
on the corresponding fractions from Ohio oil apply equally here.
At 118°-119°.5, the distillates amounted to 90 grams after the
fifteenth distillation. A vapor density determination gave 4.02 ; re-
quired for octane, CgHie. 3.94. As in the case of the Ohio products,
this fraction was purified with much care, and the following determi-
nations of carbon and hydrogen were made : —

I. 0.2013 gram of the oil gave 0.6226 gram CO^ and 0.2738 gram
H2O.
II. 0.2036 gram of the oil gave 0.6318 gram CO, and 0.2799 gram
H2O.
III. 0.2045 gram of the oil gave 0.6324 gram CO2 and 0.2762 gram
H2O.

c

Required for

84.20

I.
84.35

Found.
II.

84.61

III.
84.33

H

15.79

15.12

15.28

15.01

MABERY. — SULPHUR PETROLEUMS. 57

From the low percentage of hydrogen and higher percentage of
carbon it was evident that this oil still contained a small amount of
a hydrocarbon with less hydrogen, probably of the series C„H2„,
although the main constituent was evidently a member of the series
^,J^2:i + 2' i" further attempts to eliminate the hydrocarbon with
less hydrogen, the remainder of the oil after analysis was treated sev-
eral times with hot nitric and sulpluiric acids, and boiled repeatedly
witli sodium until there was no residue left on distillation. Although
the boiling point was not materially changed by this treatment, analy-
sis showed a decrease in the percentage of carbon to 83.91 per cent,
and an increase in hydrogen to 16.10 per cent.

As in the Ohio oil, we found that distillates collected with much
persistence between 120°-r26°; after the eleventh distillation, the
following weights were obtained : —

120^-12P 12P-122= 122^-123' 123^24^ 124^-125= 125'=-126' 126=-127° 127°-128°

Grams 85 70 80 60 30 30 70 60

The quantities between 122°-125° were not greatly diminished after
the sixteenth distillation ; and, like the distillates from the Ohio oil,
these products contained small quantities of the aromatic compounds ;
they will receive further attention with reference to their chemical
reactions. After the twelfth distillation, 110 grams of an oil collected
at 126°-128°, bar. 752 mm., which was carefully purified in the man-
ner already described. It gave, by the method of Victor Meyer, a
vapor density corresponding to octane, CgPIis ; found, 4.24 ; required,
3.95. What has been said concernin<^ the presence of the octanes
in Ohio petroleum applies also to these distillates from Canadian oil,
and they evidently require further study. The fractions collected
at 130°-142° will be considered with the aromatic series. A distil-
late persisted at 145°— 146°, as shown by the quantities collected at
the twentieth distillation : —

142^-143^ 143^-144= 144°-145' 145^-146° 146''-147= 147°-148°

Grams 30 32 50 75 25 22

Since the fractions 144°-146° resisted all attempts to separate
them into bodies with higher and lower boiling points, further study is
necessary to determine whether they represent an individual product.

Between 149°-152°, 160 grams collected at the tenth distillation, a
large portion of which distilled constant at 150°-151°, bar. 749 mm.
After treatment with nitric acid and sodium, with results similar to
those observed in the Ohio oil, this product gave, in a vapor density

58 PROCEEDINGS OF THE AMERICAN ACADEMY.

determiuation, a value required for nonane ; found, 4.56; required for
C9H20, 4.43. Distillation of the fractions from Canadian petroleum
above 100° cannot be continued under atmospheric pressure without
decomposition, probably caused by the action of air upon the hot oil.
Unless, indeed, as some attempts have shown, it will be possible to
distil them in an atmosphere of carbonic dioxide. These fractions
have been distilled several times in vacuo, and further study of them is
reserved.

From the results thus far obtained, it seems that the series C^Hgn + j
is represented in Canadian oil by the same members as are found
in Ohio and Pennsylvania oils, but the lower hydrocarbons are
present in much smaller proportions. The peculiar properties of
Canadian oil depend, at least partially, on the small quantities it con-
tains of the hydrocarbons C,jH2„_|_2- The sulphur compounds exert
an important influence. The presence of unsaturated hydrocarbons is
not yet determined. AVhether other series of bodies characterized by
their instability form important constituents of the oil can only be
ascertained by a critical study of the portions with high boiling points.

Aromatic Hydrocarbons.

Series C^Hon-e-

Benzol.

In looking for members of the aromatic series, the same methods
were followed as in fractions of the Ohio oil. At the end of the
eighth distillation, 20 grams collected at 77°-79°, 15 grams at 79°-
81°, and 30 grams at 81°-83°. In treating these fractions, which
contain benzol, with nitric acid under the conditions necessary for the
formation of uitrobenzol, avoiding loss, so far as possible, and distilling
off the unaffected hydrocarbons, the first fraction gave 2.8 per cent of
benzol, the second 4.4 per cent, and the thii-d 4.14 per cent. The
benzol calculated from these data, in the total weight of the fractions,
gave three grams as the total weight in the 64.5 kilos of crude oil
taken, equivalent to 0.0047 per cent. Practically all the benzol
was collected witliin these limits, since scarcely any nitro product was
obtained in the higher and lower fractions. The quantity of benzol in
the crude oil is probably somewhat larger than is here represented,
although it must be considered as somewhat less than the amount
contained in Ohio oil. Some loss undoubtedly resulted from distil-
lation in vacuo, as well as in the subsequent separations. The

MABERY. — SULPEUR PETROLEUMS. 59

nitrobenzol was recognized by conversion into aniline, which gave its
characteristic reaction with furfurol.

Toluol.

In the portions containing tohiol, between 107" -109° after the
eighth distillation, the distillates amounted to 40 grams, at 109°-111°
to 250 grams, and at 111°-113° to 50 grams. Theae fractions were
treated for the formation of uitrotoluol in the same manner as those
previously examined, and the fraction 109°-lir gave a weight of
unaffected hydrocarbon and nitro product equivalent to five per cent
of toluol in the fraction 107°-109° one per cent, and in the fraction
111°-113° one per cent. The higher and lower fractions gave no
nitrotoluol. Referred to the total weights of the fractions, the quan-
tity of toluol was 3.4 grams, equivalent to 0.005 per cent of the total
weight of crude oil taken. As in the case of benzol, it is probable
that the quantity of toluol is somewhat larger than is shown by these
determinations. This product was shown to be toluol by conversion
of the nitro derivative into toluidine, which gave characteristic color
reactions.

Xylols.

The single allusion to the presence of xylols in Canadian petroleum
is a general statement by Schorlemmer that he obtained reactions for
benzol and its homologues, although the only aromatic hydrocarbon he
separated was cumol in the form of the trinitro derivative.

After prolonged distillation, having found that fractions collected
within limits corresponding to the boiling points of the xylols, we have
spent considerable time in the formation of characteristic compounds
as adequate evidence of their presence in the crude oil. In the
twentieth distillation, within one degree after the twelfth, the foUow-
iucr weights collected between 136° and 143° : —

136=-137^ 137=-138= 138^-139= 139'-140^ 140^-141° 141''-142^ 142^-143°

Grams 30 40 25 40 25 47 30

On account of the close proximity of the xylols in boiling points, it
would evidently require much longer time in distillation and larger
quantities of the distillates to separate them completely. We there-
fore depended upon the formation of derivatives that are sufficiently
well characterized to warrant conclusions concerning the xylols from
which they were formed. As a qualitative test for paraxylol, a part
of the fraction 137°-138° was heated with nitric and sulphuric acids,

60 PROCEEDINGS OF THE AMERICAN ACADEMY.

aud the oil which separated was allowed to stand uutil a crystalline
product formed. After crystallization from alcohol, the nitro deriva-
tive thus obtained melted at 139°-140°, the melting point of triuitro-
paraxylol. Another portion of the same fraction was boiled during
several hours with chromic acid, aud the solution extracted with
ether. Upon evaporation, a white powder remained that sublimed
without melting, and resembled in its appearance terephthalic acid.
To determine the quantity of paraxylol according to the method sug-
gested by Levinstein, a measured portion of the fraction lo7°-138°
was shaken during thirty minutes with concentrated sulphuric acid.
The loss in volume corresponded to 10.77 per cent, representing the
other aromatic hydrocarbons. The residual oil was then agitated
with fuming sulphuric acid to dissolve the paraxylol, with a diminu-
tion in volume representing 9.02 per cent of the xylol.

The quantity of metaxylol in fraction 139°-140° was found by
treating it fii-st with dilute nitric acid, which, according to Briickner,*
should not affect metaxylol, and distilling with steam, which carries
over the metaxylol. The loss in volume, 7.5 per cent, was noted,
and the distillate was agitated first with ammonia, then with concen-
trated sulphuric acid ; the last diminution represented 8.8 per cent of
metaxylol.

Many attempts were made to prove the presence of orthoxylol by
the formation of the nitro compounds, and toluic acid. But no di- or
tri-nitro compounds could be separated after treatment with nitiic
acid except those whose melting points corresponded to metaxylol.
By the action of dilute nitric acid, which should not affect metaxylol,
an acid was formed, but not in sufficient quantity to show that it was
toluic acid. In forming the sulphonic acids and the sodium salts, so
far as could be observed, no sodium orthoxylolsulphonate was present.
It is quite possible that orthoxylol could be detected in a larger quan-
tity of product. The larger amounts of the fractions 140°-142°
therefore consisted partially of metaxylol. Referring the quantities
of meta- and para-xylol to the weights of crude oil, the amount of
paraxylol is 0.006 per cent, of metaxylol 0.003 per cent, in their
respective fractions. These numbers cannot be assumed to represent
more than an approximate estimation of these xylols, although they
are doubtless, for the most part, collected in the fractions 137°-143°.
In fractions 141°-143° the amount of sodic xylolsulphonate formed
corresponded to 0.009 per cent of xylol, and, since orthoxylol could

* Ber. der deutsch. chem. Gesellsch., 1876, p. 405. "

MABERY. — SULPHUR PETROLEUMS. 61

not be detected, this quantity should be added to the total amount of
metaxylol in the ci'ude oil. Even then 0.012 per cent is scarcely more
than a trace of this body. As in the similar determinations iu Ohio
oil, these results serve to show the very small proportion of the xylols
contained in the crude oils. In the distillates 130°-140° of the
Canadian oil there were indications of bodies capable of forming nitre
products, but much larger quantities of these fractions will be needed
for their separation.

Series C„H2„.

The fraction 68°- 69° of the twelfth distillation was carefully puri-
fied with alcoholic mercuric chloride, nitric acid, sulphuric acid, and
distillation with sodium. A combustion then gave the following
results : —

0.1889 gram of the oil gave 0.5784 gram CO2 and 0.2719 gram HgO.

Required for

CgH,4 C0H12

Found.

c

83.71 85.71

83.45

H

16.28 14.28

15.99

This fraction therefore consisted essentially of hexane, and it con-
tained no appreciable quantity of hexahydrobenzol.

The fraction 97°-98°, by the action of a mixture of nitric and sul-
phuric acids, gave a nitro compound heavier than water, equivalent to
10 per cent of the weight taken. The heavy oil turned red with sodic
hydrate, and partially dissolved, reprecipitating with acids. By
reduction with tin and hydrochloric acid, a substance with the proper-
ties of an amido compound was formed. It distilled with steam, was
soluble in acids, and was precipitated from the acid solution by sodic
hydrate. The prolonged fractional separation excluded benzol, and
furthermore the amido derivative gave no reaction with furfurol. In
its behavior toward sodic hydrate the nitro compound resembled the
unsaturated nitro compounds of the series C„H9„, but the quantity of
the distillate was not sufficient for complete verification. Another
portion of the same fraction was heated during several hours with
dilute nitric acid, the solution neutralized with sodic hydrate, and
evaporated to dryness. A portion dissolved in water gave, with ferric
chloride, the characteristic color for acetic acid, and more of the sodium
salt, decomposed with hydrochloric acid and extracted with ether, gave
colorless prismatic crj^stals resembling succinic acid.

In a distillate 95°-100°, from the Caucasus petroleum, Beilstein and

62 PROCEEDINGS OF THE AMERICAN ACADEMY.

Kurbatoff* observed tbe formation of acetic acid, considerable succinic
acid, and otlier non-volatile acids. Tlie oil remaining after the treat-
ment with nitric acid, was distilled with sodium, and a combustion gave
the following results already stated on page 56 : —

c

Required 1

C7ll,„.

84.00

'or
C,H„.

85.71

Found.

84.31

H

16.00

14.28

15.77

Since hexahydrotoluol is not affected by nitric acid, but is com-
pletely decomposed by a mixture of nitric and sulphuric acids, it can-
not be present in any appreciable quantity in this fraction.

In examining the fraction 118°-119° for hexahydroisoxylol, after
the sixteenth distillation it was heated with a mixture of nitric and
sulphuric acids. The acid was much colored from decomposition, and
crystals separated which were very sparingly soluble in alcohol. The
purified substance melted at 178°, the melting point of trinitroisoxylol.
Hexahydroisoxylol was therefore present, although it formed but a
small part of this distillate, as shown by the results of combustions
repeated from page 56. It should be borne in mind that the oil
analyzed was purified with nitric and sulphuric acids separately : —

Required for
CjH,g. CjHig.

C 84.20 85.71

H 15.79 14.28

By further treatment with the acid mixture the hexahj'dro com-
pound was sufficiently removed to prove the presence of a hydrocar-
bon C„H2„ + 2' ^s shown by the results of analysis already given.

With the limited quantity of distillates 136°-138°, which should
contain hexahydromesitylene if it is a constituent of Canadian pe-
troleum, remaining from the examination of the xylols, satisfactory
conclusions could not be reached concerning the presence of this hydro-
carbon. Since Markowuikoff has recognized it in the Russian oil, it
is very possibly present in the Canada oil. As in tlie Ohio fractions,
after removing the xylols by continued action of the acid mixture, we
obtained a small quantity of an oily nitro product that did not crystal-
lize. With a larger supply of this distillate, further attempts will be
made.

The lines of study which are now in progress on the Ohio and

I.
84.35

Found.
II.

84.61

III.

84.33

15.12

15.28

15.01

* Ber. (ler deutsch. cheni. Gesellsch., 1880, p. 1820.

MABERY. — SULPHUR PETROLEUMS. 63

Canadian oils have been ii)dicated in the preceding pages. Besides
the higher vacuum distillates sodium salts of acids have been separated
whose composition is now being determined. The facts thus far
accumulated are not sutficient to determine whether the acids exist as
such in the oils, as is maintained by Aschan, or whether they are oxi-
dation products of other constituents, as Engler and Lachowitz
believe. Since sodium salts have been obtained which contain much
nitrogen and give the characteristic odor of pyridine compounds when
heated, the possibility of pyridine carboxylic acids is suggested. In
searching for the oxygen and nitrogen compounds we have separated
bodies that give with nitric acid a brilliant red color, similar to what
was early observed iu our study of the sulphur derivatives.

The characteristic qualities of Canadian petroleum which appear in
the results of this examination may be summarized as follows: —

1. In its high specific gravity and in the proportions that distil at
different temperatures, Canadian petroleum approaches the Russian
oil more nearly than the Ohio petroleum. But the sjiecific gravity of
the distillates is lower, approaching those of Ohio oil. As indicated
by its lower specific gravity, Oil Springs oil is essentially different
from Petrolia oil. This is especially evident in the lower percentage
of sulphur, the larger quantities of the distillates, the higher specific
gravity of these distillates, and the higher bromine absorption.

2. Petrolia oil is composed principally below 150° of members of
the series C„Il2„ + 2? although iu much smaller quantities even* than in
Ohio oil. Another series is present capable of forming nitro products
resembling the nitro compounds of the series C„H2„ + 2' or the unsatu-
rated hydrocarbons C,,!!,^.

3. The aromatic hydrocarbons C„Il2„_e, benzol, toluol, para- and
meta-xylol are present in minute proportions. The hexahydro series
is represented by hexahydroisoxylol, and probably by higher members.

4. By the behavior of the distillates toward bromine, the presence
of hydrocarbons capable of forming addition products is indicated, as
well as the formation of unsaturated hydrocarbons due to cracking in
the distillates above 200° or 250°.

5. As in Ohio oil, the sulphur compounds have a tendency to
collect in the higher fractions.

In the prosecution of this work I have recnved valuable aid from
my assistants, Messrs. Cleveland, Little, and Giessen, and in portions
of the work on the Canadian petroleum, from Mr. W. H. King, a stu-
dent in this Laboratory.

64 PROCEEDINGS OF THE AMERICAN ACADEMY.

Origin of Petroleum.

In the clear and concise statement of the present condition of the theoretical
discussion concerning the formation of petroleum by Professor Edward Orton
(Geological Survey of Ohio, 1890), without including the thory of Mendelejcff
that highly heated iron or iron carbides within the earth may furnish the
world's supply, it is explained that most geologists accept the view that organic
matter of vegetable or animal origin constitutes the source, and that it was
deposited during the formation of the rock strata. Many insist on substances
of vegetable origin as the chief source, and depend upon destructive distilla-
tion as an essential agency. A small minority of the geologists, and some
chemists, especially the Germans, hold that animal remains may be accepted
as the sole source in a process of primary decomposition without distillation.

The chief difficulty in arriving at any satisfactory conclusion concerning the
formation of petroleum depends upon the fact that the principal process is
completed, and there remains scarcely a trace of the stages through which the
original substances have passed ; or indeed these stages may have been so
simple that we have before us in the oil rock all the indications that could ever
have been observed concerning the formation of petroleum. Prevailing opinions
seem to refer the genesis of the limestone oils to the decomposition of animal
remains, and that of other oils to vegetable decomposition.

The most interesting observation on the natural formation of oil that has
come to my knowledge is the experience of Mr. R. A. Townsend, of Petrolia,
who has recently returned from India, where during fourteen years he has been
engaged by the British governinent in prospecting for minerals and oils. In the
oil region of Assam, Beloochistan, and the Punjaub, the surface is bare rock,
and the anticlinals are easily located. Approaching an elevation while pros-
pecting, he found at the top a bell-shaped depression, into which he descended
to a vertical depth of 2,000 feet, and came upon beds of Tertiary oysters from
which petroleum was exuding. The excavation had been formed by a thermal
spring that had disappeared, leaving the strata bare. No oil was observed
above or below the oyster beds. In oil territory owned by Mr. Townsend in
Assam, half decayed tree trunks greasy with oil have been excavated.

One of the serious difiHculties for those who believe in destructive distillation
is an adequate source of heat. Organic matter, animal or vegetable, decora-
poses readily enough when exposed to the air at ordinary temperatures, but the
products are very different from petroleum.

In the early development of organic chemistry, the nature of the products
resulting from the destructive distillation of various forms of organic matter
•was recognized. Dippel's oil distilled from bones contains the nitrogen com-
pounds pyrrol, pyridine, and their derivatives. Reichenbach (Schweigger's
Journal, LXI. 273) identified paraffine as one of the distillation products of ani-
mal and vegetable bodies.

As already mentioned, Warren and Storer established the presence of the
hydrocarbons C„H2„ + 2, C„H2«, and C„H2r,-cas distillation products of a lime
soap prepared from menhaden oil. More recently Engler has also prepared the
petroleum hydrocarbons by the distillation of menhaden oil under pressure.
There can therefore be no question as to the ready formation of petroleum from
animal bodies when decomposed with the agency of heat. But that the same

MABERY. — SULPHUR PETROLEUMS. 65

products may be formed without the aid of high temperatures is not so easily
demonstrated. That any considerable elevation in temperature has accompanied
the formation of petroleum, at^ least in the limestones, is rendered extremely
improbable by tiie condition of the oil rock. Sections of the Corniferous lime-
stone from wells drilled at Petrolia exhibit very plainly the conditions of strati-
fication, without the slightest indication of metamorphic action. The rock
consists of alternate light and dark layers, the light portion being much more
compact than the darker strata, which are more granular, and offer greater
resistance to the drill. As our analyses show, the two varieties of rock do not
differ essentially in their composition. The darker portions of the strata evi-
dently contain more oil.

Concerning the question as to whether petroleums can be sharply divided as
to their origin, the limestone oils having their source in the decomposition of
organic matter of animal origin, and other petroleums in the decomposition of
vegetable matter, more experimental evidence is necessary.

The absence of nitrogenous organic bodies in petroleum has been suggested,
by the adherents of vegetable matter as the source, as a serious objection to
its origin in the decomposition of animal bodies. It is true that in other oils
than those found in the limestones the quantities of nitrogen hitherto found are
extremely small, as already explained (page 16). As further evidence of the
minute proportion of nitrogen in non-sulphur oils, we have determined this ele-
ment in Chinese petroleum (1), in a colorless Italian petroleum (2), in Macks-
burg, Ohio, oil, 1,900 foot level (3), and in a peculiar light yellow Berea Grit oil,
500 foot level (4), from Archer's Fork, Ohio, that is at present refined in large
quantities. The last named oil deposits parafline when cooled to 10°.

Determinations by the Kjeldahl method gave the following results : —

Trenton Limestone Oil. Corniferous Limestone Oils.

(1) 0.10 (1) 0.26 (1) 0.16

(2) 0.014 (2) 0.23 (2) 0.18

(3) 0.035 (3) 0.21 (3) 0.21

(4) 0.023

(1)

0.26

(2)

0.23

(3)

0.21

(4)

0.13

(5)

0.35

(6)

0.08

(7) 0.07

(8)

0.05

(9)

0.05

(10)

0.16

(11)

0.05

(12)

0.06

It is therefore evident that in general a higher percentage of nitrogen is a
distinctive quality of the limestone petroleums.

As mentioned above, we now have in hand certain nitrogenous bodies
extracted from Ohio petroleum which resemble derivatives of the pyridine
bases. In our earlier work on the sulphur compounds in the limestone oils
(These Proceedings, XXV. 228), there were indications that these petroleums
contain certain ethereal oils of vegetable origin beside other oils resembling
the terpenes. A question may therefore arise as to whether the limestone oils
have been derived exclusively from organic matter of animal origin.

VOL. XXXI. (n. S. XXIII.) 5

66 PROCEEDINGS OF THE AMERICAN ACADEMY.

Recently Zaloziecki (Ber. der deutsch. chem. Gesellsch., 1894, p. 2081) has
separated from the sulphuric acid extract of Galician oil, bodies which he con-
cludes indicate the presence in the crude oil of the terpenes or allied substances.

On chemical grounds it seems reasonable that changes in organic bodies
during long periods of time with exclusion of air, under enormous pressure, are
sufficient to explain the formation of petroleum. The modifications that have
been observed in various forms of chemical reactions under high pressure would
seem to indicate that organic bodies should be affected quite differently than
under ordinary conditions. I have in view some experiments on the beha-
vior of organic matter under continued high pressure, fifty to one hundred tons
per square inch, for the purpose of ascertaining what changes, if any, may
result.

RICHARDS AND PARKER. — BARIC SULPHATE. 67

II.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF
HARVARD COLLEGE.

ON THE OCCLUSION OF BARIC CHLORIDE BY
BARIC SULPHATE.

By Theodore William Richards and Harry George Parker.

Presented January 9, 1894.

It has been knowu for a long time that baric sulphate possesses the
power of carrying dowu other substances with it during the process of
precipitation. Many experimenters have investigated with more or less
accuracy the conditions which determine the amount of the occlusion
in different cases, and have sought to eliminate as much as possible
the error which the occlusion must introduce into analytical results.*
Long ago, Fresenius found that some of the imprisoned salts, which
could not be set free by any amount of washing of the freshly pre-
pared baric sulphate, might be partially dissolved by water or acid
after the precipitate had been ignited. Recently J. J. Phinney t has
shown that in some cases, perhaps in all, this purifying process is,
only partial ; hence it becomes important to discover a new method
■which may yield more satisfactory results. As Fresenius pointed out
long ago, I the dissolving of the sulphate in sulphuric acid does not
answer when barium salts are occluded.

The occlusion of baric chloride, the salt which is most generally
used as the precipitant of sulphuric acid, has long been recognized.
In most text-books upon quantitative analysis one is directed to pour
the baric chloride into the sulphuric acid, and not vice versa ; and it is
generally understood that this precaution is in order to prevent as

* A partial bibliography of the subject is given by M. Ripper, Zeitschr. fiir
Anorg. Chem., II. 36. See also Jannasch and Richards, Journ. f. prakt. Chemie,
[Neue Folge,] XXXIX. 321, XL. 236 ; E. A. Schneider, Zeitschr. fiir Physikal.
Chem., 1892, X. 425 ; Lunge, Journ. f . prakt. Chemie, XL. 239 ; F. W. Mar, Am. J.
Sci., [3.], XLL 288 ; P. E. Browning, Am. J. Sci., [3.], XLV. 399 ; J. J. Phinney,
Am. J. Sci., [3.], XLV. 468; Richards, These Proceedings, XXVI. 258 ; XXIX.
67, etc.

t Am. J. Sci., [3.], XLV. 468. t Zeitschr. Anal. Chem., IX. 52.

68 PROCEEDINGS OP THE AMERICAN ACADEMY.

much as possible the error in question. It is also generally known
that when the precipitate cakes together upon ignition, the occlusion
has been large, and the result will usually be too high. But no sys-
tematic series of determinations of the conditions of experiment,
and no exact method for wholly eliminating the error, have been de-
vised and executed ; hence the present work was undertaken with the
idea of solving these difficulties.

In the course of an extended series of experiments upon the atomic
weight of copper* finished several years ago, the attempt was made
to determine the sulphuric acid in cupric sulphate by means of the
usual analytical method. In order to correct the result for the occlu-
sion of baric chloride, the precipitate, after having been weighed, was
fused in pure sodic carbonate ; and the chlorine in the aqueous solu-
tion of the residual cake was determined, calculated as baric chloride,
and subtracted from the total weight of the precipitate.

The result in question was still vitiated by two opposite errors, the
solubility of baric sulphate, and an incorrect atomic weight of barium,
and hence it was rejected at the time. Nevertheless, the method
seemed to be of value, and the stated intention of investigating it more
closely has at last been carried out.

The objects of the present work were as follows : —

First, to discover if an accurate determination of sulphuric acid
may be made by substracting from the weight of the baric sulphate
that of the baric chloride found in it ; and

Second, to define the conditions which determine the amount of
the occlusion, and thus to show how the error may best be avoided
altogether.

A standard solution of pure sulphuric acid (containing 3.2 per cent
of H2SO4) was used as the test substance, being weighed in tightly
stoppered light glass bottles. The baric chloride used for its precipi-
tation had been purified by repeated crystallization, and made up into
another standard solution, which was used in appropriate quantities.
The sodic carbonate was also recrystallized many times ; and this sub-
stance, as well as the nitric acid and all other substances used in the
chlorine determination, were proved to be free from the halogen.

The platinized brass weights used were standardized with great
care, and the weighings were corrected to the vacuum standard. Due
allowance was made in each case for the weight of the filter ash, and
every precaution usual in ordinary analytical work was taken to in-

* These Proceedings, XXVI. 258.

RICHARDS AND PARKER. — BARIC SULPHATE. 69

sure accuracy. It is convenient to state first the experiments which
determined the true strength of the standard acid, although these
were made at a later stage of the work, when more experience had
been gained.

In order to secure a standard of comparison which should be wholly
different from the method under investigation, the strength of the sul-
phuric acid was determined alkalimetrically. Weighed portions of
the solutions were almost neutralized with weighed portions of very
pure gently ignited sodic carbonate, which had been crystallized ten
times successively in platinum vessels. This neutralization was con-
ducted in dilute solutions, so that there might be no loss of substance
from the active effervescence. The dilute solutions were then evapo-
rated to small bulk in platinum dishes, and the small excess of sodic
carbonate required to complete the neutralization was added in dilute
standard solution, methyl-orange being used as the indicator. This
method, which may easily be made to give results accurate to within
the hundredth of one per cent (one part in ten thousand) was devised
for use in the investigation upon copper.

(1.) 18.6132 grams of the sulphuric acid solution required 0.62o4
gram of dry sodic carbonate and 12.30 cubic centimeters of sodic
carbonate solution (1 c.c.= 0.001793 gram of the salt), or in all
0.6475 gram for its neutralization.

(2.) 20.3966 grams of the acid solution required 0.6761 gram of
dry sodic carbonate and 18.50 cubic centimeters of solution — in all
0.7092 gram of salt — for its neutralization.

The results of these analyses are precisely identical, each showing
the acid solution to contain 3.214 per cent of HoSO^. It was impor-
tant now to determine whether under the most favorable circum-
stances similar results might be obtained from the usual method of
precipitation. All the precautions necessary to make the occlusion of
baric chloride as small as possible, which are discussed later, were
practised, and a very slight excess only of baric chloride was used.
The precipitates after long standing were washed with as little freshly
distilled water as was necessary to free them from every trace of acid,
and the filtrates (each measuring about a hundred and thirty cubic
centimeters) were evaporated to very small bulk in platinum dishes,
allowed to stand, and filtered through tiny filters in order to collect
the baric sulphate which had been dissolved by the acid solution.
Finally, the baric sulphate was all fused with pure sodic carbonate ;
and the aqueous extract was acidified with nitric acid. The chlorine
was now precipitated as argentic chloride, which was collected and
weighed upon a Gooch crucible.

70 PROCEEDINGS OF THE AMERICAN ACADEMY.

(3.) 10.2107 grams of the solution gave 0.7804 gram of impure
baric sulphate, collected upon a washed filter and ignited after the
method of Bunsen. In the filtrate was found 0.001 1 gram of addi-
tional precipitate. Only 0.0007 gram of argentic chloride was ob-
tained from the whole of this substance.

(4.) 10.2189 grams of the solution yielded 0.7821 gram of pre-
cipitate, collected and ignited in a Gooch crucible.* The filtrates
yielded 0.0017 more, and the whole of the baric sulphate gave 0.0024
gram of argentic chloride.

If these two determinations are not corrected for the occluded baric
chloride, they give results for the strength of the acid solution which
respectively equal 3.215 and 3.221 per cent of HsSO^. When cor-
rected for this error, they are reduced to 3.213 and 3.215; and the
average of these corrected results, 3.214 per cent, is exactly equal to
that found by alkalimetry.

It is evident that these experiments show clearly enough three
facts : —

First, that, when the occlusion is small, the proposed method is
capable of yielding results which are both consistent and accurate.

Second, that the solubility of baric sulphate may easily introduce an
error of as much as one fifth of a per cent, even when this solubility is
reduced to a minimum.

Third, that the strength of the standard solution being analyzed
was very nearly 3.214 per cent.

These experiments do not show that the method is capable of giving
a true result when the occlusion is large, nor do they show the circum-
stances which determine the amount of the intruding impurity. The
first of these additional questions, which is suggested by the fact that
pure baric chloride parts with some of its chlorine when heated in the
air, will be treated next.

Throughout the following series of seventeen experiments the pre-
cipitation of the baric sulphate was conducted as usual very nearly at
the boiling point of the solutions. The precipitates were digested for
about an hour before filtering, and were each washed with nearly a
litre of boiling water. The baric sulphate was always ignited with its
filter paper ; if the whole mass of the paper is charred by gentle heat
before any active combustion is allowed, no trouble from reduction is

* When good asbestos is used, — and this is not alwayseasy to obtain, — the
Gooch crucible answers admirably for tliis purpose. Compare Jannasch and
Eicliards, Mar, Browning, and Phinney, loc. cit. Ripper must have used a
poor quality of asbestos (Zeitschr. Anorg. Chem., II. 36).

RICHARDS AND PARKER. — BARIC SULPHATE. 71

encountered. In order to prove this, two portions of baric sulphate
ignited with their respective filters were treated with a little sulphuric
acid and ignited again. In neither case was a gain in weight observed.
A comparison of experiment No. 3 with No. 4 shows the same fact ; for
in one of these experiments the precipitate was collected in the usual
way, while in the other it was ignited in a capped Gooch crucible,
where there was no chance of reduction.*

After the precipitate had been weighed, the chlorine present was
determined in the manner used in Nos. 3 and 4. The results are
given on the next page.

The reasons for the varied amounts of occlusion noticed in this
table will be discussed later. At present it concerns us only to note
that, while the variation in the uncorrected column is from 0.3198
gram (experiment No. 5) to 0.3247 gram (experiment No. 11), or 1.77
per cent, the corrected column shows an extreme difference of only
0.53 per cent (0.3192 in Nos. 5 and 12, to 0.3209 in No. 15).
The seventeen experiments may be grouped into two series ; one,
where the occlusion is large, giving a corrected average of 0.3199
(Nos. 7, 8, 9, 10, 11, 12, 13); the other, where the occlusion is
small, giving a corrected average of 0.3201 (Nos. 5, 6, 14, 15, 16, 17,
18, 19, 20, and 21). The close agreement of these averages is suffi-
cient proof that the method is capable of giving satisfactory results
even when the occlusion is very large, and hence that no essential
amount of the occluded chlorine in baric sulphate is liberated upon
ignition.

The individual variations are no doubt chiefly caused by the solu-
bility of baric sulphate in the acid solution from which it came, as is
the deficiency of nearly 0.5 per cent in the final average. It is to
be noted that in Nos. 14, 15, 16, and 17, where the average is higher
than the others, a large escess of baric chloride, which must diminish
the solubility of the sulphate, was used. The amount dissolved was
not determined in any case, as the results were intended to represent
the usual method of analysis, and were rather to be compared with one
another than with an absolute standard.

It is interesting to note, however, that when the precipitation is
conducted with reasonable care — as in Nos. 14 to 21, — the

* This observation does not agree with that of M. Ripper (Zeitschr. Anorg.
Chem., II. .36). It is possible that Ripper conducted his combustion at too high
a temperature ; at any rate, it is true, as he observes, that a slight reduction
would make much more diflference in his work than in such as ours.

72

PROCEEDINGS OF THE AMERICAN ACADEMY.

TABLE OF RESULTS.
Series I.

No. of
Experi-
ment.

Amount of Solu-
tion taken.

Mixed
Precipitate.

AgCI.

II2SO4 per 10 Grams of Solution.

Uncorrected.

Corrected.

6

12.1626

0.9262

.0025

.3198

.3193

6

11.6573

0.8913

.0032

.3211

.3203

7

18.4319

1.0306

.0091

.3223

.3202

8

11.8821

0.9096

.0071

.3215

.3197

9

10.1826

0.7866

.0143

.3244

.3202

10

10.1820

0.7844

.0137

.3236

.3194

11

10.1962

0.7881

.0152

.3247

.3201

12

10.2144

0.7870

.0149

.3236

.3192

13

10.2030

0.7887

.0133

.3246

.3207

14

10.1920

0.7800

.0039

.3214

.3202

16

10.2454

0.7853

.0032

.8219

.3209

16

10.2470

0.7840

.0039

.3214

.3202

17

10.2214

0.7828

.0046

.3216

.3203

18

10.1846

0.7789

.0039

.3212

.3200

19

10.1846

0.7788

.0042

.8212

.3199

20

10.2286

0.7850

.0057

.3223

.3206

21

10.2166

0.7812

.0052

.3212

.3196

True amount of H.iS04 present . .

.3214

amount of the error due to occlusion usually almost balances the
error due to the solution of the precipitate, so that the final uncor-
rected result closely approximates the true one. The average of these
uncorrected results is 0.3215 instead of 0.3214.

Since the amount of argentic chloride obtained from four fifths of a
gram of baric sulphate varied from 0.0005 gram (No. 3) to 0.0152
gram (No. 10), it becomes important to define the exact circumstances

> BICHARDS AND PARKER. — BARIC SULPHATE.

73

of the occlusion. Many of the experiments above were made with
this idea in view.

The two experiments numbered 18 and 19 may be taken as types of
the usual method of working. The volume of the sulphuric acid was
about fifty, and that of the baric chloride about twenty cubic centi-
meters ; the latter solution was slightly in excess, and was poured
gradually into the former.

Series II.

No. of
Experiment.

Mixed
Precipitate.

Argentic
Chloride.

AgCl per Gram of
Precipitate.

18
19

gram.

.7789
.7788

gram.
.0039

.0042

gram.

.0050
.0054

Average

.0052

In order to determine whether an excess of baric chloride increases
the occlusion, a similar series was made, using twice the amount of
baric chloride necessary for the precipitation.

Series III.

No. of
Experiment.

Mixed
Precipitate.

Argentic
Chloride.

AgCl per Gram of
Precipitate.

14

gram.
.7800

gram.
.0039

gram.
.0050

15

.7853

.0032

.0041

16

.7840

.0039

.0050

17

.7828

.0046

.0059

Average

.0050

Since this average is not greater than the last (0.0052), no occlu-
sion could have taken place after the precipitatation was complete, and
the only effect of the excess of baric chloride was to diminish the solu-
bility of the precipitate (see page 71).

In the next series the sulphuric acid was poured into the baric
chloride, instead of vice versa.

74

PROCEEDINGS OF THE AMERICAN ACADEMY.

Series IV.

No. of
Experiment.

Mixed
Precipitate.

Argentic
Chloride.

AgCl per Gram of
Precipitate.

9
10
11
22

gram.
.7866

.7844
.7881
.7832

gram.

.0143
.0137
.0152
.0156

gram.

.0182
.0175
.0193
.0199

.0187

Average of Series II. and III. . .

.0051

These results prove the necessity of the old rule, which directs that
the baric salt should always be added to the sulphate, if any proof of
this was needed.

lu all these experiments a small amount — perhaps the quarter of a
cubic centimeter — of strong hydrochloric acid was added to the sul-
phuric acid before precipitation. In order to determine if this acid
might have had some effect upon the occlusion, four experiments were
made in which the amount of this acid was very much increased. In
Nos. 12 and 13 about ten cubic centimeters of strong acid were used, and
in Nos. 23 and 24 fully twenty. The first two were allowed to stand a

Series V.

No. of
Experiment.

Mixed
Precipitate.

Argentic
Chloride.

AgCl per Gram of
Precipitate.

12

gram.

.7870

gram.
.0149

gram.
.0189

13

.7887

.0133

.0169

23

.7211

.0266

.0369

24

.7286

.0223

.0306

Average

.0264

Average of Serie

s II. and III. . .

.0051

RICHARDS AND PARKER. — BARIC SULPHATE.

75

long time, in order that the precipitate, which separates slowly from
strongly acid solutions, might have time to deposit; while the second
two were filtered within an hour, with the loss of nearly ten per cent
of the precipitate.

Thus the presence of free hydrochloric acid increases the occlusion
to an enormous extent.

In order to determine the effect of dilution, other circumstances
remaining the same, two precipitations were made from solutions like
those of Series II., diluted fourfold with water.

Series VI.

No. of
Experiment.

Mixed
Precipitate.

Argentic
Chloride.

AgCl per Gram of
Precipitate.

25
26

gram.
.7676

.7701

gram.
.0023

.0024

gram.
.0030

.0031

Average

.0031

Average of Series 11. and III. • .

.0051

Hence, the more dilute the solutions, the less is the occlnsion.

In all the preceding cases the precipitate was allowed to run down
the side of the beaker in an excessively fine stream, with continual
stirring. This procedure had been shown in preliminary experiments
to be even more efficacious as a means of diminishing occlusion than
the addition of the precipitant in definite drops, no matter how slowly.
Two experiments, in which the baric chloride was poured in with
great rapidity, close the data to be presented.

Series VII.

No. of
Experiment.

Mixed
Precipitate.

Argentic
Cliloride.

AgOl per Gram of
Precipitate.

20
21

gram.
.7850

.7812

gram.
.0057

.0052

gram.
.0073

.0067

Average , .

.0070

76 PROCEEDINGS OF THE AMERICAN ACADEMY.

As was to have been expected, the impurity is greater here than in
Series II., III., and VI., but less than in Series IV. and V.

It was not the intention of this paper to extend the investigation to
the study of the occlusion of other chlorides, which must necessarily
complicate the problem greatly. It is very interesting to note that the
amount of occlusion seems to be due rather to the amount of chlorine
present than to the amount of barium, for hydrochloric acid increased
the occlusion almost as much as an equivalent amount of baric chloride
did. This observation leads one to conclude that a careful study of
the phenomenon from a physico-chemical point of view might furnish
some satisfactory clue as to the nature of such occlusion in general ;
and it is hoped that before long a collection of suitable data upon the
problem may be obtained here.* The present paper deals only with
the solving of an analytical problem.

At first sight, the facts shown by this paper appear to be contradic-
tory to those stated by F. W. Mar,f who recommends the use of strong
nitric and hydrochloric acids for the precipitation. It must be borne
in mind, however, that Mar was precipitating the barium with an
excess of sulphuric acid, and hence that a considerable amount of
occlusion could make very little difference in his case. When one is
determining sulphuric acid by means of an excess of barium, nothing
could be worse than the addition of a large amount of hydrochloric
acid ; and this point cannot be too strongly emphasized.

Mar rightly states that a considerable excess of sulphuric acid is
necessary to secure rapid precipitation of all the barium in strongly
acid solutions. If Fresenius had realized that an excess of a common
ion in solution decreases the solubility of a precipitate, he would not
have confounded the solubility of baric sulphate in a strong solution of
baric chloride with its solubility in pure water, J or in water contain-
ing free hydrochloric acid.

The conclusions reached in the present paper are as follows : —

First, that the occlusion of baric chloride by baric sulphate may lead
to very serious error.

Second, that the amount of this occlusion is greater in concentrated
than in dilute solutions, greater in the presence of hydrochloric acid
than in its absence, and greater when the sulphate is poured into the

* For the treatment of an analogous case, see E. A. Schneider, Zeitschr.
Phye. Chem., X. 425.

t Am. J. Sci., [3.], XLI. 288, XLV. 399.
} Zeitschr. f. Anal. Chem., IX. 52.

RICHARDS AND PARKER. — BARIC SULPHATE. 77

barium salt thau when the pouring takes place in the opposite
direction.

Thir^, that under the usual conditions of careful precipitation in the
presence of a small amount of free acid, the error from occlusion is
almost balanced by the solubility of baric sulphate in acids and water,
which solubility must be considered in careful work.

Fourth, that the error due to this occlusion may be corrected with
great exactness by determining the amount of chlorine held by the pre-
cipitate, and subtracting the corresponding amount of baric chloride
from the total weight of the precipitate.

The occlusion of iron and other substances by baric sulphate is
being further studied in this Laboratory.

78 PROCEEDINGS OF THE AMERICAN ACADEMY.

III.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF
HARVARD COLLEGE.

ON THE CUPRIAMMONIUM DOUBLE SALTS.

THIRD PAPER.

By Theodork William Richards axd George Oenslager.

Presented June 9, 1894.

The work already done upon the double salts of cupriammonium *
suggested the possibility of obtaining compounds of fluorine and iodine
similar to those of chlorine and bromine.

Many attempts were made to prepare cupriammonium acetofluoride,
Cu(NH3)2FC2H302, with invariable failure. The methods and the pro-
portions of the reagents were varied in every possible way ; but
nothing could be obtained beside cupriammonium acetate and ammo-
nium fluoride. Cupriammonium fluoride itself is very difficult to pre-
pare, because of its solubility ; and it is not unnatural that the double
salt should be more so. It will be remembered, moreover, that while
cupriammonium acetobromide is very easy to obtain, the normal
chloride is almost if not quite impossible by the wet way. We should
expect the fluorine compound to be even less possible. Because of
these continued failures, the work with fluorine was discontinued.

The investigation of the compounds of iodine proved much more
fruitful, yielding the following new substances : —

(1.) Cu(NH3)3lC2Hs02.

(2.) 7 CU(NH3)2(C2H302)2.CU(NH3)2IC2H302.

(3.) 3 Cu(NH3)2 . M NH3.

(4.) 2 Cu(C2H30o)2NH4CoH302 . HjO, besides confirming the work
of Foerster t on the following substances : —

(0) Cu(NH3)o(C2H302)3.

(6) CU(NH3)2(C2H302)2 2JH20.

* Richards and Shaw, These Proceedings, XXVIII. 247 ; Richards and
Whitridge, These Proceedings, XXX. 458.

t Berichte der deutsch. chem. Gesellsch., XXV. 3416.

RICHARDS AND OENSLAGER. — CUPRIAMMONIUM SALTS, 79
(1.) Ajimon-cupriammonium Aceto-iodide,

CU(NH3)3IC2H302.

In preliminary experiments the following method was used for pre-
paring this compound. Four grams of finely powdered cupric acetate
were stirred with a mixture of ten cubic centimeters of water, twenty
of alcohol, and eight of glacial acetic acid. Ammonia gas was then
passed into the mixture until all the cupric acetate was dissolved and
the color of the solution had become deep blue. After the addition of
three and a half grams of ammonic iodide the solution was set aside to
crystallize, and in about six hours large deep blue monoclinic plates,
having a six-sided outline, separated out. These crystals were
washed with alcohol, and then dried between filter papers as quickly
as possible. The salt thus prepared is usually not very pure, hence
subsequently more alcohol was used in the preparation, with better
success. Even yet, however, there was much room for improvement.
/ After many systematic experiments, which need not be detailed, the
following method was found to yield very excellent results. Twelve
grams of cupric acetate were dissolved in fifty cubic centimeters of
ammonia water (sp. gr. = 0.90) in a flask. After cooling, thirty cubic
centimeters of aqueous acetic acid (57 per cent) were added to the
solution, then six grams of ammonic iodide, and finally fifty cubic
centimeters of alcohol. Upon boiling over the water bath the mixture
yielded a clear deep blue solution, which deposited crystals of ammon-
cupriammonium aceto-iodide upon slow evaporation in the air. The
first very small crop of crystals deposited upon cooling was not
analyzed. The second served for Analysis IV. below, and the third
crop, which was altogether the purest and best defined in crystalline
form, served for Analyses V. and VIII. The substance used in
Analyses I., II., III., VI., and VII. was prepared by earlier less satis-
factory methods ; it was undoubtedly the same substance, however.

Ammon-cupriammonium aceto-iodide consists of brilliant deep blue
monoclinic crystals. It is not very permanent in the air, although
much more so than its chlorine and bromine analogues. Upon ex-
posure for a long time the crystals become dull and dark in color,
and the substance slowly loses in weight. Water at once decom-
poses it, some of the copper going into solution, and the rest remaining
as a basic precipitate. Acids set free iodine and precipitate cuproiis
iodide, as might be expected. The only unexpected property of the
salt is the fact that it contains no crystal water, thus not maintaining a
strict analogy with the corresponding chlorine and bromine compounds

80 PROCEEDINGS OF THE AMERICAN ACADEMY.

Cu(NH3)3ClC2H302.H20 and Cu(NH3)3BrC2H302. H.O. The rea-
son for this anomaly remains obscure.

The ammonia present in the compound was determined by distilla-
tion with pure potash or soda, the distillate being titrated with stan-
dard acid. It is convenient to prepare the alkali for this purpose
directly from metallic sodium or from sodium amalgam, thus avoiding
the complications in the subsequent proceedings introduced by the
usual impurity of chlorine in commercial potash and soda. The res-
idue in the distilling flask after the ammonia had been expelled was
filtered, and to the filtrate was added a little sulphurous acid to reduce
any iodate which might have been formed. Argentic nitrate was then
added, and subsequently the argentic oxide and sulphite were dissolved
by nitric acid, the argentic iodide being collected and weighed upon a
Gooch crucible. In Analysis IV. the alkali was nearly neutralized,
and in Analysis V. the solution was just acidified by pure nitric acid
before the precipitation. In every case enough nitric acid was added
after the precipitation to insure the solution of all but argentic iodide,
and the agreement of the results is sufficiently satisfactory. The cop-
per was determined by the electrolysis of the cupric sulphate obtained
from the precipitated cupric oxide, which contained no trace of iodine;
and the acetic acid was determined by combustion.

Analyses of Cu(NH3)3lC2H302.

I. 0.2591 gram of the substance on distillation with potash re-
quired 25.34 cubic centimeters of decinormal acid for neutrali-
zation, gave on electrolysis 0.0545 gram of cojiper, and yielded
0.2042 gram of argentic iodide.
II. 0.2524 gram of the substance required 25.05 cubic centimeters
of decinormal acid and gave 0.0527 gram of copper.

III. 0.2322 gram of the substance required 23.25 cubic centimeters

of decinormal acid.

IV. 0.3960 gram of the substance gave 0.3109 gram of argentic

iodide.
V. 0.2559 gram of the substance gave 0.2004 gram of argentic
iodide.
(VI.) 0.3211 gram of the substance gave on combustion 0.0939 gram

of carbon dioxide.
VII. 0.2946 gram of the substance gave on combustion 0.0864 gram
of carbon dioxide.
VIII. 0.3062 gram of the substance gave on combustion 0.0907 gram
of carbon dioxide.

RICHARDS AND OENSLAGER. — CUPRIABIMONIUM SALTS. 81

No.

Copper.

Ammonia.

Iodine.

Acetic Acid.

I. .
II. .

III. .

IV. .
V. .

VI. .
VII. .
VIII. .

21.03
20.88

16.73
16.95
17.10

42.58

42.41
42.31

19.62
19.64
19.86

Averages .

20.96

16.99

42.43

19.71

Copper
Ammonia
Iodine
Acetic acid

Calculated for

Cu(NH3)3lC2H302.

Found.

21.15

20.96

17.03

16.99

42.19

42.43

19.63

19.71

100.00

100.09

(2.) OCTOCUPRIAMMONIUM MoNO-IODIDE AcETATE,

Cu8(NH3)ieI(QH30,)i5.

On allowing the mother liquor left over from the first method of
preparing ammon-cupriamraonium aceto-iodide to stand for a long time,
large coal-black hexagonal crystals of unknown composition were de-
posited. Mixed with these were large blue crystals, which were sepa-
rated mechanically from the black ones, and analyzed. The analysis
corresponded closely with the formula

7 Cu(NH3)2(C2H302)2 4- Cu(NH3)2TC2H302,

the complexity of which led to the suspicion that the crystals were a
mixture instead of a definite compound. Nevertheless, upon qualita-
tive testing, the smallest as well as the largest crystals were found to
contain iodine.

In external appearance the crystals, which were usually at least half
a centimeter in length, resembled those of cupriammonium acetate.
An attempt was made to measure the angles of the crystals of each

VOL. XXXI. (n. 8. XXIII.) 6

82

PROCEEDINGS OF THE AMERICAN ACADEMY.

substance, but the faces were so covered with striations that the result
was only partially satisfactory. For cupriammonium acetate the chief
prism angle was about 70°, while for the complicated salt under con-
sideration it was over 71° 30'. Even allowing a considerable margin
for possible error, it would appear that the complicated salt must be a
definite compound, and not merely cupriammonium acetate containing
occluded cupric acetate and ammonic iodide. Another evidence of the
probable definiteness of the salt is to be found in the fact that Richards
and Moulton, in a paper yet to be published, have proof of the exist-
ence of a similar compound containing aniline and bromine instead of
ammonia and iodine. The compound has no unexpected properties,
and was analyzed in the usual fiishion.

Analysis of Cu8(NH3)i6l(C2H30,),5.

I. 0.2545 gram of the substance distilled with potash required
22.68 cubic centimeters of decinormal acid, gave 0.0333 gram
of argentic iodide, and on electrolysis yielded 0.0707 gi-am of
copper.
II. 0.2120 gram of the substance required 18.52 cubic centimeters
of decinormal acid.
III. 0.2419 gram of another sample of the substance required 21.23
cubic centimeters of acid, and gave 0.0331 gram of argentic
iodide and 0.0G80 gram of copper.

No.

Copper.

Ammonia.

Iodine.

I

11

ni

27.78
28.11

15 25
14.93
14.99

7.07
7.39

Averages . .

27.95

15.05

7.23

Copper

Ammonia

Iodine

Acetic acid (by dif )

Calculated for
aboTe Formula.

28.35

15.22

7.07

49.36

100.00

Found.

27.95

15.05

7.23

49.77

100.00

KICHARDS AND OENSLAGER. — CUPRIAMMONIUM SALTS. 83

Every effort to make the normal cupriammoniura aceto-iodide free
from the acetate was unsuccessful ; when so much of the ammonia
has evaporated that Cu(NIIg)3lC2Ho02 ceases to form, the singular
double salt which has just been described always makes its appear-
ance.

(3.) Tetrammon-tricupriammonium Iodide,
3 Cu(NI-l3) X + 4 NH3= Cu8T6(NH3)i,.

This interesting substance may be prepared by a method very
closely resembling that used for preparing ammon-cupriammonium
aceto-iodide. If eight grams instead of twelve of cupric acetate are
used with fifty cubic centimeters each of ammonia and alcohol, thirty of
acetic acid, and six grams of ammonic iodide, curious black cr^'stals re-
sembling irregular triangular pyramids make their appearance in the
first place.

The new substance is similar in outward aspect to the corresponding
bromine compound,* although less brilliant. The crystalline faces are
so singularly marked and striated that an accurate crystallographic
study would not be feasible. They possess a distinct bronze lustre
which soon disappears owing to superficial decomposition. Upon ex-
posure to the air the substance loses ammonia and iodine, finally leav-
ing cuprous iodide. It is decomposed by water. Heated in aqueous
or alcoholic ammonia it dissolves, forming a deep blue solution which
upon cooling deposits bright blue needles remaining to be investigated.

From the mother liquors decanted from the black crystals may be
obtained by further evaporation at first, Cu(NH3)3lCoH302, and finally
the mixture of Cu8(NHo)i6l(C2H..02)i5, with the coal-black hexagonal
crystals already mentioned. f These latter crystals are very different
in appearance from Cu;;(NH3)ioT6; they will be investigated in the
future.

Analysis of Cu3(NHo)ioT6>

I. 0.2049 gram of the substance yielded an amount of ammonia re-
quiring 18.22 cubic centimeters of decinormal acid for neutrali-
zation, and 0.0352 gram of copper on electrolysis.
II. 0.2315 gram of the substance required 20.24 cubic centimeters
of acid, and gave 0.0385 gram of copper and 0.2901 gram of
argentic iodide.

* Richards and Shaw, These Proceedings, XXVIII. 257.
t See page 81.

84

PROCEEDINGS OF THE AMERICAN ACADEMY.

III. 0.2302 gram of the substance required 20.36 cubic centimeters of

acid, and gave 0.2872 gram of argentic iodide.

IV. 0.2307 gram of the substance required 20.35 cubic centimeters of

acid.
V. 0.2150 gram of the substance on electrolysis gave 0.0367 gram
of copper.

No.

Copper.

Ammonia.

Iodine.

I. .

II

Ill

IV

V

17.17
16.63

17.07

15.19
14.94
15.11
15.07

67.73
67.48

Averages . .

16.96

15.08

67.61

Calculated for

above Formula.

Found.

Copper

17.00

16.96

Ammonia

15.20

15.08

Iodine

67.80

67.61

100.00

99.65

(4.) Ammonic Dicupric Acetate, NH4Cu2(C2H;,02)5- HoO.

Since compounds containing three molecules of ammonia, one of
acetic acid, and one atom of chlorine, bromine, or iodine had been
proved to exist, it became a matter of interest to discover if it were
possible to prepare Cu(NH3)3(C2H.^02)2, in which the halogen is re-
placed by acetic acid. Many unsuccessful attempts were made to
obtain this compound. In the course of these experiments, however,
a new double salt of cupric and ammonic acetates was discovered.
Four grams of cupric acetate, six grams of glacial acetic acid, and
twenty cubic centimeters of alcohol were mixed together, and ammonia
gas was passed into the mixture until the green color just turns into
blue. The solution became hot from the absorption of the ammonia,
and the cupric acetate dissolved readily. After standin-z; a few hours
many small bluish green crystals were deposited, which have the com-

RICHARDS AND OENSLAGER. — CUPRIAMMONIUM SALTS. 85

position given below. The crystals are soluble in water without de-
composition, and are fairly permanent in the air.

Analyses of Cu2NH4(C2H302)5 • H^O.

I. 0.2586 gram of the substance yielded an amount of ammonia re-
quiring 5.49 cubic centimeters of decinormal acid, and an elec-
trolysis gave 0.0717 gram of copper.
II. 0.4303 gram of the substance required 9.40 cubic centimeters of
decinormal acid.

III. 0.2525 gram of the substance gave 0.0703 gram of copper.

IV. 0.2604 gram of the substance required 5.52 cubic centimeters of

acid, and gave .0725 gram of copper.
V. 0.3080 gram of the substance gave on combustion 0.2970 gram
of carbon dioxide.

No.

Copper.

Ammonium.

Acetic Acid.

I. . . . .

27.73

3.84

II

—

3.9G

Ill

27.86

IV. ... .

27.84

3.84

V

—

—

64.45

Averages . .

27.81

3.88

61.45

Calculated for

above Formula

Copper

27.75

Ammonium

8.94

Acetic acid

64.38

Water (by differ

ence)

3.93

100.00

Found,

27.81

3.88

64.45

3.86

100.00

(5 and 6.) Cupriammonium Acetate.

This salt, discovered by Foerster,* may be prepared in an anhydrous
condition by allowing an alcoholic solution of cupric acetate containing
a slight excess of ammonia to evaporate in the air. Two determina-

* Ber. der deutsch. chem. Gesellsch., 1892, XXV. 3416.

86 PROCEEDINGS OP THE AMERICAN ACADEMY.

tions of the amount of copper present in the large deep bkie crystals
showed respectively 29.41 and 29.32 per cent, the theoretical being
29.48. The ammonia was found to be 15.68, instead of the theoretical
amount 15.82 per cent, corresponding to the formula Cu(NHo)2C2H302.
The crystallized salt containing two and a half molecules of crystal
water, also discovered by Foerster, was prepared in blue feathery
crystals by the evaporation of aqueous ammoniacal cupric acetate.
Analysis showed the percentage of copper present to be 24.23 instead
of 24.38, and the ammonia to be 13.02 instead of 13.08 per cent.
The theoretical values were calculated for the formula

CU(NH3)2(C2H302)2 . 2i H2O.

All of these compounds, as well as many similar ones obtained from
other acids and the substituted ammonias, will be further studied at
this Laboratory in the near future.

RICHARDS AND MOULTON. — CUPRIANILINE SALTS. 87

IV.

CONTRIBUTIONS FROxM THE CHEMICAL LABORATORY OF
HARVARD COLLEGE.

ON THE CUPRIANILINE ACETOBROMIDES.

By J^heodore William Richards and Frederic Charles

MoULTON.
Presented June 9, 1894.

Recent investigations * have shown that double salts of cupriam-
monium containing two acids are easily obtainable in great numbers.
It becomes a matter of interest to determine whether the substituted
ammonias are capable of acting in the same way ; and the present
paper contains an account of the first experiments in this direction.

The only cupraniline compounds of which descriptions were found
are three cuprosaniline compounds described by Saglier,f and cupri-
aniline chloride (CUCI22C6H7N) described by Destrem.J Hence there
was a wide field left open for the present investigation.

The compounds described in this paper are tabulated below.

(1.) Cu(C6H,N),Br2.

(2.) Cu(CoH,N)2BrC2H.302.

(3.) Cus(aHvN)ieBri5(C2HA).

(4.) CMC,U,^\,Br,,{C,U,0,),.

(5.) Cu;(C6H,N)ieBr,3(C2H.302)3.

(6.) Cu2(C6H7N)4Br(aH302)3.

(7.) Cus(CoH,N)ioBr(C2H30)i5.

Of these the simpler ones, Nos. 1 and 2, are undoubtedly definite
compounds. Whether or not some of the others are isomorphous mix-
tures of these with cuprianiline acetate, which has not as yet been
prepared in a pure state, it is hard to say.

* Richards and Shaw, These Proceedings, XXVIII. 247; Richards and
Whitridge, These Proceedings, XXX. 458 ; Richards and Oenslager, These Pro-
ceedings, XXXI. 78.

t Comples Rendus, CVI. 1422.

t Bulletin de la Soc. Chim. de Paris, XXX. 482.

OS PROCEEDINGS OF THE AMERICAN ACADEMY.

(1.) CupRiANiLiNE Bromide, Cu(C6H7N)2Br2.

This compound is formed as a brown powder whenever cupric bro-
mide is added to an alcoholic solution of aniline ; but a much moi'e
beautiful crystalline preparation is obtained when the solution contains
acetic acid. For example, three grams of cupric bromide were dis-
solved in about forty cubic centimeters of alcoliol ; and to this were
added twenty cubic centimeters of acetic acid and fifteen cubic centi-
meters of aniline. The mixture turned green immediately, and beau-
tifully brilliant brown crystalline flakes of cuprianiline bromide were
deposited. Nothing could be easier than the preparation of this sub-
stance.

From analogy to the cupriammonium compounds, one would have
expected the cuprianiline aceto-bromide to have been formed under
conditions given above ; but this is not the case unless the solution
contains a large per cent of cupric acetate.

Among a great number of samples of cuprianiline bromide, which
have all given satisfactory analytical results, at least four colors have
been observed. A few preparations consisted of long glossy black
crystals, others consisted of dark brown shining flakes. Yet others
were of a lighter brown, and some were almost of a golden yellow.
Possibly the difference was due to variations in the thickness of the
individual crystals.

The new compound is insoluble in alcohol and glacial acetic acid.
When mixed with water it is decomposed at once into a basic bromide
of copper, aniline hydrobroraide, and a little free aniline. Alkalies
and acids of course decompose it.

In order to prepare the substance for the electrolytic determination
of the copper it was treated with nitric acid, evaporated to dryness,
and ignited. The residue was dissolved in a mixture of nitric and sul-
phuric acids ; the former acid was expelled on the steam bath, and the
copper was separated in the usual fiishion. Bromine was determined
as argentic bromide ; and the aniline was calculated from the amount
of carbon dioxide yielded by the combustion of the compound. As a
final check upon the results, the nitrogen present was determined
according to the method of Dumas.

Analyses of Cu(C6H7N)2Br2.

I. 0.1978 gram of the substance yielded 0.0.'>08 gram of copper.
II. 0.2299 gram of the substance yielded 0.0357 gram of copper.

RICHARDS AND MOULTON. — CUPRIANILINE SALTS. 89

III. 0.2009 gram of the substance yielded 0.1842 gram of argentic

bromide.

IV. 0.2530 gram of the substance yielded 0.2318 gram of argentic

bromide.
V. 0.1028 gram of the substance gave 0.1313 gram of carbou diox-
ide upon combustion.
VI. 0.1705 gram of the substance yielded 10.20 cubic centimeters of
moist nitrogen at 23° C. and 763.7 mm. pressure.

No.

Copper.

Bromine.

Aniline.

Nitrogen.

I. . .
II. . .

III. . .

IV. . .
V. . .

VI. . .

15.57
15.52

39.02
38.94

45.00

6.77

Average

15.55

38.98

45.00

6.77

Calculated for

above Formula.

Found.

Copper

15.53

15.55

Bromine

39.04

38.98

Aniline

45.48

45.00

100.00

99.53

Nitrogen

6.83

6.77

Many other similar determinations were made upon other samples,
with similar results.

(2.) CUPRIANILINE ACETOBROMIDE,

Cu(C6H,N)2(C2H30,)Br.

This compound is formed with considerable difficulty, for the reac-
tion tends to produce a compound containing either little or no acetic
acid, or else little or no bromine. For the first time it was formed as
follows. Three grams of cupric bromide were dissolved in about fifty
cubic centimeters of alcohol, three grams of cupric acetate were dis-

90 PROCEEDINGS OF THE AMERICAN ACADEMY.

solved in ten cubic centimeters of commercial acetic acid, with the
addition of twenty cubic centimeters of water; the two solutions thus
formed were mixed, and fifteen cubic centimeters of aniline were added
to the mixture. In about half an hour fine brown crystals began to sep-
arate out, and these crystals proved to be cuprianiline acetobromide.
Subsequently a great deal of time was expended in attempting to repeat
this experiment, with continued failure. At last a few of the original
crystals were sprinkled upon the surface of a mixture like the one just
described, at the time when the crystals were expected ; and a large
mass of the long sought for substance immediately separated. The
reasons of the inconsistency of the first success and the subsequent
failures remains to be discovered. It is possibly connected with vari-
ations in temperature.

The new compound is nearly insoluble in alcohol and glacial acetic
acid, and is partially decomposed by water. Potassic hydroxide re-
moves all the bromine from it only with difficulty. It is easily broken
up by concentrated nitric acid, but not until heat has been applied.

Because of the stability of the compound it was necessary to deter-
mine the bromine in a rather elaborate manner. The sample was in-
timately mixed with a large amount of sodic carbonate in a porcelain
crucible, covered with a thick layer of the same salt, pressed down
very thoroughly, and very gradually brouglit to a red heat. As the
mass contracted, more sodic carbonate was sprinkled around the edges
of the mixture, to prevent the escape of a trace of bromine. The
fused mass was dissolved in hot water, and the sodic bromide was
filtered off and determined as usual.*

Analyses of Cu(C6H7N)2C2H302Br.

I. 0.2100 gram of the first sample of substance yielded 0.0343 gram
of copper upon electrolysis.
II. 0.1768 gram of the first sample of substance yielded 0.0860 gram
of argentic bromide.

III. 0.1743 gram of the second sample of substance yielded 0.08GG

gram of argentic bromide.

IV. 0.1232 gram of the second sample of substance yielded 0.0600

gram of argentic bromide.

* This method was found to give the same results as that of Carius with
these compounds. Of course the method would not answer if the bromine were
directly combined with carbon.

Calculated for

Found.

above Formula.

I.

II. III.

Copper

16.38

16.33

Bromine

20.70

20.70 21.1

RICHARDS AND MOULTON. — CUPRIANILINE SALTS. 91

20.80

(3.) OCTOCUPRIANILINE MoNOACETOBROlIIDE,

This substance has been made repeatedly as follows. Two grams
of cupric bromide were dissolved in a hundred cubic centimeters of al-
cohol, and one and eight tenths grams of cupric acetate were dissolved
in ten cubic centimeters of acetic acid, with the addition of just enough
water to effect the solution. The two solutions were mixed, and
fifteen cubic centimeters of aniline were added. In a quarter of an
hour the desired crystals separated out from the deep gi-een solution.
The crystals resemble fine gunpowder in appearance, and possess
properties more nearly resembling those of compound No. 1 than
those of compound No. 2.

Analyses of Cu3(C6H7N)i6Bri5(C2H302).

I. 0.1158 gram of the substance yielded 0.0181 gram of copper.
II, 0.1158 gram of the substance yielded 0.0996 gram of argentic
bromide.

III. 0.2000 gram of the substance yielded 0.1738 gram of argentic

bromide.

IV. 0.2712 gram of the substance yielded 0.2354 gram of argentic

bromide.

Calculated for Found.

Cug(CeH7N),eBr,5C2H302. I. II. m. iv.

Copper 15.62 15.71

Bromine 36.82 36.58 37.06 36.94

(4.) OCTOCUPRIAXILINE Tri ACETOBROMIDE,

Cu8(C6H,N)ieBri3(C2H302)3.

Once this compound was obtained by adding 1.86 grams of aniline
to a mixture of 2.23 grams of cupric bromide and 2.0 grams of acetic
acid in alcoholic solution. The second time it was obtained by adding
a few crystals of this first preparation to a mixture similar to that used
for preparing No. 2, just as it was on the point of crystallization.
The properties of this substance resemble very closely those of the
last ; and its analysis was conducted in the usual manner.

92 PEOCEEDINGS OF THE AMERICAN ACADEMY.

Analysis of Cu8(C6Tl7N)i6Bri3(C2H302)3.

I. 0.1754 gram of the substance yielded 0.0278 gram of copper.
II. 0.1754 gram of the substance yielded 0.1318 gram of argentic
bromide.

III. 0.2445 gram of the substance yielded 0.1865 gram of argentic

bromide.

IV. 0.2200 gram of the substance yielded 0.1677 gram of argentic

bromide.

Calculated for Found.

CU8(C6U,N),cBr3(C2U302)3. I. II. III. IV.

Copper 15.82 15.85

Bromine 32.30 32.00 32.46 32.44

(5.) DiCUPRIANILINE ACETOMONOBROMIDE,

This substance is formed with comparative ease, and was often ob-
tained in the unsuccessful attempts to prepare cuprianiline acetobromide
(No. 2). Its preparation is more certain if somewhat less cupric bro-
mide— two grams instead of three — is used in the mixture described
under that head. This variation of desirable conditions was to have
been expected. Dicuprianiline acetomonobromide may also be formed
bv addint^ acetic acid to an alcoholic solution of cupric bromide until
the color of the solution has become green, and then addiug the aniline
as before. After several hours, in either case, the compound separates
as a dark bluish-black powder, insoluble in alcohol and glacial acetic
acid. It is not decomposed by water, and boiling concentrated potassic
hydrate will only remove a very small part of the bromine. At a
moderate heat it is surprisingly stable ; but upon ignition it is decom-
posed with considerable violence. When brought in contact with a
drop of hot concentrated nitric acid it is almost explosive in the energy
of its decomposition.

Analysis of Cu2(C6H,N)4(C2HA).iBr.

I. 0.1849 gram of the substance yielded 0.0310 gram of copper
upon electrolysis.
II. 0.2310 gram of the substance yielded 0.0560 gram of argentic
bromide.

III. 0.2053 gram of the substance yielded 0.0515 gram of argentic

bromide.

IV. 0.1168 gram of the substance yielded 0.2026 gram of carbon

dioxide upon combustion.

RICHARDS AND MOULTON. — CUPRIANILINE SALTS.

93

Copper
Bromine

Calculated for
Cu2(C„H,NUBr(C2H30,)3.

16.82
10.57

I.

16.77

Found.
II. III.

10.32 10.6

Carbon

. 47.61

IV.

47.31

(6.) OCTOCUPRIANILINE ACETOMONOBROMIDE,

Cu3(CeH,N)x6Br(C,H302)i5.

This, tlie last compound on the list, is formed with great ease when-
ever a large excess of cupric acetate or acetic acid is added to cupric
bromide in alcoholic solution and aniline is added to the mixture. It
separates after about twenty-four hours as a fine bluish or brownish
black powder, whose properties resemble very closely those of the pre-
vious compound.

Analysis of Cu8(C6H7N)i6Br(C2H302)i5.

I. 0.1540 gram of the substance yielded 0.0264 gram of copper
upon electrolysis.
II. 0.5098 gram of the substance yielded 0.0338 gram of argentic
bromide.

III. 0.3352 gram of the substance yielded 0.0212 gram of argentic

bromide.

IV. 0.2117 gram of the substance yielded 0.0134 gram of argentic

bromide.
V. 0.1983 gram of the substance yielded 0.3672 gram of carbon
dioxide.
VI. 0.4346 gram of the substance yielded 28.8 cubic centimeters of
moist nitrogen at 759.4 millimeters pressure and 23° C.

No.

Copper.

Bromine.

Carbon.

Nitrogen.

I. . . .

17.14

II. . . .

—

2.82

III. . . .

2.69

IV. . . .

—

2.68

V. . . .

—

—

50.51

VI. . . .

—

—

—

7.46

Averages .

17.14

2.73

50.51

7.46

94 proceedings op the american academy.

Compound No. 6,

Calculated for

CusCCeHjNH^JieBrCC.HaO,)!^.

Found.

Copper

17.18

17.14

Broiuiue

2.69

2.73

Carbon

51.04

50.51

Nitrogen

7.58

7.46

It is apparent that all the compounds described in this paper may
be arranged in a series, based upon the formulaj so multiplied as
to include eight molecules of the cuprianiline group Cu(Ccn7N)"-.
At one end of this list stands cuprianiline bromide, and at the other
should stand cuprianiline acetate. Only six of the seventeen com-
pounds necessary to fill out this scheme have been found : —

1. Cu,(CeH,N)ieBri6.

4. Cu«(C6H,N)i6(C2H302)3Bri3.

2. Cu«(C6H,N)ic(C2H302),Br8.

5. Cu8(CeH,N)i„(C2H30,)i„Br4.

6. Cu,(C6H,N)i6(CJ430,)i5Br.

These compounds exhibit a change in properties which corresponds
to the progressive change in composition. The stability of the product
and its difficulty of decomposition increases as the per cent of bromine
decreases. The violence of the reaction with a drop of strong nitric
acid increases in the same way. As the acetic acid increases and the
bromine diminishes, the crystalline form becomes less and less evident,
the first compound consisting of well marked crystals, and the last of
an impalpable powder. The average color of the substance becomes
blacker and bluer as one proceeds in the same direction. The prepa-
ration seems to depend upon two factors : in the first place, upon the
relative proportions of cupric acetate and bromide taken at the start,
and in the next place, upon the time required for the crystallization.
When the conditions of temperature and dilution are such that crystal-
lization takes place at once, a crystalline compound, rich in bromine,
separates. When, on the other hand, many hours elapse before the
substance appears, a powder very poor in bromine is formed. It is
possible that the inversion temperature of these double salts may vary

RICHARDS AND MOULTON. — CUPRIANILINE SALTS. 95

progressively, so that the use of a thermostat would be necessary to
secure absolute certainty in the yield.

The possibility that most of these compounds might be mere mix-
tures has been already suggested; but it is evident that some at least
must be definite compounds. The fact that in the case of Nos. 2 and 4
a few crystals of an earlier preparation were capable of starting the
crystallization of the same substance in larger quantity, throws valu-
able light upon this point. The fact that most of the substances fit so
closely into the series having eight atoms of copper as the basis seems
to indicate a definite law ; for the chances are decidedly against coin-
cidence in this regard. The per cent of bromine is of course the best
means of determining the agreement with the theory, and the agree-
ment is very satisfactory. Each preparation appeared under the
microscope to be perfectly homogeneous, so that if the substances are
mixtures, the components must crystallize together.

From this series of compounds one would infer the possible exist-
ence of cuprianiline acetate, but many repeated trials to prepare this
compound in a pure state ended in failure. Several different new com-
pounds were obtained during this search, one of them probably having
the formula Cu2(C6H7N)3(C2H302)4, for it contained 19.82 per cent of
copper. Another compound, which consisted, like this last, of long
light blue needles, contained only 19.10 per cent of copper ; while two
others, a brown and a black powder, contained respectively 21.80 per
cent and 17.28 per cent. All of these compounds, together with
others containing chloriue and iodine instead of bromine, and other
substituted ammonias in place of aniline, will soon be investigated
further at this Laboratory.

96 PROCEEDINGS OF THE AMERICAN ACADEMY.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY
OF HARVARD COLLEGE.

THE CHEMICAL POTENTIAL OF THE METALS.*

By Wilder D. Bancroft.

Presented by C. L. Jaekaon, May 9, 1894. t

Nearly four years ago Overbeck and Edler J published the results
of their experiments on cells consisting of two metals in a single salt
solution. Liquid amalgams of zinc, cadmium, tin, lead, and bismuth
were used as anodes, while mercury served always as kathode. The
results were as follows : the electromotive force of these cells is a
function of the metals forming the electrodes, and of the negative
ion of the salt solution ; it is independent of the nature of the
positive ion so long as this is not the same as the anode. The
values for the electromotive force coincide, within the limit of error,
with those for the corresponding, constant, reversible cells of the
Daniell type.

With the exception of the conclusion that the electromotive force is
smaller when the cell is reversible in respect to the anode, these re-
sults have been confirmed by the experiments of Paschen § and of
Gore. II The obsers^ations permit of other deductions not drawn by the
authors. If we consider cells of the type MJRXIMa and M^IRXlMg,
where Mj, Mo, M3 represent the electrodes, and RX a solution of
the salt RX, the difference of the two electromotive forces gives the
value of the cell M1IRXIM3IRXIM2. By PoggendorflPs law this is
the same as the cell MjIRXlMj. From the figures of Overbeck and

* The first part of this paper has already been publislied, Zeitschr. f. ph.
Ch., XII. 289, 1893.

t Before tliis paper could be printed there appeared the articles of Goodwin
and of Neumann. I have rewritten the part covered by their work so as to take
account of their results and the proper date of the paper in its present form is
December, 1894.

t Wied. Ann., XLIL 209, 1891.

§ Ibid., XLIIL 568, 1891.

II Phil. Mag., [5.], XXXIII. 28, 1892.

BANCROFT. POTENTIAL OF METALS.

97

Edler we can calculate the values for the combinations of zinc, cad-
mium, tin, lead, and bismuth in different salt solutions. I have done
this, and I give a few of the figures in Table I. The electromotive
forces are expressed in volts, as will always be the case when nothing
is said to the contrary. The first column gives the electrode, the
raetal to the left being the anode. The concentrations of the different
solutions are not comparable, and are therefore omitted. I have raised
the value given by Overbeck and Edler for Piilllg in Nal by 0.025
volts, and that for Bi| Hg in NaCl by 0.075 volt. There is no doubt
in my mind that these two values had been affected by secondary
causes, and that this correction was necessary iu order to make them
comparable with the other observations.

TABLE L

Electrodes.

KCl.

NaCl.

KBr.

NaBr.

KI.

Nal.

ZnCd

0.374

0.365

0 369

0.352

0.374

0.367

ZnPb

0.564

0.558

0.548

0.534

0.559

0.557

ZnBi

0.797

0.789

0.782

0.767

0.781

0.789

CdPb

0.190

0.193

0.179

0.182

0.185

0.190

CdBi

0.423

0.424

0.413

0.415

0.407

0.422

PbBi

0.233

0.231

0.234

0.233

0 222

0.232

The result is surprising. The hifluence of the negative ion dis-
appears, and, if we considered this table alone, we should conclude that
the electromotive forces of the cells MilRXIMg depended on the nature
of the electrodes only. The experiments from which this table was
compiled showed that, when Mo = Hg, the negative ion is of impor-
tance. It is clear that the metals separate into two classes, the distinc-
tion being in respect to the behavior towards the negative ion. From
the figures of Overbeck and Edler it appears further that zinc and cad-
mium, for instance, give the same value in all solutions of halogens ;
but that this is not the case for solutions of sulphates and nitrates.
The question arises whether this difference is due to errors of obser-
vation, or whether we have to do with a special property of the halo-
gens. The question cannot be settled by a reference to any existing
experiments. Gore * has made a series of observations recently ; but

* Phil. Mag., [5.J, XXXIII. 28, 1892.

VOL. XXXI. (n. S. XXIII.) 7

98 PROCEEDINGS OF THE AMERICAN ACADEMY.

his measurements are so inaccurate that no conclusions can be drawn
from them. The results of Paschen* with dropping mercury elec-
trodes show signs of the same regularities which are found in Table I. ;
but the variations in the observations are so great that the matter
cannot be said to be definitely settled. It seemed to me necessary to
make measurements in order to clear up the following points. Does
the effect of the negative ion disappear in certain cases? Is this a
characteristic of the halogens only ? What is the effect of the concen-
tration ? May one look upon these cells as the limiting cases of the
constant, reversible cells of the Daniell type?

Most of the experimental work recorded in this paper was done in
the chemical laboratory at Amsterdam, and I am much indebted to
Prof, van 't Hoff for assistance in carrying out this research.

The metals used came from the collection of the laboratory. I did
not purify them further owing to lack of time. For this reason the
absolute value of some of the figures may not be correct ; but that has
no effect on the general laws, which alone interested me. I purified
the mercury myself ; but I did not succeed in keeping the sur-
face bright very long. The salt solutions were made by weighing
out carefully the amounts of salt necessary for the concentrated solu-
tions, and the others were made from these by diluting to the required
volumes. The measurements were made with the small Lippmanu
capillary electrometer. As normal element I used a Latimer Clark
cell; its value was found by Ilerr Barendrecht to be 1.426 volts. I
compared it from time to time with a Gouy cell, and found no change
which amounted to more than a millivolt. As working cell I had
also a Leclanche with about the same electromotive force as the
Clark cell. Its variations were quite considerable, about 3%. The
measurements against mercury were made in a small test tube which
had a platinum wire melted into it; the others were made in a U tube
or in an ordinary beaker. The metals, with the exception of platinum,
were used in the form of rods or heavy wire. At first I polished them
with sand-paper ; but this proved disadvantageous, as the surface is
brought into a state of stress by this which is incompatible with accu-
rate measurements. I have found it profitable to cut away the surface
with a sharp knife, and then to wipe the metal with a cloth. There is
the danger of a piece of steel getting on the electrode; but in most
cases this will not cling firmly, and will be removed by wiping. It is
true that the surface tension of a metal is often changed temporarily

* Wied. Ann., XLIII. 5G8, 1801.

BANCROFT, — POTENTIAL OF METALS. 99

by cutting off a piece ; but experiment shows that this change passes
away more quickly than that caused by rubbing. In order to obtain
accurate figures the electrodes should not be jarred. In some experi-
ments the electrodes can be moved about without any very notice-
able effect ; but in most cases the slightest tremor has a very great
influeuce on the electromotive force. As a rule, the electromotive
force of this type of cell increases on standing, more or less quickly,
till it reaches a definite maximum, where it remains constant. It is this
maximum which I have given in the tables, because it is always the
same in different experiments, while the value first observed is very
irregular. In two cases only have I done differently. With magne-
sium one can get almost any desired value, according to the way the
measurement is made. Immediately after setting up, the value for
MglZn is about 0.56 volt. If the magnesium is left in the solution
and the liquid stirred slightly, the value becomes 0.713 volt. If the
liquid be stirred vigorously, the reading rises to 0.76 volt, and even
higher. In the last two cases there is a violent evolution of hydrogen.
I have taken the first value as the most probable, and consider it as
possibly too high. In the literature the recorded figures for MglZn
lie between 0.53 and 0.80 volt. The reason for this discrejiancy is,
as I have just said, in the varying conditions. With lead it is a little
different. One gets at first a definite value, which remains constant
quite a while, and then increases slowly to a maximum which I have
not determined. As this " stationary " value is an easy one to deter-
mine, I have measured it ; it evidently corresponds to a well defined
condition, and is therefore just as good for most purposes of compari-
son as the highest value. The probable error of the measurements is
not over 0.01 volt in most cases, and has no effect on the general
relations.

I now come to the experimental data. The concentrations are
given in chemical units (gram molecules per litre). The single values
are the averages of five to twenty observations. Table II. shows the
effect of the concentration, Table III. the effect of the negative ion
when the electrodes are any two of the metals Mg, Zn, Sn, Pb, or Bi.
The measurements of ZnlBi and CdIBi had one marked peculiarity.
The electromotive force CdIBi was at first about 0.315 volt. This
potential difference increased slowly but regularly till it reached the
value given in the table. In all the cases which I have studied Zn|Bi
and CdIBi are the only ones in which there were signs of any rational
connection between the change of the electromotive force and the time.
I wished to include aluminium in the list ; but the purest Neuliausen

100

PROCEEDINGS OF THE AMERICAN ACADEMY.

TABLE II.

Electrodes.

Concentration.

KCl.

KBr.

K2SO4.

MgZn

A

0.562

0.561

0.562

MgZn

ri<s

0.561

0.561

0.562

MgZn

1000

0 561

0.564

0.563

ZnCd

A

0.332

0.333

0.334

ZnCd

lis

0.334

0.333

0.335

ZnCd

IddO

0.334

0.332

0.333

ZnSn

1
T5

0.655

0.654

0.655

ZnSn

Tff7

0.656

0.654

0.655

TABLE III.

Electrodes.

KCl.

KBr.

KI.

K,S04.

KNO3.

KAc.

K,C03.

KjOxal.

MgZn

0.562

0.561

0.563

0.562

0.502

0.563

0.565

0.562

MgCd

0.895

0.894

0.893

0.897

0.895

0.896

0.897

0.894

ZnCd

0.333

0.333

0.331

0.334

0.332

0.332

0 334

0.331

ZnSn

0.655

0.654

0.652

0.655

0.655

0656

0.654

0.657

CdSn

0.326

0.327

0.323

0.323

0.325

0.325

0.323

ZnPb

0.526

0.528

—

0.527

0.526

Cdrb

0.195

0.194

—

0.194

0.193

ZnBi

0.829

0.829

—

0.832

CdBi

0.496

0.497

—

0.497

aluminium contains traces of iron and behaves almost like iron. In
Table IV. I give the measurements for Zn, Cd, and Pt against mer-
cury, and also the measurements for Cd|Pt in several solutions. The
figures in this table are not very accurate, with the exception of those
made against mercury in a halogen solution. While the measurements
are very easy to make in KCl, KBr, or KI solutions, with the other
salts there appear such large and apparently causeless variations that

BANCROFT. — POTENTIAL OF METALS.
TABLE IV.

101

Electrodes.

KCl.

KBr.

KI.

K,S04.

KNO3.

KAc.

KXO3.

KaOxal.

ZnHg

1.151

0.991

0.847

1.302

1.200

1.228

0.995

1.301

CdHg

0.818

0.659

0.515

0.969

0.867

0.898

0.664

0.967

Difference

0.333

0.332

0 332

0.833

0.3.33

0.330

0.331

0.334

HgPt

—

—

—

0.106

0.092

0.063

0.079

0.095

CdHg + HgPt

—

—

—

1.075

0.959

0.961

0.743

1.062

CdPt

—

—

—

1.076

0.963

0.959

0.741

1.067

I do not feel sure that there are not errors of 0.1 volt in spite of the
fact that I have taken great pains to find the real values. The two
points which are of importance may be considered as settled : the
concentration has no effect, and the negative ion of the salt solution has

TABLE V.

Cd I KCl I Hff.

ToVo

1

Too

A

lOOO

ToO'

1
Iff

0.814

0.818

0.822

0.816

0.820

0.822

0.818

0.817

0.817

0.817

0.818

0.817

0.821

0.817

0.818

0.817

0.817

0.818

0.832

0.816

0.817

0.817

0.820

0.819

0.813

0.804

0.822

0.818

0.817

0.822

0.821

0.817

0.818

0.817

0.817
0.817

0.817
0.818

0.821
0.822

0.818

Averages.
0.817

0.819

an influence both in the combinations ZnlPt, Cd|Pt, and in HglPt.
In Table V. there is a complete series of observations on the effect of
the concentration. I have made similar series with several other solu-
tions ; but as they gave the same result I do not communicate them.
When in a halogen solution one of the electrodes is platinum, both the

102 PROCEEDINGS OF THE AMERICAN ACADEMY.

TABLE VI.

Coacentr.

1
HglBaCUIPt. i

Concentr.

HglBaCljIPt.

Found.

Calc.

Found.

Calc.

27(70

0.158

0.158

Toff

O.IUS

0.194

T50'S

0.164

0.166

^V

0.212

0.213

2o~o

0.186

0.186

tV

0.220

0.221

coiicentration aucl the nature of the solution have a great influence on
the electromotive force. In Table VI. are the results for different solu-
tions of barium chloride. The figures under the heading " Calc." are
obtained from the formula

allowance being made for the dissociation. The value for the ^^^
normal solution is taken as the basis of the calculation ; Ii—2 and

TABLE VII.

Concentr.

HglKCllPt.

HglKBrlPt.

IlglKIl Pt.

Cd|KBr| Pt.

IlglKCllPt.

Found.

Calc.

Found.

Calc. Found.

Calc.

Found.

Calc.

Found.

Calc.

ToW

0.152

0.152

0.233

0.233 0.388

0.387

0.893

0.892

0.127

0.127

•505

0.164

0.164

0.245

0.245

0.400

0.399

0.905

0.904

0.138

0.139

Tffff

0.190

0.190

0.271

0.271

0.425

0.425

0.930

0.9.S0

0.105

0.165

.v

0.201

0.201

0.282

0.282

0.436

0.436

0.943

0.941

0.179

0.176

1

Iff

0.227

0.227

0.308

0.308

0.463

0.462

0.967

0.967

0.201

0.202

i

0.238

0.237

0.319

0 319

0.473

0.473'

0.977

0.978

0.212

0.213

T= 290. In Table VII. I give the figures for solutions of KCl, KBr,
and KI. The change of the electromotive force with the concentration
follows the formula

E,-E, = %RT\og

g:

using as before the value for y^ normal as a starting point. I have
"riven two series of determinations for HglKCllPt. In the first set of

BANCROFT. — POTENTIAL OF METALS. 103

experiments I found only the smaller values. Repeating the measure-
ments with other solutions I obtained chiefly the figures given in the
first column, although now and then the values for the other series
appeared. As I could not discover the reason for these sudden
clianges of 0.036 volt, I give both series, though I am inclined to re-
gard the larger values as the more probable.

The first glance at the table reveals the striking regularity of the
phenomena. Where the electrodes are any two of tlie metals Mg,
Zn, Cd, Sn, Pb, or Bi, the negative ion has no effect ; the electro-
motive force remains unaltered when the concentration changes from
10 to 1,000 litres. The influence of the negative ion is very marked
in the cells where mercury forms one pole ; but there is no sign of any
effect due to the concentration. This appears first in the combina-
tions with platinum, and then only in the solutions of the halogens.
This abnormal behavior of platinum is, as Ostwald* has pointed out,
probably due to a tendency to go over into chlor-, brom-, or iod-
platinates, where the platinum is no longer present as ion. It is worth
noticing that a similar result was not detected with mercury, though
here there is a tendency to a partial formation of complex salts.
Whether the greater accuracy of the measurements with mercury in
solution of halogens is connected with the possible existence of complex
salts, 1 do not know.

The relation between these single liquid jwlarizable cells and the
corresponding, constant, reversible cells of the Daniell type must next
be considered. According to the theory of Nernst, the potential differ-
ence between a metal and a solution of a salt of that metal is given by
the expression

P T P

TT = -^ log — X 10"^ volts, t

where tt is the potential difference, n the valency of the kation,/) its par-
tial osmotic pressure, e the quantity of electricity transported by a gram
equivalent, and P the solution pressure of the electrode. The electro-
motive force of a cell of the Daniells type Mil/)jMiX|;>2M2X|M2 will
be the algebraic sum of the two potential differences between the
metals and the solutions plus the difference of potential between
the liquids. I leave out of account a possible potential difference be-

* Lehrbuch der AUgem. Chem. (2 Aufl ), II. 897.
t Nernst, Zeitschr. f. ph. Chem., IV. 148, 1889,

104 PROCEEDINGS OF THE AMERICAN ACADExMY.

tween the metals, as this term is negligible so far as our present
knowledge goes. The electromotive force of this type of cell will be

n e \ ® PMi ° Pi)

+

where z represents the difference of potential between the solutions,
and the valency of the metals M^ and Mg is the same. If the wander-
ing velocities of the ions Mi and Mg are nearly equal, and pi and j^a

be made so, the value of z approaches zero, while the term log —
drops out entirely. The electromotive force of the cell

MiljsMiXl/jM^XIMa

is given very nearly by the expression

TT = log ~y^ X 10-^ volts,

ne ® Pmi

and is independent of the absolute concentration of the salts MjX and
MoX. Let us take as a concrete case the cell Zn|ZiiS04lCuS04|Cu,
and let the concentrations of the ZnS04 ^.nd CUSO4 always be equal.
It has been found experimentally that the electromotive force of this
cell is independent of the absolute concentration.* Suppose that, in-
stead of diluting the two solutions with pure water, we add a solution
of potassium sulphate. According to Nei-nst's theory, this will have
no influence on the electromotive force except in so far as it affects the
dissociation of the two sulphates, and thereby the concentrations of the
Zn and Cu ions. If the dilution be carried far enough, we shall come
at last, without change of electromotive force, to the cell with neither
zinc nor copper sulphate, to the cell

ZnIxZnS04 + xK2S04lxCuS04 + xK2S04|Cu,

which is the same as the cell Zn|xK„S04lCu. In other words, the
one-liquid non-reversible cells are the limiting cases of the two-liquid
reversible cells, in which the concentrations and wandering veloci-
ties of the reversible ions are equal, the dissociation being supposed
to be complete. This last clause is necessary, for if the percentage
dissociations of the zinc sulphate and cojiper sulphate were different,

* Wright, Phil. Mag., [5.], XIII. 265, 1882.

BANCROFT. POTENTIAL OP METALS. 105

equal concentrations of the two sulphates would not correspond to
equal concentrations of zinc and copper ions, and this would affect the
potential difference between the solutions. The concentration of the
K2SO4 should have no effect, and it was shown in Tables II. and V.
that this was the case. It is clear that in measurements made with
two-liquid reversible cells, there are two sources of error besides those
due to the surface conditions of the electrodes. These are differences
of concentration and differences of wandering velocities. The effects

of these two errors are that the terms log ^-^ and z do not disappear.

The determinations made with single-liquid cells are free from these
sources of error ; but the difficulties due to polarization are so great
that the variations are apt to be much larger than in measurements
made with two-liquid reversible cells. In Table VIII. I give some of
the results obtained with the two styles of cells. In the first four col-
umns are the measurements of Paschen,* myself, Overbeck and Edler,t
Ostwald,! all made with single-liquid cells. In the next three are
the figures of Wright and Thompson, § Neumann, || Braun,^ with re-
versible cells. In the eighth are the data of Magnanini,** and in the
ninth those of Regnauld,tt the former being for polarizable, the latter
for non-polarizable cells. The agreement is not as striking as one
might wish ; but it is sufficient. The values marked with a star are
not properly comparable, because the two solutions were not of the
same concentration.

Nernst's formula for the cells we have been considering is

TT = — log ^ X 10-" volts.

It is therefore necessary to discuss the nature of log P. Nernst has
not made any direct statement, so far a>i I know, about a possible con-
nection between log P and the negative ion of the salt solution.
Ostwald J J and his pupils look upon log Pas a function of the electrode

* Wied. Ann., XLIIL 590, 1891.

t Ibid., XLII. 209, 1891.

t Zeitschr. f. ph. Cliem., I. 583, 1887.

§ Phil. Mag., [5 ], XIX. 1, 1885.

II Zeitschr. f. ph. Chem., XIV. 193, 1894.
1 Wied. Ann., XVI. 575, 1882.
** Rend. Ace. Lino., VI. 182, 1890.
tt Wiedmann, Elektricitat (2 Aufl.), I. 792.
tt Lehrbuch, II. 855.

106

PROCEEDINGS OF THE AMERICAN ACADEMY.

TABLE VIII.

Elec-
trodes.

Electro-
lyte.

Pas-
chen.

W.D.B.

0 &

E.

Ost-
wald

W. & T.

Neu-
manu.

Braun.

Maji-
nani-
Di.

D=100.
Reg.
naulj.

ZiiCd

Chlorides

0 296

0.333

0.368

0.360

0.330

0.329

0.334

0.32

0.235

ZnCd

Bromides

0.293

0.333

0.364

0.340

0.315

—

0.256

0.30

0.235

ZiiCd

Iodides

0.298

0.331

0.365

0 304

0.322

—

0.262

0.20

0.235

ZnCd

Sulphates

0.350

0.834

0.430

0.401

0.3G0

0.362

0.33-37

0.36

0.307

ZiiCd

Nitrates

—

0.332

0 440

0.411

0.352

0.352

0.27-37

0.38

0.235

ZnCd

Acetates

—

0.332

—

0.373

—

—

0.336

ZnPb

Cldorides

0.512

0.526

0.561

0.610

0.591*

0.598*

—

0.51

ZnPb

Bromides

0 525

0.528

0.541

0.599

0.571

—

—

0.45

ZnPb

Iodides

0.545

0.527(1)

0.558

0.587

0.455

—

—

0.38

ZnPb

Sulphates

0.525

0 527

0.502

0.592

0.50-55*

—

—

0.51

ZnPb

Nitrates

0.526

0.589

0.598

0.585

0.589

0.440

0.51

ZnPb

Acetates

—

0.527(1)

—

0.638

0.607

0.601

0.54-58

CdPb

Chlorides

0.21G

0.195

0.192

0.249

0.260*

0.269*-

CdPb

Bromides

0.232

0.194

0.181

0.259

0.256

CdPb

Iodides

0.247

0.194(1)

0.188

0.256

0.24*

CdPb

Sulphates

0.18

0.194

0.17

0.191

0.13-17*

CdPb

Nitrates

—

0193

0.243

0.187

0.233

0.237

0.18-22

CdPb

Acetates

—

0.1941)

—

0.265

—

—

0.240

metal and the temperature only, and liold that it is independent of the
nature of the negative ion. If tliis be so, we ought to find that all
cells of the type MilpMJXftM.^XIM, should have the same value so
long as Ml and M.y remain the same, and that a change in X should
have no effect, barring secondary disturbances such as differences of
wandering velocity, of dissociation, etc. In the non-reversible cells
M1IRXIM2, where these disturbing influences are eliminated, this
should be even more noticeably true. That this is the case for certain
metals I have shown in Tables II. and III. The results of other in-

(^) Values marked thus are calculated from the other experiments, and are
not direct observation.

BANCROFT. POTENTIAL OF METALS.

107

vestigators, as given in Table VIII., show the same thing, though not
quite so clearly. The values for ZniCu in solutions of chlorides, bro-
mides, and iodides are found to be identical by Paschen, by Overbeck
and Edler and by Regnault, though the three sets differ hopelessly in
absolute value. Braun makes the bromides and iodides the same, and
puts the chlorides, sulphates, and nitrates in a group together. There
is not the same agreement among the reversible cells in which Pb
forms one of the electrodes ; but this is due in part to the insolubility
of the lead salts. With the polarizable cells things are much clearer,
though the discrepancies between the values found by different ob-
servers complicates matters very much. Ostwald finds practically the
same value for ZulPb in all solutions except acetates. Paschen makes
the bromides and sulphates the same, while Overbeck and Edler find
the chlorides and iodides identical. On the whole, we may say that
the theory of Nernst has predicted the facts with great accuracy so far.
If, however, the single-liquid cells are the limiting cases of the two-
liquid reversible cells, and if log P is a function of the electrodes and
temperature only, the electromotive force should always be indepen-
dent of the nature of the negative ion of the salt solution. That this is
not so will be seen from Table IX,

Tx\BLE IX.

Electrodes.

Electrolyte.

Paschen.

W. D. B.

0. & E.

Ostwald.

W. & T.

ZnHg

Chlorides

1.112

1.151

1.121

1.173

1.12-26

ZnHg

Bromides

0.983

0.991

0.996

1.036

0.972

ZnHg

Iodides

0.840

0.847

0.830

0.841

0.801

ZnHg

Sulphates

1.300

1.302

1.302

1.484

1.46-51

ZnHg

Nitrates

—

1.200

1.330

1,422

1.499

ZnHg

Acetates

—

1.228

—

1.451

CdHg

Chlorides

0.816

0.818

0.755

0.813

0.812

CdHg

Bromides

0.690

0.659

0.632

0.696

CdHg

Iodides

0.548

0.515

0.465

0.535

CdHg

Sulphates

0.968

0.969

0.962

1.083

CdHg

Nitrates

—

0.867

0.884

1.011

CdHg

Acetates

—

0.898

—

1.078

108 PROCEEDINGS OF THE AMERICAN ACADEMY

The variation in passing from a chloride to an iodide solution is
about 0.3 volt, far more than can be accounted for by any experi-
mental error. This necessitates a reconsideration of the Nernst
hypothesis to see where the flaw in the reasoning occurs. The
assumption made is, that, if a metal he dipped iuto a solution of one
of its salts, ions of that metal will go into solution, and the electrode
become charged negatively towards the electrolyte if the " solution
pressure " of the metal is greater than the osmotic pressure of the cor-
responding ion in the solution. If the osmotic pressure of the ion in
the solution is greater than the " solution pressure," ions will be pre-
cipitated upon the metal, which will become positive to the solution.
This same reasoning is applicable to zinc in a solution of potassium
chloride, for instance. The initial concentration of the zinc ions in the
solution is zero, and the metal will therefore send off ions until the
potential difference corresponding to equilibrium is reached. This
will not be the case when we consider mercury in a solution of potas-
sium chloride. There are no mercury ions in solution to precipitate on
the metal, and it remains an unanswered problem how the mercury is
to become charged positively in respect to the solution. Yet this
takes place, and the value as determined by the drojjping mercury
electrode method is a perfectly well defined one. This value should
be independent of the nature of the salt solution if Ostwald's assump-
tion about log P is correct. This is not the case. In this connection
I wish to say that the question as to the value of the dropping mercury
electrode as a means of measuring single potential differences does not
affect this discussion at all. It is an experimental fact that the sum of
the potential differences MJRX and RXIMj, as determined by this
method, is equal to the electromotive force of the cell Mi\RX\M.2, and
it is immaterial for the present purposes whether the single determina-
tions are wrong by a constant amount, as I am only considering varia-
tions in the values. I will now try to show what conclusions may be
drawn from the measurements of Paschen * on the potential differences
between metals and salt solutions not containing the metal of the elec-
trode as ion. He points out himself that the potential difference is
not a function of the positive ion of the salt solution. It is not a func-
tion of the concentration. Paschen inclines to the opposite view; but
I think he is wrong, and that his own results as tabulated in Table X.
will bear me out. The first column gives the nature and concentration
of the solution ; the second, third, and fourth columns give the poten-

* Wied. Ann., XLIII. 590, 1891.

BANCROFT.

POTENTIAL OF METALS.

109

tial differences between the metals mercury, zinc, and cadmium, and
the solution. Mercury is positive towards the solution, zinc and
cadmium negative.

TABLE X.

Solution.

Sol.lHg

ZnlSol.

Cdl Sol.

Solution.

Sol.jHg.

ZnlSol.

Cd |Sol.

HCl

= 11.

0.560

0.560

0.248

HBr =0.2721.

0.503

0.393

0175

= 101.

0.551

0.610

0.272

= 0.9833 1

0.490

0.423

0.202

= 100 1.

0.584

0.643

0.242

= 101.

0.493

0.567

0.238

KCl

= 0.2801.

0 524

0.525

0.260

= 1001.

0.496

0.610

0.246

= 11.

0.539

0.547

0.249

KBr =0.4021.

0.474

0.399

0.203

= 101.

0.553

0.575

0.251

= 11.

0.483

0.441

0.186

= 100 1.

0.584

0.523

0.240

= 101.

0.493

0.422

0.167

NaCl

= 0 239 1.

0.562

0.521

0,262

= 1001.

0.505

0.496

0.183

= 11.

0.556

0.512

0.266

HI =101.

0.411

0.427

0.117

= 101.

0.557

0541

0.268

= 1001.

0.417

0.515

0.159

= 100 1.

0.590

0 557

0.268

= 1000 1.

0.886

0.584

0.214

MgCls

= 0.9711.

0.546

0.525

0.252

KI = 0.795 1.

0.400

0 250

0.113

= 21.

0.547

0.531

0.277

= 1L

0.400

0.238

0.113

= 201.

0.548

0.598

0.258

= 101.

0.412

0.308

0.110

= 200 1.

0.580

0.516

0.245

= 100 1.

0.412

0.369

0.120

BaClo

= 0.809 1.

0.562

0.512

0.259

= 10001.

0.386

0.454

0.199

= 21.

0.555

0.554

0.249

K2S04 = 2.1521.

0.700

0.618

0.287

= 201.

0.553

0.583

0.281

= 201.

0.720

0.573

0.274

= 200 1.

0.586

0.566

0.240

= 200 1.

0.730

0.592

0 252

H2SO4

= 21.
= 201.
= 200 1.

0.835
0.817
0.825

0.653
0.668
0.068

0.319

0.284
0.261

The values for Sol I Hg are identical for dilutions of 11 and 101, with
the exception of KCl, KBr, K2S04, and H2SO4 ; and the variations
for K2SO4 and H0SO4 are in opposite directions, and certainly due to
experimental error. There is also no reason to assume that KCl is

110 PROCEEDINGS OF THE AMERICAN ACADEMY.

theoretically different in behavior from NaCl or BaCl2j and we
must therefore conclude that this discrepancy is also accidental. In
passing from dilutions of 101 to those of 1001, there is a distinct in-
crease in potential difference between mercury and chloride solutions.
With the other solutions the change is either non-existent or much less
marked. On the other hand, cadmium shows this behavior only with
HI and KI solutions, zinc with HCl, NaCl, HBr, KBr, HI, and KI
solutions. The solutions of HBr, KBr, HI, and KI are not the ones
where mercury shows a marked change of value with increasing dilu-
tion, so that there is no qualitative regularity in the phenomena. As
there is also no quantitative connection to be detected between the
change of concentration and the change of potential difference, and as
the experimental error is very large in the case of determinations with
dilute solutions, I see no reason to assume that there is any change of
potential difference, at any rate within wide ranges of concentration.*
I am led to this conclusion the more strongly because if we admit with
Paschen that the potential difference increases with increasing dilution,
we must admit that the electromotive force of the cell Cd|KCl|Hg is
a function of the concentration, and I have already shown that this is
not the case.

Paschen has pointed out that these potential differences are functions
of the metal forming the electrode and of the anion. This can hardly
be accounted for on the Ostwald-Nernst hypothesis. If the potential
difference between Ilg and KCl or KBr solutions be due to the
amount of mercury as ion which has gone into solution, we must say
that the amount varies as we change from KCl to KBr, or, in other
words, that the negative ion has an effect. This is quite apart from
the difficulty of accounting for the sign of the potential difference. I
do not see that the relative solubilities of mercurous chloride and bro-
mide can be dragged in to help out matters, because we do not have a
saturated solution at all, and the difference in the electromotive forces
is more likely to be connected with the difference of solubility as
cause than as effect.

There are no experimental data, so far as I know, on potential dif-
ferences at the contact surface of reversible electrodes excejjt some

* This will not hold true till the concentration of the salt becomes zero ; else
we should get always the same potential difference, that of the metal in pure
water, wliicli is not tlie case. There will certainly be a minimum concentration
beyond whicii the dissolved substance will not have the properties of matter in
mass, and the potential difference will then be a function of the concentration.

BANCROFT. — POTENTIAL OF METALS.

Ill

measurements by Neumann,* and these are not conclusive as regards
the point that they were intended to prove, owing to an unfortunate
choice of sohitions. He measured the potential difference between
thallium and solutions of tluillium salts. Most of the salts were salts
of organic acids, and Ostwald f had already found that when the nega-
tive ion was an organic radical its nature was immaterial. To settle
this question one should take negative ions which show marked differ-
ences with non-reversible electrodes, such as chlorides, bromides, and
iodides. As the negative ion has a very marked influence in these last
named cases, and as there is no reason to su^jpose that the haloid salts
form a class by themselves, the simplest assumption is that the nega-
tive ion always has an effect, and that in the cases where this does not
appear, such as tlie organic radicals, we are measuring something
else which is the same in all cases. Le Blanc | found something
similar in his studies of polarization, where beyond a certain point he
obtained the value for the primary decomposition of water.

There are certain quantitative relations connected with the change
of the negative ion which deserve to be brouglit out, and in Table XI.
are given the most probable values for the po ential differences of the

TABLE XL

Solution.

Za 1 Sol.

CdlSol.

Sol 1 Ilg.

Chlorides . . .
Bromides . . .
Iodides ....

0.589
0.507
0.436

0.255
0.174
0.104

0.562
0.483
0.410

TABLE XIL

Solution.

ZaISol.lIIg.

Cd 1 Sol. 1 IIi;.

Chlorides

Bromides

Iodides . »

Sulfates

1.151

0.990
0.846
1302

0.817
0.657
0.514

0.9G9

* Zeitschr. f. ph. Ch., XIV. 225, 1894.
t Ibid., I. 605, 1887.

t Ibid., VIII. 315, 1891.

112

PROCEEDINGS OF THE AMERICAN ACADEMY.

TABLE XIII.

Solution.

Zn.

Cd.

Ug.

KCI-KBr . . .

0.082

0.081

-0.079

KBr-KI . . .

0.071

0.070

-0.073

KCI-KI . . .

0.153

0.150

-0.152

metals Tig;, Zn, and Cd in solutions of chlorides, bromides, and iodides,
while in Table XII. are the corresponding values for the single liquid,
non-reversible cells with Zn and llg, Cd and Hg, as electrodes.*

We notice that the numerical change in passing from a chloride to
a bromide or iodide solution is the same for these three metals and that
the sign is the same for zinc and cadmium, as is shown in Table XIlI.
This enables us to formulate matters a little more clearly. The poten-
tial difference between a metal and a salt solution is the sum of two
terms, one due to the metal and the solvent, the other to the negative
ion. For certain metals in certain solutions the term due to the nega-
tive ion is independent, numerically, of the nature of the metal con-
sidered. For instance, the potential differences ZnlKCl, ZnlKBr,
CdlKCl, CdlKBr, IlglKCl, and HglKBr will be A i- a, A + b,
Ji -{- a, B -{- b, C—a, and C — b. The electromotive forces of the
cells ZnlKCllCd and Zn|KBr|Cd will be A\ = J + « — ^— «, and
^2 = -4 + b — B—b, whence we see that E^ — E^, which had already
been found experimentally. For ZnlKCllHg and ZnlKBrlHg we
shall have E]_ = A-\-a — C-f- a and En^ — A-\-b — C-{-b, and Ey will
not be equal to E.^.- ^y referring to Table III. we can now extend
our generalization and make it more precise. With the metals Mg,
Zn, Cd, Sn, Pb, and Bi in solutions of chlorides, bromides, iodides, sul-
phates, nitrates, acetates, carbonates, and oxalates, the term due to the
negative ion is not a function of the electrode. There is not much
doubt that the alkaline metals, the metals of the alkaline earths, and
the metals of the iron group belong in this same series. Ostwald's
measurements show that most organic acids may be added to the above
list of solutions. With mercury the numerical value of the term due
to the negative ion is the same as with the previous metals, but the
sign is opposite. With platinum the numerical value is no longer

* There is certainly an error in the relative positions of Sn and Pb as given
by my determinations, so I give no data for them in Table XII.

BANCROFT. POTENTIAL OF METALS. 113

the same. In whicli of these three groups copper, silver, gold, and the
other metals belong, I cannot say, though silver is probably like mer-
cury. The results in Tables XI.-XIIl. open up a whole series of
problems to be settled by future investigators. The values for the
differences of the terms for any two negative ions have to be deter-
mined with accuracy ; the behavior of the metals Cu, Ag, etc. must be
examined. The work of Magnanini * shows that other relations hold
when the dissolved salt is an oxidizing or reducing agent, and that the
value ZnlRXICd, for instance, is not a function of the metals only,
if RX is an oxidizing agent. It is also well known that in cases
where the electrode metal caimot exist in the solution as ion that the
general relations already pointed out do not hold. From the results
of Negbaur -j- and of Jones J we must conclude that the term which I
have represented by J, B, C, etc., varies with the nature of the sol-
vent. The amount of this variation is entirely unknown as yet, and it
is equally impossible to say beforehand how a change in the solvent
will affect the term due to the negative ion.

If we consider the cell ZnlZiiClolZnBralZn, the two solutions be-
ing assumed to be of the same concentration and dissociation, and the
wandering velocity of the bromine ion being further assumed to be
identical with that of the chlorine ion, we should expect an electromo-
tive force of 0.080 volt. This has not been taken into account by
Goodwin § in his determinations of the solubilities of silver chloride,
bromide, and iodide. Goodwin determined the electromotive forces of
the cells AglAgNO.lAgCl + KCllAg, Ag| AgNO.I AgBr + KBrlAg,
and AglAgNOji Agl + KIlAg. From the observed electromotive

forces the solubilities were calculated by the formula s = if ^-^^,

E ...

where log cf = --^. In this equation s is the solubility, pi the con-

centration of the Ag ions in the nitrate solutions, p^ the concen-
tration of the CI, Br, or I ions in the corresponding solutions, E is
the electromotive force of the cells, and C the integration constant,
which is equal at 25° to 0.0256, It is more than probable that a cor-

* Eend. Ace. Line, VL 182, 1890. X Zeitschr. f. ph. Ch., XIV. 346, 1894.

t Wied. Ann., XLVIL 27, 1892.

§ Zeitschr. f. ph. Cli., XIII. 645, 1894. It is only fair to Mr. Goodwin and to
myself to say that I have pointed out to liim privately the objections that I made
to his results that he might correct them himself if he felt so inclined. He thinks
that it would he better for me to make my comments in print, and I have
accordingly done so.

VOL. XXXI. (n. s. xxiit.) 8

114

PROCEEDINGS OF THE AMERICAN ACADEMY.

rection ought to be applied for a possible difference of log P in nitrate
and chloride solutions ; but as this value is not accurately determined
I will first calculate the solubilities on the basis of the assumption that
log PNO3 =log Pci.* We find, from Table XIII.,

log Pci — log PBr =0.080, and log Pci— log P, = 0.152 volt.

These values are to be subtracted from the electromotive forces
observed with AgBr and Agl in order to get the term ^called for
by the formula. In Tables XIV. -XVI. I give the results of these
calculations.

In the first column are the values iov p^ ; in the second, for p^ ; in the
third, the electromotive forces ; in the fourth, the solubilities as calcu-
lated by Goodwin ; in the fifth, the solubilities as calculated by myself
under the assumption that log PN03 = log Pci : and in the sixth, the
values if one assumes that log PnOs — log Pci = 0.03 volt. I also
give the solubilities found b}^ Kohlrausch and Rose,t and by Holle-
man % with the conductivity method.

TABLE XIV.§

Cone. Ag Ions

Cone. Cl Ions

E. M. F.

Cale. S^.

Cale. S„.

Cale. ^3.

0.0813

0.0861

0.451

1.24 X 10-5

1.25 X 10-5

2.25 X 10-5

0.0813

0.0861

0.449

1.28 X 10-5

1.30 X 10-5

2.34 X 10-5

0.04295

0.04455

0.418

1.25 X 10-5

1.25 X 10-5

2.24 X 10-5

0.04295

0.04455

0.419

1.23 X 10-5

1.22 X 10-5

2.20 X 10-5

Sohibility

AgCl at 25°, a

verage

1.25 X 10-5

1.25 X 10-5

2.26 X 10-5

Kohlrausc

1 and Rose .

1.44 X 10-5

at 25°

HoUeman

1.81 X 10-5

at 25°

It will be seen that the second column of solubilities agrees much
better with the results obtained bv other investigators than the solu-

* Tlie term log Pci denotes the value of log P for the metal under discussion
when in a chloride solution.

t Zeitschr. f. ph. Ch., XII. 824, 1893.

t Ibid., XII. 125.

§ S, and So in this table should be identical, as they are calculated from the
same data by the same formula; the variations are due to errors in calculation.

BANCROFT. — POTENTIAL OF METALS.
TABLE XV.

115

Cone. Ag Ions

Cone. Br Ions

E M. F.

Ca!c. S,.

Calc. S,

Calo. <S',.

0.0813 0.0861 0.598
0.0813 0.0861 0.603
0.0813 0.0861 0.597
0.04295 0.04455 0.570
0.04295 0.04455 0.571
0.04295 0.04455 0.570
Solubility AgBr at 25°, average
Kohlrausch and Rose ....
Holleman

7.1 X 10-1
6.4 X 10-"

7.2 X 10-"
6.4 X 10-7

6.3 X lO-'?

6.4 X 10-7
6.6 X 10-7

20.9 X 10-7
30.2 X 10-7

33.8 X 10-7
30.1 X 10-7

34.4 X 10-7

30.5 X 10-7

29.9 X 10-7
30.5 X 10-7
31.5 X 10-7
at 25°

at 25°

60.7 X 10-7
55.0 X 10-7
61.9 X 10-7
54.9 X 10-7

53.8 X 10-7

54.9 X 10-7
56.9 X 10-7

TABLE XVL

Cone Ag Ions

Cone. I Ions
— Pi-

E. M. F.

Calc. Si.

Calc. S^.

Calc. S3.

0.0813

0.0861

0.815

1.02 X 10-8

19.9 X 10-8

35.7 X 10-8

0.0813

0.0861

0.813

1.06 X 10-8

20.7 X 10-8

37.2 X 10-8

0.0813

0.0861

0.815

1.02 X 10-8

19.9 X 10-8

35.7 X 10-8

0.04295

004455

0.787

0.94 X 10-8

18.0 X 10-8

32.3 X 10-8

0.04295

0.04455

0.786

0.96 X 10-8

18.3 X 10-8

32.9 X 10-8

0.04295

0.04455

0.790

0.88 X 10-8

17.0 X 10-8

30.5 X 10-8

Solubility

Agl at 25°, av

erage

0.98 X 10-8

19.0 X 10-8

34.0 X 10-8

Kohlrausc:

1 and Rose .

. . .

60.0 X 10-8

at 18°

Holleman

395.0 X 10-8

at 28.4°

bilities calculated by Goodwin. The solubilities in the last column do
not show as good an agreement ; but I do not feel sure that this proves
that the formula by which they are calculated is wrong. It seems to
me quite as probable that these figures represent the actual solubilities
in the cells examined by Goodwin ; but not the real solubilities of

116 PROCEEDINGS OP THE AMERICAN ACADEMY

AgCl, AgBr, AgT. The solubilities of AgCl and AgBr are much
changed by continued shaking,* and I cannot 6nd that Goodwin has
taken this into account at all. I conclude, therefore, that if he had
shaken his AgCl and AgBr he would have found much smaller electro-
motive forces than those recorded in his paper. He has tried to prove
the accuracy of his formula in two ways. Having calculated the solu-
bilities by substituting the experimental data for the electromotive
forces in the formula, he reverses the operation, and, substituting the
solubilities, he calculates the electromotive forces. It is true that
there is an intervening step, but the principle is the same, also the re-
sult. If he had taken the cells AglAgNO;j|AgCl + KCll Ag, and
AglAgNO.^jlAgBr + KBr|Ag, and substituted directly in these, the fal-
lacy of such a test would have been patent. Instead of this he has
combined the two cells

AglAgCl + KCll AgNOgl Agl AgNOgl AgBr + KBrl Ag =

AglAgCl + KCllAgBr+ KBrlAg,

calculated the electromotive force of the resultant cell, and compared
this with the experimental value and with the ditference of the mean of
the two component cells. Any other formula, which had given fairly
constant values for the solubilities, would have stood the test eqxially
satisfactorily. If, instead of taking Goodwin's formula and his value for
AgBr, 6.6 X lO"*^ reacting weights per litre, one takes, for instance,
ray first modification of his formula and the corresponding value for
AgBr, 31.5 X 10~" units, one will reproduce his table exactly. One
cannot agree with him when he says, in regard to this table : " Die
Uebereinstimmung der beobachteten mit den berechneten Werten ist
eine sehr gute, wie dem ja nicht anders sein konnte, wenn die frilhere
Formel (25), nach der die Loslichkeiten berechnet wurden, liberhaupt
richtig war. Sie bestjitigt also diese Formel." f The otiier proof is not
satisfactory in the light of my experiments. Goodwin determined the
solubility of thallium bromide by the electrical and by the analytical
methods, the difference between the two being about 10 per cent of the
total solubility. This result cannot be compared with the experi-
ments on the solubilities of the silver haloids, because the conditions
were not the same. In the thallium determinations the cell used was
of the form TllTIBr + KNOglTlBr + KBrlTI. There were bromine
ions in contact with both electrodes, while with the silver salts the
bromine ions only came in contact with one electrode.

* These Proceedings, XXX. 385, 1894.
t Zeitsclir. f. ph. Ch., XIII. 651, 1894,

BANCROFT. — POTENTIAL OF METALS.

117

One other point remains to be considered, whether the potential dif-
ference at the surface of a reversible electrode is a function of the con-
centration at all. The only direct measurements are those of Neumann,*
which confirm the Nernst theory in every detail. In addition, there
are many determinations on two-liquid cells made chiefly by Ostwald's
pupils, and in all these instances there is a most satisfactory agree-
ment between the theory and the facts. On the other hand, there are
a few observations by different people which are not so easily recon-
ciled with the theory. In Table XVII. I give some measurements of
Paschen's f on cells having zinc and mercury electrodes and solutions
of ZnSOi and MgS04 of varying concentrations as electrolyte. The
first column gives the nature of the cell ; the second, the specific
gravity of the electrolyte ; the third, the concentration in grams per
hundred grams of the solution ; the fourth, the electromotive force
observed. The values for the concentrations are only approximate,
because they were not determined by Paschen directly, and I have
taken them from Landolt and Bornstein's tables.

TABLE XVII.

Electrodes.

Electrolyte.

Density.

%i°g-

E. M. F.

ZnHg

MgSO^

L042

4.

1.194

ZnHg

MgSOi

1.040

4.

1.236

ZnHg

MgSOi

L040

4.

1.180

ZnHg

ZnSOi

L433

32.9

1.249

ZnHg

ZnSOi

1.409

31.5

1.252

ZnHg

ZnSOi

1.403

31.1

1.309

ZnHg

ZnSOi

1.402

31.1

1.327

ZnHg

ZnS04

1.400

31.0

1.236

ZnHg

ZnS04

1.315

25.5

1.310

ZnHg

ZnS04

1.305

2.50

1.238

These figures lose a good deal of their value owing to the consider-
able variation in the determinations for the same solutions, and because

* Zeitschr. f. ph. Ch., XIV. 225, 1894. t Wied. Ann., XLIII. 570, 1891.

118

PROCEEDINGS OF THE AMERICAN ACADEMY.

the range of concentrations is too limited ; but two things are very
noticeable in spite of this. In the cell ZnlZuSO^IHg the electromo-
tive force does not decrease with increasing concentration of zinc sul-
phate, as it should according to the theory. The cells ZnlMglS04lHg
have the same value as the cell ZulZnS04lHg or a smaller one, while
the theory demands a larger one. The same thing is seen, though in
a less satisfactory manner, in the experiments of Damien.* He used
zinc and copper as electrodes, and his results are given in Table
XVIII. The first column shows the electrolyte; the second, the
specific gravity of the solution; the third, the percentage composition;
the fourth, the electromotive force when ordinary zinc was used ; the
fifth, the corresponding value when the electrode was amalgamated.

TABLE XVIII.t
ZnCu Electrodes.

Electrolyte.

Sp. gr. at 15^.

% '° s-

E. M. F.

Amalg. Zn.
E. M. F.

K,S04

1.036

4.5

1.035

1.067

Na,S04

1.038

10.*

1.012

1.037

(NH4)„S04

1.07.5

13 1

1.012

1.019

MgSOi

1.035

5.8*

1.047

1.059

A1.,(S04)3

1.135

5.8

1.050

1.002

ZnS04

1.061

9.2*

1.004

1.047

KCl

1.077

12.0

0.788

0.802

NaCI

1.061

8.5

0.805

0.810

NH4CI

1.039

13.

0.845

0.850

BaCla

1.110

12.

0.782

0.820

CaCIg

1.212

23.

0.743

0.741

ZnCla

1.384

37.5

0.746

0.752

As will be noticed, there are marked variations even in cases in
which no one claims that there should be any, such as between ammo-

* Ann. Chim. Phys., [6.], VI. 289, 1885. The refererence to Vol. V. in Wied,
Elektricitat, I. 734, also in Beibl., X. 185, is a misprint,
t Stars refer to liydrated salt.

BANCROFT. — POTENTIAL OP METALS. 119

nium sulphate and potassium sulphate solutions, between calcium
chloride and ammonium chloride. This weakens the conclusions
which one would like to draw from these experiments; but, making
allowance for a large experimental error, it is very curious that
ZulZnS04|Hg should give so nearly the same value as the cells with
indifferent sulphates, and that zinc chloride should be indistinguishable
electrically from calcium chloride. The experiments of Hockin and
Taylor * may be interpreted either way. They found that the com-
bination of zinc and another metal in sulphuric acid gave a higher elec-
tromotive force than the same two metals in a saturated solution of
zinc sulphate. This is not so convincing as if they had used potassium
sulphate instead of sulphuric acid, because in all except dilute solutions
free acids do give a higher value than the corresponding salts. The
reason for this variation is unknown. When it comes to the absolute
values in the zinc sulphate solution, matters are no better. With
some of the metals, notably cadmium and mercury, the zinc sulphate
appears to give the same value as any other sulphate ; with others,
there is qualitative agreement with Nernst's theory. The same re-
marks hold true of the work of Lindeck.f I have not access to the
original paper of Wolff, and the review of it | is too meagre to be of
much assistance. He investigated, among other things, the effect of
changing the concentration of the zinc sulphate in a one-liquid cell.
His results are given in Table XIX. The first column shows the
electrodes and the electrolyte ; the second, the concentrations of the
latter in specific gravities ; the third, the corresponding electromotive
forces.

In all cases there is a qualitative agreement with the theory ; that
is, the electromotive force increases with decreasing concentration of
zinc sulphate. The quantitative agreement is not so satisfactory. In
the second, third, sixth, and last cells given in the table, the variations
are much too small, while in the other cases they are too large. The
ratio of the strongest solutions to the weakest in the experiments of
Wolff lies between 100 and 1000 to 1, which corresponds to a change
of electromotive force of 0.05-0.09 volt owing to the bivalence of zinc.
Some experiments winch I made with the cell CdlCdCUIHg, the
strength of the solution being unknown, gave me 0.815, 0.821, 0.814,
average 0.817 volt, the same value which I had already found for
the KCl solution.

* J. Tel. Eng, VIII. 282, 1879. J Beibl., XII. 700, 1888.

t Wied.Ann., XXXV. 311, 1888.

120

PROCEEDINGS OF THE AMERICAN ACADEMY.

TABLE XIX.

Cells

Densities.

E. M. F.

ZnlZnSOilCu

1.438-1.001

0.965-1.066

ZnlZnSOilCuO

1.427-1.003

1.008-1.015

ZnlZnSO^IFe

1.427-1.003

0.378-0.385

ZnlZnSO^lPb

1.427-1.003

0.456-0.587

Zn|ZnClo|Cu

1.637-1.003

0.734-0.930

Zn|ZnCIo|Fe

1.917-1.003

0.385-0.390

Zn|Zn(N03)o|Cu

1.496-1.004

0.669-0.698

The simplest way to decide what effect the concentration of the re-
versible ion, if 1 may use such a phrase, has on the electromotive force,
would be to make a series of measurements on reversible electrodes by
the dropping mercury method. I have not been in a position to do
this, and I have had to find an easier, though less satisfactory, manner
of settling the question. Suppose we have electrodes of zinc and
copper in a mixture of zinc and copper sulphates, one solution. In-
creasing the concentration of the zinc sulphate or decreasing the con-
centration of the copper sulphate must diminish the electromotive
force of the cell, and vice versa if the reverse operations be performed.
Through the courtesy of Professors Trowbridge and Peirce of the
Physical Laboratory, I have been able to make the few experiments
necessary. As it was only required to find out whether there was any
change at all, there was no need of determining the absolute value of
the electromotive force. This made the experimental pait very easy.
I connected the cell with a large external resistance and a galvanome-
ter. I changed the ratio of the two components in the solution, and
noted the position of the galvanometer needle. I made measurements
with the electrodes in pure zinc sulphate solutions, in pure copper sul-
phate solutions, and in mixtures of the two in varying proportions.
Under all these different conditions I obtained the same electromotive
force, showing that it is a function neither of the relative nor of the
absolute concentrations.* Althoucrh one obtains the same value from

* This applies only to electrodes reversible in respect to the kation. I hope
to make tiie case of electrocles reversible in respect to the anion the subject of a

BANCROFT. — POTENTIAL OF METALS. 121

the different solutions, they do not behave exactly alike. With solutions
of pure copper sulphate, or from mixtures containing copper sulphate
in any quantity, the maximum value is obtained at once, and is very
constant. With solutions of pure zinc sulphate or mixtures containing
only traces of copper sulphate, the maximum Value can be obtained
only by vigorous stirring, and is very inconstant. There is, of course,
nothing surprising about this, as it is what one would have predicted.
It is a very striking and curious fact, that the Nernst formula, though
deduced from apparently erroneous assumptions, should yet give the
effect of changes of concentration in a two-liquid cell with such sur-
prising accuracy.

It will be noticed that the electromotive forces of the non-reversible
cells have nothing to do with the heats of reaction. This has always
been known ; but it acquires new significance since it has been shown
that the non-reversible cells are to be considered, as far as the electro,
motive force is concerued, as limiting cases of the reversible two-liquid
cells. In the cell ZnlHaSO^lAg there is not much doubt what reaction
takes place ; but it has nothing to do with determining the electromotive
force. An interesting example of this, which also brings up another
point, is the cell CulCuS04lPt. Here the reaction consists in the re-
phicement of copper by copper. What happens experimentally, on
closing the circuit, is that copper is dissolved from the copper electrode
and precipitated on the platinum until this latter becomes, electrically
considered, an electrode of pure copper, when further action becomes
impossible. Overbeck * made some experiments a few years ago to
determine what thickness of copper made a platinum electrode behave
like a piece of pure copper. His method was to deposit copper on
platinum electrolytically, and was open to the objection that it was
almost impossible to be certain that the copper was deposited uni-
formly over the surface of the platinum. By using the cell
CulCuSO^IPt, it would seem that this difficulty might be avoided, as
the plating is stopped automatically as soon as the minimum thickness
is reached. Suppose now we balance this cell, to some extent, by an
electromotive force less than its own. There will still be a tendency
for copper to be deposited on the platinum ; but it cannot be deposited
to the thickness corresponding to pure copper, as it must then dissolve

special communication. Since tiiis paper was written there has appeared an
article by G. Meyer, Wied. Ann., LIIL 848, 1894, which confirms my views,
tliough with certain exceptions.
* Wied. Ann., XXXL 337, 1887.

122 PROCEEDINGS OP THE AMERICAN ACADEMY.

up again under the influence of the external electromotive force. It
can precipitate only till equiUbrium is reached, and we shall have the
condition referred to by Gibbs,* of a substance present in too small
quantities to have the properties of " matter in mass." By making
the external electromotive force differ infinitely little from the electro-
motive force of the cell, it would be possible, theoretically at any rate,
to obtain an infinitely thin film of copper. It is to the separation (jf
the ion on the electrode in such small quantities as not to have the
properties of matter in mass that is due the gradual change of the
polarization, instead of having a sudden change from the initial to
the final value.

The main results of this research may be summed up as follows : —

1. The potential difference between a metal and an electrolyte is
not a function of the concentration of the salt solution, nor of the
nature of the positive ion, except in certain special cases.

2. It is a function of the electrode, of the negative ion, and of the
solvent.

3. In acpieous solutions the potential difference is the sum of the
terra due to tlie electrode and the term due to the negative ion in the
normal cases.

4. For most metals in most electrolytes the term due to the nega-
tive ion has the same numerical value and the same sign.

5. For mercury it has the same numerical value, but the opposite
sign ; for platinum, neither the same numerical value nor the same
sign.

6. For platinum in a haloid solution the change of the electromotive
force with the concentration is given by the formula

Bi-E.i = l RT nat. log ~

if the salt dissociates into three ions ; by the formula
if it dissociates into two.

E^- E^^lRT nat. log -^

* Thermodynamisclie Studien, p. 393.

JACKSON AND CALVERT. — DERIVATIVES OF BENZOL. 123

VI.

CONTRIBUTIONS FROIVI THE CHEMICAL LABORATORY OF
HARVARD COLLEGE.

ON THE BEHAVIOR OF CERTAIN DERIVATIVES OF
BENZOL CONTAINING HALOGENS.

By C. Loring Jacksox and Sidney Calvert.

Presented May 9, 1894. ^

Introduction.

It is a well known fact that in substituted aromatic compounds the
firmness with which atoms of haloijen are attached to the benzol ring:
is diminished bj the presence of nitro groups in certain positions, so
that these atoms can be replaced by reagents, which would have no
effect on them if the nitro groups had been absent. It has also been
shown that certain other negative radicals exert a similar loosening
effect : these are the two oxygen atoms in substituted quiuoues —
chloranil and bromanil are very reactive bodies — the hydroxyl groups
in phenols,* and especially in substituted resorcine,f and also probably
carboxyl.t The object of the research described in the present paper
was to determine whether such a loosening effect could be produced
by less negative radicals, and for this purpose we have taken up the
study of derivatives of benzol, in which the substituting radicals are
halogens only, so that both the radicals removed and those which
make the removal possible belong to this same class, and the latter
are distinctly less negative than any of those enumerated above. The
amount of work already done in this field is meagre. Balbiano § found
that paradibrombenzol was converted into parabromphenetol, brom-
benzol, and a little benzol, when it was heated to 190° with sodic
ethylate. The same substance with sodic methylate at 150° gave

* Armstrong and Harrow, Journ. Cheni. Soc, 1886, p. 447.
t Jackson and Dunlap, Tliese Proceedings, XXIX. 228.
t Rahlis, Ann. Cliem., CXCVIIL 112.
§ Gazz. Chim., XL 401.

124 PROCEEDINGS OP THE AMERICAN ACADEMY.

according to Blau * parabromphenol, parabromanisol, aud a little
hydroquinone dimethylether. Blauf also fouud that symmetrical tri-
brombeiizol gave with sodic methylate at 130° symmetrical dibrom-
phenol and the corresponding anisol as principal products. These
results indicate that the atoms of bromine in these di-and tribromben-
zols exert some loosening effect, as monobrombenzol must be heated
to 200° before it reacts with sodic methylate ; j but all these actions
take place in sealed tubes, and even the maximum difference between
their temperatures of reaction and that of monobrombenzol is not very
great, only 70°. We decided therefore, in taking up this subject, to
try to obtain reactions in open vessels, that is, under conditions which
would have prevented all action with monobrombenzol, and to do this
it was obvious that we must increase the number of the atoms of
halogen which are to produce the loosening effect. We selected for
our first experiment accordingly the tribromiodbenzol having the
constitution Br.H. Br.I.Br.H, as in this compound we have iodine
as the element to be removed, and this is usually less strongly at-
tached to the ring than bromine or chlorine, and the three atoms of
bromine which are to produce the loosening effect are in the most
favorable positions for this purpose (two ortho and one para). Tri-
chloriodbenzol would probably have been even better, because chlo-
rine is more negative than bromine, but we preferred the bromine
compound on account of the great difficulty of preparing the trichlor-
aniline in quantity.

Upon trying a variety of reagents with the unsymmetrical tribrom-
iodbenzol, we obtained negative results with all except two, sodic
ethylate and sodic methylate. Tliese converted it into symmetrical
tribrombenzol, melting point 119°, by replacing the iodine with hydro-
gen, the sodic etliylate acting to some extent even in the cold, more
freely when the solution in alcohoi and benzol was boiled under a
return condenser, while the sodic methylate did not act in tlie cold,
and only to a limited extent boiling. The unsymmetrical tetrabrom-
benzol Br. ILBr.Br.Br.H, which differs from the preceding compound
only in having an atom of bromine in place of the iodine, was also
partially converted into symmetrical tribrombenzol by boiling with
sodic ethylate, but there was much less action than with the iodine
compound, the tetrabrombenzol undergoing about as much substitution
when the solution was boiling, as was the case with the tribromiodbenzol
in the cold. The tribromchlorbenzol Br.H.Br.Cl.Br.H, on the other

* Monatsli. f. Chem., VII. 627. t Ibid., 630. | Ibid., 636.

JACKSON AND CALVERT. — DERIVATIVES OF BENZOL. 125

hand, was entirely unaffected by a solution of sodic etiiylate in open
vessels. These experiments show that the presence of the three bro-
mine atoms exercises a loosening effect similar to that exhibited by
nitro oToups, although much weaker in degree. They also furnish
an additional case in which the stability of the aromatic compounds
of the different halogens increases in the order iodine, bromine, chlo-
rine. This is worth noting, because Korner * states that the halo-
gens are removed from dinitrohalogenbeiizols (X 1, NOg 2, NO2 4)
in exactly the reverse order ; that is, chlorine most and iodine least
easily.

We next turned our attention to the tetrabrombenzol melting at
174°_175° which has been proved by work done in this Laboratory to
have the symmetrical constitution Br.Br.H.Br.Br.H, as in this sub-
stance only two of the bromine atoms are in the positions to the one
to be removed (ortho and para) which had proved effective in the
unsymmetrical compound. This, however, did not affect the result
materially, as this tetrabrombenzol was converted into unsymmetrical
tribrombenzol (Br 1, Br 2, Br 4) by the boiling solution of sodic
ethylate to about the same extent as the unsymmetrical tetrabromben-
zol. In separating the small quantity of tribrombenzol formed from
the large amount of unaltered tetrabrombenzol we have obtained
excellent results in both the cases just mentioned by exposing the mix-
ture for a long time to the lowest temperature at wiiich anything
sublimes. The sublimate tlius obtained, if not the pure tribrombenzol,
could be converted into it by a single repetition of this rough fractional
sublimation.

An arrangement of the bromine atoms entirely different from the
effective ortho para positions is found in the symmetrical tribrom-
benzol Br.H.Br.H.Br.H, as here all the halogen atoms are in the meta
position to each other, but, as has been already mentioned, Blau found
that one of the atoms of bromine was removed by the action of sodic
methylate at 130°. As our work just described has shown that sodic
ethylate is more active than the methylate, we thought this might act
even in open vessels, and on trying the experiment have found that
sodic ethylate in boiling alcoholic solution removes from symmetrical
tribrombenzol a portion of its bromine. The organic products of the
partial reaction were oily, but, as the object of our experiment was to
determine whether a reaction took place, and not what its products
were, no attempt was made to examine them. This is not the only

* Gazz. Chim., 1874, p. 323, note.

126 PROCEEDINGS OF THE AMERICAN ACADEMY.

case in which one of three radicals in the symmetrical position has
been removed, as Lobiy de Bruyn * has converted symmetrical trini-
trobenzol into dinitranisol or dinitrophenetol by the action of the sodic
alcoholates in open vessels. The work of Blau and Lobry de Bruyn,
therefore, establishes this as a general behavior of symmetrical com-
pounds, and such replacements are the more remarkable, because a
single negative radical has no effect upon one other in the meta
position.

In the work with unsymmetrical tetrabrombenzol, mentioned above,
it was only the fourth atom of bromine (in the diortho para position to
the others) that was replaced by hydrogen, but in some work upon the
corresponding dinitro compound C6Br^(N02)2, done some years ago by
W. D. Bancroft and one of us, this atom of bromine remained obsti-
nately unat tacked when other atoms of bromine were removed ; t in the
same way the reduction of tetrabromdinitrobenzol with tin and hydro-
chloric acid gave monobromphenylene diamine. | so that in these cases
the symmetrical atoms of bromine alone were removed, and the action
was exactly the reverse of that obtained with tetrabrombenzol and sodic
ethylate. This difference is explained by the facts that in the tetra-
brombenzol the three atoms of bromine in the trimeta position are the
loosening radicals, whereas in the tetrabromdinitrobenzol they are those
which are loosened. In this connection we decided to try the action of
sodic ethylate on tetrabromdinitrobenzol, as this experiment had not
been included in the previous work of AV. D. Bancroft and one of us,
and we found that tetrabromdinitrobenzol (Br.NO2.Br.NO2.Br.Br)
was converted by the action of a cold solution of sodic ethylate
into the tribromnitroresorcine diethylether melting at 101°, and first
obtained by W. II. Warren and one of us § by the action of sodic
ethylate on tribromtrinitrobenzol. This substance has the formula
C6Br3N02(OC2H5)2, and must have been formed in this case by the
replacement of one nitro group and the fourth atom of bromine by
etboxy groups, just as it was formed from the tribromtrinitrobenzol by
the replacement of two nitro radicals by ethoxyls. If the action with
the tetrabrom body is analogous to that with the trinitro compound,
there should be formed at the same time by a parallel reaction a substi-
tuted phloroglucine, in this case dinitrobromphloroglucine or its ethers ;

* Rec. d. Tr. Chim., IX. 208, XIII. 149.

t These Proceedings, XXIV. 288. Sodium malonic ester gave the compound
C6HBr._,(NO.j)2 CH(COOC2H5)2 ; reduction of this gave bromamidoxindol; ani-
line gave C6Br(NOo)2 (C6H5NH)3.

t See tlie following paper. § These Proceedings, XXV. 183.

JACKSON AND CALVERT, ^ DERIVATIVES OF BENZOL. 127

but although we have fouud indications that such a substance is formed,
we have not succeeded in isolating it.

Tlie principles governing the replacement of radicals attached to
the benzol ring, which we have brought forward in this introduction,
make it possible to explain the strange behavior of tribromtrinitroben-
zol with sodic ethj'late. This action consists in the two following
reactions, which take place side by side.*

C6Br3(N02)3 + 3 C2H50Na= C6(OC2H5),(NO.)3+ 3 NaBr.
C6Br3(N02)3 + 2 CaHgONa = C.BrgNOoCOaHs)^ + 2 NaNOa-

In this substance we seem to have two zones of action, one consist-
ing of the three symmetrically disposed bromine atoms, the other of
the three nitro groups also symmetrically disposed, and the action in
each molecule is confined to one of these zones, if it is carried on in
the cold. In the first or bromine zone the atoms of bromine are
submitted to two loosening influences : (a) that of the three nitro
groups in the diortho and para positions to each bromine atom, and
(b) that of the trimeta bromine atoms on each other. In the second or
nitro zone also we have two loosening influences : (c) that of the three
bromine atoms diortho and para to each nitro ji'roup, and (cl) the loosen-
ing effect of the three symmetrical nitro groups on each other. Of these
loosening influences (a) is by far the strongest, and (b) the weakest ;
it seems therefore that the combined effect of (c) and (d), of inter-
mediate strength, is about equal to that of (a) and (b) together ; con-
sequently each of these zones of action lies about equally open to the
attack of the sodic ethylate, and the two reactions take place simul-
taneously. When the solvent is alcohol alone, they run to about the
same extent. If the repellent action of the three symmetrical nitro
groups (d) is removed, that is, if tribromdinitrobenzol is used instead
of the trinitro body, the nitro groups should be much less loosened
than the bromine atoms, and as a matter of fact there has been ob-
served in this case no tendency to remove the nitro groups, the
action being confined to the bromine zone, causing the formation of
C,Br(OC2H,)2 (N0o)2H, or CgH(OC,rL),(N02)2H.

The work described in this paper has furnished us with three addi-
tional cases, in which atoms of halogens have been replaced by hydro-

* These Proceedings, XXVII. 283. Tetrabromdinitrobenzol acts in the same
way, as just stated, but for the sake of simplicity of expression the explana-
tion has been confined to the trinitro body, altliough it applies equally well to
the tetrabrom compound.

128 PROCEEDINGS OP THE AMERICAN ACADEMY.

gen under the action of sodic ethylate, but even with these we do not
feel that enough facts are known to make it possible to deduce a
general rule iu regard to the conditions, which produce this curious
reaction. The collection of such facts will be continued iu this
Laboratory.

We also tried the action of fuming nitric acid on unsymmetrical tri-
bromiodbenzol, and found that iodine was set free, and the organic
product was the tribromdinitrobenzol melting at 192°, and having the
constitution Br.NOa.Br.NOa-Br.H.

A few statistics which we have collected about the removal of radi-
cals from substituted benzols may be given here. By far the greater
number of such removals are tliose in which a negative radical stands
in the ortho position to the radical removed. Of these we have
counted over sixty cases, in only nine of which the negative radical is
a halogen. Of the cases of removal, where there is no negative radi-
cal in the ortho position, we have found four in which a negative radi-
cal is in the para position to the radical removed ; these are C^H^Brj,
CgH4ClN02, C^H^BrNOa, and C6H4(N02)2; and to these perhaps
should be added five of the nine cases mentioned above, in which the
ortho radical is a halogen, as in all of these there is a nitro group in
the para position to the radical removed.* In no case is a radical re-
moved which is only in the meta position to a single negative radical ; f
but if there are two negative radicals in meta positions to the radical
attached (symmetrical tri-compound), there are two cases in which
substitution has been observed, C^jHgBrg and CeH8(N02)3.

Experimental Part.
J3ehavior of Tribromiodbenzol.

The tribromiodbenzol (I 1, Br 2, Br 4, Br 6) was made from sym-
metrical tribromaniline by replacing the amido group by an atom of
iodine. Silberstein, J who discovered this substance, made it from the
nitrate of diazotribrombenzol. AVe have preferred to use the sulphate,
and, as our method seemed to give a better result than Silberstein's

* In the other cases where a radical was removed, which was ortho only to
halogens, these halogens were in the diortho position, and therefore probably
were sufRcient to cause the removal.

t The conversions of meta and para nitranisol into the corresponding nitran-
ilines were not counted, because they took place at such a high temperature,
200° ; but perhaps they should be added to the list given above.

X Journ. Prakt. Chem., [2.], XXVII. 119.

JACKSON AND CALVERT. — DERIVATIVES OF BENZOL. 129

to judge fi'om his statement, we give it in detail. Ten grams of sym-
metrical tribromaniline were mixed with moderately dilute sulphuric
acid in such proportion that there was one molecule of the acid to each
molecule of the tribromaniline ; the calculated amount of solid sodic
nitrite was then added in small quantities at a time, shaking the
loosely corked flask after each addition until the red fumes were
absorbed. The mixture after standing over night was filtered, and
then treated with hydriodic acid till there was no further action. The
slightly brown precipitate was washed with water, and purified by
crystallization from a mixture of benzol and alcohol, when it showed a
melting point of 104°. Silberstein gives 103°. 5. For greater safety
it was analyzed, with the following result.

0.3374 gram of the substance gave 0.6119 gram of a mixture of argen-
tic bromide and iodide.

Calculated for Cell^BrjI. Found.

Iodine and Bromine 83.24 83.32

We find the solubilities of the substance the same as those given by
Silberstein, but we do not agree with him when he says it sublimes
easily, as we have found that it sublimes much less easily than sym-
metrical tribrombenzol.

To study the action of sodic ethylate on tribromiodbenzol, 6 grams
of it were dissolved in anhydrous benzol, and mixed with 40 c.c. of an
alcoholic solution of sodic ethylate, containing 2 grams of sodium,
that is, more than enough to remove all the halogen atoms present.
The mixture was then heated on the steam bath under a return con-
denser for one hour, when the liquid had taken on a dark color, and a
brownish precipitate had appeared ; the liquid was now evaporated,
and the residue treated for half an hour with a large quantity of water,
filtered, and washed, when it weighed 4.1 grams. Upon subliming
this residue, 2 grams of sublimate were obtained in small white needles
which melted at 119°-120°, the melting poiut of symmetrical tribrom-
benzol. An analysis of the sublimate gave the following result.

0.2612 gram of the substance gave by the method of Carius 0.4695
gram of argentic bromide.

Calculated for CgllsBrs. Found.

Bromine 76.18 76.51

There can be no doubt, therefore, that the substance is tribromben-
zol, formed by the replacement of the atom of iodine in the tribrom-
VOL. XXXI. (n. s. xxtii.) 9

130 PROCEEDINGS OF THE AMERICAN ACADEMY.

iodbenzol by hydrogen. The residue which did not sublime was a
brownish yellow powder, containing many black specks. It did not
melt at 300°, but melted with blackening when held over the free
flame. It was not completely soluble in any of the common solvents,
and all our efforts to obtain from it a body fit for analysis have proved
fruitless.

The experiment was next repeated under the same conditions, ex-
cept that the mixture w^as not heated, but allowed to stand three days
at the ordinary temperature. At the end of this time it had turned
dark brown, in fact a pale brown color appeared almost as soon as
the materials were mixed ; the solvents were then allowed to evapo-
rate spontaneously, and the residue treated with water as in the pre-
vious experiment, when the aqueous filtrate gave a strong test with
starch j^aste and chlorine water for iodine. The residue insoluble in
water was sublimed at a very gentle heat, and crystals of symmet-
rical tribrombenzol were obtained recognized by their melting point,
119°-120°. The residue, which did not sublime at this gentle heat,
was crystalline and of a reddish white color, very different from the
amorphous brown pi'oduct obtained under the same conditions from
the action of hot sodic ethylate. This crystalline residue after three
recrystallizations from a mixture of benzol and alcohol showed the
melting point 104°, and was therefore unaltered tribromiodbenzol. It
seems, therefore, that cold sodic ethylate behaves like hot, but the action
is less complete. An alcoholic solution of sodic hydrate, after being
warmed for ten minutes with a solution of tribromiodbenzol in benzol,
converted it into tribrombenzol with elimination of iodine. This,
therefore, acted in the same way as the hot solution of sodic ethylate,
but the yield seemed to be larger, and the residue from the sublima-
tion was lighter in color (yellow) and free from black specks. It did
not, however, prove to be more manageable than that previously
obtained. These differences may be due to the shorter heating, ten
minutes in this case instead of one hour when the ethylate was used.
Several attempts were made to detect the aldehyd, which it seemed
probable was formed as the secondary product in the replacement of
the iodine by hydrogen, but these led to no definite result.

Sodic methjlate when boiled with a benzol solution of tribromiod-
benzol for over an hour gives a result similar to that obtained from
cold sodic ethylate ; that is, tribrombenzol was formed, but there was
a large amount of undecomposed tribromiodbenzol. The two sub-
stances were separated by careful sublimation, and recognized by their
melting points. Cold sodic methylate, on the other hand, gave no

JACKSON AND CALVERT. DERIVATIVES OP BENZOL. 131

action even after standing three days. This was proved by testing the
aqueous washings for halogens with negative results, and recovering
essentially all of the tribromiodbenzol used.

The following reagents had no effect on the tribromiodbenzol :
sodic phenylate in alcoholic solution boiling, sodic hydrate in aqueous
solution boiling, sodic carbonate in aqueous solution boiling, ziiicic
oxide and water in a sealed tube at 200° for twelve hours, argentic
acetate in aqueous solution boiling, aniline boiling, tin and hydro-
chloric acid, sodium malonic ester both cold and hot. The proof that
no action had taken place in these experiments was obtained either by
the recovery of the unaltered tribromiodbenzol, or by tests for a salt
of the halogens which gave negative results. In most cases both
methods of proof were applied.

To determine whether the removal of the iodine was due to the
loosening effect of the three atoms of bromine or to the slight attrac-
tion of iodine alone for carbon, iodbenzol was treated with sodic
ethylate under the same conditions which had produced an action on
the tribromiodbenzol ; but after heating the mixture for an hour and
a half no test for sodic iodide could be obtained, showing that there
had been no action.

Action of Fuming Nitric Acid on Trihromiodhenzol.

When tribromiodbenzol was treated with fuming nitric acid, it lost
its white crystalline appearance even in the cold, and became con-
verted into a yellow powder. If the mixture was allowed to stand at
ordinary temperatures over night, and then water added to it, iodine
appeared both in scales and in the form of vapor, as a great amount of
heat was given off. The identity of the iodine was also established
by the smell and the violet color of its solution in carbonic disulphide.
If now the insoluble portion was washed with cold alcohol until free
from iodine, and then recrystallized several times from a mixture of
benzol and alcohol, it showed the constant melting point 191°, which
indicated that the substance was tribromdinitrobenzol, and this was
confirmed by the following analysis.

0.3375 gram of the substance gave by the method of Carius 0.4740
gram of argentic bromide.

Calculated for C6HBr3(N02)2. Found.

Bromine 59.26 59.78

132 PROCEEDINGS OF THE AMERICAN ACADEMY.

The somewhat high result may be due to a trace of a substance con-
tainingr iodine. If the mixture of tribromiodbenzol and funaius nitric
acid was boiled, the organic product was the same, but no free iodine
was obtained.

Behavior of Unsymmetrical Tetrabromhenzol melting at 98°.

The action of sodic ethylate on this substance was selected for
study, because that reagent had given the best results in the work on
tribromiodbenzol just described. To prepare the unsymmetrical tetra-
bromhenzol, 20 grams of tribromaniline were dissolved in 180 c.c. of
hot glacial acetic acid, about 80 c.c, that is a considerable excess, of a
distilled solution of liydrobromic acid (boiling point 125°) added, and,
disregarding any precipitate formed, the mixture treated with sodic
nitrite in the proportion of a molecule and a half or two molecules to
each molecule of the tribromaniline. For this purpose the finely pow-
dered nitrite was slowly sifted with vigorous stirring into the solution,
which had previously been cooled so that it felt barely warm to the
hand. The white crystals which were suspended in it gradually
changed into a dirty brown solid, much of which dissolved even in the
cold. The mixture was then heated for two or three hours on the
water bath ; the solid matter at first went into solution, but later
the tetrabromhenzol, as it formed, separated, principally in long white
needles, or sometimes in part as a serailiquid brown mass. The needles
were separated mechanically, and were usually found to be pure. The
semiliquid portion solidified in a short time, and was purified by dis-
solving it in a little hot benzol, and pouring this solution into alcohol.
The tetrabromhenzol precipitated in this way had a slight reddish
color, while the mother liquor was of a dark claret-red. The precipi-
tate was easily obtained white by one or two recrystallizations from a
large volume of alcohol, to which it was well to add a little benzol.
In this way 20 grams of the tribromaniline gave 22 grams of tetra-
bromhenzol, instead of the 23.88 grams required by the theory, a yield
of 92 per cent.

Eio'ht erams of the tetrabromhenzol were mixed with an alcoholic
solution of sodic ethylate, made from 2 grams of sodium and 50 c.c.
of absolute alcohol, and a little benzol to assist the solution of the
tetrabromhenzol. The mixture was heated to boiling under a reverse
condenser for two days, when it had taken on a blackish green color ;
it was then evaporated to dryness, and washed with water. The
wash waters gave a good test for sodic bromide. The residue insol-
uble in water was sublimed at the lowest possible temperature, when a

JACKSON AND CALVERT. DERIVATIVES OF BENZOL. 133

small amount of white needles was obtained, which melted at 119°,
the melting point of symmetrical tribrombenzol. The residue which
had not sublimed at the very low temperature used made up the prin-
cipal bulk of the product, and was chiefly unaltered tetrabrombenzol.
This method of fractional sublimation has also yielded us excellent re-
sults in aoother similar case, which will be described later in this
paper. The sodic ethylate therefore acts on the tetrabrombenzol in
the same way that it does on the tribromiodbenzol, replacing by hy-
drogen the bromine atom occupying the same position as the atom
of iodine, but the action takes place with more difficulty, and is less
complete.

Behavior of Unsymmetrical Tribromchlorbenzol.

The tribromchlorbenzol was prepared by a method similar to that
used for the tetrabrombenzol. Upon adding the hydrochloric acid to
the solution of tribromaniliue in glacial acetic acid a precipitate of the
chloride was formed, but this went into solution as the amido was con-
verted into the diazo compound. Twenty grams of tribromaniliue
yielded 15 grams of tribromchlorbenzol melting at 82°. Two grams
of this substance were heated for four hours with an alcoholic solution
of sodic ethylate, prepared from half a gram of metallic sodium ; when
upon evaporation to dryness and washing with water only a very
faint test for halogens could be obtained from the wash water, and the
residue, which consisted of unaltered tribromchlorbenzol, weighed
nearly 2 grams. A similar experiment with the tetrabrombenzol, in
which the mixture was boiled for only three hours, gave a strong test
for sodic bromide and a small amount of tribrombenzol. The chlor-
tribrombenzol, therefore, if affected at all by sodic ethylate, is much
less susceptible to its action than either the corresponding iod or brom
compound.

Behavior of Symmetrical Tetrahromhenzol.

Tetrabrombenzol, melting at 174°-175°, which has been proved by
work done in this Laboratory to have the symmetrical constitution (1,
2, 4, 5), was boiled for over twelve hours with an alcoholic solution of
sodic ethylate ; tlie brown liquid thus obtained, with some long rather
dark colored crystals, which were deposited as it cooled, was evapo-
rated to dryness, and the residue washed with water. The wash
waters gave a good test for sodic bromide. The residue insoluble in
water was then extracted, with a mixture of alcohol and benzol,
filtered to remove a brown insoluble substance, and the filtrate concen-

134 PROCEEDINGS OF THE AMERICAN ACADEMY.

trated, when it deposited crystals which melted at 174°-175°, aud
were therefore the unaltered tetrabrombenzol. The mother liquor
was evaporated to dryness, and cautiously sublimed at as low a tem-
perature as possible ; the crude sublimate melted at 42°-43°, and
upon resubliming it with the same care its meltinj; point rose to 44°,
the melting point of unsymmetrical tribrombenzol (1, 2,4), which
must be formed if one of the atoms of bromine is removed from this
tetrabrombenzol. Only a small portion of the substance, however,
reacts with the sodic ethylate, by far the greater part remaining
unaltered.

Experiment with Symmetrical Tribrombenzol.

Blau* states that tribrombenzol is converted by sodic methylate into
symmetrical dibromphenol, when the substances are heated together in
methyl alcohol solution for two to three days at I20°-130°. We
accordingly tried an experiment to see whether sodic ethjdate would
liave a similar action in open vessels, as we had used this reagent in
the preceding work. Ten grams of tribrombenzol were boiled with a
solution of the necessary amount of sodic ethylate in about 200 c.c. of
alcohol for somewhat more than seventy hours ; the aqueous wash
waters from the product gave a good test for sodic bromide, and in
addition to a large amount of unaltered tribrombenzol we obtained a
very little of an oil, probably the dibromphenol ethylether, which is a
liquid. As this result, which it will be observed confirms Blau, had
given us all the information about the reaction which we wished, we
did not pursue the work further.

Hexabrombenzol gave when boiled with sodic ethylate a good test
for sodic bromide. The organic product was an oil, and a great deal
of unaltered hexabrombenzol was recovered. Hexachlorbenzol gave
a similar result.

Action of Sodic Ethylate on Tetrabromdinitrobenzol.

In the experiments described in this paper the fourth atom of bro-
mine was removed from tetrabrombenzol by sodic ethylate, while in
the work by W. D. Bancroft and one of us f upon the tetrabi-om-
dinitrobenzol, this fourth atom of bromine was not removed by
aniline or sodium malonic ester. It seemed of interest therefore to
try the tetrabromdinitrobenzol with sodic ethylate, and see whether it
behaved like the tetrabrombenzol with this reagent or in the same way

* Monatsh. f. Chem., VII. 630. t These Proceedings, XXIV. 288.

JACKSON AND CALVERT, — DERIVATIVES OF BENZOL. 135

that it had behaved with the other reagents mentioned above. For
this purpose 10 grams of tetrabromdinitrobenzol were dissolved in
benzol, and treated with an alcoholic solution of the sodic ethylate
made from 2 grams of sodium. The solution became claret-colored as
soon as the sodic ethylate was added, but this color changed later to a
reddish brown. There was no perceptible evolution of heat. The
mixture was allowed to stand at ordinary temperatures for two days,
after which it was evaporated spontaneously. The residue, after being
washed with water, was recrystallized from alcohol until it showed the
constant melting point 100° to 101°. This showed that it must be the
tribromnitroresorcine diethylether made by Warren and one of us *
from tribromtrinitrobenzol. This substance must have been formed by
the replacement of one nitro group, and the fourth atom of bromine
by two ethoxy radicals. To confirm this the wash waters of the origi-
nal product were tested for a bromide and a nitrite, and good results
obtained in both cases. If in this case the reaction has run as with
tribromtrinitrobenzol, t which we should infer from the isolation of
tribromnitroresorcine diethylether, the secondary product should be
bromdinitrophloroglucine, or its ethers. As a matter of fact, an oil
liaving the properties of a phenol was obtained by acidifying the
aqueous wash waters from the first product of the reaction, but all our
attempts to bring it into a form fit for analysis have failed, and we
have been prevented from continuing work on this substance by the
departure of one of us from Cambridge.

* These Proceedings, XXV. 183. i Ibid., XXVII. 283.

136 PROCEEDINGS OF THE AMERICAN ACADEMY.

VII.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF
HARVARD COLLEGE.

BROMINE DERIVATIVES OF METAPHENYLENE
DIAMINE.

By C. Loring Jackson and Sidney Calvert.

Presented May 9, 1894.

Introduction.

In the course of an extended investigation of the behavior of the tri-
bromdinitrobeuzol melting at 192°, carried on now for some years in this
laboratory, we took up the reduction of this substance with tin and
hydrochloric acid, since, if the bromine was not removed, new bromine
derivatives of metaphenylene diamine must be obtained, or, on the other
hand, the replacement of the bromine by hydrogen would be of interest,
because cases in which a halogen is removed from the benzol ring under
these conditions, are far from common. As, however, two of those
already observed had occurred in the work on derivatives of tribromdi-
nitrobenzol done in tiiis laboratory, we had reason to hope that the
proposed experiments might furnish us with an additional case, and
contribute something toward determining the conditions under which
this replacement of bromine by hydrogen takes place. After we had
begun our work there appeared a paper * by Schlieper on the removal
of bromine from aromatic compounds by tin and hydrochloric acid, in
which he announced his intention of making an extended research in
this field ; we accordingly wrote to Victor Meyer, under whose direc-
tion Schlieper was working, asking if we might finish our work, and
he has with great courtesy given us permission to do so.

We have succeeded in finding only the following cases in which
bromine is replaced by hydrogen in aromatic compounds by reduction
with tin and hydrochloric acid. Unsymmetrical brommetadinitrobenzol
gives metaphenylene diamine.f Tribromdinitrobenzol sulphonic acid

* Ber. d. ch. G. XXV. 552.

t Zincke and Sintenis. Ber. d. ch. G., V. 792.

JACKSON AND CALVERT. METAPHENYLENE DIAMINE. 137

gives bromdiamidobenzolsulphonic acid.* Dibromdinitroplienylma-
lonic ester gives bromamidoxiiidol,| nitrite of bromdinitrophenjlmalouic
ester gives amidoxyoxindol.J Brorauitrophenol gives amidophenol,
and bromnitrauisol gives amidoanisol.§ Metabromorthonitrobenzoic
acid and its isomeric form botb give authrauilic acid,|| and Hiibuer^
also states that he obtained small quantities of orthophenylene dia-
mine from nitroparabromauiline, but that the principal product was
the bromphenylene diamine ; this, however, seems to be the only case
observed, and none are recorded to our knowledge in the para series.
We could find no record of the removal of other halogens under these
conditions, until Schlieper published a second paper** (after our work
on tribromdiuitrobenzol was finished), in which he describes the reduc-
tion of chlor- and ioduitropheuols or anisols, when he found that iodine
was removed just as bromine was, but that chlorine was not replaced
by hydrogen. On the other hand he found that bromine could not be
removed from derivatives of ortho- or paranitrophenols, as was to be
expected from the work of previous observers. From his experiments
he infers that bromine or iodine is replaced by hydrogen under the influ-
ence of tin and hydrochloric acid, when it stands in the ortho position
to the two negative groups (OH. Br. NOj- 1, 2, 3), although he has
not yet succeeded in bringing an absolute proof that this is the constitu-
tion of the substances with which he worked.

Our experiments proved that when tribromdinitrobenzol was treated
with a mixture of tin, hydrochloric acid, and a little alcohol,tt all three
of the atoms of bromine were replaced by hydrogen, the product being
metaphenylene diamine. Schlieper's inference given in the preceding
paragraph, therefore, is not of general application, as, if this were the
case, only one atom of bromine would have been removed from the
tribromdinitrobenzol, since only one of them is in the ortho position
to both of the nitro groups, the constitution of this substance being
Br. NO^. Br. NOg. Br. H. It is to be observed also that the con-
version of unsymmetrical brommetadinitrobenzol into metaphenylene

* Baessmann, Ann. Chem. CXCI. 244.
t Jackson ami Bancroft, tliese Proceedings, XXIV. 299.
I Jackson and Bentley, these Proceedings, XXVI. 95.
§ Schlieper Ber. d. ch. G., 1892, p. 552.
II Hiibner Petermann, Ann. Chem. CXLIX. 135.
1[ Ann. Chem. CCIX. 3G0.
** Ber. d. ch. G., XXVI. 2465.

tt Sclilieper used stannous chloride and hydrochloric acid in his work, and
it is possible tliat tliis difference may have modified the result.

128 PROCEEDINGS OF THE AMERICAN ACADEMY.

diamine, observed by Ziucke aud Sinteuis, is not in harmony with
this inference of Schlieper. Although, therefore, the diortho position
is not essential to the removal of bromine by tin and hydrochloric acid,
there is no doubt that Schlieper is right in considering it the most
favorable position for this purpose, as the replacement by hydrogen
of an atom of bromine standing between two nitro groups has been
repeatedly observed in this laboratory with comparatively weak re-
ducing agents, such as sodium malonic ester, and probably sodium
acetacetic ester, or sodic ethylate.

In trying to find the cause of the replacement of all three atoms of
bromine by hydrogen in the reduction of tribromdiuitrobenzol the most
obvious theory was that the attachment of the atoms of bromine to the
benzol ring was weakened by the presence of the two negative nitro
groups, as is so frequently observed in other reactions. If this was the
case, the bromine would be removed before the nitro groups were
reduced, and tribromphenylene diamine would not lose its bromine on
treatment with tin and hydrochloric acid. Upon ti-ying this experiment,
however, we found that the whole of the bromine was removed from
the tribromdiamine as easily as from tribromdiuitrobenzol.* It follows,
therefore, that the nitro groups are not the cause of the easy removal
of the three atoms of bromine in the reduction of tribromdiuitrobenzol.
A parallel experiment with dibroraphenylene diamine showed that
this substance gave up its bromine with much more difficulty than the
corresponding tribrom compound, for whereas the tribrom derivative
was converted into phenylene diamine in a few minutes, it took over
twelve hours to bring about the same change in the dibrom compound.
It is evident from these unexpected results that the removal of the
bromine atoms depends in part upon their position toward each other,
and this is not strange, as in other cases it has been observed that
three bromine atoms, when in the symmetrical position on a benzol
ring, are more loosely attached than two in the meta position ; but
this is not the only cause of their removal, since symmetrical tribrom-
benzol is not reduced by tin and hydrochloric acid. It would seem,
therefore, that the replacement of these atoms of bromine is caused
principally by the fact that there are other radicals attached to the
ring, with little regard to the nature of these radicals, as the highly
positive amido groups apparently produce the same effect as the highly
negative nitro radicals. An additional argument for this conclusion

* Other reagents, for instance sodic ethylate or sodic hydrate, do not remove
bromine from the diamine.

JACKSON AND CALVERT. — METAPHEXYLENE DIAMINE. 139

was furnished bj experiments on the reduction of tribromaniline which
was converted into dibromaniline by the action of tin and hydrochloric
acid for a few hours. From the facts at present known we are unable
to deduce any rule in regard to the position which these other radicals
must occupy in order to cause the replacement of bromine by hydrogen,
when the compound is treated with tin and hydrochloric acid.

After we had found, as just described, that the bromine was i"e-
nioved by the action of tin and hydrochloric acid upon tribromdinitro-
benzol, we turned our attention to the study of the bromine derivatives
of metaphenylene diamine, as only one of these had been described, —
a dibromphenylene diamine, which Hollemann * stated he had obtained
by the action of bromine water on the chloride of phenylene diamine.
He gives, however, neither melting point nor analysis of his substance.
Upon trying to repeat his preparation we found that bromine water
converted the chloride of phenylene diamine (or the free base sus-
pended in water) into a tribiomphenylene diamine melting at 158°.
As the product is the same, when an insufficient quantity of bromine
is used, a portion of the phenylene diamine remaining unaltered in this
case, the statement of Hollemann must have rested on an error of
observation. A paper by Vaubehf which appeared after our work on
this subject was finished, confirms our result, as he found by a volu-
metric method that three atoms of bromine are substituted for hydro-
gen in the molecule of phenylene diamine, when it is treated with a
mixture of potassic bromide, sulphuric acid, and potassic bromate.
We have made the same tribromphenylene diamine by the action of
zinc dust and acetic acid on the tribromdinitrobenzol ; this method of
prepai-ation proves that its constitution is Br. NHg. Br. NHo. Br. H.

A dibromphenylene diamine can be obtained by the action of bro-
mine in excess upon phenylene diacetamide. The dibromphenylene
diacetamide melts at 259° to 260° ; the free base at 135°. We have
not succeeded in determining the constitution of this dibromphenylene
diamine. Nor have we obtained the corresponding monobromphenylene
diamine. A monobromphenylene diamine belonging to another series
was obtained, however, by reducing tetrabromdinitrobenzol melting
point 227° with tin and hydrochloric acid ; it melts at 93°- 94°, and
from our experiments on the reduction of tribromdinitrobenzol there
can be little doubt that it has the symmetrical constitution Br. H.
NH,. H. NH,. H. When treated with bromine three atoms are sub-

* Zeitschr. f. Chem. 1865, p. 555.

t Journ. Prakt. Chem. [2] XL VIII. 75.

140 PROCEEDINGS OF THE AMERICAN ACADEMY.

stituted for three of hydrogen in its molecule, and the tetrabrom-
phenyleue diamine is formed, which melts at 212°-213°.

The presence of the bromine in the molecule of metaphenylene dia-
mine has a strong effect on its tendency to form salts and on the composi-
tion of the salts formed. The tetrabrom and tribrom compounds do not
form salts under ordinary conditions, but tribromphenylene diamine can
be converted into its chloride by passing hydrochloric acid through a
solution of it in benzol ; this salt, however, gives up part of its hydro-
chloric acid at ordinary temperatures, the whole of it at 100°. The
dibroraphenylene diamine, on the other hand, forms salts more easily,
and its chloride is only partially decomposed at 100°, while the bromide
of this base seems to be even more stable. All three of these salts con-
tain but one molecule of the acid, the ciiloride of dibromphenylene dia-
mine, for example, having the following formula CgHaBraNHsClNHo.
The monobromphenylene diamine forms salts still more easily, and
they are also more stable. The bromide contains two molecules of
hydrobromic acid having the formula C6H3Br(NH,..Br)2. We have
also prepared the tribromphenylene diacetamide, which does not melt
at 330°, and the tribromphenylene diurethane melting at 212.°

Experimental Part.
Tribrommetaphenylene Diamine C6lIBr3(NH2)2.

The most convenient method for preparing this substance consists in
passing a stream of air and bromine vapor into a solution containing meta-
phenylene diamine, for instance that obtained by adding sodic hydrate
or carbonate to the product of the reduction of metadinitrobeuzol with
tin and hydrochloric acid until all the tin is precipitated and filtering
out the hydrate of tin. In adding the bromine an excess must be
avoided, as this turns the product black, and we have therefore found
it well to filter the precipitate out by means of cheese-cloth as fast as a
sufficient quantity of it was formed, continuing the action of the bro-
mine vapor on the successive filtrates as long as a precipitate appeared.
The crude product, which was always dark colored, was dried, and then
crystallized from alcohol containing a little benzol until it showed the
constant melting point 158°, when it was dried at 100° and analyzed
with the following results : —
I. 0.1459 gram of the substance gave by the method of Carius

0.2392 gram of argentic bromide.
11. 0.2022 gram of the substance gave 14.65c.c. of nitrogen at a

temperature of 19° and a pressure of 771.9 mm.

JACKSON AND CALVERT. — METAPHENYLENE DIAMINE. 141

Calculated for

Found.

C6HBr3(NH„)2.

I. II.

69.56

69.78

8.11

8.2

Bromine

Nitrogen

These analyses prove that the substance is tribromphenylene diamine,
and therefore our experimental result does not agree with that of
Hollemaun,* who stated that he obtained a dibromphenylene diamine
by the action of bromine water on a solution of the chloride of pheny-
lene diamine. As it was possible that this difference in the results ot
the action might be due to the fact that Hollemann used a salt of
phenylene diamine, whereas we used in our first experiments the free
base dissolved and suspended in water, we repeated our experiment
with a solution of the chloride of phenylene diamine, and obtained pre-
cisely the same result as with the free base, that is, the tribromphenylene
diamine melting at 158°. We are therefore forced to the conclusion
that Hollemann's statement is due to an error in observation.

Properties. Tribromphenylene diamine crystallized from alcohol
appears in long slender needles united longitudinally into ribbons
with the ends serrated so strongly that they look like combs. It is
white, with marked silky lustre, and melts at 158°. It is not very solu-
ble in cold but freely in hot alcohol ; somewhat more soluble in methyl
than ethyl alcohol ; freely soluble in acetone ; soluble in ether, ben-
zol, chloroform, glacial acetic acid, or carbonic disulphide ; very slightly
soluble in ligroine, but good crystals may be obtained from its solution
iu hot ligroine ; essentially insoluble in water. Strong hydrochloric
acid dissolves it easily in the cold, depositing after long standing white
transparent crystals which soon turn brown; on heating the solution
in hydrochloric acid it turned dark brown, almost black, showing de-
composition of the substance ; the formation of a similar nearly black
decomposition product was brought about at once by cold strong nitric
acid. Cold strong sulphuric acid dissolved the tribromphenylene dia-
mine at once, the solution having a purplish tint, which turned to dark
brown if it was allowed to stand for some time, or if it was stirred in a
current of air for a short time. Nevertheless on one occasion a white
solid was obtained, but before it could be collected a black spot ap-
peared in one part of it which sped rapidly through the whole mass.
Dilute sulphuric acid had no apparent effect on it in the cold, but de-
composed it when hot. These observations show that the presence of
the three atoms of bromine has destroyed in great measure the basic

* Zeitsch. fur Chem. 1865, p. 555.

142 PROCEEDINGS OF THE AMERICAN ACADEMY.

properties of the phenylene diamine ; but still it is jwssible to prepare
salts of our substance, if proper precautions are observed, and the
chloride is described somewhat later in this paper. Potassic hydrate,
even when a boiling concentrated solution was used, produced no
effect upon it.

When the tribromphenylene diamine is treated with tin and hydro-
chloric acid it is converted into phenylene diamine. The details of
this work are given later in this paper. The success of this experi-
ment suggested to us that the bromine might be removed also by other
reagents ; accordingly we heated some tribromphenylene diamine dis-
solved in benzol and alcohol with sodic ethylate, the action in one trial
being continued for four days, but at the end of this time we recovered
nearly the whole of the tribromphenylene diamine taken, and although
the filtrate after treatment with nitric acid and argentic nitrate gave a
slight precipitate, we ascribe this rather to tlie complete decomposi-
tion of a small portion of the substance (indicated by the dark brown
color of the product) than to any simple reaction. As therefore sodic
ethylate had no action, it did not seem to us worth while to try other
experiments with less energetic reagents.

Chloride of Tribromphenylene Diamine CcHBr3NH3ClNIl2. Al-
though the tribromphenylene diamine forms no salts under ordinary
conditions, we hoped that we might obtain its chloride by the method
which yielded such excellent results to Gattermann * when he applied
it to the tribromaniline. Acconlingly hydrochloric acid g.is was passed
into a solution of tribromphenylene diamine in benzol, until it ceased
to form a white precipitate, and the liquid fumed strongly. The
precipitate was then filtered out, and dried by pressing between filter
paper as quickly as possible, after which it was analyzed as follows : —
0.3096 gram of the salt lost 0.0272 gram of hydrochloric acid at 100°.

Calculated for
CeHBrsNHsClNtlj. Found.

Hydrochloric Acid 9.5 G 8.79

The residue was entirely free from hydrochloric acid. This result
is not so near the theoretical percentage as could be wished, but never-
theless proves that the salt is a monochloride of tribromphenylene dia-
mine. The poor result is undoubtedly to be ascribed to the instability
of the compound, which lost as much as 3.3 per cent of its weight from
standing four days in a desiccator over sulphuric acid, and therefore
might well have lost the 0.77 of one per cent during the drying on

* Ber. d. ch. G., XVI. 636.

JACKSON AND CALVERT. — METAPHENYLENE DIAMINE. 143

filter paper. The chloride as obtained was of a light brown color,
although it is probably white when perfectly pure, and as indicated
by the description of its analysis is very unstable, decomposing slowly
in dry air at ordinary temperatures, quickly and completely when
heated to 100°. We should judge that it was somewhat more stable
than the chloride of tribromaniline prepared by Gattermann,* as might
be expected from the presence of the second amido group. It seems tp
be soluble in strong hydrochloric acid since the tribromphenylene dia-
mine dissolves in this reagent, but it is decomposed by water.

Tribromphenylene Diacetamide, C6HBr.;(NHC2H30)2. This sub-
stance was made by treating tribromphenylene diamine with acetyl
chloride, using anhydrous benzol as a solvent. The action began even
in the cold as shown by the deposition of a white precipitate, but to
make certain that it was complete the mixture was heated under a
return-condenser for one hour. At the end of this time the white
sandy precipitate was filtered out, and purified by extraction with
glacial acetic acid, followed by recrystallization from the same sol-
vent boiling. It was dried at 100°, and analyzed with the following
results : —
I. 0.2726 gram of the substance gave 0.3561 gram of argentic bromide

by the method of Carius.
II. 0.2548 gram of substance gave 0.3358 gram of argentic bromide.

Calculated for

Found.

C(;HBr3(NIIC,H30)j.

I. II.

55.94

55.59 56.08

Bromine

Properties. The tribromphenylene diacetamide crytallizes from hot
glacial acetic acid in bunches of small white plates much longer than
they are broad, which have square ends and frequently radiate from
a centre; from alcohol in radiating groups of compound crystals made
up of needles united longitudinally. It does not melt at 330°, and is
very slightly soluble in all the common solvents. The best solvent
for it is hot glacial acetic acid, but it is far from freely soluble even
in this.

Trihromphenylene Diur ethane, C6HBr3(NHCOOC2H5)2. Five grams
of tribromphenylene diamine were boiled with about twice the amount
of chlorocarbonic ester required by the theory in a flask with a return
condenser for three hours. The crude product (which weighed 6.5
grams), was purified at first by precipitating the substance with ligroine
from its solution in a mixture of benzol and alcohol, afterward by

* Ber. d. ch. G , XVI. 636.

144 PROCEEDINGS OF THE AMERICAN ACADEMY.

recrystallization from a mixture of alcohol and ligroine. or alcohol and
water, until it showed the constant melting point 212,° when it was
analyzed with the following results : —
I. 0.3036 gram of the substance gave by the method of Carius 0.3454
gram of argentic bromide.
II. 0.1945 gram of the substance gave 0.2239 gram of argentic bro-
" mide.

Calculated for Found.

C6HBr3(NHCOOC2H5)2. I. II.

Bromine 49.08 48.42 48.99

Properties. The tribromphenylene diurethane when deposited from
alcohol forms very small crystalline masses, which under the micro-
scope are seen to be rosettes made up of what at first sight appear to
be short needles, but on closer examination prove to be groups of
finer needles united longitudinally, since they betray their complex
nature by the frayed or brushlike appearance of their ends, which re-
semble a partially untwisted cord. Its color is white, and it melts at
212°. It is soluble in cold alcohol, easily soluble in hot ; easily soluble
in chloroform, or acetone, less so in ether ; somewhat soluble in cold
benzol, easily soluble in hot, from which it crystallizes in the same form
as from alcohol, but not so well; slightly soluble in carbonic disulphide
even when hot ; essentially insoluble in ligroine. Water dissolves it to
a very slight extent when hot. Hydrochloric acid produced no apparent
effect hot or cold ; sulphuric acid did not act on it in the cold, but when
hot dissolved it ; from this solution it seemed to be precipitated un-
changed by water, if the heating had not been long continued, but long
heating with the strong acid decomposed it. A concentrated solution
of potassic hydrate seemed to have no effect on it in the cold, but when
heated turned the substance brown, and finally black with apparently
complete decomposition. The best solvent for it is alcohol diluted
with either water or ligroine.

Reduction of Trihromdinitrohenzol.

Tribromdinitrobenzol (melting at 192°, made from symmetrical
tribrombenzol) was reduced by treatment either with zinc and acetic
acid, or with tin and hydrochloric acid, but the products were different
in the two cases as described in the following sections.

Red^tction with Zinc Dust and Acetic Acid. This reduction we
found it best to carry on in an atmosphere of carbonic dioxide, as the
liquid showed a strong tendency to turn black during the process.

JACKSON AND CALVERT. METaPHENYLENE DIAMINE. 145

The zinc dust contained in a flask kept full of carbonic dioxide was
covered with acetic acid of 80 per cent, and the tribromdinitrobenzol
added in small quantities at a time. If the zinc dust and tribromdi-
nitrobenzol were mixed together before adding the acetic acid the
action was far too violent ; in fact, once the mixture took fire when
the acetic acid was added. Even when the process was carried on as
directed, the action was energetic at first, but later it was necessary to
warm the flask gently to assist the reaction. The products were in
part dissolved in tlie aqueous lic^uid, which with all our precautions
was invariably dark colored, and partly appeared as a crystalline pow-
der in the bottom of the flask. The insoluble part was recrystallized
from benzol, by which it was separated into a large fraction melting
at 158°, a small amount of substances melting at a lower temperature,
and a little unaltered tribromdinitrobenzol. The principal product
was shown to be tribromphenylene diamine by its melting point, but
to confirm this it was dried at 100° and analyzed, when it gave the
following results : —

0.2337 gram of the substance gave by the method of Carius 0.3786
gram of argentic bromide.

Calculated for Found.

CpHBralNHa)^.

Bromine 69.56 68.95

This percentage of bromine, although not agreeing with that calcu-
lated so closely as could be wished, is yet accurate enough when taken
in connection with the melting point to leave no doubt as to the nature
of the substance formed, and therefore we did not attempt to get a
number nearer to the theoretical.

The amount of tribromphenylene diamine formed by the reduction
was considerable ; in one case 10 grams of tribromdinitrobenzol yielded
4.7 grams of it, and 0.9 gram of tribromdinitrobenzol were recovered
unaltered, so that the percentage of the theoretical yield was 60.6. The
remainder of the product was contained in the aqueous solution (and to
a less extent in the small amount of lower melting material from the
benzol mother liquors) ; it undoubtedly consisted of phenylene diamine,
from which part or all the bromine had been removed, since the aqueous
solution gave tribromphenylene diamine when treated with bromine
water, a behavior which, as we have already stated, belongs to solu-
tions of phenylene diamine and its salts.

Reduction with Tin and Hydrochloric Acid. We found it best to
proceed as follows : Some granulated tin with a piece of platinum
and about 300 c.c. of strong hydrochloric acid were gently warmed in a
VOL. XXXI. (n. s. xxiii.) 10

146 PROCEEDINGS OF THE AMERICAN ACADEMY.

flask closed by a cork provided with a Bunsen valve, until the greater
part of the air had been expelled. 10 grams of tribromdinitrobenzol
mixed with enough alcohol to make a thick paste were then added, and
the reduction was allowed to go on, warming or cooling the mixture as
was necessary, until all the organic substance had dissolved. If the solu-
tion thus obtained was a strong one, beautiful white needles of a double
tin salt were deposited on cooling ; if it was more dilute, they usually
appeared after standing a day, but in some cases the substance obsti-
nately refused to crystallize. The crystals were filtered out, and de-
composed in concentrated solution with sulphuretted hydrogen, when
the filtrate from the sulphide of tin left on evaporation a crystalline
solid which after three crystallizations from a mixture of water and
hydrochloric acid was analyzed with the following results.
I. 0.2220 gram of the substance gave 31.8 c.c. of nitrogen at a tem-
perature of 26.5° and a pressure of 756.9 mm.
II. 0.2983 gram of the substance gave by the method of Carius 0.4727
gram of argentic chloride.

Nitrogen
Chlorine

The tin and hydrochloric acid had therefore reduced both the nitro
groups, and also removed all three of the atoms of bromine from the
tribromdinitrobenzol ; and that the removal of bromine had taken
place from the whole of the substance was made probable by the fact
that all the organic matter dissolved, whereas the tribroraphenylene
diamine, as already stated in this paper, is insoluble in water or
dilute hydrochloric acid. An additional confirmation of the presence
of phenylene diamine was given by the following experiment. The
base set free from the soluble chloride analyzed above was treated
with bromine water, when a precipitate was formed, wliich after recrys-
tallization from alcohol containing a little benzol melted at 158°, and
was therefore the tribroraphenylene diamine which is made in this way
from metajjhenylene diamine.

^Yhen we began work on this subject we made many analyses of the
double tin salt which crystallized out of the solution after the reduc-
tion ; but although in two different preparations we obtained numbers
agreeing with the salt C6H4(NH3ClSnCl2)2 discovered by Gudemann,*
we found that in others its composition varied so much that no certain

* Zeitsch. fiir Cliem. 1865, p. 51.

Calculated for

Found.

CeH.cNHsCl)^.

I. 11.

15.47

15.79

39.21

39.]

JACKSON AND CALVERT. — METAPHENYLENE DIAMINE. 147

inference iu regard to the composition of the base formed could be
drawn from these analyses, and therefore prepared the salt free from tin
as described above.

After we had shown by the work just described that the three atoms
of bromine are removed from tribromdinitrubenzol by tin and hydro-
chloric acid, it became of interest to determine whether their removal
was due to the effect of the two nitro groups upon them. This seemed
a probable supposition, because it has been repeatedly observed that
atoms of bromine can be easily removed from the benzol ring when they
stand in the ortho position to nitro groups, and work done in this labora-
tory has proved that this tribromdiuitrobenzol is a very reactive body,
but on the other hand it did not seem likely that the bromine atoms
would be removed before the reduction of the nitro groups, as all previ-
ous experience seems to show that the latter are much more easily at-
tacked than the former. It was obvious that this point could be decided
by trying the action of tin and hydrochloric acid on tribromphenylene
diamine, for none of the bromine would be removed from this substance
if this action depended on the presence of the nitro groups. Accord-
ingly five grams of tribromphenylene diamine were mixed with tin and
hydrochloric acid and a piece of platinum added to increase the action.
In the cold no change was observed, and therefore the mixture was
warmed on the water bath, when, after a vigorous reaction had gone on
for a short time, all the organic matter dissolved. The brown clear
solution deposited crystals as it cooled, but disregarding these the whole
was treated with an excess of sodic hydrate and extracted with ether
repeatedly. The residue from the ethereal extracts was an oil, which
solidified after it had been scratched with a sharp glass rod, and then
melted at about 62°. There was little doubt therefore that the sub-
stance was metaphenylene diamine, which melts at 63°, and this view
was confirmed by the following observations. The substance gave a
white salt with hydrochloric acid soluble in water, but nearly insoluble
in hydrochloric acid. To a solution of this chloride bromine water
was added, taking care to avoid an excess ; this produced a white pre-
cipitate, which after recrystallization from benzol melted at 158°, and
was therefore tribromphenylene diamine. It follows from this experi-
ment (which was repeated several times) that tribromphenylene dia-
mine gives up all its bromine to tin and hydrochloric acid, and therefore
that the removal of the bromine atoms does not depend on the
loosening effect of the two nitro groups.

Action of Tin and Hydrochloric Acid on Trihromhenzol and on Tri-
bromaniline. After we had shown, as just described, that all the bro-

148 PROCEEDINGS OF THE AMERICAN ACADEMY.

mine can be removed from tribrommetaphenylene diamine by the action
of tin and hydrochloric acid, it occurred to us that the comparatively loose
attachment of the bromine to the benzol ring might be due to the fact
that the three atoms of bromine were in the symmetrical position, since
Blau has shown that one of these atoms in tribrombenzol can be replaced
by the methoxy radical, and we have confirmed this observation. This
hypothesis seemed the more probable because dibrommetaphenylene
diamine is reduced under these conditions much more slowly than the
tribrom compound. If this view is correct, tin and hydrochloric acid
should remove bromine from the symmetrical tribrombenzol, and ac-
cordingly we tried the following experiment. Ten grams of tribrom-
benzol were mixed with granulated tin and strong hydrochloric acid, a
piece of platinum added to promote the action, and enough alcohol to dis-
solve a large part of the tribrombenzol. The reason for adding the alcohol
was that we thought the action of the reducing agent on the tribrom-
metaphenylene diamine might be in part due to its tendency to dissolve
in the acid, or still more to the solubility of the products of the reduc-
tion in the acid, and we hoped by the addition of the alcohol to establish
similar favorable conditions in the case of the tribrombenzol. The mix-
ture was heated under a return condenser on the water bath for a week.
At the end of this time enough water was added to precipitate any brom-
benzols which might be present, and the heavy crystalline precipitate
thus obtained was filtered, and washed until it was free from stannous
chloride. It was then dried and weighed, when 9.5 grams were ob-
tamed, so that the loss was no greater than would be expected from
the rough way in which the experiment had been carried on. This
substance melted at 120°, and was therefore unaltered tribrombenzol
(melting-point 119.6°). It follows from this experiment that tin and
hydrochloric acid do not remove bromine from symmetrical tribrom-
benzol, and, therefore, that the conversion of tribrommetaphenylene
diamine into metaphenylene diamine by these reagents is not due to
the symmetrical position of the bromine atoms, although it may have a
subsidiary effect in assisting this conversion.

An experiment similar to that just described was tried with 10
grams of tribromaniline. The conditions were exactly the same as
those used for tribrombenzol, given at length in the preceding para-
graph. Upon adding water to precipitate the organic substances, after
the reduction was completed, only a very small precipitate was formed,
which in time changed to long needles. This precipitate weighed less
than 0.5 gram, and as it melted at 118° was undoubtedly unaltered
tribromaniline. The filtrate from this precipitate was treated with

JACKSON AND CAI^VERT. — METAPHENYLENE DIAMINE. 1-19

an excess of sodic hydrate, the precipitate thus formed dissolved in
hydrochloric acid, and reprecipitated with sodic hydrate in excess to
remove the hydrates of tin as completely as possible, and then extracted
with alcohol. The alcoholic extract deposited long white needles,
which melted at 78"-79°, and were therefore probably dibromaniline.
To establish the nature of the substance more fully, we tried the fol-
lowing experiments. The amido group was replaced b^ bromine,
when a product was obtained melting at 44°, the melting point of
the unsymmetrical tribrombenzol, and this when treated with nitric
acid was converted into the mononitrotribrombenzol melting at 93°.
There can be no doubt therefore that the tin and hydrochloric acid
have removed one of the atoms of bromine from the ortho position to
the amido group, replacing it by hydrogen. A quantitative experi-
ment gave the following results, — 10 grams of symmetrical tribroraani-
line yielded after reduction for six hours 5 grams of dibromaniline, from
which were obtained 5 grams of unsymmetrical tribrombenzol, and from
this 5.2 grams of tribromnitrobenzol.

Dlhrommetaphenylene Dicetamide, C6H2Br2(NHC2H30)2.
"Whereas free metaphenylene diamine gives a tribrom derivative,
when treated with bromine water, we have obtained only the dibrom
compound from metaphenylene diacetamide, even when a considerable
excess of bromine was used. The samples analyzed were prepared by
the method given below, but a better method discovered later is de-
scribed after the analyses, — Metaphenylene diacetamide was dissolved
in common acetic acid with the aid of gentle heat, and treated at first
with bromine, and afterward with bromine water, until a slight excess
had been added. The new substance separated from the solution as it
was formed, and after standing a short time it was filtered out, and
purified by crystallization from hot glacial acetic acid until it showed
the constant melting point 259°, when it was dried at 100°, and ana-
lyzed with the following results : —

I. 0.2471 gram of the substance gave 19 c.c. of nitrogen at a tem-
perature of 29° and a pressure of 754.3 mm.
11. 0.21 45 gram of the substance gave by the method of Carius 0.2294

gram of argentic bromide.
III. 0.2033 gram of the substance gave by the method of Carius 0.2198
gram of argentic bromide.

Calculated for Found.

CeHjBroCNHCjHaO),. I. II. III.

Nitrogen 8.00 8.33

Bromine 45.71 45.52 46.01

150 PROCEEDINGS OF THE AMERICAN ACADEMY.

The substance can be obtained more expeditiously in a pure state by
avoiding the preparation of solid phenylene diacetaniide as follows.
A strono- ethereal solution of phenylene diamine is treated with acetic
anhydride until a portion gives a white or slightly yellow precipitate
with bromine (if unaltered phenylene diamine is present, the precipitate
will be brown). When this is the case, an excess of bromine water
is added to the solution, and the mixture stirred vigorously until
the precipitation is complete. The product is essentially pure without
recrystallization.

Properties. The dibrommetaphenylene diacetamide crystallizes by
slow evaporation from an alcoholic solution in very small, rather
short prisms. If on the other hand the crystallization takes place by
cooling, obscurely crystalline masses are obtained in which it is hard
to make out any definite form ; they seem to be prisms coated with
needles. Its color is white, and it melts at 259° to 260° with decom-
position, as is shown by the fact that the melted mass turns black and
puffs up to many times its original volume. On account of this de-
composition it is necessary to keep the melting tube in the oil bath as
little as possible in determining the melting point. If the substance is
allowed to remain in the bath while its temperature is raised, a melting
point as low as 250° may be obtained. It is very slightly soluble in
cold alcohol, and only a little more soluble in hot ; very slightly sol-
uble in benzol, chloroform, or carbonic disulphide ; somewhat more
soluble in acetone, or glacial acetic acid ; essentially insoluble in ether
or ligroine. The best solvent for it is glacial acetic acid or alcohol,
although it is soluble in the latter only sparingly and with great diffi-
culty. It is very slightly soluble in boiling water, insoluble in cold.
Strong hydrochloric acid saponifies it quickly when the two are gently
heated together ; strong sulphuric acid dissolves it in the cold, but
without decomposition, as the unaltered substance is obtained by dilut-
inor and neutralizing the acid, it decomposes the substance when hot;
if dilute it has no apparent effect when cold, but when heated
saponifies and afterward decomposes it; strong nitric acid dissolves it
quickly when cold, but the substance is recovered unaltered on neu-
tralization ; when hot it decomposes it. A strong solution of potassic
hydrate saponifies it when hot.

Dibrommetaphenylene diamine, C,;H2Bro(NH2)2. — This substance
was obtained by removing the acetyl groups from the preceding com-
pound as follows : 10 grams of the dibrom phenylene diacetamide were
heated with from 30 to 40 c.c. of commercial strong hydrochloric acid
for half an hour in a flask with a return-condenser ; at the end of this

JACKSON AND CALVERT. — METAPHENYLENE DIAMINE, 151

time the solid was completely dissolved, and the liquid had taken on a
dark brownish red color. After it was cool, the solution was treated
at first with sodic carbonate, and finally with sodic hydrate in excess,
which precipitated a brown solid. This was filtered out, washed until
free from alkali, and purified by crystallization from boiling water con-
taining a little alcohol until it showed the constant melting point 135°.
In this recrystallization long heating of the solution should be avoided,
as this seems to decompose the substance with precipitation of a black
compound. The pure substance was dried at 100° and analysed with
the following result : —

0.3872 gram of the substance gave by the method of Carius 0.5473
gram of argentic bromide.

Calculated for Found.

CcHoBr, (NH2)2

Bromine 60.15 60.16

Properties. — The dibromphenylene diamine crystallizes in fine
needles which are white, if the substance is perfectly pure, but usually
show a brownish color, which is due to an impurity so slight that it
has no effect upon the analysis. It melts at 135°, and is easily solu-
ble in alcohol whether hot or cold. It is very soluble in acetone ;
easily soluble in ether, somewhat less soluble in benzol, or chloro-
form, if cold, easily soluble, if hot; slightly soluble in cold carbonic
disulphide, soluble when hot ; very slightly soluble in cold ligroine,
more soluble when it is hot ; nearly insoluble in cold water, slightly
soluble in hot, but it is decomposed, if boiled for some time with
water. The best solvent for it is dilute alcohol. It shows a much
stronger tendency to form salts than the tribromdiamine, as it dis-
solves in either hydrochloric acid or hydrobromic acid, and the solu-
tion leaves the salt on spontaneous evaporation ; cold strong sulphuric
acid turns it purple, and then dissolves it, but upon standing the sul-
phate separates from this solution in glistening white or pinkish plates,
which seem to belong to the monoclinic system ; when hot, sulphuric
acid decomposes it ; strong nitric acid dissolves it, and then decom-
poses it even in the cold forming a brown solution. A cold solution
of potassic hydrate has no effect upon it, but, if the mixture is heated,
decomposition sets in as shown by the appearance of a black color.
We selected for further sturly the chloride and the bromide of the
base with the results described below.

We have also studied the action of tin and hydrochloric acid on the
dibromphenylene diamine. For this purpose 3.75 gram of the dibrom-
phenylene diamine were warmed with tin, hydrochloric acid, and pla-

152 PROCEEDINGS OF THE AMERICAN ACADEMY.

tinum under the conditions already described when speaking of the
tribromphenylene diamine. The action in this case was very slow ;
whereas the tribrom compound dissolved after treatment for a short
time, it was necessary to carry on the action for over twelve hours in
order to bring the dibromamine into solution. The product yielded
an oily free base, smelling like metaphenylene diamine, which, how-
ever, we did not succeed in obtaining in a solid state. It gave an
easily soluble chloride, so there can be no doubt that bromine was
removed. We are unable to determine with certainty the cause of the
very slow removal of the bromine from the dibromphenylene diamine ;
it is possible that it is only mechanical, as it was observed that this
substance showed a tendency to form a coating on the surface of the
tin, or on the other hand it may be that the symmetrical position of
the three atoms of bromine in the tribromphenylene diamine has the
effect of loosening their attraction for the benzol ring. This latter
explanation seems to us the more probable.

Chloride of Dibromphenylene Diamine, CsHgBrgNHsClNHo. This
salt can be made by dissolving the diamine in strong aqueous hydro-
chloric acid and allowing the solution to evaporate spontaneously, but
we did not use this method in preparing the salt for analysis, because
we feared that it might undergo a partial decomposition, if deposited
from a solution containing water. We therefore prepared it by pass-
ing hydrochloric acid gas through a solution of the diamine in benzol,
until the liquid fumed strongly, and no more solid matter was depos-
ited. The precipitate thus obtained was pressed repeatedly between
filter paper, and dried for one hour over sulphuric acid, when it gave
the following result on analysis : —

0.2173 gram of the salt gave by the method of Carius 0.3713 gram of
the mixture of argentic chloride and bromide.

Calculated for Found.

CgHjEroNHaClNHj

Chlorine and Bromine 64.63 64.28

The chloride of dibromphenylene diamine is much more stable than
the corresponding tribrom-compound, since, whereas the chloride of
tribromphenylene diamine lost all its hydrochloric acid when heated
to 100° for a short time, the chloride of dibromphenylene diamine
showed a loss under the same conditions of only 5 per cent instead of
the 12.07 per cent, which represents the loss, if the whole of its
hydrochloric acid had been given up. Further the salt of the dibrom
compound can be made by the action of a strong aqueous solution of

JACKSON AND CALVERT. METAPHENYLENE DIAMINE. 153

hydrochloric acid upon the free base, while, although the tribrom base
dissolves in strong hydrochloric acid, we have not succeeded in obtain-
ing the solid chloride from this solution. It must not be inferred from
this comparison, however, that the chloride of dibromphenylene dia-
mine is especially stable, for this is not the case, since it loses nearly
4 per cent of its weight by standing in a desiccator, and is also
decomposed by water.

Bromide of Dibromphenylene Diamine, C(.H2Br2NH3BrNH2. —
This substance was made by adding dibromphenylene diamine to dis-
tilled hydrobromic acid (boiling point 123°) until it ceased to dissolve
easily in the cold. A good deal of trouble was encountered in obtain-
ing the solid salt from this solution, as heating decomposed it, and none
of the organic solvents miscible with water gave a precipitate with it ;
spontaneous evaporation indeed gave the salt, but as we did not know
its properties at the time of the preparation we were afraid of decom-
position during such long standing, and accordingly proceeded as fol-
lows : The aqueous solution was covered with a rather thick layer of
ether, and the whole stirred vigorously for about a quarter of an hour ;
under this treatment the bromide crystallized out in the lower part of
the ether, when it was separated, washed with ether, and dried in a
desiccator over sulphuric acid. It then gave the following results on
analysis : —

I. 0.4433 gram of the substance gave by the method of Carius
0.7191 gram of argentic bromide.
11. 0.3014 gram of the substance gave 0.4884 gram of argentic
bromide.

Calculated for Found.

Cr.HoBroNIIjBrNHj I. II.

Bromine 69.16 69.04 68.96

The sample used in the first analysis had been dried for only three
hours, that for the second had stood in a desiccator over sulphuric
acid for three days, showing that the salt is stable under these circum-
stances. The bromide is therefore more stable than the chloride, as
that lost nearly four per cent of its weight in a desiccator, which
amounts to one third of the acid that it contains.

The bromide of dibromphenylene diamine forms white needles,
which turn brown quickly, when exposed to the air. It is decomposed
by water, but is soluble in ethyl, or methyl alcohol, the solution
apparently decomposing on standing ; insoluble in ether, benzol, oi
ligroine.

154 PROCEEDINGS OP THE AMERICAN ACADEMY.

Reduction of Tetrabromdinitrohenzol. The tetrabromdinitrobenzol
was prepared essentially by the method given by Bancroft and one of
us * but a few improvements in the process should be mentioned here.
By usinc a large amount of glacial acetic acid the yield of tetrabrom-
benzol was materially raised. The details of this process will be
found in our paper, " On the Behavior of Certain Halogen Deriva-
tives of Benzol." In converting the tetrabrombenzol into the dinitro
compound we found that the quantities of sulphuric acid and nitric
acid used could be reduced to 50 c.c. of each for 10 grams of tetra-
brombenzol, if the tetrabrombenzol was pure ; if this was not the case
the best method was to treat the tetrabrombenzol with the acid mix-
ture, which had already been used once with tetrabrombenzol ; the
product of this action of the residual acid was then treated with one
half the amount of fresh sulphuric acid and nitric acid needed for the
complete conversion of the tetrabrombenzol into the dinitro compound,
and in this way a nearly pure product was obtained at little expense
of fuming nitric acid. With pure tetrabrombenzol the yield was very
nearly quantitative, 100 grams giving 121 grams of a product melting
at 220°, which after one recrystallization melted at 227°, the correct
melting point for tetrabromdinitrobenzol ; as 100 grams should give
122.8 grams, this makes the yield of this crude product over 99 per
cent.

To reduce the tetrabromdinitrobenzol it was mixed with alcohol,
tin, and hydrochloric acid, and the mixture heated on the water bath
in a flask with a return condenser, more hydrochloric acid being added
from time to time. When the volume of liquid became unpleasantly
large, it was evaporated in an open dish to a convenient volume, and
returned to the flask with fresh alcohol and acid. After the organic
matter had gone completely into solution, which usually happened after
boilino' for about 70 hours, the liquid was again concentrated and ren-
dered strongly alkaline with crude sodic hydrate. The hydrate of tin
was filtered out with cheese cloth, and dried as thoroughly as possible
by sucking air through it, after which it was shaken repeatedly with
rather large amounts of alcohol. The alcoholic extract thus obtained
was freed from alcohol by distillation, followed by evaporation on the
water bath, when an oil and an aqueous solution were left, which were
separated, if possible, mechanically ; if not, by shaking with ether.
The oil obtained directly or from the ether solidified after a short
time. The aqueous filtrate from the hydrate of tin with the liquid

* These Proceedings, XXIV. 289.

JACKSON AND CALVERT. METAPHENYLENE DIAMINE. 155

separated from the oil was shaken with ether repeatedly, and the oil
thus obtained added to the main portion. The aqueous solution after
shaking with ether gave a very slight precipitate on the addition of
bromine, but the amount of tetrabromphenylene diamine obtained in
this way was so small that this treatment is hardly worth while. The
principal product, after it had solidified, was purified by dissolving is
in rather dilute alcohol and saturating the solution witli suphuretted
hydrogen, when it was warmed gently, and allowed to stand in a
corked flask for some time ; after filtering off the sulphide of tin it
was evaporated to dryness, and the residue recrystallized from pure
ligroine (boiling from 60°-70°) until it showed the constant melting
point 93° -94°, when it was dried in a desiccator and analyzed with
the following result : —

0.2117 gram of the base gave by the method of Carius 0.2113 gram of
argentic bromide.

Calculated for

C6H3Br(NH2)2. Found.

Bromine 42.79 42.48

Properties of 3TonohrommetaphenyIene Diamine. It crystallizes from
a mixture of benzol and ligroine in white spindle-shaped forms often a
centimeter long ; from benzol alone in bundles of small needles united
longitudinally ; from alcohol in not very well developed prisms appar-
ently belonging to the monoclinic system and often twinned. It melts
at 93°-94°, and is readily soluble in acetone, alcohol, ether, or chloro-
form ; in benzol, or carbonic disulphide it is slightly soluble in the cold,
more freely when hot ; nearly insoluble in cold ligroine, difficultly solu-
ble in hot ; it is moderately soluble in cold water, as shown by the fact
that such a solution gives a good precipitate with bromine water ; in hot
water it is even more soluble, the solution as it cools depositing oil drops,
which later change to needles. The best solvent for it is a mixture
of benzol and ligroine, but crystallization from pure ligroine is neces-
sary to get a perfectly white product. It shows a tendency to separate
from its solutions at first as an oil, and also to become decomposed,
when in solution, forming a brown product.

The monobromphenylene diamine shows marked basic properties.
Hydrochloric acid or hydrobromic acid gives with it a salt soluble in
water, a description and analysis of the bromide formed in this way is
given below; moderately strong sulphuric acid also gives a soluble salt,
which forms white crystals when the solution is allowed to evaporate
spontaneously ; dilute nitric acid seems to have little action on it in the
cold, but when hot converted it into a brown substance. Sodic hydrate

156 PROCEEDINGS OF THE AMERICAN ACADEMY.

seemed to have no action even when warmed with it for several min-
utes on the water bath. The constitution of this monobromphenylene
diamine is probably NH2. H. NHg. H. Br. H, as the tribromdinitro-
benzol made from symmetrical tribrombenzol loses all its bromine when
reduced under the conditions used in obtaining the monobromphenylene
diamine from tetrabromdinitrobenzol.

Bromide of Monobromphenylene Diamine, C6H3Br(NH3Br)2.

Monobromphenylene diamine was mixed with distilled hydrobroraic
acid (boiling point 124°-125°), and the liquid heated to boiling. The
solid was then brought into solution by the careful addition of just the
proper amount of water to the boiling solution, after which more of the
hydrobromic acid was added, until crystallization began, when the mix-
ture was allowed to cool ; the crystals thus obtained were washed
with hydrobromic acid and then with ether, dried by pressure with
filter paper, and washed with ether a second time. After drying over
sulphuric acid the salt was analyzed with the following result : —
0.2474 gram of the substance gave according to the method of Carius
0.3945 gram of argentic bromide.

Calculated for
CeHsBrCNIIjBr),. Found.

Bromine 68.77 G7.86

This analysis leaves much to be desired, but the result is near enough
to show that the formula given above is correct. The salt forms
colorless transparent glassy crystals, which look like cubes under the
microscope ; freely soluble in water, slightly in hydrobromic acid.

Teirahromphenylene Diamine, C6Br4(NH2)2.

This substance was made by dissolving the monobromphenylene
diamine, melting at 93°-94'^ in ether and adding a slight excess of
bromine ; the product, after recrystallization from a mixture of chloro-
form and ligroine, showed the constant melting point 212^-213°, when
it was dried and analyzed with the following result: —
I. 0.2357 gram of the substance gave by the method of Carius 0.4166

gram of argentic bromide.
11. 0.2493 gram of the substance gave 0.4432 gram of argentic bro-
mide.

Bromine

Calculated for

Found.

C6Br,(NH,)2,

I. IT.

75.47

75.22 75.67

JACKSON AND CALVERT. METAPHENYLENE DIAMINE. 157

Properties of Tetrahromphenylene Diamine. It crystallizes in very
small white needles, which melt at 212°-213°, and are freely soluble
in benzol, chloroform, ether, acetone, or carbonic disulphide ; slightly
soluble in cold alcohol, freely iu hot ; almost insoluble in cold ligroine,
slightly soluble iu hot. The best solvent for it is a mixture of chloro-
form or benzol with ligroine. The tetrabromphenyleue diamine has
nearly, if not quite, lost the basic properties of a diamine, as was to
be expected since it is loaded with so many negative bromine atoms.
Hydrobromic acid had but little apparent action on it, even when hot,
although the liquid deposited after long standing a few transparent
crystals, which may have been the bromide.

158 PROCEEDINGS OF THE AMERICAN ACADEMY.

VIII.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF
HARVARD COLLEGE.

A REVISION OF THE ATOMIC WEIGHT OF ZINC.

FIRST PAPER : THE ANALYSIS OF ZINCIC BROMIDE.

By Theodore William Richards and Elliot Folger Rogers.

Presented April 10, 1895.
Table of Contents.

PAGE

IntroducUon 158

Balances and Weights 102

The Specific Gravity of Ziucic Bro-
mide 102

PAGE

Preliminary Analyses of Zincic

Bromide 163

Final Series of Determinations . . 170

The Atomic Weicht of Zinc ... 179

Introduction.

In an account of a recent investigation on the occlusion of gases by
the oxides of metals * it was shown that zincic oxide, in common with
cupric and magnesic oxides, has the power of retaining important
quantities of oxj'gen and nitrogen gases, even at very high tempera-
tures. Hence it was evident that all determinations of the atomic
weight of zinc depending upon the conversion of the metal into the
oxide through the ignition of the nitrate must be influenced by a con-
stant error, whicli has the tendency to make the results lower than
the true value. In consideration of this fact, it becomes very impor-
tant to review all of the results thus far obtained regarding the atomic
weight of zinc, in order to determine how seriously the error in ques-
tion may influence our accepted value. A chronological listf of the
accessible data is given below.

* These Proceedings, XXVIII. 200, Richards and Rogers.

t Much assistance in preparing the list has been obtained from the works of
Clarke, Meyer and Seubert, and others. The results have been recalculated
with the assumption of the following atomic weights: 0 = 16, C = 12.002
CI = 35.456, Aq = 107.98, H = 1.0075.

RICHARDS AND ROGERS. — ATOMIC WEIGHT OF ZINC. 159

1809. Gay Lussac, Mem. d'Arceiiil, II. 174 65.55

1811. Berzelius, Pogg. Ann., VIII. 1«4 65.57

1842. Jacquelaii), An. de Chim. et de Phys. [3.], VII. 204 , 6G.24

1844. Favre, An. de Chim. et de Phys. [3.], X. 163 .. . 65.99
1844. Erdmann, Berz. Jahresber., XXIV. 132, and Pogg. Ann.,

LXII. 612 65.05

1883. Pelouze and Fremy, Chem., p. 55 65.07

1883, Baubigny, Comptes Rendus, XCVII. 906 .... 65.41

1883. Marignac, Archiv. Sci. Phys. et Nat. [3.], X. 5, 193 . 65.30

1885, Van der Plaats, Comptes Rendus, C. 55 ?

1887. Reynolds and Ramsay, J. Ch. Soc. Trans., LI. 854 , 65.67

1888. Morse and Burton, Am. Chem. Journ., X. 311 . . . 65.27

1889. Gladstone and Hibbert, J. Ch. Soc. Trans., LV. 443 . 65.44

The various results have been reached : —

1. By converting a known weight of metallic zinc into the oxide
(Berzelius, Jacquelain, Erdmann, Morse and Burton).

2. By the evolution of hydrogen from acids by metallic zinc, the
hydrogen being either measured or burned, (Jacquelain, Favre, Van
der Plaats, Reynolds and Ramsay).

3. By the conversion of a salt of zinc into zincic oxide through
ignition (Favre, zincic oxalate ; Pelouze, zincic lactate ; Baubigny,
zincic sulphate).

4. By the determination of the electrolytic equivalent of zinc
(Gladstone and Hibbert).

5. By analysis of a haloid salt of zinc (Marignac.)

The work of Morse and Burton by the first method is so far supe-
rior to the previous determinations made in the same way, that the
the older ones may be wholly neglected. The two or three possibil-
ities of infinitesimal error, such as the chance that the zinc might con-
tain impurities taken from the glass used for its distillation, may be
wholly neglected when compared with the great error due to the
occlusion of nitrogen and oxygen. As the amount of this error is
dependent upon the pliysical condition of the zincic oxide, it is im-
possible to make an acccurate correction except by the actual deter-
mination of the gas in the oxide remaining from the determinations.
From our own experiments * it would appear that a gram of zincic

* These Proceedings, XXVIII. 200.

160 PROCEEDINGS OF THE AMERICAN ACADEMY.

oxide obtained from the nitrate usually contains about 0.00057 gram
of occluded gas ; upon this basis Morse and Burton's result would
become 65.458 instead of 65.2G9.

In considering the results obtained by the second method the results
of Favre and Jacquelain may be rejected at once. Of Van der
Plaat's results it is necessary to state only that some error must have
crept in while recording his data, for it is inconceivable that 6.6725
grams of zinc should yield only 1.1424 litres of hydrogen. The
work of Reynolds and Ramsay was much more careful and detailed ;
but the results varied very widely. Hydrogen evolved from very
pure zinc was measured, with many precautions ; but after the
rejection of thirty-four experiments eleven more gave a value of
65.24, and still five more gave 65.47 as the atomic weight of zinc.
Since the publication of their work many investigations have shown
that the density of hydrogen is greater than the value assumed at
that time. If a litre of the gas weighs 0.9001 * gram, and the atomic
weight of hydrogen is taken as 1.0075 (O = 16.000) the atomic
weight of zinc deduced from Reynolds and Ramsay's experiments
becomes about 65.63.

The two older results obtained by the third method are worthy of
no further mention. Baubigny's work upon the ignition of zincic sul-
phate is very interesting, but probably incomplete. It will be remem-
bered that the value for copper obtained by the same method was
too low, t owing, probably, to the occlusion of sulphuric acid by the
cupric sulphate. It is not impossible that a similar error may have
crept in here, for the conditions were similar ; but it is probable that
it is here counterbalanced by the retention of sulphur trioxide by the
zincic oxide. In a series of experiments made in this laboratory pure
zincic oxide obtained from the carbonate (see p. 163), was ignited to
constant weight in an oxidizing atmosphere at a temperature above the
fusing point of gold ; t it was then dissolved in dilute sulphuric acid
which left no residue upon evaporation, and very gradually brought
again to the same high temperature. In no case were we able to
expel all of the sulphuric acid which we had added. Three experi-
ments are appended : —

* Lord Rayleigh and others.
t llichards, These Proceedings, XXVI. 275.

I Four grams of pure gold melted in fifteen minutes from the time of turn-
ing on the air blast in tlxe furnace.

RICHARDS AND ROGERS. — ATOMIC WEIGHT OF ZINC. 161

Weight of
Zincic Oxide before.

Weight of
Zincic Oxide after.

Gain.

Grams.

Grains.

Grams.

1.03009

1.03032

0.00023

0.80243

0.80265

0.00022

1.03447

1.03473

0.00026

Average for 1 gram ZnO . .

. . . 0.00025

Gladstone and Hibbert deposited silver in one cell by means of a
voltaic current, while zinc was being dissolved from an amalgamated
plate in another. Although the experiments are interesting, it would
appear from the results of Vanni * and others that the possibility of
side reactions makes the strict applications of Faraday's law for the
determination of atomic weights of rather doubtful efficacy. One
would exi^ect the method adopted to give a result larger than the true
one.

Marignac's work upon the chloride of zinc and the double chloride
of zinc and potassium is even less satisfactory than his investigations
of other chlorine compounds ; it merits no further notice.

It is evident from these statements that the three least unsatisfac-
tory determinations are all vitiated to a greater or less extent by con-
stant errors ; the work of Morse and Burton by one which tends to
lower the result ; the work of Gladstone and Hibbert by one which
may tend to raise the result ; and the work of Baubigny by two which
tend to counteract one another. One would expect the atomic weight
of zinc to prove in the end equal to about 65.4.

It seemed very desirable to obtain a series of determinations which
should be wholly different from any of these ; and for this reason
zincic bromide was chosen as the starting point of the present re-
search. Additional advantages presented by the use of this substance
are the fact of its ready and accurate analysis, and the fact that a
determination of the ratio 2 Ag : ZnBr, would bring the element into
a series of elements which have been determined by Stas and others
with great precision in this way.

* Berichte d. d. Ch. G., XXIV., Ref. 882.

VOL. XXXI. (n. S. XXIII.) 11

162 PROCEEDINGS OF THE AMERICAN ACADEMY.

Balances and "Weights.

The preliminary determinations were made upon a long-armed
Becker balance with the help of very carefully standardized platinized
brass weights. The final determinations were made upon the admi-
rable Troemer balance procured for the research upon copper, and sub-
sequently used for those upou barium and strontium. The weights
used in these final determinations were the same as those used in the
researches just cited ; they were compared with one another at the
beginning and at the close of the research, with satisfactory results.
All weighings were made by substitution, the tare weights being ves-
sels as nearly as possible similar to those being weighed ; and all
were reduced to the vacuum standard by the usual formula.

The Specfic Gravity of Zincic Bromide.

In the course of recent investigations upon the atomic weights so
many of the usually accepted specific gravities of hygroscopic sub-
stances have been found to be seriously in error, that it seemed advis-
able to redetermine the constant which influences the reduction to
vacuum of the present results.

Pure zincic bromide was dried for a long time at a temperature of
200°, and then fused and heated for a sliort time at 300°. The pyc-
nometer in which this drying was effected was then stoppered and
cooled in a desiccatoi-. Carefully dried toluol having a specific grav-
ity of 0.8646 at 20°, when referred to water at 4°, was used as the
liquid to be displaced. Toluol is convenient for the purpose, because
so few inorganic substances are soluble in it, and because its volatility
is not so great as to cause serious loss during the weighing, but is
great enough to allow of the rapid drying of the exterior of the appa-
ratus. The first sample of zincic bromide was made from very pure
hydrobromic acid and ordinary pure zinc, and was distilled in an
atmosphere of carbon dioxide ; the second was made from the purest
electrolytic zinc and the purest bromine. Boih samples gave a per-
fectly clear dilute solution in water after the expulsion of the toluol
on the steam bath after the experiment. The decanted toluol left
upon evaporation on the steam bath only a trace of residue, which
was insoluble in water. Water decanted from this residue gave no
trace of precipitation with argentic nitrate ; hence it is evident that
zincic bromide is insoluble in toluol. The liquid contained in glass

RICHARDS AND ROGERS. — ATOMIC WEIGHT OF ZINC.

163

increases about 0.001 of its apparent volume for each degree of tem-
perature ; and the small appropriate correction is applied below.

Specific Gravity of Zincic Bromide.

No. of
Experiment.

Weight of
ZuBrj.

Tempera-
ture.

Weight of
Toluol displaced.

Water at 40° corre-
sponding to Toluol.

Specific Gravity
of ZnBrj at 20'.

(1)

Grams.

3.8856

11.2394

Degrees.

19.8
20.3

Grams.
0.7960
2.303

Grams.
0.9206
2.664

4.220
4.218

Average 4 910 1

The value 4.22 is used in the work which follows. In this connec-
tion it may be of interest to compare the sj^ecific gravities of the sub-
stances recently determined here.

Specific Gravities compared with Water at 4°.

Substance.

Old Values.

Former
Experiments.

New Values.

Temperature.

Anhydrous Ba CI2

3.85

f Quincke ]
J Favre and 1
j Volson [
[ Scliroeder J

3.86

24°

BaBfo

4.23

Scliife

4.79

24°

Sr Bn,

3.96

Bodeker

4.22

24°

Zn Br„

3.64

Bildeker

4.22

20°

Crystallized Ba -

3.05±

f Joule and ]

3.10

24°

CI22H2O

Playfair 1

■ Schiff f

Scliroeder J

■

PRELmiNARY ANALYSES OP ZiNCIC BrOMIDE.

Preparation of Zincic Oxide. The zincic bromide used for the
first series of experiments was made by the action of pure hydrobromic
acid upon pure zincic oxide. For the preparation of the oxide "pure"
zinc of commerce was dissolved in pure dilute sulphuric acid, and the
solution was allowed to remain over an excess of the metal for several
weeks. The filtered solution was acidified with sulphuric acid, warmed,
and treated with well washed hydric sulphide, until a considerable
mass of pure white precipitate had formed.

164 PROCEEDINGS OF THE AMERICAN ACADEMY.

The strongly smelling filtrate from this zincic sulphide, now freed from
all traces of the most usual metallic impurities, was treated with chlo-
rine water to oxidize any iron or manganese which might be present,
and precipitated fractionally with pure sodic carbonate. The first
fraction of the precipitate containing traces of iron and manganese was
thrown away.

After a thorough washing the second fraction, containing most of the
zinc, was dissolved in pure nitric acid, keeping the carbonate in excess.
After filtration, the addition of a little ammonic carbonate, and another
filtration, the greater part of the zinc present was precipitated by means
of ammonic carbonate. When it had been subjected to a thorough wash-
ing the basic zincic carbonate was ignited in a double platinum crucible
over an alcohol lamp. The oxide thus obtained was washed repeat-
edly with water, for often an impurity which is held by a wet precipitate
may be washed out when the precipitate has been partially decomposed
or disintegrated by heat. Artus * has observed that the sodic carbo-
nate occluded in the quantitative precipitation of zinc may be easily
removed in this way, and the present experience showed that occluded
zinc chloride could be waslied away with equal ease. The zinc oxide
thus obtained was almost white, with a very faint tinge of yellow ;
upon solution in nitric acid it gave absolutely no opalescence with
argentic nitrate. Its method of preparation made the presence of
non-volatile impurities almost impossible, for it was precipitated from
a solution containing nothing but zinc, nitric acid, and ammonia.

Preparation of Hydrohromic Acid. This substance was made by
the action of bromine on water in the presence of phosphorus. The
bromine was purified by the well known method of Stas,| having been
dissolved in a saturated solution of potassic bromide holding zinc oxide
in suspension, and distilled from this solution after long standing. The
bromine, thus freed from chlorine and iodine, was collected under
water and redistilled. Red phosphorus was now purified by very fine
pulverization under water and by repeated washing of the powder with
pure water. According to Stas this method removes every trace of
chlorine which may be held by the substance. Our own experience
has not been uniformly favorable in this respect, but upon this occasion
and several others both the qualitative tests for the absence of chlorine
and the quantitative analysis of the hydrobromic acid made with the
assistance of well washed phosphorus were satisfactory. The bromine

* Berzelius, Jahresbericht, XXIII. 132.
t Mem. Acad. Belg., XLIII. Pt. II. 90, 38.

RICHARDS AND ROGERS. — ATOMIC WEIGHT OF ZINC. 165

was allowed to act upon the phosphorus and water with the usual pre-
cautious in an apparatus made wholly of glass. The acid thus formed
was distilled in five fractions, of which the first consisted chiefly of
water and a trace of bromoform. Only the last fraction of the distil-
late containing perhaps a third of the bromine taken, was used in the
work ; and this was redistilled with the further rejection of the first
and last portions. The remainder was analyzed quantitively with pure
silver in order to test its freedom from chlorine, with very satisfactory
results. The silver was prepared, weighed, and dissolved with great
care, the precipitate was collected upon a Gooch crucible, and all
weighings were reduced to the vacuum standard.

Purity of Hydkobromic Acid.

No. of
Experiment.

(3)

(4)

Weight of Silver.

1.86058
1.72320

Weight of
Argentic Bromide.

3.23884
2.99983

Percent of Silver in
Argentic Bromide.

57.446
57.443

Average 57.444

Stas found 57.445

Preparation of Zincic Bromide. For experiments 5 and 6 zincic
bromide was made by simply dissolving in a platinum dish the pure
oxide in the pure acid described above. For experiment 7 similar
zincic bromide was sublimed in a wide glass tube in a current of pure
dry carbon dioxide. The next experiment, No. 9, was made with
similar zincic bromide prepared wholly in glass and not sublimed.
A large portion of the substance was then prepared by exact neutrali-
zation and evaporation in a platinum dish, a strip of pure zinc was
added to precipitate a trace of platinum, — which had been dissolved
because of the presence of traces of oxidized nitrogen in the oxide, —
and the whole was subjected to crystallization. The mother liquor
served for analysis 8, and the pure white crystals for the third series
of preliminary determinations (Experiments 10, 11, 12, and 13).

Preparation of Silver. — This substance was prepared by the method
described in a recent paper upon the atomic weight of barium.*

Pure argentic chloride was reduced by means of pure sodic hydrate
and invert sugar, and the metal was washed and fused in a gas flame

* These Proceedings, XXVm. 22; XXIX. 64.

166 PROCEEDINGS OF THE AMERICAN ACADEMY.

upon charcoal. The himps thus obtained were dissolved in pure nitric
acid and precipitated by electrolysis. The beautiful crystals of elec-
trolytic silver were rapidly fused upon cupels of sugar charcoal * in the
flame of illuminating gas, and cooled in a reducing atmosphere. Such
silver is essentially identical with the mucli more carefully prepared
metal used in the final experiments, and gives every evidence of being
pure within two or three parts in a hundred thousand. f A solution of
argentic nitrate through which over fifty grams of such silver had
passed by electrolysis, was found upon suitable treatment to yield only
half a milligram of baric sulphate ; hence the silver could not have
contained more than two parts of sulphur in a million.

Preparation of other Materials. — Nitric acid was repeatedly dis-
tilled, the last portion of the distillates being used. It is needless to
say that it was wholly free from halogens. The sulphuric acid used for
analytical purposes was distilled with great care ; that used for desic-
cators was boiled with a little aramonic sulphate. The water used in
the preliminary determinations was twice distilled in a tin condenser.
In order to avoid the introduction of chlorine, carbon dioxide was
prepared at first from acid sodic carbonate and sul[)huric acid, later
from dilute nitric acid and marble. Nitrogen mixed with argon was
prepared by passing a mixture of air and ammonia over red-hot copper.
Both gases were very thoroughly washed.

Method of Analysis. — The necessity of driving out every trace of
water from the substance to be analyzed was the precaution upon
which most labor was expended. The low boiling points of zincic
bromide makes it possible to distil the substance easily in a hard glass
tube. Accordingly, for the first series of very crude experiments pure
zincic bromide was distilled in a tube provided with bulbs, which were
sealed off when filled. In order to obviate the introduction of an error
from the additional weight of the atmosphere of carbon dioxide in
which the distillation took place, the bulbs were heated to about 120°
at the moment of sealing. The bulbs were weighed after scratching
them upon each end with a file ; and after the zinc bromide had been
dissolved out the glass was weighed alone. The bromine present was
weighed as argentic bromide, and the atomic weight of zincic was
calculated from the ratio of argentic and zincic bromides. The more
trustworthy results ranged from 65.40 to 65.54, with an average of
65.47 ; but it was clear that the method admitted of too many possi-
bilities of error to yield satisfactory results.

* These Proceedings, XXIX. 65. t Ibid.

RICHARDS AND ROGERS. — ATOMIC WEIGHT OF ZINC. 167

3

hr

C-

■^3

X

u

a

•r-

a>

2S

'U

+j

?

—

"51,

br,

w

3

,4j

^

o

c«

c:

c:

■:g

^

0)

0/

o

^n

a

"m

o

c

c

o

CO

as

hr

a
o

dj

r

M

0;

0)

-

= > — , 2

o H

w tS *"" "" O)
^ "5) — *^ "!!^

Sis-go

^ ^ Si

0)

01 3

S 5

.S 08 3

C nj bO

■S « 2

3 ^ «*-i

•3 C- O

^ fcc

rt .5

O 2 OJ
3 X -M

2 =•

^ o S

O; O

•3 0) en

3 r- S

^ i .2

2 2 «

£^ *" a

-w O

>> 3 5;

on >

168 PROCEEDINGS OF THE AMERICAN ACADEMY.

Hence the method which answered well in the recent analysis of
strontic bromide* was adopted here. The pure recrystallized or
sublimed zincic bromide was placed in a platinum boat and kept for
some time in an atmosphere of pure dry nitrogen charged with pure
hydrobromic acid. It was found that in this way all the water could
be expelled from the salt without the introduction of a trace of oxy-
bromide ; indeed, zincic bromide which by rapid heating in the air had
been partly decomposed could be speedily brought back to its normal
condition by fusion in the atmosphere of dry dilute hydric bromide.
The presence of the insoluble oxybromide is easily detected by dis-
solving the bromide in large amounts of water ; iu every case the
bromide used in the analyses below gave an absolutely clear solution.
Baric and strontic bromides ignited in the same way give solutions
which are absolutely neutral to methyl orange and phenol phthalein ;
hence it is most likely than the zincic bromide, which does not admit
of similar alkalimetric testing, is also quite normal. For the details
the paper upon strontium must be consulted, but a sketch of the appa-
ratus will probably suffice. (See preceding page.) In the present
case hydrogen could not be added to the nitrogen lor fear of reducing
some of the zincic bromide ; but no trouble was experienced from cor-
rosion of the boat. When the substance had been kept for some time
in a state of tranquil fusion, and had just solidified, the boat was
quickly slid into a weighing tube which was in its turn placed in the
automatic desiccator tube shown below. After the tube and boat

had been heated to about 200° for some time in a current of pure dry
air, the desiccator tube was raised to a vertical position, the stopper
being thus allowed to fall into place.

After weighing, the pure zinc bromide was dissolved in water and
precipitated by means of a slight excess of very carefully weighed pure
silver in very dilute solution. The argentic bromide was collected
upon a Gooch crucible, the shreds of asbestos carried through (0.05
to 0.20 milligram) were collected upon a very small fine filter, and the
total weight of the argentic bromide thus obtained gave one ratio upon
which to base the atomic weight of zinc. In the third series the filtrate

* These Proceedings, XXX. 369.

RICHARDS AND ROGERS.

ATOMIC WEIGHT OF ZINC.

169

was all evaporated to very small bulk, and the excess of silver precipi-
tated by hydrobromic acid and weighed upon a Gooch crucible in the
same way as the other portion of argentic bromide. By subtracting
this excess from the total silver originally weighed out, the weight
of silver equivalent to the zincic bromide could be easily found. Great
care was taken to exclude daylight, and to carry out all the precautions
necessary in accuiate work. The data are given below.

SECOND SERIES.
The Ratio of Zincic Bromide to Argentic Bromide.

No. of

Weight of

Weight of

Atomic Weight of

Analysis.

Zincic Bromide.

Argentic Bromide.

Zinc.

Grams.

Grams.

(5)

1.69616

2.82805

65.469

(6)

1 98198

3.30450

65.470

(7)

1.70920

2.84949

65.487

(8)

2.35079

3.91941

65.470

(9)

2.66078

4.43751

65.400

Average . . .

. . 65.459

THIRD SERIES.
The Ratio of Zincic Bromide to Silver and Argentic Bromide.

No. of
Analysis.

Weight of
Zincic Bromide.

Weight bf
Silver.

Weight of
Argentic Bromide.

Atomic Weight
of Zinc from
Agj : ZnBrj.

Atomic Weight

of Ziuc from
2AgBr : Zn Brj.

(10)

(11)

(12)
(13)

Grams.
2,33882
1.97142
2.14985
2.00966

Grams.
2.24063

1.88837
2.05971
1.92476

Grama.
3.90067
3 28742
3.58539
3.35074

65.409
65.444
65.396
65.472

65.400

65.434
65.402
65.463

Av

jrasre . . ....

65.430

65.425

The second series of results, excepting the last determination, is
undoubtedly affected by the presence of water in the zincic bromide ;
for the methods of drying and transference had not been perfected.
The third series was much more carefully made, but even here there
was a possibility of the retention of a small amount of water in some
of the analyses, hence this result also is probably somewhat too high.

170 PROCEEDINGS OF THE AMERICAN ACADEMY.

The results of the four analyses of the third series give the follow-
ing figures for the per cent of silver in the bromide : 57.443, 57.443,
57.447, and 57.443. The mean of these results is 57.444 ; Stas hav-
ing found 57.445. It will be remembered that the hydrobromic acid
from which the zincic bromide was made gave precisely similar results.
This identity proves that the analytical work was without fault, and
that argentic bromide does not possess the slightest tendency to occlude
zincic bromide when precipitated from dilute solutions. The analysis
of the hydrobromic acid proved that the material was free from chlorine
and iodine.

Accordingly the rather large variations in the results must be due
wholly to the original condition of the samples of zincic bromide. It
remained therefore, to make a final series of determinations upon zincic
bromide from which water should have been absolutely excluded ;
and since one of us was unfortunately called away, this series was
made by the other alone.

Final Series of Determinations.
By Theodore William Richards.

One determination of the final series. No. 17, was made with the
old zincic bromide in the new apparatus. The others were all made
from new material prepared from electrolytic zinc and pure bromine,
instead of from zincic oxide and hydrobromic acid. The electrolytic
zinc was prepared with great care in the following manner : An excess
of " pure" zinc was treated with somewhat dilute pure sulphuric acid
at 80° until upon dilution a marked amount of basic salt was formed.
About three hundred grams of zinc had been dissolved ; and this di-
luted solution was allowed to stand for many hours in contact with
the zinc and the basic salt. After filtering, clean pieces of zinc were
added to the solution; and no further metallic precipitate formed upon
the zinc. The solution was then decanted and treated with a small
amount of sulphuric acid and much hydric sulphide. After some time
the pure white precipitate was separated by decantation and filtration,
and the solution was oxidized by an excess of pure chlorine. To this
solution was then added enough very pure sodic carbonate to form a
slio-ht precipitate, and the mixture was allowed to stand several days
with occasional stirring. The pure white precipitate, which must
have contained any trace of iron remaining, was filtered off, and the
sulphate of zinc was crystallized three times successively from hot
water.

RICHARDS AND ROGERS. — ATOMIC WEIGHT OF ZINC. 171

The solution of the last crystals was allowed to stand for two days
over several grams of the purest crystalline electrolytic ziuc in a
large platinum dish. At the end of that time the solution contained
some basic salt, but the dish showed no sign of a metallic coating.
The solution was filtered, treated with an excess of freshly distilled
ammonia, and electrolyzed. A thin rod of very pure zinc served as
the negative pole, and a platinum wire as the positive. Six decom-
posing cells were run simultaneously on a shunt from a fifty-volt
dynamo which was being used for charging a storage battery. The
current in each decomposing cell varied from one to one and a half
amperes, — if a much stronger current was used the cells became too
warm. As Ramsay and Reynolds * have suggested, it is advisable to
remove the remarkably beautiful crystals from time to time as they
grow ; for this purpose a bent five-pronged glass fork made from a
heavy rod was found very useful. The crystals were washed with
ammonia until the washings were absolutely free from sulphuric acid,
then with ^^ure very dilute hydrobromic acid, and finally with much
pure water. About forty grams of pure zinc thus formed were treated
with an excess of pure bromine, which had been shaken with an alka-
line bromide in aqueous solution, dissolved in concentrated calcic
bromide, precipitated by water, and distilled under dilute pure hydro-
bromic acid. The Jena glass flask in which the combination took
place was cooled during the reaction. The red solution was filtered
in a glass funnel through asbestos, and the excess of bromine, together
with any trace of iodine which may have been present, was driven off
by leaving the Jena flask upon the steam bath for some time in a very
much inclined position. The diluted colorless solution was evaporated
to small bulk, and the greater part of it was subjected to fractional
crystallization by cooling to zero. A portion which had crystallized
twice successively, from water was labelled (A), and another portion,
the extreme mother liquor remaining from two crystallizations was
labelled (C). Sample (B), the intermediate fraction, was not used in
these experiments, as (A) and (C) were proved to be identical.

Before analysis both (A) and (C) were subjected either to distillation
or to sublimation. Tlie sublimation was carried on in the lower part
of a platinum retort, to which had been fitted closely a glass adapter
for conducting the current of pure dry carbon dioxide. The substance
to be sublimed was contained in a small platinum crucible fitted with
a wire handle, by which it could easily be raised, lowered, or removed.

* Loc. cit.

172

PROCEEDINGS OF THE AMERICAN ACADEMY.

Co,

The adapter was so arranged that the current of gas came as closely
as possible in contact with the crucible, and so that any zinc bromide
which might condense in a liquid form upon the glass, and thus run
the risk of taking alkali from it, must return to the crucible and be
redistilled. The sectional drawing will give a clearer idea of the
arrangement.

Two powerful Bunsen burners supplied heat from below, impinging
upon a porcelain dish which fitted closely to the bottom of the retort

and protected the platinum. The
gases from the flame were diverted
by a large diaphragm of asbestos
board. By means of this arrange-
ment it is possible to sublime about
half a gram of zincic bromide an
hour ; the crystals are exceedingly
beautiful, and give every evidence
of great purity.

Instead of being sublimed, some
of the pure salt was distilled in a
current of carbon dioxide. For this
purpose a medium sized tube of the
hardest glass was drawn out so as
to serve for a small retort, and this
was encased in a larger hard glass
tube, from which it was separated
by several pieces of platinum foil. A platinum boat, into which was
directed the drawn out and turned over point of the inner tube,
served as the receiver. Here again a diagram must assist the
explanation.

Sca/e J nd/iwai size

Scale, y nataral size

The zinc bromide thus distilled possessed a peculiarly brilliant white
lustre ; in no case did the boat lose or gain the twentieth of a milli-
gram in weight during the distillation.

RICHARDS AND ROGERS. — ATOMIC WEIGHT OF ZINC. 173

The attempt was made also to distil the bromide in a vacuum, but
the reduction of the pressure lowered the boiling point too nearly to
the proximity of the melting point for convenient manipulation.

In this connection it may be well to state the atomic weight of zinc
obtained from these specimens, in order to show their identity. The
details of these figures are given later.

From the substance used in the preliminary determinations

(Expt. 17) 65.410

From new substance not crystallized from water but distilled

in carbon dioxide (Expt. 14) 65.403

From extreme mother liquors from crystallization (C) sub-
limed in carbon dioxide (Expt. 18) 65.404

From purest crystals (A), twice crystallized from water and

sublimed (Expt. 15, 19) 65.404

From purest crystals (A) twice crystallized from water and

twice distilled in carbon dioxide (Expt. 16) ... 65.398
Average Zn = 65.404

Siloer. — The silver used in the final determinations was repeatedly
purified by the methods already described. Finallj^ the beautiful
electrolytic crystals were fused in a small crucible of pure lime in a
vacuum. In this way the metal may be obtaine.l in the purest possi-
ble state, for if it is distilled according to Stas there is always danger
of impurity from the oxygen and illuminating gas or hydrogen used in
the oxygen blowpipe.

Other Materials. — The acids were purified in the usual fashion and
the greater part of the water used was only distilled twice, rejecting
the first portions. For experiment 16, all the water used was dis-
tilled three times, once over potassic permanganate. Of course the
platinum condenser which has been already described served for all of
these distillations.*

Phosphoric pentoxide was sublimed in a stream of pure oxygen.
Since the presence of oxides of nitrogen in carbon dioxide might assist
the partial decomposition of zincic bromide, nitric acid was rejected as
a means of decomposing marble, and very dilute hydrochloric acid
was used instead. The gas was purified by passing through a solu-
tion of sodic hydric carbonate, long tubes containing argentic nitrate,
and much pure water. Since the last tube containing water gave
absolutely no test for chlorine after over a hundred litres of the gas

* These Proceedings, XXX. 380.

174 PROCEEDINGS OF THE AMERICAN ACADEMY.

had passed through it, one may safely assume that the purification was
sufficient. The gas was dried by means of sulphuric acid and phos-
phoric peutoxide.

Method of A nalysis. — Perhaps the best method of explaining the
method of analysis is to give a detailed description of a single deter-
mination; and for this purpose Analyses 15 and 19, in which both
silver and argentic bromide were weighed, will best serve.

The very pure sublimed zincic bromide was pressed into a platinum
boat ; and the boat was placed in a tube of hard glass, which had been
ground into another tube designed to contain a weighing bottle. The
apparatus consisted essentially of a combination of the two pieces of
apparatus shown upon pages 167 and 1G8; it was devised for a
research upon the atomic weight of magnesium now being carried on
by Messrs. Richards and Parker, and it will be described in full when
that investigation is pixblished. With the help of this apparatus
it was possible to heat the zincic bromide to any temperature below
its boiling point in an atmosphere of pure dry air, pure dry carbon
dioxide, or pure dry carbon dioxide charged with hydrobromic acid ;
and these gases could be changed at will merely by the opening and
closing of stopcocks. When the heating had been continued for the
desired length of time, it was possible to push the boat into the weigh-
ing bottle and to stopper the weighing bottle very tightly in a perfectly
dry atmosphere, without the least chance of the absorption of moigture
from the outside air. All the apparatus which could possibly come
into contact with bromine or hydrobromic acid was made of glass,
with ground glass joints and glass gridirons for convenient refilling.

The zincic bromide was heated very gradually at first in an atmos-
phere of carbon dioxide which had been dried by passing over sul-
phuric acid, fused zincic bromide, and phosphorus pentoxide. If
heated very gradually in this way, zincic bromide may he almost
wholly dehydrated without loss of bromine ; but a basic bromide is
certain to form if the heating is rapid. When all of the apparent
water had been expelled from the substance and its containing tube,
dry hydiic bromide was added to the carbonic dioxide, and the tem-
perature was gradually raised to the fusing point of zmcic bromide.
The bromide was kept at a temperature just above its melting point
for about an hour ; during this time perhaps a tenth of the substance
sublimed in the exit end of the " combustion " tube, — rendering the
drying tube — which had been ground on to protect the exit — un-
necessary. It was assumed that at the end of an hour the fused zincic
bromide must be as free from water and from basic salt as it was

RICHARDS AND ROGERS. — ATOMIC WEIGHT OF ZINC. 175

possible to obtain it ; accordingly, the temperature was allowed to
fall to about 200°, and the current of dry hydrobromic acid was
stopped. Soon air — dried by means of sulphuric acid, fused potash,
and phosphorus pentoxide — was substituted for the carbon dioxide,
the temperature being allowed to fall to about 150° to avoid any possi-
ble decomposition of the ziucic bromide; and this current of air was
continued for several hours until long after every trace of hydrobro-
mic and carbonic acid had been swept away. In order to " make
assurance doubly sure " the whole length of the tube, weighing-bottle
and all, was heated to 1 00° or more several times, in order to prevent
any possible occlusion of acid. When all was in readiness, the warm
boat was pushed by means of a long glass rod into the weighing bot-
tle, and the bottle was stoppered with the help of the same rod. The
apparatus was then pulled to pieces, the closed bottle was transferred
at once to a tight desiccator, and after a suitable rest of several hours,
it was weighed with all possible care. The various weighings are
tabulated below.*

Weight of boat beforehand .
Weight of boat + bottle . .
The same + ZnBr2 after seven )

liours' cooling . . . . )
After two hours more . . .
After two hours' standing in )

balance )

Weight of boat after experiment
Boat + bottle afterward . .

Gain of boat

Gain of boat + bottle . . .

Average tare of boat + bottle )

+ ZnBr., <j

Average tare of boat + bottle .

Weight of ZnBrj in air . . .

Correction to vacuum 20° and

762 mm

Weight of ZnBro in vacuum .

Common Weights
Rigbt-hand pan,

7693
,9757

2389

2389

,7693
9757

Tare :

Standard Weights,

Left-hand Pan by

Substitution.

Grams.

7.76937

0.00372

5.26754

5.26754

5.26757

7.76938
0.00375

Standard
Weights

7.76932
0.00372

5.26747

5.26747

5.26750

7.76933
0.00375
0.00001
0.00003

5.26748
0.00373

5.26375
.00074

5.26449

Before having been treated with water, the bottle and boat were
allowed to stand for twenty hours in somewhat moist air. During

* For detailed method see These Proceedings, XXVIII. 5.

176 PROCEEDINGS OF THE AMERICAN ACADEMY.

that time they gained only two tenths of a milligram, showing that
the diffusion of moisture through the stopper was very slow. In order
to prove the efficiency of the drying and fusing apparatus, the bottle
with its contents were returned to it, and the stopper was removed by
means of a wire while a current of dry air was allowed to pass through
the whole apparatus. This current was continued for half an hour,
the boat and bottle being warmed to about 120°. After suitable cool-
ing the substance was found to have lost one tenth of a milligram,
being one tenth heavier than it was in the first place. These weights
were not included above in order to avoid complications ; they do not
belong necessarily to the analysis, but merely serve to show that the
apparatus was wholly sufficient for its purpose.

The boat and its contents were conveyed with great care to a large
Bohemian beaker, and the weighing bottle was rinsed out many times
with the purest water. When the zincic bromide had wholly dis-
solved, the perfectly clear solution was transferred through a large
funnel to the glass stoppered Erlenmeyer flask intended to serve for
the precipitation. The boat was easily washed by allowing it to rest
in the funnel, and of course beaker and all were rinsed with the most
scrupulous care. The total volume of the thus diluted zincic bromide
amounted to about two hundred and fifty cubic centimeters.

The silver to be used for the precipitation weighed 5.04328 grams
in the air, or 5.04313 grams in vacuum. It was dissolved in ten cubic
centimeters of nitric acid diluted with water in a large flask provided
with a bulb tube, and the absolutely clear solution was freed from
lower oxides of nitrogen by standing upon the steam bath. The con-
tents of the bulbs, usually containing only a few hundredths of a milli-
gram of silver, were washed back into the flask, and the whole was
diluted to about half a litre.

The addition of the argentic nitrate to the zincic bromide took place
in the dark-room,* and great pains were taken to prevent exposure of
the materials to anything but non-actinic light from this time until the
final weighing of the argentic bromide. Of course every trace of sil-
ver was washed into the flask containing the zincic bromide. After
the whole had been very thoroughly shaken and allowed to stand for
several days, 0.21 cubic centimeter of hydrobromic acid (one cubic
centimeter was equivalent to a milligram of silver) failed to produce
an evident precipitate, even when the solution was lighted by a bril-
liant condensed beam of yellow light. On the other hand, 0.60 cubic

* See previous papers upon barium and strontium.

RICHARDS AND ROGERS. — ATOMIC WEIGHT OP ZINC.

177

centimeter of an equivalent solution of silver produced an evident
cloudiness. 0.40 cubic centimeter more also produced a cloud. Still
0.20 cubic centimeter gave a very faint indication of opalescence, after
a long time, and finally 0.21 gave no trace. On titrating back vpith
hydrobromic acid, it was necessary to add 1.21 cubic centimeters more
before the point was reached where an extra 0.20 produced no precipi-
tate. The total amount of silver solution added was then 1.41 cubic
centimeters, or 0.21 to precipitate the first amount of hydrobromic acid
added, 0.21 to show that all the bromine was precipitated, and 0.99 to
correspond to the true amount of silver to be added in order to attain
the end point ; while the total amount of hydrobromic acid added to
reach the end point in the other direction was 1.42 cubic centimeters,
an amount which must be subtracted from the amount of silver added.
Hence, since the true amount of silver corresponding to the zincic
bromide is the mean between the amounts found by titrating in oppo-
site directions,* we obtain it as follows : —

Amount of Silver Required.

Grams in
Vacuum.

Titrating with argentic nitrate .
Titrating with hydrobromic acid

5.04313 + 0.00099
5.04313 -f 0.00141 -

= 5.04412
- 0.00142 = 5.04132

Average : silver required ....

. . . . 5.04362

Since it is hardly probable that the end point may be obtained by
this method more nearly than the tenth of a milligram, the last figure
is omitted below.

If, as is probable, the action of silver and hydrobromic acid in pre-
venting the solution of silver bromide resembles the action of silver
and hydrochloric in precipitating solutions of silver chloride (Stas),
the difference between the end points (1.00 milligram of silver) cor-
responds to six times the amount of argentic bromide dissolved. That
is, this amount must be about '/^ or 0.29 milligram in 0.85 liter of a
solution containing two or three cubic centimeters of nitric acid and
about five grams of zincic nitrate.

In order to precipitate all the bromine, four milligrams more of
argentic nitrate were added to the solution and the whole was very

* These Proeeedings, XXVIII. 24, XXIX. 74, XXX. 384.

VOL. XXXI. (n. 8. XXIII.) 12

178

PROCEEDINGS OF THE AMERICAN ACADEMY.

vigorously shaken. After a day or two the mixture was filtered
througli a weighed Gooch crucible and the precipitate was shaken and
allowed to stand with many successive portions of water containing
five milligrams of argentic nitrate to the litre. It was finally trans-
ferred quantitatively to the crucible, washed with pure water to
remove the traces of argentic nitrate and dried at 160° in a porcelain
drying oven in a stream of air purified by ()assing over sulphuric acid
and potash. The two litres of filtrate were allowed to stand until the
shreds of asbestos had settled ; and these were collected upon three
thicknesses of the best filter paper, ignited and weighed. The main
body of the precipitate was transferred to a porcelain crucible, weighed,
fused in a porcelain oven, and weighed again. Below are the data.

Goocli crucible alone .
Gooch crucible + AgBr

Argentic bromide in air
Correction to vacuum .

Chief mass of argentic bromide in vacuum

Porcelain crucible + AgBr

The same after fusion

Loss on fusion ....
Crucible + ash + asbestos
Crucible

Ash + asbestos .
Ash of three filters

Asbestos

Total weight of AgBr

Subtract amount corresponding to }

1.42 + 0.20 = 1.62 c.c. HBr solution . . |

Argentic bromide corresponding to z
bromide

incic )

Common
Weights.

Grams.
18.4812
27.2633

25.0757
25.6755

15.2857
15.2853

Standard
Weights by
Substitution.

Grams.
0.22840
y.01052

.37781
.37700

0.00021
0.05284
0.05226

0.00058
0.00012

0.00046

Standard
"Weights
corrected.

Grams.
0.22840
9.01050

8.78210
0.00039

8.78249

0.00021

0.00046

8.78274
0.00282

8.77992

Thus 5.2G449 grams of zincic bromide correspond to 5.0436 grams
of silver and 8.77992 grams of argentic bromide. If bromine and
silver are taken as 79.955 and 107.93 as usual, the atomic weights
of zinc computed from" these ratios are identical, — 65.404. Since
this is the case, of coui-se the per cent of silver found in argentic bro-
mide Itojo X 100 is identical with that obtained by Stas, or 57.445.

RICHARDS AND ROGERS. — ATOMIC WEIGHT OF ZINC. 179

In no one of the other determinations were both silver and argentic
bromide determined, hence the others were simpler in execution,
although essentially the same in principle. The end point in analy-
sis 17 was probably not determined more nearly than two tenths
of a milligram, while that in analysis 19 was determined by the
nephelometer * much more accurately even than in the detailed deter-
mination. Indeed this value was so accurate that the hundredths of a
milligram may have some significance, although the weights taken
were so large. The data of all these final determinations are given
below.

FINAL DETERMINATIONS.

The Ratio of Zincic Bromide to Silver.

Number of
Analysis.

Weight of
Zincic Bromide.

Weiglit of
Silver.

Ratio.
ZnBr,

2Ag

Atomic Weight
of Zinc.

(14)
(15)
(16)

Grams.
6.23833
5.26449
9.36283

Grams.
5.9766
5.0436
8.9702

104.379
104.380
104.377

65.403
65 404

65.398

Average

104.379

65.402

The Ratio of Zincic Bromide to Argentic Bromide.

Number of
Analysis.

Weight of
Zincic Bromi(ie.

Weight of
Argentic Bromide.

Ratio.
ZnBr,
2AgBr '

Atomic Weight
of Zinc.

(17)
(18)
(19)

Grams.
2.65847
2.30939
5.26449

Grams.

4.43358
3.85149
8.77992

0.599622
0.599606
0.599606

65.410
65.404
65.404

Average

0.599611

65.406

From the first ratio, if 0= 16
From the second ratio, if 0 = 16
Average, if 0 = 1 6

If 0 = 15.96
If 0=15.88

Zn = 65 402
Zn = 65.406
Zn = 65.404

Zn = 65.240
Zn = 64.912

* These Proceedings, XXX. 385.

180 PROCEEDINGS OF THE AMERICAN ACADEMY.

la this determination the only serious possibility of error is that the
zincic bromide may have contained a lingering trace of water in spite
of the elaborate precautions adopted to secure its absence. It is
hoped that before long a determination of the zinc as well as of the
bromine in zincic bromide may be made in this laboratory, furnishing
evidence upon this point. However, since the figure given above
aorees closely with the probable corrected result of Morse and Bur-
ton's investigation, as well as with the somewhat incomplete deter-
minations of Baubigny and Gladstone land Hibbert, the value 65.40
may be safely adopted for the present as the most probable value of
the atomic weight of zinc.

TABER. — LINEAR TRANSFORMATION. 181

IX.

NOTE ON THE AUTOMORPHIC LINEAR TRANS-
FORMATION OF A BILINEAR FORM.

By Henry Taber.

Presented April 10, 1895.
§ 1.

In the BnUetin of the New York Mathematical Society for July,
1894, I have shown that not every proper orthogonal substitution can
be generated by the repetition of an infinitesimal orthogonal substitu-
tion. That is to say, if we designate a substitution of the orthogonal
group as of the first or second kind according as it is or is not the
second power of a substitution of the group, there are then proper
orthogonal substitutions of the second kind ; and whereas any substi-
tution of the first kind can be generated by the repetition of an infini-
tesimal substitution of the group, no substitution of the second kind
can be generated thus. Nevertheless, by the repetition of an infini-
tesimal substitution of the orthogonal group, we can obtain a substitu-
tion of the first kind which shall be as nearly as we please equal to
any proper substitution whatever of the second kind.

I also pointed out in this paper that for the orthogonal group every
substitution of the first kind was the mth. power of a substitution of the
group, for any positive integer m ; and that every substitution of the
second kind was the (2 m -j- l)th power of a substitution of the group.

It follows of course at once that an exactly similar theory exists for
the group of linear substitutions which transform automorphically a
symmetric bilinear form with cogrediant variables. An exactly similar
theory exists also for the group of linear substitutions which transform
automorphically an alternate bilinear form with cogrediant variables,
and for the group of linear substitutions which transform automorphi-
cally a general bilinear form (neither symmetric nor alternate) with
cogrediant variables, as remarked in a note at the conclusion of the
above mentioned article.

On the other hand, any linear substitution of the group of linear
substitutions which transform automorphically a bilinear form with

182 PROCEEDINGS OF THE AMERICAN ACADEMY.

contragrediant variables can be generated by the repetition of an infin-
itesimal substitution of the group. For, if the two sets of variables
of the bilinear form

are contragrediant, and if

(a.-i, To, . . . X,) = (<^ 55 tl' ^2, • • • Q,

(rji, rjo, . . . 7j„) = {f^yui/.„'-- y,),
in which ^ denotes the transverse or conjugate to ^, we have

(.^-1 O (^ ^ t^l, t^2, . . . 4 ^ 7?!, 770, . . . 77,.)

= (fi ^ ari, 0-2, . . . ar„ i;^ yi , 2/2 , . . . y„) ;

and the necessary and sufficient condition that the transformation shall
be automorphic is

or

fi(/) = </)f2.*

Let now 8 denote the identical substitution. Since it is assumed
that a reciprocal of </> exists, that is, | 0 | =^ 0, a polynomial ;^ = y (^)
in ^ can be found such that

<^ = e^

* In this paper I employ the notation of Cayley's " Memoir on the Linear
Automorpiiic Transformation of a Bipartite Quadric Function," Philosophical
Transactions, 1858, with these exceptions, namely, the identical substitution will
be denoted by S, and the linear substitution or matrix transverse or conjugate
to the linear substitution or matrix ^ will be denoted by ^. Cayley denotes the
bilinear form 2,- 2s ar,Xsjjr {>', s = 1, 2, . . . n), as above, by

(n (J xi, xo, . . . x„^ yi, y2> ■ • ■ yn),
the symbol XI denoting the matrix

a.2i a.22 . . .

that is, the square array of coefficients of the form.

The determinant of the linear substitution (p will be denoted by | <j) |

TABER, — LINEAR TRANSFORMATION. 183

where e^ denotes the convergent infinite series

Since ^ is commutative with fi, ;( is also commutative with Q ; con-

1 l^

sequently for any positive integer w, — ;^ and therefore e" are com-
mutative with fi. Whence it follows that if

1
^ = e'" ,
we have

and

^m _ gX _ ^_

That is, any linear substitution which transforms automorphically the
bilinear form

with contragrediant variables is the ?nth power of a linear substitution
which also transforms this form automorphically. By taking m suffi-
ciently great, the coefficients of the linear substitution ;^ can be made
as nearly as we please equal to zero, and thus the linear substitution

ij/ = e"' may be made as nearly as we please equal to the identical
substitution. But however great m may be, we have, nevertheless,

* The infinite series

eX = S + X + 2X' + 3!X^ + • • • + ^ X-" + • • •

is convergent for any linear substitution x ; and we liave

(eX)-i = e- X.

and if m is any positive integer,

If X and x' are commutative, we also have

gX gX ^ gX + X .

For any linear substitution <j) whose determinant is not zero a polynomial
X = f{^) in <p can be found such that

184 PROCEEDINGS OF THE AMERICAN ACADEMY.

whence it follows that any linear substitution of the group of linear
substitutions which transform automorphically a bilinear form with
contragrediant variables can be generated by the repetition of an
infinitesimal substitution of the group.

§2.

If the two sets of variables of the bilinear form

are transformed by linear substitutions transverse or conjugate to each
other, so that

(a:i, x^, . . . a:,,) = (<^ ^ ^i, ^2, • • • L),

(j/uy^, ' ■ ■ y„) = (i^Vi^Vi^ • ' ■ Vn),

where </> denotes the linear substitution transverse or conjugate to <^,
we have

(^n<t>^$i, |o,.. . ^„^ 771,770, . . .r]„)

= (n^Xi,x., . . .x„'^yi,yr,, . . . y„).

The necessary and suflFicient condition that this transformation shall
be automorphic is that <f> shall satisfy the equation

The class of linear substitutions that satisfy this equation, that is,
the linear substitutions which transform the bilinear form in the man-
ner described, do not form a group ; but they can be separated into
substitutions of the first or second kind according as they are or are
not the second power of a substitution of this class.* And any sub-
stitution of the first kind can then be generated by the repetition of
an infinitesimal substitution of this class, whereas no substitution of
the second kind can be generated thus.

* If <pn.(p = a, then <p- a (p"^ = <p {<p a (p) (p = <p a <p = n.

TABER. — LINEAR TRANSFORMATION. 185

For assuming that the determinant of the form is not zero, that is,
I O I 4^ 0, it follows that | </> | 4^ 0. Consequently a polynomial
X =^ f i4>) c^" be found such that

Let

then

J. „'*"
<p ^ e

From the identity

e He ^= ilj
we have also

Since
therefore

is a polynomial in <f), and consequently commutative with &i2.
Whence we have *

sa + Qx'i r?n n» / / — i ^
e =e e = (fxp '^ = o.

Conversely, if ^ O and fi ^9^ are commutative and such that

then if

</) = e ,

from the preceding identity it follows that

* See note, page 183.

t For any positive integer m we have

Therefore

t Since

nf{4>) n-i = a (Sm cm <^'") n-i = s^c^ n<^'«n-i = Sm c^ {a <f> n-'^)'"

= 2„c„(.^-i)'»=/(<|.-i).

186 PROCEEDINGS OF THE AMERICAN ACADEMY.

In particular, it & CI — — CI &, (f) = e^ satisfies the preceding
equation.
Let now

then

n e^ — ^ n {& n + n &) n-^

= i (o/(</.) o-i + "/(0-') n-')

= i (/(fi «/) 0-1) +/(fi <jf>-i o-i))

Again, let
then

Since 60 0 and 0i fi are polynomials in ^, they are commutative.
Therefore

J 2 __ g2,>fi _. g2 9„n+2e,n ^ e2 9„n g2 9jn _ - g2 e, n

g2 9on ^ g,>n + n,? -- g_
2

«A

=

-<

),n ^

: e

-^n^i

>

•A

fii/'

=

: e

' m " ^1

n

,^.«

= 0,

•A

m

(.^.

")"

—.

g2flifi :

= ./>-^.

smce
If now

and

Consequently, any linear substitution, as ^^ which is the second
power of a linear substitution satisfying the equation

(p ii (p ^= i2j

is the wth power for any positive integer m of a solution of this
equation. By taking m sufl[iciently great, we can make the coefficients

* See last note, page 185.

TABER. — LINEAR TRANSFORMATION. 187

2

of — ^1 fi as small as we please, and thus we may make the substi-
tn

tution i// as nearly as we please equal to the identical substitution.

Whence it follows that any linear substitution of the first kind which

.satisfies the equation

(that is, any linear substitution which is the second power of a solution
of this equation) can be generated by the repetition of a linear substi-
tution which is also a solution of this equation and which is infinitely
near the identical substitution.

Any linear substitution </> satisfying the equation

is of the first kind if — 1 is not a root of the characteristic equation of
cf) (that is, (f) is then the second power of a substitution satisfying this
equation). For if — 1 is not a root of the characteristic equation of
<t>, we may put

n-^ Y = - 8 + 2 (S + (/))- 1 = (8 - <^) (8 + </>)- ^

and we then have

Yt2-^=:n(o-^'Y)n-^

= n (8 - <f>) (8 -\- <^)-in-i

= o (8 - <^) fi-i • fi (8 + </))-^ fi-i

= (8 — n (t> n-^) {8 + n cjiQ- ^)~^

= (8-<^-0(s + c/>-0-^

= (cf> - 8) (ct> + 8)-'^ = - n-^ Y.

From the expression for Q,~ ^ Y we also obtain

(0-1Y + 8) (<^ + 8) =2 8;
and consequently, since | 0~-^ Y + 8 | 4= 0,

<^ = -8 + 2(8 + 0-iY)-i= (8 + n-^Y)-i(S-n-iY).
If now & r=y (()-i Y) is a polynomial in fi~^ Y such that
8 + fi- 1 Y = e^

188 PROCEEDINGS OF THE AMERICAN ACADEMY.

then, if d' =f(— fi"^ Y) we have

>>

8-n-^Y = e^' i*
aud consequently

Wherefore, if

en = -&-{-y=: -/(n-i y) +/(- q-^y),

= -q/(q-iy)q-i + q/(-q-iy)q-i
= -/(YQ-^)+/(-YQ-^)

and consequently, if

, pien -iue
ij/ = e- =z e ^ ,

we have
and

that is, <^ is the second power of a solution of the equation

(fi Q cf> = a.

If a linear substitution satisfying this equation is sufficiently near to
the identical substitution, — 1 is not a root of its characteristic equa-
tion. Therefore an infinitesimal substitution satisfying this equation
is of the first kind. But the repetition of a substitution of the first
kind gives a substitution of that kind. Whence it follows that no
substitution of the second kind, which satisfies the equation, can be

* For

Therefore,

i?'=/(-n-iT)
=f(n- n-iT. n-i)

= n (5 + n-i r) n-i

TABER. — LINEAR TRANSFORMATION. 189

generated by the repetition of an infinitesimal substitution satisfying
this equation.

Let <^ be a linear substitution of the first kind of the class we are
now considering. That is, let

li/ Q li^ = i2,
and let

The roots of the characteristic equation of (f> are then the squares of
the roots of the characteristic equation of ip. Consequently, if — 1 is
a root of the characteristic equation of ^, ^/ — 1 is a root of the char-
acteristic equation of i}/ ; that is,

I xj/ — V^ 8 I = 0.
But then

I ./.-I - V^T S I = 1 12 (,/,-i - V^^l 8) Q-i I = I ,/. - V^S I = 0 ;

and since

^-1 _ y'lTi 8 = - V^ ^-^ ("A + V^ 8),
we have

I ^ + V^ 8 I = 0 ;

that is, — V — 1 is then also a root of the characteristic equation of {j/.
It is convenient at this point to introduce a term which has been em-
ployed by Sylvester. Thus, following Sylvester, I shall say that the
nullity of the linear substitution 4> is m, if all the (m— l)th minor? of
the matrix or determinant of $ are zero, (that is, the minors of order
n — m + 1, if n is the number of variables,) but not all the mth
minors (the minors of order n — m). If now the nullity of ij/ — ^/—IS
is tn, then, since

^j, — ^ZTT s = Q. (,/.-! — V^ 8) n-\
the nullity of

— V^TiA"' (<A + V^^ 8) = ./.-I — V^^ 8

is also m ; therefore, since | i//"^ | 4= 0, the nullity of i/r-f- V — 1 8 is m.
Whence, by the "corollary of the law of nullity," the nullity of

^ + 8=(ip- V^^l 8) (^ + V^ 8)

190 PROCEEDINGS OF THE AMERICAN ACADEMY.

is then 2 m. Similarly, if p is any positive integer, and if tte
nullity of

(^ _ ^3T ^p = Q (^- 1 _ ^zn: ^)P Q-i

is m, the nullity of

is also m ; and therefore the nullity of {ip + V — 1 S)^ is ?«. But then
the nullity of

is 2 m.

We have therefore the following theorem by which we may ascer-
tain the existence of substitutions of the second kind. If ^ is a linear
substitution of the first kind which satisfies the equation

</) f2 (^ = i2,

then if — 1 is a root of the characteristic equation of <f), the nullity of
any positive integer power of ^ + 8 is even.

That substitutions of the second kind actually exist may be shown
by considering the form

« (^i Vx — ^'2 y-i) + ^ ^2 y\^

which is transformed automorphically if we impose upon the a:'s the
substitution whose matrix is

-1 1

0 -1

and upon the ^s the transverse substitution ; that is, if we put

3"l = — Cl "T ^25 X2 ^ t2»

and

3/1 = — y]\-, 1)2 = rix — V2-

This linear substitution does not satisfy the preceding conditions,
and is therefore of the second kind. It follows that there are one or
more bilinear forms for any number of variables which are trans'
formed automorphically in the manner we have considered in this
section by linear substitution of the second kind.

By definition, no solution of the second kind of the equation

TABER. — LINEAR TRANSFORMATION. 191

in an even power of a solution of this equation. But if ^ is a solution
of the second kind, for any positive integer m, we can find a linear
substitution satisfying this equation whose (2/?2+l)th power is equal
to '}>. Thus, employing the notation of pages 185, 186, let i^Q be a
polynomial in (/> such that

and, as above, let

^0 Q =: I- (^ O + Q {>), OiQ^ ^{&Q — Q&).
Then, if

i// 12 i/^ = O,
and

Corresponding to any linear substitution ^ of the second kind satis-
fying the equation

^ i2 (^ ::= Q,

can always be found a solution cf)^ of the first kind whose coefficients
are rational functions of a parameter ^, such that, by taking ^ suffi-
ciently small, the coefficients of ^^ may be made as nearly as we please
equal to the corresponding coefficients of </>.

§3.
If the two sets of variables of the bilinear form

(12 ^ a-i , a-2, . . . a:„ ^ yi , 5/2, . . . «/«)'

of non-zero determinant, are transformed by a linear substitution whose
product is equal to the identical substitution ; thus, if

(Xi, X^, . . . .T„) =r (0 ^ ^1, 4, . . . 4),

(yi, 1/2, " ■ y,,) = (</)~^ ^ vu V2,--- vn\

we have

and the necessary and sufficient condition that the transformation shall
be automorphic is

^~' i2 ^ = i2.

Any one of the class of linear substitutions which satisfy this equa-
tion is the mth power of a linear substitution of this class, and can be

192 PROCEEDINGS OF THE AMERICAN ACADEMY.

generated by the repetition of an infinitesimal substitution of this class.
For let X —j (</>) be a polynomial in 0 such that

Let

that is,
then
And since

^ = Q-i0i2,
and consequently

we have
Therefore

^ Q =/ (0) — dO..
If now m is any positive integer, and

j/^ r= e'" = e'" ,
then

y ,^i o S i fi *

\J/ ^ = e '" = e ' ;

and we have identically

- , -lad 'on

if/-^ Qil/ = e '" Qe'" = 12.

We also have

if/"" = e'» ^ = <j>.

Consequently any linear substitution <f) which satisfies the equation

<^- ^ Q (^ — Q

is the Twtli power for any positive integer m of a linear substitution if/
which also satisfies this equation ; and since by taking w sufficiently
great we can make ij/ as nearly as we please equal to the identical
substitution, 4> can be generated by the repetition of an infinitesimal
substitution which also satisfies this equation.

* See note, page 183.

HOLMAN. — THERMO-ELECTRIC FORMULA. 193

Investigations on Light and Heat, made and published wholly oe in part with
Appeoprution prom the Rumfoed Fund.

X.

CONTRIBUTIONS FROM THE PHYSICAL LABORATORY OF THE
MASSACHUSETTS INSTITUTE OF TECHNOLOGY.

XLIV. —THERMO-ELECTRIC INTERPOLATION
FORMULA.

By Silas W. Holman.

Presented November 13, 1895.

In this paper are collected the several well known types of for-
mulae for expressing the thermal electro-motive force of a couple as a
function of the temperature of its junctions. Two new formuke are
also proposed. All then are tested against the most reliable experi-
mental data upon the subject, and their relative merits discussed.

The Existing Formula.

Consider a simple closed electric circuit composed of two different
metals, each homogeneous in matter and temper, the metals beino- in
contact at two points. For simplicity, assume the metals to be in
the form of wires joined at their ends. Let one junction be at a tem-
perature of /i°, the other of c°, on the ordinary centigrade scale. Let
2^e be employed as a suggestive symbol to denote the resultant elec-
tro-motive force in the circuit, induction being excluded from consid-
eration. Then 2*e is a function of h and c which involves constants
dependent upon the nature of the metals, and which may be repre-
sented by

The discovery of the natural expression for / (h, c) is not only of
scientific importance, but is urgently needed in the development of
the art of pyroraetry. At present even a satisfactory empirical
formula for interpolation is lacking, the best still being probably that
of Avenarius and Tait.

VOL. XXXI. {y. s. xxiiT.) 13

194 PROCEEDINGS OF THE AMERICAN ACADEMY.

The existing; formuloe are the five foUowincr.
Ordinary or parabolic :

^'^e = at + ht^ + ct'^ -}- (1)

This is, of course, merely a series in ascending powers of t, where
one junction is at any temperature t° C, and the other at 0° C, a, b,
and c being constants. A more general form for the case where the
cold junction is at any constant temperature, ti° , is

2,\e = a{t-t^) + b (t- - ti") + c (t' - t{) +

These expressions may, of course, be inverted, giving t as a function
of 2e.

Avenarius

2^ e = (h — c) {a + b (h + c)}, (2)

in accordance with the foregoing notation.
Thomson :

2^ = a(r,-x.)-[r„-Ii + I-«|, (3)

where t is the absolute temperature, t„ being that of the " neutral
point."
Tait:

2^^e=(k'-k) (r,-r,)^r„-Z^Y (4)

Both of the last two, by the substitution of t + 273 for t obviously
reduce to the Avenarius form.
Barus :

e^ + e, = 10 ^+ «'' + 10 ^' + «''•• (5)

where e,, represents the thermal emf. of the hot junction and e^ that of
the cold junction. In view, however, of the existence of the Thom-
son effect, these symbols can strictly be interpreted only as having
the meaning that e^, — e<, — ^c ^-

Note. — With regard to the Avenarius, Thomson, and Tait expres-
sions it may be remarked that they are not only mutually equivalent,
but that if t^ or r^ becomes 0° C. they reduce at once to the ordinary
parabolic form of two terms :

%:=at-\-bt^.

They are all, therefore, forms which must apply if the latter purely
empirical expression for the same temperature ranges applies, and
with the same closeness, so that it is unnecessary to test more than

HOLMAN. — THERMO-ELECTRIC FORMULA. 195

one of the first four expressions against any one set of data. Also
the fact that the Avenarius and Tait equations approximately con-
form to the observed data does not necessarily in any material degree
strengthen the hypotheses which are adduced to show that these equa-
tions are a natural expression of the law.

Without attempting here a further analysis of the components
making up the resultant emf. 2'!e, which is the measured emf, of the
thermo-couple, tlie proposed interpolation formulaa will be merely
developed and applied. It may, however, be suggested in passing,
that there seems to the writer to be little hope of arriving at a close
approximation to the natural law except through an expression which
shall contain separate terms representing the temperature function of
the component arising at the contact of the dissimilar metals, and that
arising from the inequality of temperature of the ends of each (homo-
geneous) element (Thomson emf.). The parabolic and Avenarius for-
mulre would comply in part with this requirement on the supposition
that the emf. at contact varied as the first power, and the Thomson
emf. in both wires as the square of the temperature. And looked at
from that point of view, the neutral point would seem to have an
explanation materially different from that usually accorded to it.

The Proposed Formula.

Exponential Equation. — The significance of this proposed expres-
sion may be thus stated. Suppose the cold junction of the couple be
maintained at the absolute zero of temperature, r = 0°, and its emf.
to be consequently zero. Let the other (hot) junction be at any tem-
perature t\ absolute. The proposed equation is based on the assump-
tion that the total emf. of the couple would then be representable by

e' ^=z m r".

where m and n are numerical constants. If then the cold junction
were raised to any temperature t° there would be introduced an
opposing emf. e" , which would be expressible by

The resultant emf. S^e would then be e' — e", and therefore expressible

by

2* e = m T^ — m T^. (6)

If in any instance, as is frequently the case in measurements, the tem-
perature of the cold junction is maintained constant while that of the

196 PROCEEDINGS OF THE AMERICAN ACADEMY.

hot junction varies, then w t" becomes a constant, and it will be
convenient to denote this constant by /3 when t = 273° abs. =: 0° C. So
that for this special case where the cold junction is at 0° C. and the hot
junction at f C, we have

2^e=:mT"-/3. (7)

This expression is not advanced as a possible natural form of the
function y(y^,c). It is essentially empirical, and is not designed to
account separately for the several distinct components entering into 2 e.
The fact that it closely fits the experimental data arises chiefly from
the well known adaptability of the exponential equation to represent
limited portions of curved lines. The equation also leads to certain
inferences which appear inconsistent with the known thermo-electric
laws, and fails to explain some known phenomena.

The evaluation of the constants m, ?i, and [i is unfortunately
attended by considerable labor. No application of the method of
least squares readily presents itself, but by a method of successive
approximations the values can be obtained with any desired degree of
exactitude. Only two measured pairs of values of 2j e and t are
necessary for this approximation method, the third required pair
being furnished by SqC = 0 and < = 0; although, of course, by the
employment of three pairs of values well distributed in the data, a
more closely fitting equation might frequently be obtained. The
calibration of a thermo-couple for pyrometric work can thus be
affected by the employment of but two known temperatures, and this,
on account of the uncertainty of our knowledge of high melting points,
is of great importance in high temjierature work.

Let ^0 = 0^ C, <', and t" be the selected observed temperatures
from which to compute the constant, so that tq = 273°, t' —t' -\- 273°,
t" = t" 4- 273° abs. And let Sj" e = 0, 2o' e, 2u" e, be the correspond-
ing observed emfs. of the couple. Then, by substituting these in equa-
tion (7), and combining the three expressions, or their logarithms,
we easily deduce

^-//y_/ (8)

_ log (2^" e + ^) - log (2;' g + jg) . (^9)

n =

log T — log

HOLM AN. — THERMO-ELECTRIC FORMULA. 197

By means of these the uumerical values of the constants may be
calculated from those of /, /', I'o e, etc., as follows : —

1. Assume as a first approximation some value of n, say n =1,
unless some better approximation is in some way suggested. Substi-
tuting this value in (8), compute the corresponding value of ^.

2. Using this as a first approximation, substitute it in (9) and com-
pute the corresponding value of n.

3. Using this value as a second approximation to n, insert it in
(8), and compute a second approximation to j3.

4. With this compute a third approximation to n, and so continue
until consistent values of (3 and n are found to the desired number of
figures. Then compute m by (10).

The rate of convergence is not rapid, but after one or two approxi-
mations have been made an inspection of the rate will enable the
computer to estimate values of n which will be nearer than the pre-
ceding approximation, and thus hasten the computation.

"Where an equation is to be computed to best represent a progres-
sive series of observed values of t and 2 e, this method is of course
open to some objections, since it incorporates in the constants the acci-
dental errors of the selected observations from which the constants are
deduced. This difficulty can be sufficiently overcome by computing
residuals between the equation and the data, and amending the equa-
tion if necessary to give them a better distribution.

Logarithmic Formula. — A very simple expression for interpola-
tion is of the general form

2o e = m r,

where m and n are constants. This serves fairly well for a short
range, t" — t\ when t' — 0° is not less than one third of «" — t'.

The convenience of the expression arises from two facts : first, that
its two constants are very easily evaluated either by computation or
graphically from the logarithmic expression (whence the name)

log So e = n log t + log m ;

second, that its logarithmic plot is a straight line, since this expression
is the equation to a straight line if we regard log Sq e and log t as the
variables. If, therefore, a series of values of 2 e and t are known for
a given couple, points obtained by plotting log t as abscissas and log
2 e as ordinates should lie along a straight line. Thus a couple may
be completely " calibrated " for all temperatures by measuring 2 e and
t for any two values of t (suitably disposed). The constants m and n

198 PROCEEDINGS OF THE AMERICAN ACADEMY.

may be computed, or a plot of log 2 e and log t may be made, and a
straight line be drawn through them. Graphical interpolation on this
line will then of course yield the values of lug t and hence of t corre-
sponding to observed values of 2 e, and vice versa, and, if desired, the
constants m and ?i. The expression for ^ as a function of 2 e is, of
course,

t = n/(%er, or t= (^)»-

This formula is well adapted to pyrometric work not of the very
highest grade of accuracy, and has been advantageously emplo^-ed in
connection with the Le Chatelier thermo-electric pyrometer in a
method to be described in a later article.

Test of Formulae.

This will be made by applying the several formula3 to the experi-
mental data of Barus, Holborn and Wien, Chassagny and Abraham,
and Noll. These investigators employed modern methods of ther-
mometry and of electrical measurement. Temperatures are either
made in or reduced to the scale of the hydrogen (C. & A.), or of the
air thermometer (B., H. & W., N.). Constants for the formulae will
be deduced, and the residuals or deviations of the data from the equa-
tions (i. e. 8 = data-equation) will be computed for the observed
Doints. For discussion these deviations will be expressed in per-
centages, viz. 100 ^, rather than in microvolts or degrees. This is
preferable because the process of measurement of the emf., and to some
extent at least of the temperature, is such as to yield results of a
nearly constant fractional or percentage precision at all temperatures
rather than of a constant number of microvolts or degrees. Thus by
comparing percentages we eliminate a complication arising otherwise
from the increasing value of 5 as < increases. Incidentally there are
also other well recognized advantages frequently attending the com-
parison of percentages rather than of absolute quantities.

The Barus Data. — Taking the data in the order of priority, those
of Barus will be first employed. The measurements to be used con-
sist of very elaborate and painstaking direct comparisons of several 20
per cent irido-platinum thermo-couples with several porcelain bulb air
thermometers used under the constant pressure method.

Quotations of, or rather interpolations in, his original data* are

* Barus, C. U. S. Geol. Surv. Bull., No. 54 (1889). Phil. Mag., XXXIV. 1
(1892).

HOLMAN. THERMO-ELECTRIC FORMULA.

199

given by Barus * later, as a basis from which to deduce constants for
his jii'oposed equation

Barus's numerical values for the constants are :

eg = 45680 microvolts.
P = 4.6515
F' = 2.849 "

Q = 1.106 • 10-*.
Q' = 3.01 • 10-3.

These constitute his "equation 3," for which e^ corresponds to
20° C. The data and the deviations which I have computed for it, viz.
8 = data-equation, are given in Table I. The last column gives the de-
viations expressed in percentages, viz. 100 -, where E—e + e^-}- 1880.
This value of jE is adopted to make the percentages comparable with
those in subsequent discussions. The number 1880 is 1730 -f- 150,
which are the. values of e^ and Sq" e of the next two pages.

TABLE I.
Barus's American Jouenal of Science Data.

f«

e -J- f 0 ^'^■
Observed.

e -|- Co Computed
from " Equ. 3."

s

mv.

ioo|

Per Cent.

°c.

0

-150

100

+680

653

+27

+1.11

200

1650

1657

-7

-0.20

300

2760

2788

-28

-0.60

400

3950

3994

-44

-0.80

600

6560

6551

+9

+0.11

800

9310

9273

+37

+0.34

1000

12200

12140

+60

+0.43

The lines ^5 and CD on the diagram (page 212), constructed with
percentage deviations as ordinates and temperatures as abscissas show
clearly that the deviations are systematic. Upon inspection of this plot

* Amer. Jour. Sci., XLVIII. 332 (1894). See also XL VII. 366 (1894).

200 PROCEEDINGS OF THE AMERICAN ACADEMY.

it appears that the data may be separated into two groups, one includ-
ing 0°-300° ; the other 400°-1000°, which appear to have entirely
distinct forms of systematic error. This division corresponds to two
distinct groups of data, one extending from 0° to 300°, the other
including the second group and extending from 350° to 1075°. The
latter were given in the Bulletin as the final results of the high tem-
perature comparisons of the irido-platinum couple with porcelain bulb
air thermometers. The detailed statement of the 0°— 300° comparison
I have not seen. Although the discrepancy between the two sets of
systematic deviations is not extremely large, yet it has seemed to me
that it was beyond the limits of concordance in the higher tempera-
ture work, and that it would be better for the present purpose to deal
solely with the 350°-1075° data. Two points regarding Barus's work
should be noticed : one the strikingly high degree of concordance be-
tween individual observations even with different thermometer bulbs
and different thermo-couples ; the other the remark in which Barus
notes a possibility of being able still further to reduce the "stem
error " entering into the result, which so far as I am aware has not
yet been done.

The high temperature air thermometer comparisons (Bulletin,
Series I., II., III., IV., and V.) of Barus are so numerous (108) and
80 distributed that the labor of utilizing them simply for deducing con-
stants and testing an equation would be excessive. Also they are too
concordant to permit interpolation on a direct plot without a sacrifice
of some of their precision. For the purposes of discussion, therefore,
I averaged them in nine groups. The first group contained all where
the emf, lay between 3,000 and 4,000 microvolts ; the second group
between 4,000 and 5,000 mv. ; and so on by steps of 1,000 microvolts,
except that the seventh group covered 2,000 mv. from 9,000 to 11,000.
These groups were not exactly equal in number of observations, and
therefore in weight, nor is the arithmetical average a strictly legiti-
mate value where the function is not linear ; but, as easily seen by
inspection of the originals, the errors thus introduced are negligible.
In Table 11., columns one and two give the direct values of the
averages. Column three reduces 22oC to SJe by adding 150 micro-
volts the value of 2^e being elsewhere given by Barus as — 150
microvolts.

HOLMAN.

THERMO-ELECTRIC FORMULA.

201

TABLE II.
Barus's Air Thermometer Comparisons, Series I.-V.

t"

2,^e

4e

Avenarius.

Exponential.

Logarithmic.

5
Data Eq.

Per Cent

= 100^

S
Data Eq.

Per Cent

S
Data Eq.

Per Cent

= 100^
e

°c.

mv.

mv.

mv.

mv.

mv.

0.0

(-150)

0

0

0.0

+23

+1.3

378.5

3679

3829

-84

-4.0

-66

-1.2

-33

-0.60

440.3

4508

4658

+18

+0.28

+18

+028

+30

+0.47

522.0

5486

5636

-4

-0.07

-25

-0.34

-36

-0.50

588.4

6404

6554

+70

+1.1

+33

+0.40

+9

+0.12

672.1

7550

7700

+110

+1.5

+60

+0.64

+26

+0.28

7456

8530

8680

+82

+0.92

+26

+0.25

-9

-0.10

840.1

9898

10048

+101

+1.0

+49

+0.41

+26

+0.22

946.6

11306

11546

+9

+0.07

-19

-0.14

-13

-0.10

1010.7

12475

12625

-45 '

-0.32

-45

-0.32

-10

-0.07

A verag

e percent

age devif

itions .

. 0.93

0.53

The Avenarius equation applied to these data yields
2^6 = 9.104 < + 3.249-10-='^2_

Microvolts. Eange 350° to 1075° C.

Computing from this equation values of SJe for the successive
values of t in column one, and subtracting them from the data in
column three gives the deviations between data and equation. These
are expressed in microvolts in column four, and in percentages in
column five, the percentage being reckoned in terms of Cf as deduced
by the exponential formula. Objections may be felt to this use of e^
(here as throughout the subsequent tables) as a basis, since e^ involves
Co, which is an extrapolated value, certainly not exact, and possibly
wide of the truth. Such a criticism is valid, but inasmuch as the values
of ^0 employed are nearly equal, and as the percentage deviations are
used merely for purposes of expressing relative accuracy, the possible

202 PROCEEDINGS OP THE AMERICAN ACADEMY.

error involved is nearly auuulled. Hence, although it would be better
to compute 8 t, and express this as a percentage of the absolute tem-
perature T, the added labor did not seem justified by the small gain.
The exponential equation applied to the Barus data yields

2;;e =: 0.7G91t^-^''^— 1730, or
e^ = 0.7691 T^-'^^, and yS = 1730 mv.
Range of data 350° to 1075° C.

[N.B. This equation was deduced with the value 0°C. == 273° .7 abso-
lute, whereas in all subsequent tables 0°C. = 273°. 0 absolute is em-
ployed as a more probable value. The numerical values of the
constants are therefore subject to a slight modification, but as for the
present purpose we are concerned only with 8, which would not be
sensibly changed, the recomputation is not worth while.]

Columns six and seven, Table II., give 8 and its percentage value
for the exponential equation.

llie Barus Equation, — • The excessive labor involved in the evalua-
tion of the constants P, Q^ P', and Q' of Barus's proposed equation
detracts so seriously from its usefulness that I have also allowed it to
deter me from computing them for the above tabulated values. The
comparison of tlie values of 8 for his " Klquation 3," and for an ap-
proximate exponential of my ow'n based on the same data, is, however,
decidedly in favor of the latter.

The logarithmic equation applied to the Barus data yields

2^6 = 2.665^1220^

or its equivalent, ^

log 2^6 = 1.220 log t + 0.4-2570.

The deviations are given in the last two columns of Table II.

Holborn and Wien Data. — This important comparison * of the
rhodo-platinum thermo-couple with the porcelain' bulb air thermometer
up to high temperatures was performed under the auspices of the
Reichsanstalt at Berlin, and appears to be on the whole the most im-
portant and reliable contribution to this subject in recent years. The
experimental work was evidently conducted with great care, and
although not showing the concordance of results, nor the multiplica-
tion of observations of Barus's work, yet in respect to stem-exposure
correction, to exposure of the thermal junction, and to direct measure-
ment of the coefficient of expansion of the bulb, it is probably more

* Holborn and Wien. Zeit. f. Iiistk., XII. 257, 296 (1892). Also in full in
Wied. Ann., XLVII. 107 (1892).

HOLMAN. — THERMO-ELECTRIC FORMULAE. 203

free from systematic error. It is to be regretted that the results were
not more thoroughly discussed, and that neither a chemical analysis
nor even a statement was given to indicate the reliability of the stated
percentage composition of the various alloys used. For when closely
examined, the data seem to indicate a definite relation between the
composition and the emf., as was shown by a relation discovered be-
tween the constants in my exponential equations for the various alloys.
The deviations were only such as might be attributed to uncertainty of
composition, but as no measure of the latter was given, a statement
of the relations and interesting inferences from them is not warranted.
It is also unfortunate that an analysis, or at least a definite state-
ment of the percentage purity, was not given for the gold, copper, and
silver whose melting points were observed. The assertion that the
gold showed on qualitative analysis only a trace (" Spur ") of copper,
and the silver a " trace " of iron, is hardly definite. The value of the
whole work would have been enhanced by these additions far more
than in proportion to the comparatively small labor demanded by
them, and such completeness is naturally to be expected in work
emanating from this source. It is to be hoped that a continuation of
this research is in progress, and that additional high melting points
may be measured.

Table III., columns one and two, quotes the interpolated mean
values of several comparisons expressed in international microvolts
and degrees centigrade. With regard to these data it should be stated
that below about 400° they were not supposed to be of as high ac-
curacy as above that point. Also, that, owing to unavoidable cir-
cumstances, the data below 300° were obtained with only a single
air thermometer bulb, and similarly those above about 1300° with
one bulb only, but a different one, while the data intermediate be-
tween 400° and 1300° are the mean of observations with the two
bulbs. This fact may partially account for the erratic character of
the residuals above 1300°, where the deviations are so great and so
distributed (see diagram, page 212) as to render these observations of
very little service. Direct comparison with the air thermometer was
made with one 10 per cent rhodo-platinum couple "A" only.

The parabolic formula applied to these by Holborn and Wien, when
corrected as to decimal points,* is

* The equation at both references, and stated to be in microvolts and
degrees, is erroneously printed as

'■t-f{e) = 1.3.76 e - 0.004841 e2 + 0.000001378 eV

204

PROCEEDINGS OF THE AMERICAN ACADEMY.

^:=: 1.376-10-1 (2l>e) -4.841-10-« (^'oef + 1.37 8-10 -'\2'oey.
Range -80° C. to +1445° C.

The residuals are given in Table III., columns three and four.

TABLE III.

HOLBORN AND WiEN. AlR THERMOMETER COMPARISONS, AlLOV A.

t

2,«e

H. and W. Eq.

Avenarius.

Exponential.

Logarithmic. 1

5

6

S

5

Centigr.

-qC

= Data

Per Cent

r= Data

Per Cent

= Data

Per Cent

= Data

Per Cent

-Eq.

= 100^-

-Eq.

= 100^

-Eq.

= 100-

— Eq.

-100^

mv.

e

mv.

e

mv.

e

mv.

e

-80

-361

0

0

0

0

0

0

0

0

0

0

+82

+500

-84

-4.6

-107

-5.1

-69

-3.6

+40

+2.20

154

1000

-147

-6.1

-166

-7.2

-122

-4.8

+11

+0.50

220

1500

-135

-5.0

-199

-7.1

-140

-4.7

-27

-1.00

273

2000

—150

-4.5

^142

-4.3

-85

-2.6

+16

+0.50

329

2500

-130

-3.6

-124

-3.3

-73

-1.9

+11

+0.30

379

3000

-60

-1.4

-66

-1.5

-24

-0.57

+45

+1.00

431

3500

-30

-0 60

-41

-0.90

—6

-0.12

+45

+0.90

482

4000

0

0.00

-14

-0.22

+8

+0.15

+41

+0.80

533

4500

+ 10

-fO.17

-1

0

+9

+0.15

+27

+0.50

584

5000

0

0.00

0

0

-2

-0.03

+1

+0.01

633

5500

0

0.00

+9

+0.13

-4

-0.06

-13

-0.20

680

6000

+10

+0 11

+28

+0.38

+5

+0.07

-15

-0.20

725

6500

+30

+0.40

+58

+0.74

+26

+0.33

-1

-0.01

774

7000

-10

-0.12

+35

+0.42

-5

-0.06

-39

-0.47

816

7500

+20

+0.22

+78

+0.90

+33

+0 37

-8

-0.10

862

8000

0

0.00

+69

+0.74

+19

+0 20

-24

-0.26

906

8500

-20

-0.20

+72

+0.74

+ 19

+0.20

-25

-0.25

952

9000

-55

—0.55

+44

+0.43

-.10

-0.10

-53

-0.52

996

9500

-88

-0.80

+29

+0.26

-24

-0.22

-67

-0.60

1038

10000

-88

-0.80

+29

+0.25

-20

-0.18

-57

-0.50

1080

10500

-100

-0.85

+22

+0.20

-23

-0.20

-54

-0.43

1120

11000

-100

-0.80

+31

+0.25

—7

-0.06

-22

-0.19

1163

11500

-140

-1.10

-6

-0.05

-33

-0.26

-51

-0.40

1200

12000

-96

-0.74

+26

+0.20

+9

+0.07

+2

+0.02

1211

12500

-96

-0.70

0

0

-3

-0.03

-8

-0.06

1273

13000

0

0.00

+84

+0.60

+94

+0.67

+111

+0.80

1311

13500

+36

+0.24

+84

+0.57

+110

+0.80

+140

+0.90

1354

14000

+24

-0.17

+10

+0.07

+.58

+0.38

+107

+0.70

1402

14500

-60

-0.40

-141

-0.90

-65

-0.41

+ 2

+0.01

1445

15000

-72

-0.45

-231

-1.40

-128

-0.80

-38

-0.23

ad. for

Otol

445.

1.12

1.15

0.77

1.46

ad. for

431 to 1

445.

0.39

0.43

0.25

0.38

ad. for

431 to ]

241.

0.43

0.36

0.15

0.34

The Avenarius Formula applied to the Holborn and Wien data
with constants deduced from t = 584° and 1273° becomes

2'oe= {t — to) {7.2188 + 0.00 229 94 {t + ^o)}, or

= 7.2188^ + 2.2994-10-^ <l

Range 0° to 1445° C.

HOLMAN. — THERMO-ELECTRIC FORMULA.

205

The deviations in microvolts and percentages from this equation
are given in columns five and six, Table III.

Tlie exponential equation fitting these data most closely, and coin-
ciding with them as nearly the same points as the others, viz., at
about 584° and 1250°, is

%e = 0.57 674 t^""'' ~ 1310, or
e = 0.57G74Ti"^y8= 1310.
Range 0° to 1445° C.

The deviations in microvolts and percentages are in columns seven
and eight, Table III.

The logarithmic equation applied to the Holborn and Wien data
on A yields

2,',e=: 2.1682 ^i-^i^!, or

log %e = 1.2156 log t + 0.36 610.

The deviations are given in the last two columns of Table III.

Holborn and Wien not only compared the ten per cent rhodo-plati-
num couple A directly with the air thermometer, but compared
with A seven other couples in which one element was platinum,
and the other a rhodo-platinum alloy, the percentage of rhodinum
being stated respectively as, for Ci and C2, 10 per cent (these two I
have combined under C), D, 9 per cent, E, 11 per cent, F, 20 per cent,
G, 30 per cent, H, 40 per cent. For the present purpose I have com-
bined these data, which were differences of emf. between C and A, D

TABLE IV.

Designation of
the Alloy.

Nominal Per-
centage of
Khodium.

Expon. Eq. Constants.

m

n

mv.

D
C
E
A
F
G
H

9
10
11
10
20
30
40

1.36 71
0.95 596
0.81 734
0.57 689
0.22 865
0 06 599 0
0.06 303 4

1.250
1.310
1..336
1.377
1.522
1.708
1.720

1517
1485
1469
1305
1167
956
977

206

PROCEEDINGS OF THE AMERICAN ACADEMY.

and A, etc., with the corresponding emf. of A, and thence have de-
duced the exponential equations for each of the alloys. Table V.
gives the percentage deviations of these alloys from the exponential
equation (data — equation), and Table IV. shows the values of m, n,
and yS for those equations.

TABLE V.

HOLBORN AND WiEN. — COMPARISON OF AlLOTS.

t

D

A

C

E

F

G

H

Average.

154

—

-4.8

—

-4.0

—

-5.0

-4.4

-4.5

273

-0.9

-2.6

-2.7

-2.2

-2.3

-3.5

-2.3

-2.6

379

+2.1

-0.57

-0.80

-0.^0

-M.3

-0.60

-0.25

-0.67

482

+2.6

+0.15

-0.03

0.

-0.60

+0.22

+0.46

-0.03

584

+0.9

-0.03

+0.19

+0.08

+0.13

+0.30

-0.03

+0.02

680

-0.30

+0.07

+0.13

+0.13

-hO.12

-0.10

-0 26

0.

774

-0.19

-0.06

-0.08

-0.02

-0.12

-0.42

-0.46

-0.19

862

+0.23

+0.20

+0.20

+0.20

+0.09

+0.15

-0.11

+0.12

952

+0.20

-0.10

0.

-0.11

-0.15

+0.09

-0.24

-0.08

1038

+0.12

-0.18

-0.10

-0.21

-0.27

+0.04

-0.14

-0.16

1120

-0.15

-0.06

-0.09

-0.02

+0.10

+0 01

+0.22

-0.01

1200

+0.02

+0.07

0.

+0.10

0.

+0.04

0.

+0.03

1273

—

+0.67

+0.70

+0.80

+0.80

+0.43

+1.00

+0.73

1354

—

+0.38

+0.60

+0.67

+0.51

+0.30

+1.00

+0.58

1445

—

-0 80

-0.38

-0.50

—

-0.90

-0.40

-0.60

a

d. 400-1200

0.52

0.15

0.09

0.10

0.12

0.16

0.21

0.07

a

d. 400-1445

0.25

Direct from
Air Th.

0.21

0.24

(0.24)

0.23

0.36

Chassngny and Ahraham Data* — The apparently very careful
measurements of these observers cover a range of 0° to 100° C. with

* Chassagny et Abraham. Ann. de Chitn. et de Phjs., XXVII. 355 (1892).

HOLMAN. — THERMO-ELECTRIC FORMULA.

207

observations at 25°, 50°, aud 75° only. The range is too short, and
the intervals are too great to render the work of much service in
testing a genei-al formula, but if its accuracy is as high as about
0°.01, as it appears to be, this in j^art offsets the disadvantage.
Measurements of 2o e and t were made with four thermo-couples
with the results shown in Table VI. (international microvolts and
degrees centigrade on hydrogen scale).

TABLE VL

Couple.

sTe

2'o'e

STi'e

Fe-Cu

1093.2

864.9

604.8

315 5

Fe-Pt Rh

895.1

708.9

496.1

259.1

Fe-Ag

1123.0

885.6

617.4

821.1

Fe-Pt

1685.1

1278.9

859.9

432.1

T/ie Avenarius equation was applied to these data by Chassagny and
Abraham in the form

%e — at ■\- ht'.

They evaluated the constants from the 50° and 100° data. With
these they computed the temperatures which the equation would
yield by insertion of the observed values "^e and 2o^e. These
values are given in Table VII., columns two and three.

The exponential equation applied to these data for Fe - Pt be-
comes

2f,e= 105.096 f

6525.3 [Range 0° to 100° C.].

The values of t corresponding to the observed values Sf e and s'lf e are
given in Table VII. It has not seemed for the present purpose worth
while to make similar computations for the other couples, as they would
not materially affect the inferences to be drawn.
The logarithmic equation yields

2^6= 19.2946 «*'^^^;
log %e = 0.97 059 5 log ^ + 1.28 543 6.

The deviations are given in the table.

208

PROCEEDINGS OF THE AMERICAN ACADEMY.

TABLE VII.

Avenarius.

Exponential.

Logarithmic

Couple.

t

&t

t

ht

&
my.

100 a,v

Per Cent.

it

mv.

Fe-Cu

24.°88

+0?12

o

0

o

Fe-Pt Rh

24.885

0.115

Fe-Ag

24 87

0.13

Fe-Pt

24.87

0.13

24.80

+0.20

-2.6

-0.037

+0.52

-6.7

Fe-Cu

75.13

-013

Fe-Pt Rh

75.135

0.135

Fe-Ag

75.135

0135

Fe-Pt

75.135

0.135

75.15

-0.15

+2.5

+0.032

-0.26

+4.3

TTie Noll Data. — A contribution of much permanent value to the
data on thermo-electrics has recently been made by Noll,* who has
measured Sq e and t for thirty-two couples over a range in most cases
of 0° C to 218° C. The metals employed (including carbon) were
usually of a high and stated degree of purity, and consisted of eigh-
teen different substances, two of which were alloys (German silver
and brass) , and the remainder samples of different degrees of purity
or hardness of the pure substances. The couples contained, as one
element, for the most part, either copper or mercury. Temperatures
were reduced to the air thermometer scale.

The Avenarius formula was applied to fourteen of the more impor-
tant of them by Noll. The deviations are given in Table VIII.

The exponential equation I have applied to the same data. It has
not seemed essential to reproduce here the entire series of data, and
the deviations of both equations. They are, therefore, presented in a
somewhat more digested form. Table VIII. gives the constants for the
exponential equation (those for the Avenarius may be found in Noll's
article), the mean deviations (= data -^ equation) for each series, and
the mean percentage deviations (= 100 8/e). (See remark as to use
of e under " Barus Data.") Table IX. groups the percentage devia-
tions under their nearest values of t for exhibiting their systematic

* Noll, Wied. Ann., LIII. 874 (1894).

HOLMAN.

THERMO-ELECTRIC FORMULA.

209

character. The fact that the experimental method brought the obser-
vations all very nearly to the respective temperatures t given in the
table renders this grouping possible. I have taken the liberty of
correcting a few obvious numerical errors, and of dropping a very
few values evidently containing a mistake.

TABLE VIII.
Noll's Data on Pure Metals.

3

Av fct. Deviation.

Couple.

m

n

mv.

Aveuarius.

Expon.

Au-ITg

4.6954-10-3

2.136

750.4

±0.27

±0.17

Ag-Hg

2.8637-10-3

2.206

677.8

0.33

0.15

Ni-Cu

8.2333-10-1

1.511

3950.2

0.30

0.17

(Ccl-Cu

3.7617-10-11

4 94

40.7

0.48

3.40)

Br-Cu

2.4969-10-1

1.366

531.1

0.14

0.13

, Zn-Hg

8.2890- 10-4

2.420

651.6

0.15

0.18

Pb-Cu

1.7674-10-^

1.800

429.0

0.05

0.07

Cui-Hg

4.6726-10-3

2.130

768.4

0.12

0.11

[Fe-Hg

1.0913 10 + 2

0.7220

6264.2]

Co-Hg

8.3295-10-3

2.166

1575.2

0.26

0.22

Pti-Cu

2.1475-10-3

2266

711.1

0.08

012

Pto-Cu

1.1095-10-3

2.353

599.0

0.19

0.25

Sni-Ca

4.2021-10-2

1.667

482.8

0.21

0.09

Mg-Cu

2.0449-10-2

1.782

448.7

0.15

0.17

Al-Cu

7.5643 10-2

1.590

565.3

Oil

0.12

G.s.-Cu

2.0454-10-1

1.684

2589.9

0.06

0.05

Average

omitting Cd-Cu ar

d Fe-Hg

±0.17

±0.14

It may, perhaps, not be out of place here to caution those who
would make use of Noll's data to their full accuracy that his original,
and not his interpolated, numbers should be resorted to. The approxi-

VOL. XXXI. (n. S. XXIII.)

14

210

PROCEEDINGS OF THE AMERICAN ACADEMY.

mate linear interpolation which he has employed is not as accurate as
his experimental data demand.

The logarithmic equation applied to the Cu-Hg couple as typical of
the Noll data yields

2^6 = 2.57 434^12250.
or log 2j,e = 1.2250 log t + 0.41 066 5.

The residuals to this expression are given in Table IX.

TABLE IX.
AvENARius Equation. — Data minds Equation in Per Cent.

15=

57=

100=

138=

181=

198=

217'

Au-Hg

-0.10

-0.40

0

+0.05

+0.60

+0.90

0.

Ag-Hg

-0.13

+0.20

0

+0.21

-0.50

-1.30

0.

Ni-Cu

-0.26

-0.15

0

0.

-0.22

-0.31

-1.17

Br-Cu

—

-0.10

0

+0.20

+0.17

+0.35

0.

Zn-Hg

—

-0.03

0

+0.09

+0.41

+0.40

0.

Pb-Cii

—

-0.03

0

-0.11

. +0.12

-0.04

0.

Cui-Hg

+0.02

-0.16

0

+0.01

+0.15

+0.38

0.

Co-Hg

—

—

0

+0.42

0.

—

-0.50

Pti-Cu

—

-0.12

0

-0.06

+0.09

+0.21

0.

Pto-Cu

—

-0.13

0

' -0.11

0.

+0.07

-0.21

Sni-Cu

—

-0.20

0

0.

+0.34

+0.40

+0.30

Mg-Cu

—

+0.11

0-

+0.22

-0.37

+0.11

0.

Al-Cu

—

+0.16

0

+0.18

-0.18

+0.15

0.

G. s.-Cu

—

-0.01

0

+0.16

+0.11

—

0.

Average

-0.12

-0.07

0

+0.09

+0.05

+0.16

-0.12

HOLMAN. — THERMO-ELECTRIC FORMULA.

211

TABLE IX. — Continued.
Exponential Equation. — Data minus Equation in Per Cent.

15°

57=

100°

138=

181°

198°

217°

Au-Hg ,

Ag-IIg

Ni-Cu

Br-Cu

Zn-Hg

Pb-Cu

Cui-Hg

Co-Hg

Pti-Cu

Pt2-Cu

Suj-Cu

Mg-Cu

Al-Cu

G.s.-Cu

+0.01
-0.18
-0.26

-0.19

-0.12
+0.22
-0.16
-0.06
-0.80
-0.10
-0.27
-0.05
-0.15
-0.56
-0.04
+0.18
+0.16
+0.07

-0.06

0.

0.
-0.01
+0.02

0.
+0.01
+0.01
+0.03

0.
+0.04

0.

0.
-0.01

-0.06
+0.05
+0.07
+0.12
+0.14
-0.25

+0.48
-0.06
-0.10
-0.29
+0.15
+0.12
+0.08

+0.09
-0.03
0.00
+0.20
+0.10
+0.04
-0.03
-0.01
+0.17
-0.16
+0.12
-0.52
-0.24

+0.12

0.

0.
-0.26
+0.65

0.

0.

0.

+0.70
+0.06
+0.17
+0.18

-0.80
-0.50
-0.72
-0.13
-0.04

0.00
-0.25
-1.00
-0.31

0.

0.

0.

0.
+0.03

Average

-0.16

+0 08

+0.01

+0.03

-0.02

+0.14

-0.27

Logarithmic Equation. — Data minus Equation in Per Cent.

Cui-Hg

+2.3

+1.5

0.

-0.40

0.

+0.50

+0.80

Discussion of the Deviations.

Plots are given in the following diagram with temperatures as
abscissas and percentage deviations between the data and the sundry-
equations as ordinates, i. e. 100 8/e where 8 = data-equation. Inspec-
tion will show that with one exception (viz. the logarithmic equation
applied to the Barus data) these plots, whether the equation is the
ordinary parabolic, the Avenarius, the Barus, the exponential, or the
logarithmic, have the same general form, which may be imperfectly

212

PROCEEDINGS OP THE AMERICAN ACADEMY.

dsecribed as follows. If the equation be made to conform to the data
at 0°C. and at two higher points, a and b, then the deviation will be of
the negative sign from 0 to a, positive from a to b, and negative above b.

B^fuS.Au.J Set.

B*fWS.

V /OQ0° Aveit.

s^^-J-.G ' exPO/r.

OfiOINAT£S or ALL PlOTS
AR£ PERCSNTAGe D£y/ArtO/fS.

HolboRh &Wf£tf.

H&W

A V£N — —

£XPOt*

LOGAP

V . H- I

CffAssAony ^Abhaham.
AvefJ

CXPON

\ /

/

The slight departures from this general form are clearly due either to
accidental errors, or to failure to make the equation conform to the
data at all three points, or at suitable ones. The evidence is there-
fore conclusive that /br all of the expressions the deviations are sys-
tematic and not purely ^'■accidental" in character.

HOLMAN. — THERMO-ELECTRIC FORMULiB. 213

One of two inferences is therefore warranted : —

1. That neither the parabolic, Avenarius, Barus, exponential, nor
logarithmic equation is the natural expression of the function.

-2. Or that the scale of temperature to which the vaUies of t are
referred in the foregoing investigations departs from the normal scale
by an amount and system roughly indicated by the above residual
plots.

The latter inference, suggested by Chassagny and Abraham in the
interpretation of their results, does not seem to possess much weight,
notwithstanding the urgent need of renewed elaborate experimental
investigation of the relation between the hydrogen, air, and thermo-
dynamic scales of temperature.

As to the relative usefulness of the various expressions for purposes
of interpolation and extrapolation some further inspection is necessary.
The Barus equation 3, line OD, shows slightly smaller deviations on
the plot than do the Avenarius and exponential, lines EE and FF.
This, however, is due to the fact that the data against which 3 is
tested are mean interpolated values, and hence have a sensibly
less variable error than those against which the other equations are
tested. An approximate exponential equation showed less deviations
than 3 against the same data. There seems, therefore, to be no
advantage in this equation sufficient to oflfset the difficulty of evalua-
tion of its constants.

Applied to the Barus data from 350° to 1250°, the exponential
equation shows deviations considerably less than one half as great as
those of the Avenarius, while those of the logarithmic equation are so
small as to lie far within the range of the variable errors, and they
moreover show no clear evidence of systematic error between these
limits of temperature. For interpolation in the Barus data, therefore^
the logarithmic equation is far preferable, and must be conceded to be
representative of the data. For extrapolation it is undoubtedly better
than the Avenarius, which (as would the exponential in less degree)
would certainly give above 1000° extrapolated values of 2 e too
large, or of t too small. The advantage due to its simplicity is also
to be noted.

Applied to the Holborn and Wien data from 400° to 1450° the
exponential equation shows (line KK^ the same sort of superiority to
both logarithmic (line LL) and Avenarius (line IT) that the logarith-
mic shows to the others with the Barus data, but in a still more
marked degree. Within the limits 450° to 1450°, in fact, the dis-
tribution of the residuals to the exponential is such as not to warrant

214 PROCEEDINGS OF THE AMERICAN ACADEMY.

of itself alone any inference of systematic departure, especially when
the mean line MM from all the couples is considered. It will be
noted as an important confirmation of both the exactness of the
electrical measurements in the investigation and the applicability of
the exponential formula through a considerable range of alloys (and
therefore of values of m and n) that this mean line 3IM is almost
identical in form with the line KK iov alloy A. Relatively to the Hol-
born and Wien formula (line HH), the exponential possesses a simi-
lar advantage, with also the merit of greater simplicity of form.

It may therefore be affirmed that for interpolation between 450**
and 1450° in the H. and W. data the exponential equation is abun-
dantly exact. For extrapolation above 1450° it would not be entirely
safe, although presumably better than the others, since the departure
between 0° and 450°, and the similarity of the form to others, makes
a systematic departure sufficiently certain.

Applied to the Chassagny and Abraham data, 0°-100°, and to
the Noll data, 0°-218°, (see diagram.) the Avenarius and exponen-
tial formulas show about equal deviations, but with the advantage
slightly on the side of the former. In the case of the Noll data, the
line indicates that the systematic error is slightly greater for the
exponential than for the Avenarius expression. The average devia-
tions in Table IX., on the contrary, show that for each individual
equation the concordance is greater for the exponential than the
Avenarius. This discrepancy is due to the fact that, in order to
eliminate local accidental errors, the equations (both Avenarius and
exponential) are not all made to coincide with the data at the same
temperatures, so that the process of averaging by which the data for
the Noll plots is obtained is not numerically rigid. This does not,
however, sensibly affect the general form of the curve. Tlie greater
ease of computation of the numerical constants of the Avenarius
expression, and its applicability where both t and t^ change, ought not
to be overlooked. For extrapolation the exponential would be safer,
for the reason that it has been above shown that for long ranges its
systematic error is less.

The logarithmic equation fits the Noll data very badly, as shown by
the deviation in Table IX. (not plotted), and also is much less close to
the Chassagny and Abraham data than are the others.

The GENERAL CONCLUSION as to applicability, then, seems to be
that, while the Avenarius expression may be equally good or better
than the exponential ybr interpolation over short ranges, yet for inter'

HOLMAN. — THERMO-ELECTRIC FORMULAE. 215

polatlon over long ranges and for extrapolation above the observation
limits the exponential is decidedly preferable. The exponential form
is also preferable to the remaining expressions with the exception
noted.

The logarithmic form, although closely applicable to the Barus data,
is of more doubtful general value, yet on account of its great con-
venience it may find application in industrial pyrometry, as will be
elsewhere indicated. Although failing below 300° or 400°, it may
probably be applied to the irido- or rhodo-platinum couple between
400° and 1200° C. with a maximum error not exceeding about 5°.
If extended to cover 400° to 2000° the error might rise to 15° or 20°.

More in detail it may be briefly noted by way of summary : —

That the logarithmic equation fits the Barus data between 400°
and 1250° with scarcely sensible systematic error, and within the
limits of variable errors of the data.

That the exponential equation similarly fits the Holborn and Wien
data within the limits 400° to 1445°.

That when made to coinci^le with the data at about 450° and 1200°
tlie systematic deviations of the exponential equation from the Barus
data, and of the logarithmic equation from the Holborn and Wien
data, are in general of opposite sign and of roughly equal magnitude.

Barus Melting and Boiling Point Data.

From the foregoing demonstration of its applicability, it seems
proper to apply the logarithmic formula to the Barus thermo-electric
data on melting points.*

Whether the extrapolation above 1000° by the logarithmic formula
is entitled to any great weight may be questioned, but there is no
obvious reason why it is not more reliable than by any of the others.
I have employed the equation given on page 202, which repre-
sents very closely Barus's high temperature air thermometer com-
parisons, calculating thence the temperatures t corresponding to the
values of 1^ e given by Barus for the various points, assuming Barus's
value S-y'e = 150 mv. The results are given in column three of
Table X. Column four quotes the most reliable previous determina-
tions of the same points by other observers. As to which of the two
columns of results best represents Barus's work there can be little
doubt from the above evidence that below 1000° it is the second, that

* Amer. Jour. Sci., XLVIII. 332 (1894).

216

PROCEEDINGS OF THE AMERICAN ACADEMY.

is, the one computed from the logarithmic equation. These combine
both his own air thermometer and melting point work. Above
lOOC^ the logarithmic values are probably slightly too high.

TABLE X.
Barus Melting and Boiling Points.

Computed
by Eq. 3.

Computfd
hy Log. Kq,

Data by other Observers.

Mercury (B. Pt.)

o
357

o

359

o

356.76

Callendar and Griffiths.

Zinc

420

423

417.57

"

Sulphur (B. Pt.) .

446

449

444.53

"

Aluminum . . .

638

641

635

Le Ciiateher.

Selenium (B. Pt.)

694

697

Cadmium (B.Pt.)

782

782

Zinc(B. Pt.) . .

929

926

930

Deville and Troost.

Silver ....

98G

985

968
954

llulboni and Wien.
Violle.

Gold

1091

1090

1072
1035

Ilolborn and Wien.
Violle.

Copper ....

1096

1095

1082
1054

Ilolborn and Wien.
Violle.

Bismuth . . .

1435

1441

Nickel ....

1476

1485

1450

Carnelly and Williams.

Palladium . . .

1585

1597

1500

Violle.

Platinum . . .

1757

1783

1775

Violle.

Remark.

Review of the laborious researches which have been devoted to
the direct comparison of thermo-electric elements with the air ther-
mometer, mainly for the purpose of advancing the art of pyrometry,
has enforced the conviction that, at least for the immediate future,
this end would be better served by accurate gas thermometer meas-

HOLMAN. — THERMO-ELECTRIC FORMULAE. 217

urements of melting points of metals. Each such determination made
upon a reproducible metal of known high purity under proper repro-
ducible conditions fixes an enduring and reproducible reference point,
a pyrometric *' bench mark." And there are enough inexpensive
metals, together with a possible system of simple alloys, to give points
of sufficient frequency. These would then afford a convenient means
of obtaining accurately known high temperatures for purposes of
study of all high temperature phenomena, and particulai'ly for cali-
brating thermo-eleetric, electrical resistance, optical, or other second-
ary pyrometric interpolation apparatus, — for it must be remembered
that all such apparatus is necessarily secondary, the gas thermometer
being inevitably the primary.

On the other hand, comparison with the air thermometer of a
thermo-couple, or of a resistance pyrometer, or the study of any pro-
gressive thermal phenomenon, while it possibly may result in the educ-
tion of a natural law, is very unlikely to lead to anything more than
the establishment of an approximate equation with constants charac-
teristic only of the individual materials actually employed, and not
transferable to other, although similar materials. Such results are
obviously of a much more ephemeral character than the melting point
measurements. Even when any pyrometer thus tested is applied to
the establishment of melting points, it must at best yield results in-
ferior to direct application of the gas thermometer, except in cases
where the latter is hampered by want of sufficient quantity of the
metal to be experimented upon, — a condition which need only affect
such costly substances as gold and platinum.

Stated broadly, the great need of the art of pyrometry is convenient
methods of producing, or of recognizing vphen produced, a series of
accurately known high temperatures. The analogous problem has
been partially solved for thermometry at temperatures up to 300° C.
by the investigation of boiling points of certain chemically pure sub-
stances under controlled pressures.

Rogers Laboratory op Physics,
Massachusetts Institute of Technology,
Boston, September, 1895.

218 PROCEEDINGS OP THE AMERICAN ACADEMY.

Investigatio>s on Light and Heat, made and pdblished wholly or in part with
Appbopriation from the Rujjford Fund.

XI.

CONTRIBUTIONS FROM THE PHYSICAL LABORATORY OF THE
MASSACHUSETTS INSTITUTE UF TECHNOLOGY.

XLV.— MELTING POINTS OF ALUMINUM, SILVER,
GOLD, COPPER, AND PLATINUM.

By S. W. Holmax, with R. K. Lawrence and L. Barr,

Presented November 13, 1895.

The following melting points are offered as provisional only, but
with the belief tliat they are more reliable than previous data. The
absolute values dejiend in jiart upon the assumption of 1072° C. as
the melting point of pure gold, the recent determination of Holborn
and Wien at the Reichsaustalt. Should that datura, however, be
shown to reijuire revision, the validity of the present measurements
would not be impaired, but new values of the melting points could be
readily computed from them, which would be consistent with the better
value for gold.

Al Ag

660° 970°

The Pure Metals used were of a high degree of fineness, except
unfortunately the platinum.

The gold contained less than 0.01 per cent total impurities, these
being, if any, probably minute traces of silver and platinum. It was
obtained as part of a specially prepared lot from the United States
Assay Office in New York through the courtesy of Professor H. G.
Torrey, upon whose authority the above statement is made. The
purity was at least as great as the best " proof " metal used at the
United States or London mints.

The silver was from the same source, and equally pure.

The aluminum was manufactured and given by the Pittsburg Reduc-
tion Company, of Pittsburg, Penn., and was stated by Mr. Alfred E.

Au

Cu

pt

[1072° C]

1095°

1760°

Assumed.

HOLMAN, LAWRENCE, BARR. — MELTING POINTS. 219

Hunt, President of the company, to contain but 0.07 per cent of
impurity, consisting entirely, of" silicon.

The platinum was the ordinary platinum wire supplied by Carpen-
tier of Paris with his Le Chatelier thermo-electric pyrometers. It
presumably contained 0.5 per cent or more of impurity.

The copper was electrolytically produced, and was from the Lake
Superior region. It was kindly given by Mr. Maurice B. Patch of
the Buffalo Smelting Company, Buffalo, N, Y., who stated that it
showed by analysis 99.D9+ per cent of Cu, and contained no Ag, As,
or S, and only 0.0002 per cent of Fe.

The Less Pure Metals. — Partly for the purpose of testing the
effect of impurities, other samples of gold and copper were employed
with the results stated later. These were : —

Dentists' Gold. — This was a gold foil employed by dentists, fjur-
chased as being of good quality.

Ingot Copper. — This was also from Mr. Patch of the Buffalo
Smelting Company, who gave its analysis as

Cu 99.825

Ag . 0.032

As 0.003

S 0.022

Fe 0.003

O 0.116

100.001

This was the company's " regular run " of copper.

Oommercial Electrolytic Copper. — A sample of commercial elec-
trolytic rolled sheet copper, furnished by a friend, and not assumed to
be of unusual purity. It was probably Montana copper.

Commercial Hard-drawn Copper Wire. — This was from a lot
purchased for electrical testing purposes, which showed a specific
resistance of 0.1440 international ohms per meter-gram, or an elec-
trical conductivity of about 98.3 per cent referred to Matthiessen's
copper.

Methods and Apparatus. — The method consists in measuring the
thermal electro-motive force of a couple composed of one wire of
platinum and the other of a 10 per cent rhodo-platinum alloy. One
junction is immersed in the melting or solidifying metal, and the other
surrounded by ice. The wire was that furnished by Carpentier of

220

PROCEEDINGS OF THE AMERICAN ACADEMY.

Paris (through Queen & Co. of Philadelphia) with the Le Chatelier
2)yrometer.

The emf. was measured in microvolts (international) by the Pog-
gendorfF null method modified for rapid and convenient working.
The disposition of apparatus is shown in Figure 1. i? is a battery of
sufficiently steady emf. (A single Samson-Leclanche cell was en-
tirely satisfactory.) In direct circuit with this were two water rheo-
stats, W, in series ; an ammeter, A, which was a Weston voltmeter
(No. 395) with the calibrating coil only in use ; and a manganine

wire resistance, a, b, c, d,
divided into sections,
each of accurately known
resistance. T is the
thermo-couple connected
through a sensitive gal-
vanometer, G, and key
to any desired sections
oC the coil o, b, c, d.
Tiie water rheostats were
of about 100 olims and
8 ohms respectively, and'
the vertical motion of
theirplungers thus served
to give a coarse and fine
adjustment to the re-
sistance in the circuit.
The current could thus be promptly and closely adjusted. The volt-
meter was one of the type having a " calibrating coil," that is, one
having a connection by means of which the usual high resistance series
coil could be cutout, leaving its resistance about 117 ohms. Any of
the Weston voltmeters with a special connection made to effect that
result would answer equally well. The voltmeter was preferred to a
mil-ammeter as probably more reliable. The instrument was carefully
and repeatedly calibrated throughout its scale by an application of the
Poggendorff method, measuring by the Clark cell the drop of poten-
tial in a known resistance through which a current was passing in
series with the ammeter, and at the same instant reading the ammeter.
The calibrations at different times checked at the same point, with an
average deviation of only a few hundredths of one percent. A test
for temperature error showed a change of but 0.1 percent for a change
of 15° C. ; so that, as the temperature during the work was constant

HOLMAN, LAWRENCE, BARE. — MELTING POINTS.

221

nn^ sM

Fig. 2.

within a few degrees, no correction was needed. The raanganine coil
Figure 2, consisted of about 16 feet of No. 20 wire, had a total resistance
of about 8.8 ohms, and was divided into
nine sections by copper potential wires
leading into diflferent points along the coil.
These sections were so designed that by
suitably shifting the connections along a, b, c,
etc., any thermal emf. which was to be meas-
ured could be balanced by a current which
would deflect the ammeter to a point be-
tween 90 and 140 divisions (readable to
tenths), — corresponding to currents from
0.006 to 0.009 amperes roughly. The coil
was immersed directly in kerosene, and, as
its temperature coefficient was but 0.001
per 1° C, the correction became very small.
The relation and actual resistance (inter-
national ohms) of the whole coil and its
several sections were repeatedly determined against a standard ohm
by the differential galvanometer, and checked by a modified Wheat-
stone bridge arrangement. These data were reliable probably well
within 0.05 per cent throughout.

In the thermo-couple circuit, the sensitiveness necessary in the gal-
vanometer to give the smallest emf. to 0.1 per cent was easily com-
puted to be only about 7. 7-10^ (mm. defl. at 1 m. per ampere or
d/c). The instrument as actually used exceeded this requirement,
averaging about 5-10^. Its resistance, all in series, was 14.3 ohms.

The cold junction c of the thermo-couple was fused together in an
oxyhydrogen flame. The wires, insulated from each other by having
one strung through a very fine glass tube, were run down another
tube of about ^ inch inside diameter, and 8 or 10 inches long. This
tube was fused together at the bottom and top, as well as at some
intermediate points, and when in use was always packed in a double
vessel of cracked ice, as shown in Figure 3.

The intermediate junctions from which the copper leads went off to
galvanometer and key were soldered. They were kept at an equal
temperature by the device of enclosing them in a stoppered glass tube,
which was packed with hair felt into a one-inch hole in a five-inch
cube of cast iron. This arrangement was entirely satisfactory, but
seems to possess no material advantage over making the junction of
the copper leads with the Pt and Ft Rh serve as the cold junction, and

222

PROCEEDINGS OF THE AMERICAN ACADEMY.

immersing this in ice as in Figure 3, except that the latter makes a
rather more bulky mass to insert in the ice.

The wires were also fused together at the hot junction except when
this was unnecessary on account of their being immersed in metal.
It may be noted here that, as a null method was employed, the total
resistance of the thermal circuit, or any variation in it, was without
effect other than a corresponding change in sensitiveness.

As the hot junction was to be immersed in vapor of sulphur as one
of the known temperatures, the following apparatus was designed for
this purpose. It is substantially the sulphur boiling point apparatus

Fig. 3.

of Griffiths, and is shown in Figure 4. A glass tube, A, similar to the
Victor Meyer vapor density tube, 16 inches long and with a two-inch
bulb, was provided with an asbestos jacket and hood, B. B. The
upper few inches of the tube were wound with a spiral wire spring, S,
which rendered this part efficient as a condenser. The top was closed
with a layer of asbestos. Two overlapping diaphragms of asbestos
were inserted in the tube at D and E. The couple passed downward
throuo-h a glass tube to the asbestos tubular hood, (7, which served as
an umbrella to shed the dripping cooler sulphur, and as a radiation
screen. The hood, however, had openings top and bottom for the
free circulation of the vapor. An asbestos diaphragm, H^ upon which
the bulb rested, reduced the chances of superheating.

HOLMAN, LAWRENCE, BARR. — MELTING POINTS.

223

For the melting metals, after trial of several devices, the one
shown in Figure 5 (of exactly half size) was settled upon as proving
very satisfactory. The crucible G (usually of fire clay) is supported
by clay blocks in the double-walled fire-clay furnace, F. A carbon
block, E, channelled to fit the crucible, forms its cover, and a carbon
diaphragm D inside the crucible serves to support some powdered
carbon shown by the dotted mass. The object of these carbon parts
was to prevent oxidation of molten metals, and they proved very
eflfective in the case of aluminum, silver, and copper. G G was an

Fig. 5.

asbestos diaphragm supporting a non-conducting layer of fibrous
asbestos, A, A. The temperature was controlled by the blast lamp B.
The clay crucible was one inch in diameter outside, and the amount of
metal employed ranged from 11 grams (gold) to 35 grams (copper).
Larger amounts might be advantageous, but with 30 to 35 grams
it was easily possible to obtain a constant indication for five minutes
during the melting or solidifying of copper. No difficulty whatever
was experienced with this arrangement with silver, gold, or copper.
With aluminum, however, a peculiar action occurred, the cause of
which in the time available for investigation could not be determined

224 PROCEEDINGS OF THE AMERICAN ACADEMY.

beyond doubt. The phenomenon was that after a few minutes of
constant temperature at the melting point the indication of the thermo-
couple fell off with increasing rapidity. On withdrawing the couple,
cleaning it, or clipping it off and restoring it to jilace, the melting
metal meanwhile being untouched, the indications returned to their
original high value. The apparent explanation was the formation of
a slag between the wires, but this was not entirely satisfactory. The
use of a plumbago crucible in place of the clay, and an entirely fresh
lot of aluminum, did not remove the phenomenon, and gave the same
initial readings, which it could not be doubted were the ones corre-
sponding to the melting point. The fusion of the aluminum was, how-
ever, the least sharply defined of all the metals used.

The fusion of platinum was, of course, differently effected. For
this the two wires of the couple were laid close together on a piece of
lime. An oxy-hydrogen flame was then directed upon their ends, and
the platinum fused into a globule which with care could be made to
travel slowly up the wire. There was no difficulty in ol)taining
steady temperatui"es for a sufficient period to make the necessary read-
ings, and check results to 0.1 per cent were obtained on different
days.

The galvanometer, keys, coils, and all junctions of dissimilar metals,
were, so far as possible, covered with boxes of wood, pasteboard, or
asbestos, to maintain uniformity of temperature, and thus minimize
local thermo-electric disturbances. With this precaution no sensible
trouble from that source was experienced.

The procedure is as follows. To take the observation for vapor
of sulphur, for instance, the hot and cold junctions are exposed as
described. After a sufficient time the main circuit is closed, the ther-
mal circuit is connected to a suitable part of a, b, c, d, and the rheo-
stats, W, are adjusted until on pressing the key no deflection occurs
in the galvanometer, G. At this instant, A is read, which gives the
current c in the main circuit. The adjustment is disturbed and re-
made a number of times, and the resulting readings should check to
the nearest tenth of a division of A, provided the metal has reached
a steady state of temperature.

By this adjustment the drop of potential c r due to the current c
amperes in the part r ohms of the resistance a, h, c, <f, spanned by
the thermal circuit, is made equal to the total resultant eraf. in the
thermal circuit. The latter, which will be denoted by 2^ e or S^ ^j is
the algebraic sum of the thermal emf. proper of the junctions of the
Thomson emf. in the wires, and of any " stray " or local thermal emf.

HOLMAN, LAWRENCE, BARE. — MELTING POINTS. 225

in the circuit. Tlie last was found to be negligible throughout the
work.

To observe the melting point, the furnace containing the metal is
heated more or less rapidly until the meltiug point is approached.
The blast lamp is then adjusted to give a slowly rising temperature.
The thermal circuit, with the couple previously fused into the metal,
is connected to a suitable section of a, h, c, d. The rheostats are con-
tinually adjusted for zero deflections of the galvanometer, G, and the
corresponding readings of A are taken. These will show gradually
increasing values, but the rise will presently be interrupted by a
series of constant readings, after which the readings, will again steadily
increase. This period of constant, or nearly constant, readings of A is
that in which the latent heat of fusion is being absorbed, and its dura-
tion is frequently several minutes. The temperature at that time is,
of course, that of the melting point. The reverse process, starting
with the metal in a molten state and cooling it gradually, shows a
similar period of solidification.

No difference was discovered between the ascending and descending
readings when a sufficient amount of the metal and a slow rate of
heating and cooling were employed. With small amounts the steady
reading was more or less masked by phenomena which were clearly
due to inequality in distribution of temperature throughout the mass
of mixed liquid and solid metal. In the case of aluminum, however,
something more than this irregularity was observed, as elsewhere
stated, but the time at command did not permit a study beyond the
point of satisfying ourselves that the point observed was unquestion-
ably the true melting point.

This work was done chiefly as the thesis work of Messrs. Lawrence
and Barr. The efficient assistance of Mr. C. L. Norton contributed
materially to its progress and success.

The computation of temperatures t from the observed electro-motive
forces 2 6 involves a knowledge of the function connecting the two,
i. e. of the function

ne=f(t) or t = F(lU)'

This problem has been elsewhere discussed by one of the authors of
this paper.*

In that article two interpolation formulae were developed. They
were respectively of the following forms, applying to the case where

* Holman, These Proceedings, ante, p. 193.
VOL. XXXI. (n. s. xxiii.) 15

226 PROCEEDINGS OF THE AMERICAN ACADEMY.

one junction of the couidIo is kept at 0° C, and the other is at any
other temperature t C, or t = ^ + 273° absolute ; m and n are constants,
different for the two expressions; 2*e denotes the resultant thermal
emf. of the circuit, viz. that which is the object of direct measurement.
The first, called the exponential equation, is

2o e = m t" — (3 (where /3 = ?7i tq = m X 273").

The second expression, called the logarithmic equation, is

^qC =:: m f\ or log 2o e = ?« log t + log m.

Both formulas have been applied to (he data of the present investi-
gation given in Table I., with results shown below. The Avenarius
formula has also been applied for purposes of comparison.

To evaluate the constants tn and 7i of the exponential equation (for
method, consult the paper referred to) it is necessary to have values
of 2o e at three known temperatures. Of these, however, one may be
2o e = 0, at T = 273°, i. e. with both junctions in ice. It therefore
remains to fix upon two other temperatures between which to inter-
polate, or, in other words, two other temperatures which shall be
assumed as known. In looking over the ground it seemed that the
boiling point of sulphur, being so high and so accurately determined
by Callendar and Grifiiths,*

444.53 -f 0.082 (H — 760),

was pre-eminently one of these points. The other must be much
higher, and the melting point of pure gold seemed to be almost, if not
quite, the only one upon which reliance could be placed.

Apart from freedom from oxidation, and its conveniently high
point of fusion, gold seemed the more suitable because its melting
point had recently been so carefully measured by Holborn and Wien,
and because the metal could be obtained of the necessary purity.
Add to these considerations the fact that its melting point in a state of
at least fairly high purity has been measured by more experimenters
than that of any other high melting metal, so that it serves as an
excellent connecting link between their work, and we have claims
which no other substances can at present offer. The fusion point of
gold was therefore chosen as the second reference or calibration tem-
perature. As to the figure to be assumed as the melting point of
gold there is room for differences of opinion. The claims of the work

* Phil. Trans. CLXXXII, 119, 157 (1891).

HOLMAN, LAWRENCE, BARE. — MELTING POINTS. 227

of Holborn and Wien, supported to some extent by considerations
advanced by Barus,* lend much weight to the conclusion that Violle's
value of 1035° is considerably too low. Granting this, and in the
absence of sufficient basis for the assignment of weights to the work
of divers other investigators, the simplest and best step seemed to be
to adopt provisionally, without modification, Holborn and Wien's
value,

1072°.

These two points settled upon, the constants m and n could be com-
puted as elsewhere described, and the equation transposed to deduce
other values of t from observed values of % e. Representing m tq by
jB, a constant, the equation for the temperature as a function of Sq e
takes the form

=v/^

l+^-273<

which is, of course, easily solved by logarithms.

The data given in Table I. yield the values m = 0.3901, n = 1.488,
jS — 1645, in international microvolts and degrees centigrade, so that

2^ = 0.3901 T^-"^- 1645, or <= 1.488 i/?iUlI^ _ 273

V 0.3901

From these the temperatures of column six have been computed.

The constants of the logarithmic formula have been computed from
the same data for sulphur and gold, the method being sufficiently
obvious. The equation becomes

2U= 2.49655 f-^\

The corresponding melting and boiling points are given in Table I.,
column seven.

Substitution of the same data in the Avenarius equation yields

2^ e = (t^ — Q {9.7385 + 0.00 484 49 (t^ + Q}.

The corresponding melting and boiling points are given in column five.

Provisional Values of Melting Points.

In the paper referred to it was shown, 1st, that the logarithmic
expression fitted the Barus comparisons of the irido-platinum couple
with the air thermometer within the limits 400° to 1200" C. with no

* Am. Jour. Sci., XLVIIL 336.

228

PROCEEDINGS OF THE AMERICAN ACADEMY.

sensible systematic error ; 2d, that the exponential equation simi-
larly fitted the Holborn and Wien comparison of the rhodo-platinum
couple with the air thermometer within the same limits ; 3d, that the
exponential equation diverged systematically, although slightly, from
the Barus data, and the logarithmic from the Holborn and Wien data,
by about equal and opposite amounts both inside and outside these
limits, but much more markedly between 0° and 400° than at higher
points.

TABLE I. '

Melting Points.

Date.

Subst.

Mh-ro-

volts.

Temperatures.

Assumed
as Cor-
rect.

From
Aveu.
Eq. ta-

From
Exponent
Eq. t,.

From
Log Eq.

2

Provis-
ional
Values.

3-?

HoO

885.8

99°64

4-10

H2O

890.4

100.57

888.1

0

87.4

0
91.7

0
107.3

0
99.5

[100.10]

S-15

CioHg

2213

218.3

3-23

CioHs

2224

218.9

3-25

^10^8

2216

218.2

2218-

[218.5]

206.6

211.4

222.4

216.9

3-?

S

5287

444.7

3-22

s

5289

445.2

3-29

s

5287
5288

444 5
[444.8]

4-24

Cu

16463

—

1095.

1095.0

1096.5

1095.5

1095

4-29

Au

16002

[1072]

—

—

—

—

[1072]

4-29

Ag

14093

—

975.

972.

969.

970.5

970

5-2/3

Pt

30313

—

1695.

1735.

1783.

1759.

1760

5-3

Al

8638

—

665.5

662.5

656.2

659.4

660

Aven. 2o e= {t^ - h) j 9.7335 + 0.00 484 49 {t^ + h) {.
Exp. 2oc = 0.3901 t1-*88 -1645.
Log. 2oe = 2.49655 /1-2563.

HOLMAN, LAWRENCE, BARR. MELTING POINTS. 229

Inspection of columns six and seven, Table I., will show that the
computed boiling points of water and napthalin by the exponential and
logarithmic equations depart widely from the known temperatures in
opposite directions, by about equal amounts, and in the directions
according with the departures from the Barus and Holborn and Wien
data. Also, that the difFereuces between the computed melting points
intermediate between sulphur and gold differ but slightly by the two
formulae, thus confirming the former conclusions. It is obvious, thei'e-
fore, that although either of the two formulae would yield fairly good
interpolations for Al, Ag, and Cu, yet that a mean between the two
would probably quite nearly offset against each other the systematic
errors of the respective equations. This is also true in the dangerous
process of extrapolation for the platinum melting point, where the
chances of error in the result seem to be probably very much reduced
by averaging. The means of the melting points computed by the
exponential and logarithmic equations are, therefore, regarded as the
nearest available approximations, and the round numbers of column
nine are adopted as provisional values to represent the results of
the work.

Comparison of the results of the Avenarius formula, column five,
will show that they depart widely from the others in the direction
which would have been anticipated from the conclusions of the pre-
vious paper, thus further strengthening those inferences.

In addition to the foregoing, the melting points of three other
samples of copper, and one other of gold, were measured. The gold
was dentists' gold " foil," purchased in Boston. This is usually
classed as " very nearly pure," but its analysis was not known. No
special interest, therefore, attaches to it beyond the indication that it
gives of the sign and order of magnitude of the error (about — 4°)
which would be introduced by the use of such gold in the calibration
of the Le Chatelier pyrometer, or in similar ways.* The melting
point was found to be 10G8°.

The four coppers yielded the appended results.

* Hohnan, Calibration of the Le Chatelier Thermo-electric Pyrometer.
See These Proceedings, j^ost, p. 234.

230

PROCEEDINGS OF THE AMERICAN ACADEMY.

TABLE II.

2oe
Microvolts.

Melting
Points, (J.

Purity of
Metal.

Description.

16463

16448
16456
16446

1095.0

1094.3
1094.7
1094.2

%

99.99+

99.83
Unknown.
Unknown.

Electrolytic. Probably Lake Superior
copper, Buffalo Smelting Co.

Ordinary ingot. Same source.

Electrolytic. Probably from Montana.

Commercial bard drawn wire from
Washburn and INIoen Co. Sp. Elect.
Conductivity (referred toMatthiessen
value) 98.3 per cent.

The concordance of these results on various coppers, together with
the completely satisfactory behavior of the metal in fusion, and the
ease and cheapness of obtaining the metal of a very high grade of
fineness, suggest the decided availability of copper in a direct study of
high temperatures or melting points by the gas thermometer. A
large mass of the metal could be employed, and a constant and uni-
form temperature for a protracted period thus secured for the bulb of
the gas thermometer, or for other apparatus immersed in the molten
or solidifying material. There are unfortunately too few substances
which fulfil even these requirements. An added merit lies in the
nearness to the gold melting point, enabling the two to be satisfactorily
connected by some means of relative measurement.

It also appears that the use of good commercial copper would
introduce sensibly less error (3° less) into the calibration of the
Le Chatelier pyrometer than the use of the '' dentists' gold " above
tested, which is as good metal as would readily be obtained in the
market by most observers.

ReHability of the Results. —The points involved are : —

Instrumental errors.

Purity of the metal.

Was the observed point the real melting point?

Validity of the interpolation equation.

Error in the assumed melting point of gold and boiling point of
sulphur.

The investigation was planned and the apparatus arranged with the
intention of reducing the combined instrumental errors below one

HOLMAN, LAWRENCE, BARR. — MELTING POINTS. 231

teuth of one per cent iu the measurement of 2e above 200°C. Tests,
check measurements, and a discussion of the sources of error, unne-
cessary to detail here, have given satisfactory demonstration that an
even higher accuracy than this was attained. As far, therefore, as
constant or variable instrumental errors are concerned, it is believed
that no error beyond 0°.5 to 1° C. exists in the results, while probably
this estimate is large.

The error from impurities must have been exceptionally small, as
the analysis of the metals indicates. Some impurities from alloying
with the platinum and rhodium of the thermo-couple must have en-
tered during the experimenting, but as results at difFei'ent stages of
the work checked those obtained upon the first use of the metal, and
as reuewals of the metal made no difference in readings beyond the
limits of other variations (about 5 parts in 10,000), the error from
this source must have been negligible.

In the case of platinum the metal at command was unfortunately
not of known composition, nor was it possible at the time to obtain
any whose purity was known. An analysis of the wire used may
perhaps be obtained later, and it is hoped to carry out further meas-
urements with the better platinum now obtainable through the recent
advances made in its manufacture in Germany and England.

The aluminum was of very high grade, but it is thought that still
better may be obtained, and the peculiar occurrence attending its
melting point measurements should be further investigated.

The actual effect of the small impurities cannot be numerically esti-
mated, but must have been inconsiderable except for platinum, where
the error probably has the positive sign.

As to the third point, there was no reasonable doubt left in the
minds of the observers that the observed temperatures were sensibly
the melting points. Except as noted for aluminum, the readings with
rising and falling temperatures did not exceed about one part in one
thousand. Also entirely independent observations on separate days,
and with renewals of the metals in some cases, were equally concor-
dant. The average difference was much less than the error of read-
ing the ammeter. As an example of the concordance, and at the
same time as showing the homogeneity of the thermo wire, three
calibrations in sulphur are quoted in Table III.

232

PROCEEDINGS OF THE AMERICAN ACADEMY.

TABLE III.

Date.

2 e Microvolts.

Computed Tempera-
ture of Sulphur.

2 e reduced to
445^

March ?
March 22
March 29

5287
5289
5287

444.73
445.18
444.53

5290
5287
5298

Between these observations a considerable length of the wires was
necessarily clipped off. Reduced to a common temperature of 445°,
the maximum difference is but G microvolts in 5290, or 0.11 per cent,
while the average deviation of a single observation is but 0.02 micro-
volts, or 0.04 per cent, and of the mean but 0.02/V3 = 0.012 micro-
volts, or 0.024 per cent. At higher temperatures the discrejjancy was
even smaller.

The validity of the interpolation formulae has been ah-eady dis-
cussed. A statement of the extreme error which may have been intro-
duced into the results by this source should however be added. This
is believed to be for aluminum less than ± 2°, for silver less than ±2°,
for copper less than ±0.5°, and for platinum less than ± 10°.

Comparison with the temperatures computed by the Avenarius
equation show errors by the latter to be about 1.5 times as great for
water and naphthaliii, and of the same signs. It is therefore much
less reliable, especially for the platinum temperature, and no weight
is attached to its results.

Melting Points hy various Authorities. — A collection of these is
given in Table IV. Except in the case of the Barus data, the results
are set down directly as given by their authors. A further discussion
of these with reference to the purity of the metals used, and the char-
acteristic errors of the methods employed, would doubtless prove
instructive, and might partly remove or account for some of the
apparent discrepancies, — a task which will perhaps be undertaken
later.

HOLMAN, LAWRENCE, BARE. — MELTING POINTS. 233

TABLE IV.

Metals.

Authority.

Al

Ag

Au

Cu

Pt

H, L., and B. . . .

1895

Th.-el.

o
660

o
970

o
[1072]

o
1095

o
1760

Violle

1879

Sp.Ht.

954

1035

1054

1775

Ledebur . . . . .

1884

Sp. Ht.

960

1100

Le Chatelier ....

Th.-el.

635

[1035]

Callendar

[945]

1037

Erhard and Schertel .

954

1075

Barus, by Log. Eq * .

1894

Th.-el.

641

985

1090

1095

1783

" by Eq. 3 . . .

«

986

1091

1096

1757

Holborn and Wien . .

1892

Th.-el.

968

1072

1082

Mean of independent f
i. e. exclud. H., L. & B

ibsol. meas'ts,
,Le C, and C.

641

964

1008

1083

1779

N. B. — Values in brackets [ ] are those assumed by the observers,
and upon which their other values depend to a greater or less extent.

* See discussion by Holman, ante, p. 193.

Rogers Laboratory "of Physics,

Massacuusetts Institute of Technology,

Boston, October, 1895.

234 PROCEEDINGS OP THE AMERICAN ACADEMY.

Ikvestigations on Light and Heat, made and pcblished wholly ob in pabi with
Appkoprlation from the Rumford Fund.

XII.

CONTRIBUTIONS FROM THE PHYSICAL LABORATORY OF
THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY.

XLVI. — PYROMETRY : CALIBRATION OF THE

LE CHATELIER THERMO-ELECTRIC

PYROMETER.

By Silas W. IIolmax.

Presented November 13, 1895.

In a discussion of thermo-electric interpolation formuLe * the
author has shown that the resultant thermal emf. S^e of a closed
metallic circuit of two metals, with one junction at 0° C. and the other
at t° C, could be expressed to 0.5 per cent or better above 300° C. by an
expression of the form ^'^e — m f or log SoC = n log t + log m, where
m and n are constants depending on the metals of the couple. Similar
demonstrations show that with the cold junction constantly at 20° C.
the emf. may be expressed by 2^0^ = m t" nearly enough. This fact,
together with the known fact that the D'Arsonval galvanometer gives
readings nearly proportional to the currents, and therefore to 2 e on
a circuit of constant resistance, led to the thought that the calibration
of the Le Chatelier thermo-electric pyrometer might be simplified by
the use of a logarithmic instead of a direct plot. A study of a series
of six calibrations made at different times by four different sets of
observers, with two distinct pyrometers, and in nearly every case
with different suspension wires and adjustment, confirmed the deduc-
tion and led to the following method. It will be seen that the method
requires calibration at only two known temperatures, instead of several,
as formerly, — a saving of labor and a gain in accuracy. The method,
although suggested by the facts referred to, is not merely a deduction
from them. It should rather be regarded as an empirical process
based on experiment. It is not rigidly exact, viewed from either a
mathematical or experimental standpoint, but is merely an approxi-

* Thermo-electric Interpolation Formulffi, ante, p. 193.

HOLMAN. — PYROMETRY. 235

matiou holding withiu the limits of variable error of the instrument,
and probably well withiu the limits of uncertainty of the assumed
values of the melting points employed in calibration.

For clearness, the usual method of calibration will be first stated,
and tlien the proposed method.

Usual Method.

The cold junction is kept at a temperature usually about that of the
room, and measured by a mercurial thermometer. The spot of light
is made to read zero with the circuit open. The hot junction is
then exposed successively to several known high temperatures, and
the scale readings are taken. These temperatures are, ordinarily,
the boiling points of naphthaliu (C10H3), and of sulphur, the melt-
ing points of aluminum, gold, and platinum, — selected as being
the most satisfactory in manipulation, and reliable in value. The
values assumed for these points differ, following the judgment of
the experimenter, Violle's figures 1775° C. for platinum and 1035°
or 1045° C. for gold, and Le Chatelier's 625° or 635° for aluminum,
being most commonly accepted. From measurements and considera-
tions elsewhere * discussed, the author recommends the following as
provisional numbers in preference to the foregoing values.

Crafts,t Holman and Gleason.J
Callendar and Griffiths. §
Holman, Lawrence, and Barr.*

C10H3

218.7 + 0.0625 (H-760)

s

444.53 + 0.082 (H-760)

Al

660.

Ag

970.

Au

1072.

Cu

1095.

Pt

1760.

A plot is then made with deflections (scale readings) as abscissas
(horizontal) and temperature differences (t — c) between the hot and
cold junctions as ordinates (vertical). The points thus obtained lie
along a curve which in the upper part approaches somewhat closely to
a straight line. But it is nowhere exactly straight, and the error of
assuming it so may amount to 10° or 15° or even more. On the
other hand the points lie so far apart, if plotted upon a scale com-

* Holman, Lawrence, and Barr, ante, p. 219.

t Crafts, Bull, de la Soc. Chim., XXXIX. 196, 277 (1883).

t Holman and Gleason, Proc. Amer. Acad., XXI. 237 (1888).

§ Calendar and Griffiths, Phil. Transact., CLXXXII. 119, 157 (1891).

236 PROCEEDINGS OF THE AMERICAN ACADEMY.

mensurate with the sensitiveness of the instrument, that it is impossi-
ble to draw a curve through them which shall be much more reliable
in the upper parts (the portion most frequently needed) than the
straight line. Although this statement is perhaps counter to others
that have been made on the subject, it is the author's opinion, arrived
at after prolonged experience with the instrument.

In subsequent temperature measurements this curve is, of course,
used for interpolation in the customary manner ; that is, any deflec-
tion being observed, and at the same time the temperature c' of the
cold junction, the plot is entered with this deflection as abscissa, and
the corresponding ordinate of the curve is read off. This ordinate
will be approximately {t — c') the temperature of the hot minus that of
the cold junction. Hence the desired unknown temperature t is ob-
tained approximately by adding c' to this ordinate. ,

Proposed Method.

Observations. — The scale is adjusted so that the spot reads zero
with the circuit open. The cold junction is kept at a temperature
about that of the room, and measured by a mercurial thermometer.
In careful work this temperature should be kept constant, or nearly
so, either at 20°, or, better, at 0° C. The other junction is tlien ex-
posed successively to two known temperatures, preferably boiling sul-
phur and melting copper (for reasons to be given), or, if temperatures
upwards of 1200° C. or 1300° C. are chiefly to be measured, to melt-
ing copper and melting platinum. Deflections and temperatures of
the cold junction are taken.

Correction for Gold Junction. — If the cold junction is always kept
within about one degree of a constant temperature c°,e. g. 20° (or
0°), both in calibration and subsequent work, these observed deflec-
tions require no correction. If that is not the case, then the deflec-
tion should be corrected to obtain what it would have been had this
been done. This is readily effected as follows. Let a be the
change of deflection per degree at 20° C. This may be found nearly
enough by observing once for all the deflection with the cold junction
at some observed temperature, anywhere from 10° to 15°, and the hot
junction at an observed temperature from 30° to 40°. The deflection
divided by the temperature difference will, of course, be a. Then
if the observed temperature of the cold junction in a measurement
is, e. g., 16° instead of 20°, the correction would be (20 — 16) a = 4 a,
to be subtracted from the observed deflection, since that was ob-
viously too large. These corrections are small, and can easily be

HOLMAN. — PYROMETRY. 237

made mentally, since a is seldom more than one twentieth of a small
division of the scale.

Calibration, and Computation of Unknown Temperature. — Let d^
represent the corrected deflection for the sulphur temperature t^ (com-
puted from the reduced barometric pressure H Sit the time by the
foregoing formula) and let d^ be that for the assumed melting tem-
perature t^ of copper or the metal used. Plot two points with log d
as abscissa and log t as ordinate respectively. Draw through these
two points a straight line. Then within the limits of error of the ap-
paratus, the ordinate of this line corresponding to any abscissa log d
will be log t, the log of the desired temperature of the hot junction, d
being any observed deflection corrected to 20° or 0° as above. That
is, the number corresponding to this logarithm will be t.

By far the most convenient way to effect this plotting and subsequent
interpolation, instead of employing logarithm tables and the ordinary
co-ordinate or plotting paper I'uled with equidistant lines, is to use
plotting paper with a logarithmic ruling, that is, with lines spaced at
distances progressively diminishing according to the logarithmic law,
like the divisions of the ordinary slide rule. By suitably numbering
the axes of such a sheet the desired quantities can be plotted directly
without the aid of the logarithm tables. Subsequent interpolation
then becomes exceedingly easy, since it is merely necessary to enter
the plot with the corrected deflection as abscissa, and read off the cor-
responding ordinate on its numbered axis, this giving directly the un-
known temperature t. Unfortunately such paper (which would be
useful in many physical and practical problems) is nowhere on sale so
far as the author is aware.

In default of such paper, and instead of using the plot of logarithms
for continuous interpolation, a direct reading plot of deflections and
temperatures may be made from it once for all. On the logarithmic
plot look out the ordinate for every ten (or twenty) divisions of the
scale, i. e. corresponding to log 10, log 20, log 30, etc. (smallest
divisions). Look out in the log tables the numbers corresponding to
these ordinates, thus obtaining <,g, <oq, etc. Make a new plot, with 10,
20, 30, etc. as abscissas, and <,g, <.,g, ^„q, etc. as just found, as ordi-
nates. These points can be taken at such short intervals that a relia-
ble smooth curve may easily be drawn through them, and this is then
available for subsequent direct interpolations.

In the absence of logarithmic plotting paper, log plots may very
conveniently be made by a straight edge, graduated with a logarithmic
scale. Four of such rulers (two of single and two of double scale)

238

PROCEEDINGS OF THE AMERICAN ACADEMY.

may be cut from an ordinary boxwood slide-rule by a skilful car-
penter. They are very useful in a variety of physical work, but less
so than the ruled paper.

Proof of the Method. — The method rests for demonstration on the
fact that logarithmic plots thus made for the six calibrations above
cited, each containing observations on naphthalin, sulphur, aluminum,
gold, and platinum, and using the above temperatures, were straight
lines well within the limits of variable error of the observations with
but one exception, when the platinum point showed a wide divergence,
presumably due to mistake, or possibly to faulty setting up of the
instrument.

Apparatus and Procedure in Calibrating : Sidphur. — For this boil-
ing point, in order to obtain as good results as the instrument is

capable of, the boiling point tube
shown in the sketch is satisfactory,
and a duplicate serves well for
naphthalin. The glass tube is about
an inch in diameter, and twelve
inches long, terminating below pref-
erably in a bulb about two inches
in diameter. The tube, to within
two inches of the top, is wrapped
with asbestos cloth of a thickness
of at least a quarter of an inch, to
prevent radiation and superheating.
An asbestos cover closes the top.
The clear part of the tube serves
as a condenser, which may be made
more efficient by a spiral coil of
jl wire, S. It is easy to regulate the

rate of boiling so that the visible
line of condensation in the tube is nearly steady in position. The
bulb rests on a diaphragm of asbestos board, having an opening about
half an inch smaller in diameter than the bulb. The whole apparatus
is supported in a clamp stand. The flame plays upon the bulb where
it is exposed by the hole, but the diaphragm (six or eight inches in
diameter) prevents the hot gases from passing up around the tube,
which would cause superheating. Two overlapping diaphragms, I)
and E, about an inch above the sulphur and half an inch apart, pre-
vent spattering and radiation from the liquid to the thermal junction.
The wires of the couple extend downward through the cover to a

HOLM AN. — PYROMETRY. 239

point an inch or less above the diaphragm. The sulphur should
extend from the bulb to a point half an inch or more into the tube.
It will usually be found necessary to pour out the liquid sulphur at
the close of a calibration to avoid breaking the tube at the next heat-
ing. If there is trouble from breaking of the tube during boiling
(none with good glass) a wire copper gauze wrapped about the bulb
will be found a preventive. The boiling should be maintained for a
few minutes after the sulphur vapor has risen above the top of the
asbestos to insure uniformity of temperature. It is advantageous to
put the wires through a small hood, or umbrella of asbestos at a
point about an inch above the junction, so shaped as to shed the drip
of cooler liquid sulphur which will run down the wires. A special
form of this hood is shown at G in the figure. Both this and the dia-
phragms in the tube are, however, unnecessary in ordinary work.

For comparatively rough work it is sutficient to boil some sulphur
in an ordinary or small test tube (about one fourth full) and hold the
junction at the surface of the boiling liquid, taking pains not to have
it touch the tube, and also not to overheat the tube. This is the cus-
tomary procedure, but is not quite worthy of the instrument.

The barometer and its temperature should be observed at the time
and place of making the calibration, although the errors from omis-
sion of this are negligible in rough work.

Copper. — The high grades of commercial copper are good enough*
for this use ; such, for example, as are used in the best rolled sheet
copper and wire. This, of course, is easily obtainable and inexpen-
sive, and may probably be relied upon as having a melting point not
much, if any, over 1° C. below that of pure copper. Electrolytic copper
of a much higher grade of purity is preferable, and is not now rare or
expensive.

A fire clay crucible,! 1^ to l.V inches deep, and about 1 inch in
diameter, is held in place inside a small furnace by pieces of fire-clay,
as shown on page 240. A double-walled furnace of the design and
proportions shown % (the figure is half size) is most satisfactory. It
is about 4^ inches outside diameter, and 3^ inches high. The cru-
cible contains upwards of fifty grams of copper. Above this, within

* See results by H., L., and B.

t Buffalo Dental Co., BuflEalo, N. Y.

X Wiesiiegg, also Carpentier, botli of Paris, supply such a furnace ; or
one may readily be made by any manufacturers of fire-clay apparatus. The
Fletcher furnaces may also be used but are not so convenient for the purpose
because the products of combustion escape at the top.

240

PROCEEDINGS OF THE AMERICAN ACADEMY.

the crucible, is a diaphragm, Z), made from a plate of battery car-
bon. It lias a central perforation large enough to permit the passage
of the clay tube carrying the thermo-couple. Upon this diaphragm is
placed powdered charcoal. A block, E, of battery carbon is turned to
fit the rim of the crucible somewhat closely to serve as a cover. G
is a diaphragm of asbestos board which deflects the gases, and upon
this at A is placed fibrous asbestos to reduce the heat conductivity.
The cover belonging to the furnace is not used. Simpler details
might serve sufficiently well, but they must effect the same result.

namely, to produce by slow combustion an atmosphere of carbon
dioxide in the crucible to retard the oxidation of the copper. As
above arranged, the powdered charcoal does this, and the carbon plates
do not burn.

The thermo-couple projects as shown into the metal, being of
course thrust in during the first melting. It is obviously indifferent
whether the wires are joined or not. The end of the wires will
require clipping subsequently, to remove the short section injured by
action of the copper. The uniformity of the wire, however, renders
this a source of no sensible error.

The apparatus being arranged as shown, and the copper liquefied

HOLMAN. — PYROMETRY. 241

(it should not be unduly superheated), the blast is slightly reduced so
that the crucible may cool slowly, and the j^yi'ometer is read con-
tinuously. When the solidifying point is reached, the reading will
become constant for some minutes (dependent on the rate of cooling)
and will then begin to descend. This "steady point" occurs, of
course, during the absorption of the latent heat, and is the desired
reading corresponding to the solidifying point. The blast is then
increased, and the corresponding constant reading in the ascending
temperatures gives the melting point. This should, with copper, be
sensibly coincident with the solidifying point, and will be found so
unless complications are introduced, through too rapid heating or
cooling, or through the formation of slag upon the top of the copper.
It is well to withdiaw the junction, clean it if necessary, and take
check readings.

The advantages of copper over gold, as demonstrated elsewhere,*
are that it is more easily and cheaply obtainable in a state of purity
insuring a reliable melting point, and that, by using a large mass, an
unmistakable steady reading of considerable duration can be obtained.
The latter considei'ation is of special weight where the galvanometer
is exposed to jarring, or the observer is not experienced. The
operation is no more laborious, and, except possibly for a skilled
observer, is less time-consuming, and more satisfactory than the cus-
tomary method with gold.

In Le Chatelier's original method of employing gold (or other
metals) this same furnace was used, but with its cover on. The
crucible, without a cover, was filled with magnesia or lime, and the
thermal wires, passing through a perforation in the furnace cover,
terminated with their junction in the centre of the crucible, surrounded
by the magnesia. Before its insertion the junction was carefully
wrapped witli a small piece of wire or foil of gold, or of any metal
under test. This being arranged, the blast was controlled to raise
the temperature at a slow and steady rate. The spot of light on the
pyrometer scale was then watched closely. It would be seen to ad-
vance steadily until a certain point was reached, then to stop abruptly
for a moment, then to start upward again almo'st with a bound and
go on rising. This temporary stopping was due to the time required
for the absorption of the latent heat of fusion. A similar stopping
point may then be generally observed on allowing the furnace to cool
slowly. These two points, the "melting and freezing points," should

* See former reference to H., L., and B.
VOL. XXXI. (n. s. xxhi.) 16

242 PROCEEDINGS OF THE AMERICAN ACADEMY.

in general agree. With too rapid rates of heating or cooling, they
are likely to show an erroneous disagreement. Under favorable con-
ditions, namely, with the galvanometer in a place free from jarring,
and with a steady air-blast of quite sufficient capacity, this method
yields perfectly good results, although the inexperienced observer will
generally miss the reading on the first one or two trials. But in
most places, notably in connection with industrial plants, it is not in-
frequently difficult or impossible to secure the necessary steadiness of
support or of blast. Failing these, the stopping point sought is very
likely to be so masked by irregularities in the motion of the spot as
to introduce much uncertainty into readings, if not entirely to prevent
satisfactory calibration by this method.

Another way is to have the two thermal wires held together at
the ends merely by the wrapping of gold, etc., and under slight ten-
sion, so that as soon as the metal melts the wires will draw apart and
open the circuit. The spot will then reach a highest point correspond-
ing to the melting point, and then suddenly drop to zero. In prac-
tice this method, although convenient in some cases, has been found
not to yield satisfactory results on the whole.

With some metals, particularly those melting below gold, the author
has found the following arrangement to work well, being more rapid
and under better control. A small crucible of fire-clay, lA or 2 inches
long, and ^ to j^- inch inside diameter, with straight sides and a flat
bottom, is employed.* Into this is inserted the clay tube carrying the
wires, a packing of asbestos being used around it, if it does not pretty
well fit the crucible. The junction remains free about half an inch
from the bottom and not touching the sides. The crucible is then
gradually heated in the flame of the blast lamp, and the stopping
point observed as before.

Platinum. — For this we depend on the direct fusion of the plat-
inum wire of the couple just at the junction. f A convenient way is to
fuse the ends together, then to bend the wires so that the two lie
closely side by side, and are nearly straight, but not touching. Then
lay the end on a smooth surface of quicklime. A small flame from
the oxyhydrogen blowpipe is then directed upon the junction. The
globule at the end will fuse, and may gradually work up along the

* Made to order by Hall and Sons, Tonawonda St., Buffalo, N. Y.

t It may be of interest to others, as it was to the writer, to learn that the firm
of Heraeus of Hannover, Germany, whicli has made platinum of exceptional
purity for the Reiclisanstalt, has an agency in New York, Charles Englehard, 41
Cortland St., N. Y.

HOLMAN. — PYROMETRY. 243

wires. By watching ttie surface appearance of tlie globule closely, it
is easily possible to control the position and effect of the flame so as
to maintain the mass just on the point of fusion and get a fairly con-
stant reading. Heavily clouded dark glasses must be used in watch-
ing the platinum. Neglect of this involves risk of serious permanent
injury to the sight, and, if persisted in, of even greater danger.

3Iodijications of the Le Chatelier Thermo-electric Pyrometer. — By
an experience extending over four years in the use of this instrument
in its original form, and after tests and study of many other forms
of pyrometer, the writer is convinced both that the thermo-electric
method is by far the most promising one as the basis for industrial
pyrometer, as well as for much measurement in scientific research,
and that the Le Chatelier form of the instrument is the most generally
satisfactory one now obtainable on the market. Yet for general in-
dustrial services the Le Chatelier instrument falls short of the require-
ments in two ways. First, it is not sufficiently simple of manipulation
in the setting up, and in the calibration, to be put into the hands of
any but a fairly well trained observer. A trained chemist of works,
an educated superintendent, or any man of such caliber may be ex-
pected to take the instrument as sent by the makers, mount it, and
operate it successfully ; and under such observers it is doing important
service in many places. The mere reading can be done by any fore-
man after a few words of explanation. Secondly, the conditions of
support requisite for the galvanometer, although often obtainable, are
by no means commonly to be found in those portions of industrial
works where it would be otherwise advantageous to locate the instru-
ment. These facts are restricting the introduction of an instrument
otherwise excellent in its performance and marking a distinct onward
step in the art of practical pyrometry.

The urgent need of a still more universally practicable pyrometer
cannot be too strongly nor too often emphasized. In a large number
of industrial processes of great magnitude the employment of a reli-
able pyrometer would certainly result in a notable advance in quality
of product, in prevention of waste, in improvement or perfection of
the process itself, and in the discovery of new methods and new
products. If a galvanometer as reliable, as simple, and as portable
as the Weston magnetic voltmeter could be directly connected to a
rhodo-platinum thermal couple, there would result, as I have for some
years urged, an almost ideal pyrometer for technical work. Thus far,
unfortunately, the Weston instruments do not yield in this combination
sufficient sensitiveness to admit of their being put forward as meeting

244 PROCEEDINGS OF THE AMEIilCAN ACADEMY.

any but very special needs. Nevertheless, at least one such combina-
tion is said to have been successfully used for a certain purpose. The
writer has made repeated efforts to obtain a suitable instrument from
the Weston Company, and from some others, without success.

Rogers Laboratort of Physics,

Massachusetts Institute of Technology,

Boston, September, 1895.

NOTE.

A portable form of D'Arsonval galvanometer, combined with a rhodo-
platinum thermo-couple and graduated in degrees up to 1500° C, is now
announced by Keiser and Schmidt of Berlin.* Its gi-aduatiou is based
upon the investigations of Holborn and Wien at the Reichsanstalt.

The author desires, also, to call attention to the convenience of
maintaining a constant calibration curve for the Lc Chatelier pyrom-
eter by the insertion of an adjustable resistance in the circuit. By
this means the unavoidable slight changes of resistance of the circuit,
of sensitiveness of the galvanometer, and even of electro-motive force
of the couple, can be readily offset. Tlie procedure would, of course,
be to adjust the resistance until the sulphur or the copper reading was
the same as in the original calibration. S. W. H.

* Zeitschrift f iir Instrumentenkunde, XV. 285 (August, 1895).

HOLMAN. — CALORIMETRY. 245

Investigations on Light and Heat made and pubushed whollt or in part with
Appropriation from the Romford Fund.

XIII.

CONTRIBUTIONS FROM THE PHYSICAL LABORATORY OF
THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY.

XLVII. — CALORIMETRY : METHODS OF COOLING
CORRECTION.

' By Silas W. Holmax.

Presented November 13, 1895.

The following general method of treating the cooling correction is
applicable to nearly all non-continuous calorimetric processes, such as
the measurement of specific heats by the '' method of mixture," of heat
of combustion by the Berthelot or Mahler bomb, etc. It is reliable in
ordinary practice within the limits of error imposed by the thermom-
etry and by irregularity of conditions as to surroundings.

Following the brief statement of the " General Method " is that of
a " special case " of that method, and of a " Modified Method." These
are supplemented by a demonstration of the "Theory of the Methods,"
and a " Critique."

The procedure in the modified method is supposed to be new. Be-
yond that fact this paper claims the attention merely of those who
desire a concise working statement of a method of cooling correction
which has stood the test of practice.

General Method.

Procedure. — Start with the water of the calorimeter at any con-
venient temperature, e. g. about that of the air. If there is a water
jacket around the calorimeter, let that be stirred thoroughly once for
all, or, better, continuously. Record its temperature at the beginning
and end of the measurements as a possibly useful check.

Let the water of the calorimeter be stirred thoroughly, continu-
ously, and at a uniform rate. Record times and temperatures of the

246

PEOCEEDINGS OF THE AMERICAN ACADEMY.

calorimeter each half-miuute until the completion of the measurements
(quarter-minute readings are sometimes preferable). After five or
more minutes, at a noted time, drop in the hot substance, or start the
calorimetric operation, whatever it may be, continuing the tempera-

ar

n/Wf! "» ATwurts. ,
Fig. 1.

ture reading throughout the operation, and for at least five (better
ten) minutes after its completion. The plot shows a typical set of
observations.

A few readings may unavoidably be missed, where there is but one
operator, viz. at the time B and in the steep part oi B C ; but

HOLMAN.

CALORIMETRY.

247

these are unimportant, siiice the former can be obtained by prolonging
the line through the data preceding the omission, and the steep part
oi B G can be drawn in by judgment accurately enough because of
the fact that it is of so short duration, and must be nearly straight.

fflT-

/rj'

tsr

tea-

,e -5

Its-

Q ..-.^.^^

43

SO S5

Fig. 2.

4t00

The Correction, — The significance of the terms given in this sum-
mary will appear on reading the " Theory of the Methods." From a
plot, or from the data direct, find the initial rate rj and the tempera-
ture ti at the time mi ; also the final rate r^ and temperature t^ at m^.

248 PROCEEDINGS OF THE AMERICAN ACADEMY.

From the plot or data compute the mean calorimeter temperature for
the time interval m^ — ???i by the expression

where n is the number of half-minute temperature readings t^, b, c
. . . , to- Compute also

" ^" and 6* = ^ + ^2. ov = ^y + ti.

to — ti a a

Then the corrected rise of temperature in the calorimeter will be

t^ — ti + a (T— 6) (^2 — Wi),

or if the area method is adopted,

tn — ti -\- a A.

Special Case. — "Where B C h sensibly straight, and C J/(Fig. 1)
is of short duration relatively to B C, then ij is to be found by pro-
longing UD (Fig. 2) to its intersection in /with the vertical through
C, the point where the operation ceased, or where £ C begins to de-
flect from a straight line. From C onward the exchange has been
at the rate oi D E ; hence the ordinate of 1 may be taken as the tem-
perature which would have been indicated had all the remaining heat
been distributed throughout the calorimeter instantly when the opera-
tion ceased. This may be called to. The gain by exchange during
the passage from J5 to C is obviously about ^ {r^ -f r^ (jn^ — nti),
where jrio is the time corresponding to t^. The corrected rise of tem-
perature is then

Whether this special method is close enough can be determined for
any series of measuremfnts by solving a typical example both by it
and by the general method.

Modified Method.

Procedure. — By computation from approximate or estimated values
of the quantities involved, or from a preliminary trial, determine
nearly what the rise of temperature in the calorimeter is to be and
call this A t. Start with the water of the jacket at, or a degree or two
above, the temperature of the air of the room. In the calorimeter,
start with the water A t° below that of the jacket, or enough more than

HOLMAN. — CALORIMETRY. 249

At to allow for unavoidable rise during preliminary manipulations, —
in short, so that ti° shall be as nearly as may be A t° lower than 6°.
The object of this is to have <2 so nearly 0 that t-j shall be nearly
zero. On the plot of observations, Figure 1, this would cause ^ ^ to
incline upwards and D E to become nearly horizontal.

Then proceed precisely as in the general method, both as to manipu-
lation and computation.

Theory of the Methods.

A typical set of readings obtained under the general method, cor-
rected for tliermometric error, and plotted with times as abscissas and
temperatures as ordinates, would lie along a line A B C D E, Figure 1.
This curve should be convex upward all the way from C to D, there
becoming straight. If any concavity shows itself between the maxi-
mum M and E, or if the straight line through the points between D
and E cuts the curve anywhere between C and D, then the data
should be rejected. For the inference, almost without exception, is
that the stirring was insufficient to keep the thermometer down to the
average calorimetric temperature. This may be due either to the
bulb being badly located, so that it is unavoidably heated to excess, or
merely to ineffective stirring. The difficulty should be removed, and
the observation repeated.

The line consists of four portions, viz. : 1. The preliminary read-
ings, A to B. 2. The readings during the operation, B to ?ome inde-
terminate point C. 3. Readings G to D while the calorimeter and
contents are becoming equalized in temperature. 4. Final readings
after this equalization has become sensibly completed, and the calo-
rimeter and contents are cooling, or heating, by exchange with their
surroundings. In many cases, as in specific heat determinations, the
second and third parts are one ; that is, the operation is the equaliza-
tion of temperature between the calorimeter and the substance intro-
duced into it. In the case of a combustion and many other operations
these two parts are to be distinguished.

The method is based upon Newton's law of cooling. Test observa-
tions have shown that this law holds with sufficient accuracy for both
open and closed calorimeters, that is, within the uncertainty attribu-
table to varying conditions of surroundings and to thermometric
errors.

From the best representative straight line drawn through the ob-
served points between A and 7? is deduced the initial rate of exchange,

250 PROCEEDINGS OF THE AMERICAN ACADEMY.

ri (positive for a rising, negative for a falling temperature), expressed
iu degrees per minute.

For the data between D and E, i. e. through the observations for
the final few minutes, which will lie sensibly along a straight line, is
deduced similarly the final rate /v

Let ti be the temperature at the point B, the beginnino- of the
operation. Call this the initial temperature. Then rj is the exchange
rate at t{. Let mj be the corresponding time. Let f^ be the tempera-
ture at the earliest minute or half-minute at which the line has become
straight, that is, at which the fourth stage has been sensibly attained,
e. g. the point 1). Call this the final temperature of the calorime-
ter, and 711.2 t^^e corresponding time in minutes. Then u will be the
rate at t^.

The average temperature of the calorimeter between ???i and ni^ is
next found as follows. Drawing through the points the be-t re|)re-
sentative line B C M D, read off its ordinates at each half-minute
(better each quarter-minute) from Wi to m,2. Instead of a plot the
corrected data may be used directly. Call these temperatures a (=<i),
b. c, d, . . . , n {= (2), then the average temperature will be

?2 — 1 L2 2 J 71 — 1

This is, of course, the well known process. A slightly better one
is to read off the temperature at the end of the first qua7-t€r-imnute
after mj, then at intervals of a /<«//-minute (viz. at |, |, etc.) up to m^,
which must therefore be selected at a point an even number of quarter-
minutes from ???|, as there must be a quarter-minute interval at each
end. The average of these also gives T.

By Newton's law of cooling, the rate of gain or loss of heat by a
body through exchange with surroundings is directly proportional to
the difference of temperature between the body and the surroundings.
In calorimetry, provided that the heat capacity of the calorimeter and
contents is not changed materially during the process (by the insertion
or removal of substances), the rate of gain of temperature through
exchange is proportional to the rate of gain of heat, so tliat the cool-
ing correction may be applied directly to the temperature, instead of
to the heat.

Under this condition, the rate of exchange of temperature when the
calorimeter is at any temperature t will be

HOLM AN. CALORIMETRY. 251

where 6 = some constant temperature (representing that of the sur-
roundings, which may not be uniform or known) and a = the rate per
degree difference between 0 and t.

The two quantities a and 6 are unknown, but obviously can
be computed from the two pairs of observed values, r^, ti, and rg, t^.
For

r^ = a (6 — tfi) ;

'-i + 1,

a

Both a and 6 are next computed numerically from the data. 6 is
best found from the smaller value of r, for the reason that it influ-
ences the result most largely through its subsequent combination with
values near that one.

The gain of temperature by exchange, then, in any short interval
of time, A, during which the temperature is t, will be a (^ — t) A.

The rate of gain of temperature is a (^ — t), or ad — a t, so that, as 6
is constant, r varies directly as t. Hence the average rate will be
proportional to the average value of t, that is, to T, and will be
a 6 — a T, OT a (0 — T). The total gain will therefore be this quan-
tity multiplied by the duration (???2 — '^i) of this average rate, or
a (6 — T) {rrio, — m^^ and the exchange (or " cooling ") correction will
be this with reversed sign, viz. :

-a {6 - T) (w2 - OTi), or a{T-e) {m.,- m^).

The corrected rise of temperature of the calorimeter will then be

h-h^ a {T-6) (m.,-m,).

Obviously (T—6) {m^ — in^) is the area ^0 Oi^ZT minus the area
B G H B. This difference. A, in proper units, may therefore be
measured on a plot by the planimeter or otherwise, and the corrected
temperature rise will then be

h — tx ■{■ a A.

Critique of the Methods.

Three assumptions, beyond that of Newton's law of cooling, are
involved in the employment of this method, unless otherwise pro-
vided for in the computations into which the corrected rise of tem-
perature is introduced. First, that the thermometer indicates the

252 PROCEEDINGS OF THE AMERICAN ACADEMY.

surface temperature of the calorimeter. Secondly, that the tempera-
ture distribution throughout the entire contents of the calorimeter
is uniform both when r^ and 1\ are taken, so that these are really
rates of exchange with surroundings, and not resultants in which the
desired rate is more or less modified by redistribution within the
calorimeter. Thirdly, that the heat capacity of calorimeter and con-
tents is the same both when rj and 1\ are taken.

Proper location of thermometer, efficient stirring, and the preven-
tion of direct contact with the calorimeter of any hot or cold object
within it, readily insure the fulfilment of the first with sufficient
closeness.

Wherever the tem25erature of the calorimeter is varying, uniform
distribution of temperature throughout its contents is obviously
impossible if these are heterogeneous, especially where there is a
large metallic mass surrounded by the water, as the bomb in the
Berthelot or Mahler combustion apparatus. Such objects will neces-
sarily lag more or less behind the water as the temperature of the
latter falls or rises by external exchange. The effect of this is to
falsify the apparent rate r^ or rg, and still further to vitiate the
results through the usual assumption that all parts are at the same
temperature. A form of cooling correction, taking this into account
and applying to a certain class of cases, has been described by Pro-
fessor Rowland, but the only method which could entirely eliminate
the error would be to arrange the calorimetric process so that the
rate of exchange should be sensil)ly zero or constant during the
entire measurement. This means that the jacket temperature must
be so controlled as to be always at the temperature of the calorimeter,
or at a constant difference from it. "While this at first sight appears
wholly impracticable, and is so for most cases, I am disposed to think
that it might be feasible and helpful in certain investigations, such as
the stiady of the specific heat of water, and the mechanical equivalent
of heat, using electrical methods of heating, combined with thermo-
electric means of detecting and controlling the temperature difference
between the calorimeter and jacket.

The " Modified Method " described above deals with this error in
another way. It makes the final rate r^ as nearly as possible zero,
and hence secures the greatest possible constancy of temperature
during the portion C D Eoi the curve when the rate is about r,. The
advantages gained are as follows. The apparent rate r„ is sensibly
the true rate, since it will in general be much less than the rate of
redistribution of heat amongst the contents of the calorimeter. This

HOLMAN. — CALORIMETRY. 253

is of vital importance to freedom from constant or systematic error,
because the duration of the rate r.^ is inevitably greater, often much
greater, than of any other rate. If, as in the usual process, therefore,
it is a large (often the largest) and the least accurately known rate,
it is obviously productive of the most serious part of the error of the
coolinf correction. The modified method does not increase the lapse
of time preceding the arrival of the final steady condition, D, which
must in all methods, directly or indirectly, be the basis from which to
deduce t^. And as it does increase the percentage reliability of r,,
and reduces its amount, it diminishes the total amount of the cor-
rection, thus in a twofold way improving the result. The gain for
operations of such a nature that the duration of C B is several or
many times that of i? C is large, being approximately in proportion
to those dui-ations. Into that category would fall many of the ordi-
nary calorimetric processes, all those resembling the measurement of
specific heats of liquids which cannot be mixed with that of the calo-
rimeter, the use of the bomb for heats of combustion, and almost all
processes involving the presence of considerable metallic or glass ap-
paratus within the calorimeter. The method is of much less advan-
tage where the duration of C D is brief.

In certain cases there is a slight offsetting increase of error in Vi,
namely, when the contents are heterogeneous when both r^ and To are
taken. This occurs in the case of the bomb, but not that of the spe-
cific heats of liquids or solids. For then r^ is a larger rate, and is in-
accurate from the cause under consideration. Inasmuch, however, as
this rate has during the operation B C only a relatively short ex-
istence, and its error as affecting the value of a is divided by (/o — ^i),
the error introduced by it is small. It should be noted that if the calo-
rimeter is used as a secondary apparatus, not as a primary or absolute
one, the systematic parts of the above errors are further reduced.
Thus, for instance, in using the bomb, the heat capacity of bomb,
calorimeter, and entire contents may best be ascertained by burning in
it a known mass of some pure substance of well determined heat of
combustion (e. g. naphthalin), and computing backward from that heat
to find the total capacity. This makes the apparatus a secondary one,
dependent upon the assumed heat of combustion, but it is much safer
than computing the capacity from the assumed specific heats and the
masses of the component parts, or than measuring it by the addition
of warm water in the well known manner.

Departure from the third assumption above named cannot be dis-
cussed in detail, since the effect varies widely with the change of vol-

254 PROCEEDINGS OP THE AMERICAN ACADEMY.

ume and exposed surface, as well as of heat capacity of the contents of
the calorimeter in each case. Whether the change of capacity is
sufficient to demand allowance can be usually determined by inspec-
tion of the case, and a special method of treatment provided. The
usual effect is to make the apparent value of r^ smaller relatively to r^
than would correspond to Newton's law.

Rogers Laboeatokt of Physics,

Massachusetts Institute of Technology,

Bostou, May, 1895.

KICHAEDS. — DEVELOPMENT OF ^CIDIA. 255

xiy.

CONTRIBUTIONS FROM THE CRYPTOGAMIC LABORATORY
OF HARVARD UNIVERSITY.

XXXIV. — ON SOME POINTS IN THE DEVELOPMENT
OF ^CIDIA.

By Herbert Maule Richards.

Presented by T. W. Richards, November 13, 1895.

Very little has beea published in the last twenty years regarding
the development of the a3cidium stage of the UrediueiE, and very
little more is known than at the time when De Bary * wrote his text-
book. Very recently Neumann t has given an account of the develop-
ment of the a^cidium, but his researches do not throw a great deal
of light on the question. He is able to trace the origin of the basidia,
or foot cells, as he calls them, only in a very general way, and as
regards disputed points as to the formation of the peridium and the
growth of the hymeniura there is much to be said. Other authors
have considered the question of the sexuality of the fecidium, but
hardly any of them have taken it up from the developmental stand-
point of the organ itself. Reference will be found to these papers
in a genei'al discussion at the end of this article.

In obtaining material for the study of the development of the
cecidium it is necessary to consider several things. Not only must the
material be plentiful, but the a^cidium had best be a large one, and,
above all, the tissues of the host of such a character as to allow of suc-
cessful manipulation. For the latter purpose nothing could be better
than an aquatic host, the tissues of which are usually loose, and the
chlorophyll, at least in the more or less submerged portions, not so
abundant as to interfere with a clear view of the hyphfe of the fungus.
Notliing was found more favorable than an ^Ecidium on Peltandra
lui'Iulata, which may be obtained in quantity in the neighborhood of

* Morphologie der Pilze, 1884.

t Ueber die Entwickl., der Aecidieii, etc., Hedvvigia, 1894, Heft 6.

256 PROCEEDINGS OF THE AMERICAN ACADEMY.

Boston duriug the spring. Most of the work to be described was
done on this form, but ^cidia on Hoastonla ccendea^ Ranunculus sep-
tentrionalis, Anemone neinorosa, and Sainbacus Canadensis, were also
studied. Peridermium elatinum on Abies balsaineu, and Roestelia lacerata
on the fruit of Amelanchier Canadensis., were also examined, and some
particulars regarding the later stages of the a;cidium seen in the other
species wei'e confirmed.

In all the work the methods of manipulation employed w^ere the
familiar ones. Sections were made both free-hand and with the
microtome. With the microtome sections the material was stained
either in toto or on the slide ; Delafield's hajmatoxylin being found
most favorable for the former, while for the latter and also for the
hand sections eosin, Hofmauu's blue, or Mayer's acid ha^malum
was used. Lactic acid was usually employed instead of potash for
distending and clearing the sections, because it does not disorganize
and render so exceedingly transparent the tangled masses of hyphaj
as does the latter. Most of the actual work of examination and
almost all of the drawings were made while the sections were still in
aqueous media, for any method of mounting soon renders the speci-
mens so transparent that it is impossible to see the details. For final
mounting, glycerine was usually preferred to balsam, although for the
study of the nuclei specimens stained with ha^matoxylin and mounted
in the latter were found the best. As is usual in the studying of
structures of this sort, where certain^ differentiated portions must be
disengaged from tangled masses of filaments, it was found that a toler-
ably thick section, properly dissected and macerated, was often more
instructive than even a complete series of thinner ones.

^CIDIDM ON PeLTANDRA.

This form, which is the oecidium of Uromyces Caladii, being the one
which was most thoroughly worked up, will be taken up first, to be
followed by a discussion of the other forms named. The first material
used for this examination was collected by Professor Thaxter, who
kindly placed it at my disposal ; quantities of the same ajcidium have
been collected by myself several times since, and the material killed in
both picric and chromic acids. In selecting the portions of the plant
for sectioning, petioles were usually taken, as they were easier to cut
and also contained less chlorophyll than the leaves. By taking pieces
of considerable length and cutting longitudinal sections the young
aecidia could be traced with ease.

RICHARDS. — DEVELOPIMENT OF vECIDIA. 257

For the earliest stages, pieces of tissue in which with a hand
lens nothing but the spermogonia could be detected wei-e taken.
In sections of these, the youngest stages of what De Bary called
" primordia" were seen, which consist of a large number of branching
hyphie, which are massed together and finally become compacted into
an irregular sized ball. Even the youngest conditions may be dis-
tinguished from developing spermogonia, both by the fact that they
are embedded more deeply in the leaf, and also because the hyphaa
are more loosely arranged than in a spermogonium of equal size.
After this massing together of the hyphte to form the primordium
has kept on for a certain time, changes are seen to take place. The
hyphoe already septate become very much more so, and the small
cells so formed begin to enlarge very much. The result is that a
large part of the primordium now consists of a pseudo-parenchyma
of cells of very irregular size and shape, and with walls of very vari-
able thickness. Finally, the whole primordium changes to this pseudo-
parenchyma with the exception of a weft of tightly woven hyphse
around the periphery, which merge at the base into a mass of some-
what more loosely compacted threads. A careful search was made
in these young stages to determine if anything in the nature of a
trichogyne could be found, but in all of the sections examined nothing
to which the function of a trichogyne could be ascribed was seen.
Occasional hyphse were seen protruding out of the stomata, but they
did not connect with any of the primordia, and showed no evidence
of any specialization.

During the formation of the pseudo-parenchyma, the young a?cid-
ium increases very much in size, and consequently displaces the cells
of the host plant. The primordium, starting in a large intercellular
cavity, fills it, and then the hyphte, pushing out, soon surround the
neighboring cells, and often cut them off from their fellows. Such
isolated cells are usually completely broken down by the hyphaj, and
apparently absorbed- In this case practically no distortion accom-
panies the growth of the hyphge or tecidia.

Earlier than the formation of the pseudo-parenchyma no sign is
seen of the hymenium, nor can the hyphae which are to form it be
distinguished. At about this stage, however, there arises at the base
of the primordium a definitely differen tilted hypha, which may in
respect to its further development be called the fertile or sporogenic
hypha. It is to be distinguished from the ordinary sterile hyphse
which surround it by its highly granular and somewhat more refrac-
tive contents. It absorbs aniline stains slowly, but to a great amount,

VOL. XXXI. (n. s. xxiit.) 17

258 PROCEEDINGS OP THE AMERICAN ACADEMY.

finally becoming much more deeply colored than the surrounding
threads. It is not, however, very easy to demonstrate, and many
sections must be examined before obtaining the fortunate ones vvliich
will show it.

At first the fertile hypha presents no very characteristic form, but
as time goes on it becomes twisted sometimes almost spirally (Fig. 1).
Although usually simple in the beginning, later it may fork, or even
branch several times. This fertile hypha may be traced down in
among the mass of threads at the base of the primordium, and there
is nothing to indicate tliat it originates from any specialized organ,
but simply from the unditfereutiated mycelium. The subsequent
development of it which leads to the formation of the hymenium is
subject to some variation. In the smaller ajcidia the fertile hypha
simply begins to bud out at the tip into short projections which are
young basidia (Fig. 2), but in the larger lecidia it may branch so that
the hymenium may arise from several points (Fig. 3). In this
species the contents of the fertile hypha very soon {)ass into the
buds of the young hymenium, leaving it empty and quite indis-
tinguishable from the surrounding hyphaj. Following the develop-
ment of one of the smaller ^cidia, it is seen that the buds rapidly
increase in number and assume a more definite position. The larger
older ones are usually in the centre of the cluster, and the young
buds are formed around the periphery (Fig. 4). The young basidia
are cut oflT from the central mass as fast as they are formed, and
after enlarging somewhat show other cross septa, which are the first
indications of what may be called the spore mother cells. To make
room for the developing hymenium, the pseudo-parenchyma usually
splits at this juncture. A rift is formed in the centre of the primor-
dium, and, the cells around it collapsing, there is left a large cavity
into which the young basidia push up (Fig. 2). As the basidia in-
crease, they finally compress the pseudo-parenchyma, until in the
mature ascidium it is seen only as a narrow covering outside of the
peridium, made up of the collapsed and almost disorganized cells
which formerly filled the whole primordium.

The development of the hymenium in the larger secidia is essen-
tially the same, except that in place of a single compacted mass of
basidia they arise over a somewhat larger area or from several distinct
points (Fig. 3). Soon, however, the basidia form a solid mass, and,
although there is a certain amount of intercalary formation, most of
the growth of new basidia takes place around the periphery of the
hymenium.

RICHARDS. DEVELOPMENT OF iECIDIA. 259

The septatiou of the basidia which started in the older ones is con-
tinued by the younger ones near the periphery, and the next definite
change that is to be noted is the beginning of the peridium. The
formation of the peridium commences by the differentiation of the
terminal cells cut off from the older basidia. Their walls become
definitely thicker and their contents very vacuolate, presenting quite a
distinct difference from the spore mother cells below (Fig. 3). As the
other young spore chains push up, their terminal cells also become
similarly altered and the formation of the peridium progresses rapidly
from the centre outwards. The peridial cells so formed increase
more rapidly in size than the spores below, and owing to this and the
interpolation of new basidia, which grow up from the base, their
connection with the chains of cells from which they originated is
often lost. Sometimes, however, this connection is retained quite
clearly even when the aecidium is well along in its development, due
perhaps to the fact that in such cases the young basidia have grown
up more nearly simultaneously than usual and that there has been
no intercalary formation of basidia. At times the enlargement of the
peridial cells is more rapid than need be to cover the hymenium
as it enlarges, in which cases they may be pushed out of place so
as to give the appearance of a double layer of cells.

This metamorphosis of the terminal cells continues to the very mar-
gin of the hymenium, and there the same modification affects the whole
peripheral row of spore chains, so that the aecidium is now completely
incased in a covering of modified sporei which may be called the perid-
ium. This last point is shown better in other aecidiathan the one now
under discussion, and will be referred to later. Nothing was seen
that would lead one to suppose that the peridium originated around
the periphery and grows up until it meets in the centre, nor did it
appear that it is derived in any way from the pseudo-parenchyma
layer, for the line of demarcation between this and the peridium is
always apparent (Fig. 3).

Concerning the formation of the spores there is not much that is
new to be said. They are cut off basipetally from the basidia or
foot cells, and the cells thus formed are being constantly pushed up
by the formation of new ones below. The upper row forms, as has
already been described, the peridium ; from the others the spores are
produced. In these cells may be seen a double nucleus, or rather two
nuclei almost always in close proximity to each other (Fig. 5, a). These
have already been described by Rosen,* and also by Dangeard, but as

* Cohn's Beitrage, VI.

260 PROCEEDINGS OF THE AMERICAN ACADEMY.

my own observations as to the fate of these nuclei do not entirely cor-
respond with Rosen's view, there remains something to be said about
them. The cells which are first cut off from the basidia are not the
spores themselves, for from their proximal end is always cut off a
small usually wedge-shaped cell (Fig. 6, «, b) the interstitial cell of
De Bary. The larger upper cell becomes the spoi'e, the small lower
one eventually disappearing.

In the basidia there are always two and often more nuclei. When
a spore mother cell is about to be cut off one of these migrates to the
tip and there divides, a septum is soon formed below it, and the spore
mother cell is formed. Rosen states that when the interstitial cell
is about to be cut off, each of these twin nuclei divides, and from the
four daughter nuclei so formed two migrate to the lower end of the
cell, where they are cut off, and the other two remain in the spore
proper. In my own observations in by far the majority of cases the
nuclei separated without dividing, and were separated as single nuclei,
one in the spore and one in the interstitial cell (Fig. 6, a). The
nucleus in the spore always divided at least once, — in a considerable
number of cases, indeed, three nuclei were seen in the ripe spores
(Fig. 5, b, c), — while the nucleus in the interstitial cell sometimes
divided, but quite as often remained single, as long as the interstitial
cell persisted. As the sjjore ripens, the wall thickens and becomes
rough, the interstitial cell disappears, and the spore may be said to be
mature. The double nucleus remains quite distinct in the spore to
the oldest stages.

In speaking of the double nuclei in the spore it should be men-
tioned that they are found in all of the other parts of the recidia, in
the hyphne, the pseudo-parenchyma, and the peridium. This ubiquity
tends to deprive them of so great a significance as some authors have
placed upon them.

It is not necessary to follow the development of the tecidium any
further, the final rupturing of the peridium, etc. having been often
described in other forms.

-^CIDIUM ON HOUSTONIA C^RULEA.

Most of the material used for the examination of this species was
collected by Professor Thaxter, who kindly allowed me free use of
it. A small amount collected by myself at Sharon, Mass., was killed
in picric acid, and was valuable for some of the younger stages. This
aecidium is found both on the leaves and corolla of its host. The

RICHARDS. — DEVELOPMENT OF ^CIDIA. 261

leaves, owing to their comparatively loose structure, and to the fact
that the chlorophyll was mainly disorganized by the fungus, proved
the best for use.

In the mature aecidium nothing of note is to be observed which
differs from the tecidium already described. The layers outside of
the hymenium are better developed and the peridial cells and spores
smaller than in the Kcidium on Peltandra. The whole ajcidium
is more compact. The intercalary cells are very prominent and in
them, perhaps on account of tlieir large size the double nuclei are
usually easily seen.

The fertile hypha appears at about the same stage as noted before.
When it is traced back it is seen to merge into the rest of the
mycelium, and at first can only be told by its more granular and
refractive contents. Perhaps the best case observed was seen in
Figure 7, where at the base of a young primordium is seen a very
much twisted highly granular hypha which has already sent out a
branch from which the hymenium will develop, as in Figure 8. In
the somewhat older fecidia the young basidia may be seen to radiate
very distinctly from a common centre, which marks the position pre-
sumably of the end of the fertile hypha (Fig. 9).

The subsequent development of the spores and peridium accords
well with that already described in the previous case, with the excep-
tion of the fact that the pseudo-parenchyma does not split as in the
aecidium on Peltandra, but is gradually pushed aside as the hymenium
develops.

-^ciDiuM ON Ranunculus septentrionalis.

On the lower sides of the leaves and on the petioles of Ranunculus
this species of secidium may be found occurring in clusters, with the
older fecidia in the centre and the younger around them. A con-
siderable quantity of the material was collected by myself in the
vicinity of Cambridge in the spring of 1894.

The mature fecidia present no great peculiarities, with the excep-
tion of a very considerable variability in the size of the peridial cells.
Ordinarily they are large, much larger than the spores, but occasionally
specimens are found with very much smaller cells. In such cases, it is
quite easy to trace their connection with the spore chains below : they
differ but little from the spores in size, but have the characteristic
walls of peridial cells. It is possible that this condition may be
explained by a delay in the formation of the peridium in such cases.
The hymenium having attained its full size, there was not the same

262 PROCEEDINGS OF THE AMERICAN ACADEMY.

necessity for the rapid growth of the peridium to cover it. Inter-
stitial cells are present, and do not show anything unusual. The
usual double nuclei were easily demonstrated in all the various parts
of the iecidium.

This iecidium proved more favorable than some to demonstrate the
connection of the basidia with the fertile hyphte. In one fortunate
macerated preparation a somewhat tortuous hyplia was seen, from
which branches arose that terminated directly in the basidia (Fig. 11).
The character of the leaf was not favorable for the study of the
younger stages, as the compacted condition of the tissue and the
presence of much chlorophyll obscured the young primordia. Never-
theless, several interesting stages were seen that are worthy of note
since they help to substantiate the observations made on the other
secidia. In Figure 10 a section of a young a^cidium is shown where
the fertile hypha has pushed its way into the already fully developed
pseudo-parenchyma. From the fertile hypha, in which many nuclei
could be made out, there have arisen the first buds from which event-
ually the basidia and spores will be formed. Unfortunately this
otherwise excellent section was marred by the fact that the hypha
was cut off only a short distance below the placp where it had begun
to bud out. Its connection with the hypha below, as indicafed by the
dotted lines, was only inferred by the similarity of their contents and
general appearance. At this early stage no septa could be seen that
cut off the buds from the parent hypha. Other stages of a similar
sort, some younger and others older, were found, but none showed as
clearly as the one just described. A very much younger condition,
where the fertile hypha is but a simple thread with a rounded end
and rich in granular contents, was perhaps the earliest condition
observed. In some of the larger aecidia, which however never ap-
proached the size often attained by the occidium on Peltandra, ther
were apparently more than one point of origin for the basidia.

-^ciDiuM ON Anemone.

This form, which is ^cidium punctatum Pers., is in many respects
unfavorable for examination, but, as it was apparent that it presented
some variation from the type already described, some work was under-
taken on it. The gecidia do not occur in clusters, as in some other
cases, but are more or less scattered over the under side of the leaf;
but they have a certain advantage over other fecidia in that the
spermogonia being so very superficial there is never any danger of
confusing the young stages of these two organs.

RICHARDS. — DEVELOPMENT OF ^CIDIA. 263

The basidia are comparatively short, aud.the layer below them,
which may be called the sub-hymenial layer, is much more closely
compacted than in any of the forms jareviously described in this paper.
It is also noticeable that the base of the hymenium, which iu other
cases, at least in the younger stages, is often somewhat rounded, is
here practically flat, since the basidia almost all terminate on the same
level. The primordium progresses iu its development witliout much
variation from the general rule, except that the amount of pseudo-
parenchymatic tissue is relatively smaller, and the surrounding weft
of hyphre relatively larger than iu the three cecidia already described.
It was evident from the appearance of the young hymenium, iu which
the basidia develop much more nearly simultaneously than in the
cases already noted, that there was some variation in the development
of this form from the types previously examined.

Below the young basidia in the sub-hymenial layer there could be
distinguished fertile hyphse, from which the basidia are budded out,
and sterile hyphas that wound in among them (Fig. 12). The former
could be separated from the latter on account of their power of stain-
ing deeply. This layer of sporogenous hyphjE which underlies the
pseudo-parenchyma gives rise to the basidia more nearly at the same
time than they arise, for instance, in ^cidium Caladii, although even
here it is apparent that the younger basidia are the ones nearer the
periphery of the hymenium. The fertile hyphaj arise from the
branching of hyphag tliat make their way up from below, and appar-
ently the sporogenous layer may arise from more than one of such
hyphfe. This course of development much more nearly corresponds
to that given by Neumann * than any of the others do, but is not
exactly as he describes it in the aecidia he investigated.

It was in this species especially that hyphfe were seen protruding
through the stoma, but examination failed to show any connection
of such hyphae with the primordia, and it was probably due more to
the very great crowding of the mycelium in the leaf than to any
other cause. The subsequent development of the spores and perid-
ium failed to reveal any noteworthy peculiarities. This tecidium is
especially interesting as showing a variation from others in the matter
of its development, but a discussion of this point will be reserved
until a few words have been said about some other fecidia which
afforded interesting facts, especially regarding the formation of the
peridium.

* Loc. cit..

264 TROCEEDINGS OF THE AMERICAN ACADEMY.

Aecidiuji on Sambucus.

In this jEcicIiiim there is a considerable amount of distortion of the
tissues of the host, but at present only the Eecidiuiu itself will be
considered. It did not afford a very good chance for studying the
younger stages, but, as nearly as could be made out the development
corresponds more nearly to that of the tecidium on Ranunculus
than to any other.

The peridium begins in the usual way from the metamorphosis of the
terminal cells in the older spore chains, and spreads over the top of
the hymenium, as already described. The fact that the side walls ot
the peridium correspond to tlie outer row of spore chains is excellently
shown in this species. The peridial cells, wliich become smaller and
thinner walled as they approach the base of the hymenium, are seen
finally to merge into a foot cell which exactly resembles the other
basidia. In the youngest peridial cells even the interstitial cells were
observed, although De Bary in his text-book says that they are not
present in the peridium (Fig. 13). Very soon, of course, owing to
the enlarging of the peridial cells and to the thickening of their walls,
the interstitial cells disappear sooner than they do between the spores.

Passing to the vEcidium known as Peridermiwni elatinum^ we find an
interesting variation in the formation of the primordium. What may
be properly called the primordium in this case is but poorly developed,
consisting of sim{)ly a mass of loosely arranged pseudo-parenchyma
of rather large cells, and with none, or practically none, of the hyphal
weft about them. It is not improbable that this form might prove
a very excellent one for studying the young stages of the hymenium,
and it is greatly to be regretted that the material at hand was too
advanced for that purpose. The formation of the peridium, however,
showed admirably. In Figure 14a young a^cidium is shown, in which
two peridial cells have just begun their differentiation from the spore
chains. The mature aecidium, aside from the necessary reduction of
the mass of collapsed pseudo-parenchyma outside of the peridium,
showed no great peculiarities.

The last ascidium to be considered is Roestelia. It was hoped
that, owing to the long period of formation that the Roestelia! un-
dergo before they emerge through the tissues of their hosts, it n-ight
be coupled with some longer preparatory growth of the a^cidium
itself. Such, however, did not prove to be the case, for in all of the
species examined, although the origin of the Roestelia was very deep
in the tissues of the host, and consequently it was a long time in

EICHARDS. DEVELOPMENT OF iECIDIA. 265

breaking through, the spores and peridium began to form as early as
in the other a;cidia. The subsequent develoj^ment was simply a
multiplication of the spores and peridial cells. In the species chiefly
examined, Rcestelia lacerata on Aiitelanchier Canadensis, there were
however some facts made out that should be mentioned.

The mass of pseudo-parenchyma in the primordium is elongated
vertically, and appears still to keep on forming, while the hymenium
is developing, until it almost reaches the surface of the tissue in which
it lies. Although excellent young material was obtained, it was
impossible, owing to the very compact and obscure nature of the
primordium and the very small size of the hyphie, to make anything
out of the youngest stages even by all the various methods tried. The
young basidia are very narrow, and those which form the peridium
swell abruptly into the peridial cells, which at first, though large, are
very thin-walled. The intercalary production of basidia is more
marked here than in any a?cidium previously examined, and owing to
this the connection of the peridial cells with the spore chains is rapidly
lost. The cells of the peridium, which are at first about the normal
shape, lengthen very much, and become of the elongate almost rhom-
boidal shape found in the mature a?cidium.

General Discussion of the ^cidium Stage.

In comparing the results of the foregoing observations with the
accounts of the development of the oecidium already published certain
differences of considerable importance will be noticed. As regards
the formation of a primordium with a mass of pseudo-parenchymatic
cells in the centre, there appears to be a general unanimity of opin-
ion, although Neumann* figures the formation of the hymenium at a
much earlier stage than seems to be the case in the tecidia described
here or in those mentioned in previous accounts. In Plate XVI.
Fig. 3 of his paper he shows very definite indications of a hymenium
before the appearance of the compact pseudo-parenchyma. The
same is true of his figure of the aacidium on Ficaria ranunculoides,
but if one may judge from the drawing of the form on Eaphorhia
Cyparissias, the condition is more nearly as I have seen it myself,
where there is a definite pseudo-parenchyma before the formation of
the young basidia. Concerning the formation of the sporogenous
hyphse, there seems to be but little better knowledge in the pub-

* Loc. cit.

266 PROCEEDINGS OF THE AMERICAN ACADEMY.

lished accounts than there was at the time De Bary wrote his
text-book. All that De Bary says is, that after the formation of the
primordium with its pseudo parenchyma, " the hymenium now
makes its appearance at the base of the secidium," etc.* He also
leaves in doubt the question of the manner in which the hymenium
increases in area, whether it is by the peripheral or intercalary forma-
tion of new basidia. Plowright f throws no light on either question,
leaving them just as De Bary has. There is really nothing which
deals with the question further, until we come to Neumann's paper
already mentioned. Even here the matter of the origin of the basidia
is passed over somewhat vaguely, being referred to as being cut off
from the hypluB. In the question of the enlargement of the hymenial
surface, it will be seen that his conclusion and the one put forward in
this paper do not entirely coincide. He expresses the opinion that
all of the basidia arise practically simultaneously, and that one cannot
speak of any after-growth of new ones. This is certainly not the con-
dition found in the aecidium of Peltandra, or of forms like it, although
in the aecidium on Anemone something more nearly approaching it
is seen. The opinion of De Bary that the hymenium usually starts
from a comparatively small centre is borne out by my own observa-
tions. The opinion expressed here, which seems to have basis ia
the f;xcts observed, is that the formation of new basidia, while mainly
peripheral, is also to some extent intercalary, and that in most a3cidia
they are not formed simultaneously.

As regards the origin of the basidia Neumann adds but little, and
there have been practically no suggestions in this direction except
a very brief accoimt by Massee t of the occurrence of an oogonium
which after fertilization by an antheridium, after the manner of the
Peronosporeoe, grew out into the basidia. It is unnecessary to do more
than refer to this publication, and to add that, so far as the observa-
tions recorded here go, nothing of such a nature was discovered.
There are, however, certain definite fertile hyphiE which rise from the
mycelium and push their way into the primordium. It has also been
seen that these hyphaj are often twisted, even spirally bent, suggesting
possibly the " Woronin's hyphse " seen in the development of certain
Pyrenomycetes. From these hyphae are given off buds or branches

* DeBary, Conip. Morph. of the Fungi and Mycetogoa, Garnsey and Balfour's
trans., page 274.

t British Uredinene and Ustilagineae, 1889, p. 22.
j Annals of Botany, 1889.

RICHARDS. — DEVELOPMENT OF ^CIDIA. 267

which give rise to the basidia. At this poiut there is apparently
some difference in the development of different species, and even in
the tBcidia of the same species. In the tecidia on Peltandra and
Iloustonia, at least among the smaller ones, the basidia arise with but
little branching on the part of the fertile hypha, while in the form
described on Ranunculus the fertile hypha seems to branch to some
extent, and in the aecidiura on Amemone it forms apparently an
extended growtli before giving rise to the hyraeniura. It is probable
that in some cases, particularly in the larger iecidia, more than one
fertile hypha takes part in the formation of the hymenium.

In the formation of the spores, there is first formed what may
be called the spore mother cell, from which by division is formed
the spore itself and the small interstitial cell, in the manner previously
described. Plowright * speaks of the spores as being formed in a
mother cell, and that by subsequent growth and thickening of the
walls the spore and the mother cell fuse together. Such, however,
can hardly be considered the state of affairs, for the intercalary cell
is plainly cut off in the manner described by Rosen and also given
in this paper.

Regarding the peridium, the view advanced here is probably more
different from the generally accepted idea than any other particular
mentioned in this paper. Almost everywhere one finds the statement
that the peridium originates along the sides and "arches over" the
hymenium. De Bary has said that the cells of the peridium do not
possess intercalary cells, but this is hardly the case in the tecidium
herein desci'ibed on Sambucus, although in other cases the thickening
of the wall might begin before it was time for the intercalary cell to
be formed, in which case its production might be interrupted. It
might be objected that the union of the terminal cells with the pe-
ripheral spore chains to form the peridium is hard to account for, but
when one considers the condition of affairs in the young gecidium it
is not so difficult to understand. The walls of the young peridial
cells are very delicate, and, being under a relatively considerable
pressure, are literally forced together. It is true that in their subse-
quent development the peridial cells on the side increase mainly in
length, while the peridial cells on the top of the secidium increase
in all directions ; but that is hardly an unsurmountable objection. It
may be that it is the pressure under which the cells of the peridium
are formed which induces the excessive thickening of the wall.

* Loc. cit., p. 22.

268 PROCEEDINGS OF THE AMERICAN ACADEMY.

It has been the view of some writers that the joeridium is the result
of a development from a circle of paraphyses, and the condition of the
SBcidium of Phragmidium is cited in support of this view. But, as
De Bary * has said, the paraphysal envelope of this secidium is not
well understood, and no light has been thrown on this matter since
he wrote regarding it. Whatever may be the condition in this
a^cidium, however, it is safe to say that the peridium of the ordinary
aecidium is the result of a metamorphosis of the outer layer of spores.

As to what interpretation is to be placed upon the lecidium, and
what place it occupies in the life history of the Uredineoe, so much has
been said, and so little comparatively is known, that a re-discussion
cannot bring out much that is new. What has been described in this
paper, although perhaps not of the most positive nature, certainly
rather suj^ports than contradicts the view held by De Bary. In the
secidia here investigated, it would seem that the aecidium is not to be
regarded at the present time as the result of a sexual process, but that
it has a definite spore-bearing hypha, possibly corresponding to the
arch i carp.

Some of the more recent writers in text-books have entirely disre-
garded De Bary's view, and speak of the Uredineai as forms of
fungi with three chlamydosporic stages, in some cases comparing the
chains of a;cidiospores and their interstitial cells with the chlamydo-
spores found in Chlamydomucor. This seems to carry it to some-
what of an extreme, for one certaiidy does not associate with
chlamydosporic conditions such relatively complicated sporocarps as
are found in the Urediuete.

Vuillemin f has recently published an account of the a?cidiospores,
in which he maintains to have found the equivalent of a sexual stage
in the union of the double nuclei found in these spores. The question
of the sexuality of such a process, should it occur, ran hardly be dis-
cussed here, for the idea itself naturally suggests its own objections,
particularly in the light of the ubiquity of the double nuclei through-
out the Uredinete. It is interesting to note in this connection a fact
already recorded, that triple nuclei have been not infrequently seen.

To sum the whole question up, it seems to the writer that the view
expressed on this question by De Bary, where he likens the iEcidium
to the sporocarp, has more basis in fact than any other theory, but that
at the present time there is no sexuality in the process of the forma-

* Loc. cit., p. 246.

t Comptes Rendus, 1893, Vol. CXVI. p. 1464.

RICHARDS. — DEVELOPMENT OF ^CIDIA. 269

tion of tbis sporocarp. Hartig* has said that it is probable that the
tEcidium is the result of a preceding sexual act, and is therefore a true
sporocarp, like the perithecium and apothecium of the Ascomycetes.
This is stating the case strongly, but even if the secidium is not the
result of a sexual act, it does not preclude the comparison with the
sporocarp, for it has been abundantly shown that the sexuality in
the formation of the latter is often much reduced among many fungi,
the Ascomycetes in particular.

* Diseases of Trees, Trans. Somerville and Ward, p. 153.

Crtptogamic Laboratory, Harvard Unitersitt,
May, 1895.

270 PROCEEDINGS OP THE AMERICAN ACADEMY.

EXPLANATION OF PLATE.

^cidium of Uromyces Caladii on Pellandra.

1. Section of primordiiini showing fertile hyplia. X 320.

2. Older stage, where the fertile hypha has begun to bud. X 320.

3. Young stage of a large iecidium sliowing several points of origin of hyme-

niuni. The peridium has begun to form. X 320. a. Showing the
formation of the peridium at a somewhat later stage. X 320.

4. More advanced stage of same. The pseudo-parencliyma has ruptured.

X 320.

5. a. Single ripe spore showing double nucleus, b. Spore vvitli tliree nuclei.

X 1250.

6. Young spore just cut off from the basidium showing nuclei, h. Spore;

older, but not yet mature. X 1250.

jEcidium on Iloustonia arrulea.

7. Section of primordium with fertile hypha. X 320.

8. Older stage, showing connection of fertile hypiia with young basidia. Not

quite median although showing the fertile hypha. X 320.

9. Section of young aecidium, showing a soniewliat older hymenium. X 320.

^'Ecidium on Raiiunctdus.

10. Section of primordium with fertile hypha already budding. X 820.

11. Section of an almost mature aacidium, in which the connection of the fertile

hypha with some of the basidia is plainly shown. X 320.

12. Young basidia from ascidium on Anemone (yEcidhim punctalum). Tlie

sterile hyphse, the contents of which are not so granular, show among
the fertile ones. X 560.

13. Margin of a section of aecidium on Sambucus, showing origin of the

peridium. X 560.

14. Young fficidium of Peridermium elatinum. X 320.

All the drawings were made with the aid of an Abbe' camera and reduced
about one third in reproducing Figures 3, 3rt, 4, 5, 6, 7, 9, and 14 were drawn
from microtome sections, the rest from hand sections. All were treated with
lactic acid, and stained before drawing.

JC

H

HALL. CONDUCTIVITY OF MILD STEEL. 271

lNVBSTiaATI0N8 ON LiGHT AND HeaT, MADE AND POBLISHED WHOLLY Oa IN PART WITH
APPBOPaiATiON FROM THE ROMFOUD i'VVD.

XV.

ON THE THERMAL CONDUCTIVITY OF MILD STEEL.

By Edwin H. Hall.

Presented January 8, 1896.

Some years ago,* the author gave an account of certain experi-
ments which he had made, by the method of Forbes, upon the thermal
conductivity of cast iron and cast nickel, and stated that he lioped
" soon to make trial of a method for estimating thermal conductivities
somewhat different from any heretofore used." This method has
now been tested, and it appears to be a good one, although certain
changes in the apparatus used, and presently to be described, might
be made with advantage.

It is well known that early experiments upon thermal conductivity
were made with thin plates or sheets of metal. The results obtained
were absurdly erroneous, owing mainly to the difficulty of determin-
ing accurately the difference in temperature of the two faces of the
plates. Later the method was by various devices made to give better
results, but none satisfactory, and for a long time it has been discarded,
replaced generally by the method of Forbes, steady flow of heat in
a long bar exposed to the air, or that of Angstrom, waves of heat
sent along the bar at regular intervals. None of the methods have
been very simple, or easy of application, and the results obtained in
the case of the metal most studied, iron, have been as various as the
difficulty of the investigation would lead one to expect.

Ev^ery one who has engaged in the experimental study of heat con-
duction, or has examined the curiously indirect methods employed in
this study, must have regretted the apparent impossibility of using
successfully the old method of thin plates. About four years ago, it
occurred to the author that one might determine the difference in
temperature of the two faces of a thin plate by making these faces

* These Proceedings, Vol. XXVII. p. 262, 1892.

272 PROCEEDINGS OP THE AMERICAN ACADEMY.

the junctions of a thermo-electric element. Suppose, for example,
that a plate of iron is coated on its two broad faces with a layer
of copper, and that copper wires lead from these copper sheets to a
galvanometer. If one copper side is exposed to water at t°, and the
other copper side to water at t'°, heat will flow through the whole
plate, and the difference in temperature of the two surfaces where
the iron joins the copper sheets will give rise to a thermo-electric cur-
rent, which may be made to give an accurate measure of the difference
of temperature of the two faces of the iron plate. No attempt need
be made to determine the difference in temperature of the outer sur-
faces of tlie copper sheets. This is an outline of the most distinctive
feature of the method now to be described. The chief obstacle has
been, and still is, the difficulty of getting a sufficiently even tem2:)era-
ture over the whole of each surface.

Figure 1 shows with tolerable accuracy a vertical section through
the conducting disk and the apparatus most closely connected with it,
the parts being represented at one third their actual size. Certain
unessential changes have been made in the drawing, in order to illus-
trate as many features as possible in this one figure.

(Sis the conducting disk of iron, or, rather, "open-hearth" boiler-
plate steel. The following information concerning its quality and
composition was furnished bj' INIr. Parker, of Parker, Field, and
Mitchell, Cambridgeport, who made the disk : —

Tensile strength 62,300 lbs. per sq. in.

Elastic limit 34,800 '' «

Phosphorus .0035.

Carbon .001.

The analysis was made by the Central Iron and Steel Works of
Harrisburg. The metal is very soft and tough, as the composition
would indicate.

The mean thickness of the disk, which was carefully measured at
many points, is 0.295 cm., the smallest thickness found being 0.290
cm., and the largest, 0.299 cm. The diameter of the disk is 9.80 cm.

The parts c and c' are the coatings of copper deposited electro-
lytically upon the steel. The author hesitated at first to apply
copper in this way, fearing that galvanoplastic copper might be
seriou.cily different in thermo-electric quality from ordinary copper
wire, which must be attached to the coatings ; but all his own experi-
ence and that of Professor B. O. Peirce, who has had occasion to

HALL. — CONDUCTIVITY OF MILD STEEL.

273

\4 M RLr^^

VOL. XXXI. (n. s. XXIII.)

Fig. 1.

18

274 PROCEEDINGS OF THE AMERICAN ACADEMY.

iuvestigate the same matter recently, go to show that this appre-
heusiou was groundless, that different specimens of copper wire now
purchased in the American market differ very little among themselves
in thermo-electric quality, and differ very little from galvanoplastic
copper.

The ordinary method of depositing copper, from a sulphate solution,
is not immediately available for a steel plate, because the solution
attaclis steel. On the other hand, the method which uses a cyanide
solution is comparatively slow, and requires much time for a thick
coatiug. Accordingly, the disk here used was subjected to the
depositing current in a cyanide solution for a sufficient time to cover
the surface completely, aud was then transferred to a sulphate solu-
tion, where the process was continued for many hours, until the coating
was about 0.05 cm. thick on each side of the disk.

The letters rr in the Figure mark a ring of steel, similar in quality
to the disk. It was turned a bit too small to receive the disk when
both were at the same temperature. The ring was then heated, and
the disk, smeared on the edge with a i)itcliy cement, was slipped into
it. The contraction of the ring in cooling made a joint so tight that
under severe tests no leakage through it has been detected.

The parts h h and lih' are rings of hard rubber. These rings do
not extend quite to the copper coatings of the disk, but leave a line
of the steel ring rr, about 0.07 cm. wide exposed, above and below
these coatings. These exposed edges have caused trouble through
rusting, under the action of the water flowing through the apparatus.
No varnish protects them effectually. In future experiments care
will be taken to have no iron or steel similarly exjjosed.

The parts hh and h'V are hollow brass cylinders about 10 cm. in
external diameter, fitting into grooves in the hard rubber rings h h aud
/^'A' respectively, a small rubber tube at the bottom of each groove being
used to make, under pressure, a water-tight packing of the joint. At
first these brass rings fitted directly into the ring r^-, the rings hh
and h'h' not being- used. But it soon appeared that heat, carried to
or from the ring r r by the rings h h and b'b' , each of which in the
first arrangement exposed a broad surface to the water, marie the
difference in temperature between the top and bottom of the disk
much greater near the edge than elsewhere. For it is to be noted
that the temperature difference between top and bottom of the disk
was in most parts only a small fraction, jierhaps 5%, of the maximum
temperature cliflFerence between the water above and the water below
the disk.

HALL. — CONDUCTIVITY OF MILD STEEL. 275

The parts dd and d'd' are disks of brass, the edges of which press
upon the rings hh and h'h'. These disks are perforated near the edge
by numerous holes for flow of water. The disks are packed water-
tight, or nearly so, in the brass rings b b and b'b', small rubber tubes
being used in the joints.

The parts g g and g'g' are blocks of hard rubber secured by the nuts
n and n' to the brass tubes t and t'. The object of these blocks is to
guide the flow of the water across the surfaces of the disk in order to
make the temperature over each surface as nearly as possible every-
where the same.

The block (j'g' is not fastened to the disk d'd' upon which it rests,
but is carried by the tube t' in such a way that when t' is rotated g'^
turns with it. It was the practice at certain stages of the work to
rotate l' briskly back and forth through a number of degrees, in order
to facilitate the escape of air-bubbles from the space above and around
g'g'. Numerous small holes drilled through the upper and outer edge
of g'g' were intended to make, and probably did make, this operation
more effective,

Most of the water (80 to 85%) coming up through t' passed down
and out by way of the double funnel f'f, but in order to remove the
air-bubbles a certain amount of overflow was maintained through seven
small tubes leading obliquely upward" and outward from the water
space at the edge of g'g'. One of these overflow tubes is shown in
the figure at the left. Each of them was connected with a small
rubber tube, and the outer end of each rubber tube was about 2 cm.
higher than the top of hh. Whenever the tube t' and the block g'g'
were being agitated to remove air-bubbles, as before described, the
main outlet o' was closed, so that all the water which passed under
the disk had to escape by the overflow tubes. This method of getting
rid of air-bubbles was adopted after the problem had been studied by
means of a kind of dummy apparatus, in which the disk »S'was replaced
by a plate of glass. There is reason to believe that it was fairly
effective, but it was cumbrous, and probably a different arrangement
will be used in future work. No similar device appeared to be
necessary for the space immediately above the disk, as air-bubbles
originally there would naturally be carried upward through the tube
t by the escaping water. A single overflow tube, q, led out from the
double funnel //, by which the water was brought to the upper
surface of the disk, and through this tube a small overflow was
maintained to carry off bubbles that might otherwise have caused
trouble. Another function of q was to serve as a kind of gauge

276 PROCEEDINGS OF THE AMERICAN ACADEMY.

indicating roughly the amount of pressure exerted by the water within
the apparatus, some parts of which were not well calculated to bear a
great disruptive force.

The copper wire w, about 0.016 cm. in diameter, leads from a point
on the upper copper coating of the disk through a hole in h, where it
is held in a water-tight joint between a hard rubber peg and a soft
rubber sheath surrounding the peg, to a wooden shelf, where its end is
gripped between two copper washers under the head of a screw, all of
which objects are shown in the figure. A thicker copper wire, about
0.06 cm., one end of which is pinched between the upper washer
and the head of the screw, leads off toward a galvanometer. Another
thin copper wire, w', similar to w, starting from a corresponding point
on the under copper coating of the disk, passes through // and is
fastened at its outer end exactly as the wire w is fastened. It is
necessary to consider these details, for one can easily imagine a
method of connecting the fine wires with the thick wires that would
give rise to very disturbing thermo-electric forces and make the whole
investigation useless.

At first the copper wires w and w' were soldered to the copper faces
of the disk, as little solder as possible being used, with the hope that
the thermo-electric forces at the top and bottom of the very thin layer
of solder used would be very small in comparison with the large
thermo-electric forces due to the copper-steel junctions. But tins hope
was not justified. The disturbance arising from these soldered junc-
tions was very serious. The fact appears to be that each copper wire,
exposing on the whole a surface of many square millimeters to the
water, kept its point of attachment with the solder much warmer or
much colder, as the case might be, than the same point
would have been kept by direct contact with the water.
The effect is similar to that observed when a soldered
— junction between two copper wires, one thick and one
thin, is held a short distance beneath the surface of
Fig. 2. water differing considerably in temperature from the
air above it. In such a case, represented by Figure 2,
the submerged part of the thin wire, and also its point of attachment
with the layer of solder, has very nearly the same temperature as the
water ; but the submerged part of the thick wire does not so nearly
attain this temperature. The result is a difference of temperature
between the two points of attachment with the solder, and a conse-
quent thermo-electric effect, which an unwary observer might attrib-
ute to difference in the thermo-electric quality of the two wires. If

HALL. CONDUCTIVITY OP MILD STEEL. 277

in such a case the junction is pushed deep beneath the surface, the
observed effect is diminished and perhaps disappears.

Accordingly it was necessary to get rid of the soldered junctions
upon the disk. Therefore the device was adopted of attaching the
copper wires to the copper surfaces by means of copper electrolyti-
cally deposited. The details of the operation, where one junction only
is to be dealt with, are as follows : Cover the surface of the disk, in
the neighborhood of the spot where the junction is to be, with a thin
layer of some non-conducting substance, paraffine for instance, and
then scrape off this coating from a patch about 0.5 cm. long and
0.1 cm. wide. Coat also, with a non-conducting substance, the copper
wire, and then scrape one end of it bare for a distance of about 1.0 cm.
Lay the bare part of the wire along the middle of the bare strip
on the copper disk and press it down flat in that position, the end of
the wire projecting a little beyond the limits of the bare strip. While
holding the wire firmly in place, attach it near each end of the bare
strip by melting at eacli spot a bit of paraffine. A few drops of sul-
phate of copper * are now placed upon the bare part of wire and disk,
above which they stand in a little mound, as in Figure 3. Another
copper wire, the lower end of which dips a little distance into the
mound of liquid, serves as the anode in the electrolysis. The exposed
surfaces at the bottom of the liquid are the cathode, the current
being cariied off, not through the thin wire, but through the body of
the disk, to another wire not here shown, the intensity of current
aimed at being 0.02 ampere per sq. cm. of the exposed surface of the
cathode. A run of three or four hours makes the junction sufficiently
strong for proper use, though perhaps it is safer to let the action go on
longer. After the deposit is completed the protruding end of the
fine copper wire is cut off and the paraffine surrounding the junction is
scraped away. When several junctions upon one side of the disk are
to be formed, it is easy to make the deposit upon all of them simul-
taneously.

In the final use of the apparatus the copper wires must be kept
from contact with the copper surface of the disk at other points than
the legitimate junctions. Accordingly the wires and the disk were
coated with a thin layer of asphaltum varnish, baked on at a tempera-
ture above that to which the apparatus was to be exposed in the con-
duction experiments.

* Of specific gravity 1.10, with 20 drops of strong H2SO4 added to 200 cu.
cm. of the solution.

278 PROCEEDINGS OF THE AMERICAN ACADEMY.

There are thirteen of these fine wire junctions on each side of the
disk, those on the upper side being arranged and nnmbered as shown
by Figure 4. On the under side are corresponding jnnctions, 1' just
beneath 1, 2' just beneath 2, 3' just beneath 3, etc. The method of
combining these junctions and the method of estimating the mean dif-
ference of temperature between the two copper-steel surfaces, by
means of the thermo-electric currents sent through these junctions to
a galvanometer, will be described later.

The parts ^1 andj^, in Figure 1, are copper- German-silver junctions.
The German-silver wire g s, about 0.015 cm. in diameter, is contin-
uous from one junction to the other. At each junction the German-
silver wire is joined to a copper wire, about 0.010 cm. in diameter.
From each junction the two wires, separated by a stri[) of plastic non-
conducting material ("Jeukin's packing "), y, are led into and along
a plug, p, consisting of two half-cylinders of hard rubber held firmly
together by clamps k and k. Within this plug the fine copper wire is
attached in a soldered joint to another copper wire, W, about 0.06 cm.
in diameter, which leads away toward a galvanometer. The loop of
wire exposed to the water near each junction is coated with varnish
to prevent disturbance due to chemical action. Shellac, baked on,
seems best for this purpose. The water, as the arrows indicate,
passes j\ on entering, and jz on leaving, the upper chamber of the
apparatus. The thermo-electric current from these junctions shows,
when interpreted, the difference of temperature of the incoming and
outgoing streams.

The amount of water flowing through the apparatus per second,
about 24 grams, was measured by noting the time required to fill a
flask of known capacity, F, in Figure 5. This time was about 75 sec-
onds, and it was observed at stated intervals by means of an ordinary
watch. The devices used for making this flow nearly constant, for an
hour or more at a time, will be described later.

JJ, in Figure 1, is a double-walled copper jacket, supported by
means of lugs, /, I, upon the ring bb, a. thick piece of rubber under the
end of each lug preventing metallic contact with b. The water, after
passing the junction yg on its way out, entered this jacket near the
bottom, flowed around and upward and escaped by the orifice 0 into a
descending tube, the lower end of which was kept at any desired level,
usually lower than any point shown in the figure. Thermometer 7\,
passed by the stream on its way to^i, and T2, in the jacket near the
exit, gave a rough indication of the change of temperature suffered by
the water between these two points. It was usually about 1° C. when

280 PROCEEDINGS OF THE AMERICAN ACADEMY.

it was a fall, aud somewhat less when it was a rise, as one would
expect from the fact that the jacket was usually warmer than the air
surrounding it. The mean temperature of the jacket probably differed
in most cases less than 1° C. from the temperature of the double fun-
nel, yy, which it almost completely sun-ounded. The space around
h h and // h' beneath the jacket, and the lower part of the jacket itself,
was protected l)y a thick layer of cotton wool. The flat top of the
jacket and the plugs and wires above it were protected in the same
way. Under these conditions, the cooling or heating of the water by
radiation or absorption at its envelopes betweenji and^o '^^'^.s probably
exceedingly small. As a further precaution against error from this
source (and from certain other possible currents due to chemical ac-
tion at jx and^aj ^oi" instance), the stream passing through the upper
chambers of the apparatus and over the upper surface of the disk was
alternately made warmer and colder than the stream passing across
the under surface of the disk.

The water-jacket J J was open to the air at A to allow the escape
of air-bubbles, which would otherwise have accumulated to such an
extent as to affect the flow of water at 0. Another outlet, at E,
served as an overflow when the water came too fast to be delivered
wholly through O. The streams from 0 and E were joined in the
flask F, Figure 5. These minutiae are of much importance, as the
need of great uniformity in the rate of flow is imperative.

The general arrangement of the apparatus, with the exception of
the electrical appliances, is shown in Figure 5. H^^ He,, //;, and H^
are very effective contrivances * for heating a stream of water by the
combustion of gas, which is supplied at constant pressure through two
gasometers G^ and G>>, one of which is shown in section. The water,
which came from a large tank at the top of the Laboratory, with
great regularity of flow when uninterrupted by obstructions in the
apparatus, and at a nearly constant temperature, passed through the
three heaters, H^, H^, and H.^ without division. Between H^ and IT^
the stream divided, one part, about one half, going through Hi to the
standpipe S^, the other part going without further heating to the stand-
pipe »S'i. These standpipes. from which a continual small overflow was
maintained, had the double function of maintaining a constant head for
the flow of the two streams through the conduction apparatus, and sift-
ing out much of the air, which the heating had driven out of solution

* Made by the Buffalo Dental Manufacturing Company, No. 347 of their
Catalogue for 1895.

HALL. — CONDUCTIVITY OF MILD STEEL. 281

in the water. "With the arrangement shown in Figure 5 the warmer
stream of water enters the upper chamber of the conduction apparatus,
and the cooler stream enters the lower chamber. This arrangement
was, however, used alternately with another, in which the water en-
tered S^ after leaving H^, the warmer stream going in this case to the
lower chamber of the conduction apparatus. One stream was usually
about 10° C. warmer than the other.

It is now time to describe and discuss more fully the electrical de-
vices and apparatus used for measuring, 1st, the difference in tem-
perature of the two sides of the steel disk, and, 2d, the difference in
temperature of the incoming and outgoing water of the upper cham-
ber of the conduction apparatus.

Calibration of the Copper-Steel Thermo-electric
Elements.

It has been stated that there were thirteen copper wires leading off
from each copper coating of the steel disk. Each pair of wires, one
above and the corresponding one below, represents a copper-steel
thermo-electric element, consisting of a piece of steel about 0.3 cm.
thick between two pieces of copper. To calibrate these elements in
situ, that is, to determine by direct trial upon the disk the e. m. f.
corresponding to a given difference of temperature between the upper
and lower surfaces of the iron, is apparently impossible. Accordingly,
two slender bars, one about 1 square millimeter in cross section, the
other about one third as large in cross section, and each 7 or 8 cm.
long, were cut from the same plate of steel from which the disk had
been made, and tests were made with these,* copper wires having been
soldered to the ends. The apparatus used for these tests is shown in
Figure 6. J3^ and B2 in this figure are vertical brass tubes through
which streams of water at any desired temperature may be made to
run. T^j and T^ are Baudin thermometers, remarkably similar in size

* Experiments, in wliich a thin bar of steel, of the same quality as those here
described, was subjected to longitudinal and to torsional stress, showed its
thermo-electric quality to be very little affected by such conditions. Fearing,
luiwever, lest the disk, which had been subjected to certnin processes which the
thin bars had not suffered, might differ from them in tlierrao-electric quality, I
have recently cut from the rim of a duplicate disk, made at the same time as
the other, two narrow strips, which I have compared directly with tlie thinner
of the two bars used in the copper-steel tests. Tliese strips proved to be so like
the bar in thermo-electric quality that it was difficult to make out with cer-
tainty any difference between them.

282

PROCEEDINGS OF THE AMERICAN ACADEMY.

:b.

T^j6

and reading nearly alike at all temperatures, each graduated to 0°.2 C.
The steel bars b and h\ and the copper wires soldered to them, extend

through the rubber stoppers rj
and Tn, so that the ends are
exposed to the streams of water
for a distance of about 1.3 cm.

To prevent disturbance from

*~ chemical action upon the steel
and copper, the parts exposed
to the water had been dipped
in a solution of copal in ether.
With this protection of the junc-
' tions it appeared from tests

made at various temperatures
that there was no chemical ac-
tion of sufficient magnitude to
affect materially the results of
the experiments made to deter-
mine their thermo-electric be-
havior. The tests for chemical
action were made by exposing
both ends of one bar to water
at the same temperature, some-
times with the bar in situ, as in
Figure G, and sometimes not.
The difference of temperature used in the thermo-electric tests was
commonly about 4° C. The thermometers were read, by means of a
telescope, to one tenth of one division, that is, to 0°.02. To eliminate
disaoreements of the thermometers, and various other possible sources
of error, the streams were regularly interchanged, other conditions
remaining the same. Thus, if a set of observations had been made
with Ti reading 28° and T2 reading 24°, another would immediately
be made with 7\ reading about 24° and T^ about 28°.

The following course of reasoning will show how this method of
using the thermometers tended to eliminate their errors from the
result. Let us suppose that 7\ is used at 24° and 7T, at 28°, then
7\ at 28° and T^ at 24°, and that afterward 7^ is used at 66° and
7; at 70°, then T, at 70° and T. at 66°. Let [^Joq be the error of
Ti at 20°, etc. Then we have in the various cases supposed : —

HALL. — CONDUCTIVITY OF MILD STEEL. 283

True Difference of Error

Temperature.

1st case 28 + [^ejes - 24 + \_E{\., [^o],s - [E^^^

2d case 28 + IE{\._, - 24 + [^2]-24 [^J^s - \.E.{\^,

Error of mean = ^ ([^i],3 - [^i].,,) + \ {^E;],, - \_E.;\^,)

Similarly, the error of the mean at the higher temjjeratures is

i ([^iJtO - [^1]C6) + h ([^2]70 - IE,-],,).

If the "error of the mean" had been equally great at all observa-
tion stages on the thermometers, it would have affected the absolute
value of the result obtained for the sensitiveness of the copper-steel
thermo-electric element, but would have left unaiFected the ratio of
the sensitiveness at one temperature to that of any other temperature.
Calibration of the thermometers showed that the " error of the mean "
was not the same everywhere on the thermometers. The variation,
though small, was perceptible, and might, if neglected, make the in-
terval called 4° at one stage one half of one per cent greater or less
than the interval called 4° at another stage. This error would not be
very serious in view of the general character of the experiments if it
had its full effect ; but, in fact, the conclusion as to the conductivity
of the steel depended upon the ratio of the sensitiveness of the copper-
steel element at any temperature to that of the copper-German-silver
element at or near the same temperature. Both kinds of element
were calibrated by means of the same thermometers, and therefore
the chance that final error of any magnitude could result from neglect
of the errors of the thermometers is exceedingly small. Accord-
ingly, such errors were neglected, the thermometers being used as if
correct.

The question of course arises whether the two junctions differed in
temperature just as much as the two thermometers did. It is no doubt
true, that, when temperatures much above those of the room were used,
each of the steel bars, several centimeters of which were exposed to
the air, was a little cooler at each end than the water flowing past it.
This difference was probably small, and must have been about the
same at each end. The disturbance, or inaccuracy, caused by it
should be smaller with the slender bar than with the other. Accorfl-
ingly the slender bar was used most of the time, and all the results
but one recorded in the following table were obtained with it. On
the last day of the observations on copper-steel junctions one of the
junctions on this bar broke. Then the thick bar was substituted for

284

PROCEEDINGS OF THE AMERICAN ACADEMY.

the Other, and the result given for the temperature 74°. 9 was thus
obtained. This result agrees well enough with the others to indicate
that there was no large difference in behavior of the two bars at the
high temperatures. Previous trial had shown that they gave nearly-
equal effects at low temperatures. The one value here given for the
thick bar was used with the others, obtained with the thin bar, in
plotting the calibration line, Figure 7, for copper-steel junctions.

lloo

1800

(^of Jter—^trman-Sififer

Zd' ' ' 30' ' ' ' ' w' ' ' j'o- ' ' 6cr' ' iff 8<r

»ic

tJOo^JSer- pteet

8ia

I0-' ' ' ' za' " ' ' 3ir' ' ' ■ifl- ' so'' ' ' 'iff' ' ' 10' so*

The electric currents were measured by means of a fairly sensitive
astatic galvanometer, the reduction factor of which was frequently
determined by use of a current of known strength.

Copper-Steel Junctions.

Date.

Temperature
of Koom.

Mean Temperature
of Junctions.

E. M. F.

of Temp.

per 1° C. Diff.
of Junctions.

Sept. 25,

1894

21

22.6

994 xlO-« volts.

Nov. 9,

((

12

12.7

1019

u

« 16,

a

14

11.8

1019

<(

Dee 8,

«

13

11.2

1034

<(

Nov. 9,

u

13

36.1

965

((

Dec. 8,

<(

13

40.6

949

u

Nov. 16,

((

15

71.4

868

li

Dec. 8,

((

13

69.5

873

((

« 12,

(?)-

18.5

74.9

866

it

HALL. — CONDUCTIVITY OF MILD STEEL. 285

"With these data the line given in Figure 7 was, obtained. It is a
straight line, the inclination of which , indicates a neutral point in the
neighborhood of 400° C.

Calibration of Copper-German-Silver Junctions.

It has already been stated that junctions of copper and German-
silver were used to measure the difference of temperature of the
incoming and outgoing water of the upper chamber. These junctions,
held firmly in the plugs p, />, of Figure 1, were tested in the same
apparatus and by the same method that had been used with the
copper-steel junctions, the ends of the plugs fitting into the short
side tubes occupied by rubber stoppers in Figure 6. The results
presently to be recorded were obtained with two pairs of junctions
called respectively No. 1 and No. 2. The same piece of german-
silver wire was used in both pairs. The copper wires of No. 2 were
not the same as those of No. 1, but all were taken from the same
spool, that is, all had once been parts of one continuous wire.

The following method was sometimes used in making the junc-
tions. The German-silver wire was heated in melted parafRne to the
neighborhood of 230° C, the object being to forestall any change of
its properties which might otherwise take place in the heat of solder-
ing. In order to make sure that this temperature was not exceeded
in the soldering, a bath of melted paraffine of about the same tempera-
ture was used to heat the soldering "iron." This device was aban-
doned after a time, experience seeming to show that it was unneces-
sary, and the junctions of Nos. 1 and 2, just mentioned, were not made
in this way. The junctions of No. 1 were not coated with copal to
protect them against chemical action of the water, but rosin was used
in making this pair of junctions, and this left a partial coating which
gave considerable protection. Experiments had appeared to show
that a protective coating was not necessary. A test for disturbance
due to chemical action with this pair of junctions was satisfactory, no
evidence of such an efi^ect being found. In making No. 2, however,
the junctions were carefully coated with shellac, dried on at a tem-
perature of about 80°. All of the conduction experiments of which
the results will be given in this paper were made with No. 2, except
those of August 13, which were made with a similar pair.

286 PROCEEDINGS OF THE AMERICAN ACADEMY.

Copper- German-silver Junctions.

Junct

ion;

3. Date.

Mean Temperature
of Junctions.

E.

M. F. in Volts per 1° C. Diff.
of Temp, of Junctions.

No.

1

Dec. 27, 1894

9°6^

9.6 -9°.2
8.4)

1668

X 10-

")

" 28, "

1677

a

y-1679 X 10-

Jan. 1, 1895

1691

u

)

Dec. 28, 1894
Jan. 1, 1895

39-n39°.5
39.3)

1858
1838

1 1848 "

Dec. 28, 1894
Jan. 1, 1895

o

73.61^30^
73.2 )

2007
2022

1 2015 «

No.

2

Aug. 29, 1895

o

25.4

1746

ii

((

(I k(

43.4

1849

a

«

(( ((

59.1

1936

it

August 31, 1895, further observations were made on No. 2 at 25°. 4
and at 59°. 2, but they were comparative only, as the reduction factor
of the galvanometer was not determined on this day. If we assume
the effect at 25°. 4 to have been the same as on August 29, we get

o

No. 2 August 31 25.4 1746

" " 59.2 1951

It is found by interpolation that No. 1 would give for

25.4

1772

43.4

1871

59.1

1949

The differences between the results obtained with the two pairs of
junctions are probably within the limits of e.\;perimental error, and it
appears better to use the results from both pairs in plotting the cali-
bration curve than to use those from No. 2 only. No great accuracy
is claimed for the absolute values obtained in any of the calibration
tests. The calibration curve of No. 1 is shown as No. 1 in Figure 8.
It is very slightly convex upward. The observations of August 31,
reduced to absolute measure by means of the assumption already
stated, are used, with those of August 29, to give data for the calibra-
tion curve of No. 2. This curve, No. 2 of Figure 8, is very slightly

HALL. — CONDUCTIVITY OF MILD STEEL.

287

concave upward. Line No. 3, a mean between No. 1 and No. 2, is
practically a straight line, which, if continued, would indicate a neutral
point in the neighborhood of the absolute zero. In subsequent work
this stiaight line is assumed to be correct, and data needed for use in
the conductivity calculations are taken from it.

The Electric Circuits.

Figure 1 shows that the fine copper wires connected with the upper
and under copper coatings of the steel disk led out to connections
with heavier copper wires on a wooden shelf surrounding the apparatus
carrying the disk. Some slight thermo-electric effect could be pro-
duced by warming one of these connections, but as the wires from
both faces of the disk were connected here in the same way, the con-
nections being all equidistant from the disk, the thermo-electric forces

Fig. 9.

existing in them must have neutralized each other in the ordinary use
of the apparatus. The copper wires leading away from these connec-
tions were all of the same length, size, and quality. They terminated
in a kind of commutator, or switch-board, shown in Figure 9. The
numerous small circles in this figure indicate small holes containing
mercury. Holes 1, 2, 3, 4, etc. are connected, by means of the wires
just mentioned, with junctions 1, 2, 3, 4, etc., respectively, on the
disk. A connector, G, brings the four junctions 1, 4, 7, 10, which
are (see Fig. 4) all at the same distance from the centre of the disk,
into multiple arc, and takes the currents coming from them to one

288 PROCEEDINGS OF THE AMERICAN ACADEMY.

wire leading, through the mercury cup, W, to a galvanometer. Con-
nector C" does the same for junctions 1', 4', 7', and 10'. The same
connectors can evidently be used to bring junctions 2, 5, 8, 11, and 2',
5', 8', 11' into action, or 3, 6, 9, 12, and 3', 6', 9', 12'. The central
junctions, 13 and 13', were used by themselves.

The apparatus, beyond this switch-board, does not require detailed
description. The resistance coils used were reduced to legal ohms
by comparison with tliose of an Anthony standard (Queen & Co.).
Thermometers at two or three points along the circuit gave such
account of the temperature as was necessary. The galvanometer was
a Thomson astatic, with a concave mirror. The strength of current
required to produce a deflection of 1 cm. on the scale was about
12 X 10"^ amperes.

The circuit containing the copper-German-silver junctions, ^i and j\^
of Figure 1, is simpler and requires little further description. The
galvanometer of this circuit had a plane mirror, and was used with tele-
scoi^e and scale. Its sensitiveness to a given current was about the
same as that of the first galvanometer. The circuit included, at a
point corresponding to the switch-board of the other circuit, mercury
cups at which the connection could be broken.

In addition to the mercury cups already mentioned, each circuit
included a mercury commutator, by means of which the course of the
current in the galvanometer could be reversed. These mercury cups
and commutators were a source of much disturbance, and toward the
last of the work solid metallic connections were in some cases substi-
tuted for them. The difficulty was thermo-electric. If two copper
wires run into the same mercury cup at different levels, or if the two
wires are of different size or diflferent external condition, there will
be at the copper-mercury junctions, unless the temperature of the
room is more than usually uniform and constant, thermo-electric
effects which cannot be neglected in delicate measurements of electric
current. There was a certain approach to uniformity in these dis-
turbances, which made it possible to eliminate them in great part
from the result by the device, already described, of running the
warmer stream of water alternately above and beneath the disk.
The full measure of this source of error was, however, not realized
until the last week or two of the work, and it may be that the dis-
cordance between the results of August 13 and 15 and those ob-
tained later was due to lack of care in protecting the copper-mercury
junctions from changes of temperature on the days mentioned.

HALL. — CONDUCTIVITY OP MILD STEEL. 289

Greater vigilance, and probably greater accuracy, marked the later
work.

The difficulty here discussed can be very much reduced by a more
sparing use of gas flames in the room.

The Observations.

As an example of the observations in the main experiment, showing
their character and their arrangement, the last set made will now be
given, some unimportant details being omitted.

The letters L and R refer merely to the position of the commuta-
tors. A change from L io R reverses the current in the galva-
nometer concerned. In each circuit the first observation is made
under Z, the second under R, the third under L, etc.

The note " 1, 4, 7, 10," means that the junctions of the outer circle
of the disk (see Fig. 4), both above and below, are in use, connected
as in Figure 9. After this set of junctions, those of the next circle,
2, 5, 8, 11, are used, etc.

Each group of seven or five observations in the first circuit was
made between two groups of three observations in the second circuit.
Before the first, third, fifth, etc., groups of three in the second circuit,
the main outlet of the water from the lower chamber of the apparatus
was closed for a few seconds to wash out, through the overflow tubes
(Fig. 1), the air-bubbles that might have accumulated beneath the
disk. This operation did not greatly affect the current in the secona
circuit. The course of operations gave the first circuit about two
minutes for recovery after the disturbance before observations in this
circuit were renewed.

The water had been running through the apparatus for about an
hour, as usual, before the observations began.

VOL. XXXI. (n. 8. XXIII.) 19

290 PROCEEDINGS OF THE AMERICAN ACADEMY.

Junctions. Copper-Steel Circuit. Copper- German-silver.

i. Jl. L. H.

1,4,7,10 36.28 ^^^^x 46.98

30.76 ^
30.80 V

36.20 .. .. f ^_^^ 46.90

36.14

36.15 -— / 46_82

30.80 j

46.84

^^•1^ Q, «, I 46.90

35.11

35.66
35.66

46.76
1,4,7,10 35.98 g^^^s 46.70

40.85 y 6.09

46.90 40.92]- 5.94

46.94 )

2,5,8,11 35.80 g^^^^x 4g9Q 40.85 J- 6.07

^^•"^^ 31.04 > 4.73

35.76 46.95 )

35.78 ^^•'^^-' 46.87 40.86 1 6.05

40.78 i 6.07

3,6,9,12 35.15 ^^,^. 46.86 ^^•'^J

31.58 V3.54

q, fi, ( ' 46.90 )

''■''} 46.96 40.88 1 6.05

lo o^ -o 47.02 •)

13 34./ 8 o, oO ^7nA 40.95 U.06

34.80 f^-^n2.88 ^'-^^ ^

31.91 ^

34.84

46.85 )

46.91 41.00 I 5.88

3,6,9.12 35.12 „, ,^ x 46.70 .noaKoi

35 10 ^1-^^) 46.88 40.98 [-5.81
"' '^ 3.54

35.14

S)10 "51.00; 46.75 .

46.81 41.10 ;. 5.68

2, 5, 8, 11 35.70 X 46.78 )

33.68 ''■'-'],., 46.74 ^'-'^'^-ee

30.98 >4.73 ^

30.96 j 46.80 ->

,^ „^ 41.14-5.61

46. /O j

41.30 ^ 5.43

^.^^ 30.66 y 5.25 ^^^^

3<'>-90 ^^„^ ( 46.65 )

^^,, 30.70 ,,^. 41.18 U.47

35.90
35.90
35.98 — — ; 4gg5 -xx.xo^-

Time occupied, 5:39 to 6:35 =56 minutes

Mean, 5.85

HALL. — CONDUCTIVITY OF MILD STEEL. 291

The following observations accompanied those just recorded : —

t = temperature of water running from lower chamber.
tx = " " at thermometer T'j, Fig. 1.

>S' = number of seconds required to fill flask F\ Fig. 5.

^ -— 44 4« " U p U

The capacity of each flask was about 1800 grams. The numbers in
brackets in the column headed S^ indicate the number of grams of
water escaping per minute through the overflow tubes of the lower
chamber. In the column aS is a similar note for the upper chamber.

Time.

t

<i

U

S'

iS

5:33

22°.0

32°.8

31°.9

5:35

70.5
[300]

^2-H72.8
73.0 i

6:02

22°.0

32°.5

31°.7

72.5

6:37

21°.8

32°.2

31°.4

6:40

71

[290]

^^•H73 5

73.5 r^-^

[50]

Means 21°.9 32°.5 31°.7 70.8 72.9

The 50 grams per minute noted as the overflow from tlie upper
chamber is disregarded, not because it is inconsiderable in comparison
with the main stream, which carried about 1400 grams per minute,
but because it was taken off without having been brought near the
disk, so that its temperature must have been practically unaffected by
its admission to the chamber.

Assuming, for our present purpose, that no corrections need be
applied to the thermometer readings, we see that the water escaping
from the jacket is about 0°.8 colder than it was when it entered the
upper chamber, that the difference of temperature of the streams in
the two chambers is about 10°, and that the temperature of the disk
must be about half way between those of the streams, 27°, let us say.
The main object in noting the difference between ^j and U is to get
information as to the effectiveness of the jacket that surrounds the
upper chamber. If, as in this case, the temperature of the jacket
differs only 1°, or thereabout, from the temperature of the chamber,
the protection afforded must be nearly perfect. Apparently the differ-
ence was little, if any, more than 1° in any case.

292 PROCEEDINGS OP THE AMERICAN ACADEMY.

The Calculations.

If the deflections obtained from the junctions at different points on
the disk were all equal, or nearly so, the calculation of thermal con-
ductivity from the observations would be exceedingly simple. Un-
equal as these deflections are, it would still be allowable to take their
mean as representing the mean difference of temperature between the
upper and lower surfaces of the steel disk, if all junctions represented
equal areas of the disk ; but this is not the case. The process of find-
ing what the mean deflection would be, if all parts of the disk were
equally represented, is indicated by the following formula, in which
A stands for the mean deflection desired ; 8 for the deflection given by
a particular pair of junctions ; cr for the area of cross-section repre-
sented by this pair of junctions *, S for the area of one face of the
disk : —

A = — — = — 7^ .

If there were many pairs of junctions, but only one pair at a given
distance, r, from the axis of the disk, each pair would represent a cer-
tain ring-shaped portion of the disk, the area of which at either sur-
face is 2 7rr X dr. Writing this expression in place of cr, and calling
the radius of the disk E, we get

-^ I 8 X r X dr.

"We can approximate to this integration by means of a graphical pro-
cess, making use of the data given by the observations actually made.
From the series of observations given on page 290, we get the follow-
ing values ; —

r 8

0 2.88

2 cm. 3.54

3.5 " 4.73

4.5 " 5.33 [= i (5.40 + 5.25)]

The observed value of S for r = 0 is to be increased a little to make
allowance for the fact that the resistance of the circuit, including
junctions 13 and 13', was somewhat greater than the resistance of the

HALL. — CONDUCTIVITY OF MILD STEEL. 293

circuit wheu the other junctions were in use. Making this slight
change, and then taking the values of r as abscissas, and the values
of 8 as ordinates, we get the four determining points of curve A in
Figure 10. Taking various ordinates from this curve, and multiplying
each by the corresponding value of r, and, for convenience in plotting,
by 0.2, we get the determining points of the curve a. The area un-
der this curve is evidently

0.2/8 Y.r'>idr,

0

from which the value of A is at once obtained. In this case A = 4.52.
The mean of the observed values of 8 in the series under discussion
is 4.12. All this is for the case where the water above the disk was
warmer than that below the disk, on September 2.

The same day another series of observations was made, with the
warmer stream beneath the disk. The curves B and h (Fig. 10)
have to do with this second series of observations. They give for the
value of A, 4.60.

It is assumed in the subsequent calculations that A is equal to the
deflection which would be obtained from any pair of junctions on the
disk, one above and one below, if the difference of temperature were
made, for all values of r, equal to the difference between the mean
temperature of the whole upper face and the mean temperature of the
whole lower face of the steel disk. The validity of this assumption
will be discussed later.

In the calculation presently to be made, the subscript 1 will have
reference to the circuit containing the junctions on the disk ; the sub-
script 2 will have reference to the circuit containing the copper-
German-silver junctions.

The test for " sensitiveness " of the galvanometers, which has not
been described at length, was made by running the same current
through the two simultaneously. The " reduction factor," by which
the deflection in centimeters must be multiplied to give the strength
of this current in amperes, might be given for each galvanometer in
numerical terms ; but it will be evident that, for the calculation in
hand, it is necessary to know only the ratio of the two reduction fac-
tors, which is the reciprocal of the observed simultaneous deflections
of the two galvanometers in the sensitiveness test.

In finding the value of m, allowance is made for the change of
capacity of the measuring flask with change of temperature of flask
and water.

294 PROCEEDINGS OF THE AMERICAN ACADEMY.

We have : —

h = the thermal conductivity in c. G. s. units.

Q = number of heat units flowing through disk in one second ;

T = thickness, in centimeters, of the steel disk ;

S = area, in square centimeters, of horizontal cross-section of the

steel disk ;
t = mean difference in temperature of sides of steel disk ;
m =z mass, in grams, of water traversing upper chamber per second ;
Ai = mean deflection in circuit 1, as before described;
Ag = mean deflection in circuit 2, in conduction test;
Jij^ = reduction factor of galvanometer 1 ;
^2 = reduction factor of galvanometer 2 ;
di = mean deflection of galvanometer 1 in sensitiveness test ;
do = mean deflection of galvanometer 2 in sensitiveness test ;
[i?i X c?! = -ff 2 X c?2j and so Rx -=r R« ^= d^ ■— d(\ ;
Ex = e. m. f., in volts, of copper-steel element per 1° C. difference of

temperature ;
^2 = 6. m. f., in volts, of copper-German-silver element per 1° C.

difference of temperature ;
Vx = resistance, in legal ohms, of circuit 1 in conduction test ;
To = resistance, in legal ohms, of circuit 2 in conduction test ;

Then

Ao X i?o X ?-o

Sxt A, X 7?i X rx

^^ £x

„ R. To Ex\ T
= I «^ X ^ X ^ X ^- X ^ J X -.

( ^2

Using this formula with the data for series A and B of September 2,
we have : —

HALL.

CONDUCTIVITY OF MILD STEEL.

295

A (warm water above.)

Ji (warm water below. j

log.

log.

m = 24.57 1.3904

m - 24.58 1.3906

A2= 5.85 0.7672

A. = 4.91 0.6911

d^ - 10.48 1.0204

c/i = 10.48 1.0204

7-2 = 28.15 1.4495

r^ = 28.14 1.4493

E^ = 985 X 10-8 [26°. 9] 6.9934

El = 985 X 10-8 [27°.l] 6.9934

log.

1.6209

log.

1.5448

Ai= 4.52

0.6554

Ai = 4.60

0.6625

f/2= 8.38 '

0.9191

(/, = 8.80

0.9191

7-1 = 16.58

1.2196

7-1 = 16.58

1.2196

E2 = 1796 X 10-8 [320.1J

5.2543

2.0484

£2 = 17.39X10-8[21°.8]

5.2403

2.0415

T __ 0.2952
S ~ 77 9

1.5725
3.5786

1.5033
3.5786

1.1511

1.0819

1.1511 = log. of 0.1416 = log. of ka 1

1.0819 = log. of 0.1208 = log. of h

1

Meaa =

0.1312 = k

Thus, according to the observations of September 2, the conduc-
tivity of the steel at 27° C. is 0.1312.

The difference between k^ and k^, is large in this case, but it has al-
ready been explained that neither is intended to be complete by itself.
The mean of the two is necessary to eliminate certain errors which have
been discussed. It may well be doubted whether this elimination is per-
fect when the difference between k^ and kf, is so large. It will presently
appear that the difference was exceptionally large on September 2.

The following table includes all the results obtained with the appa-
ratus in its final condition. It has already been stated that the results
of August 13 and August 15 do not agree with those found later, and
a possible reason for this has been given, namely, lack of sufficient
precaution, in the earlier experiments, against error from disturbing
thermo-electric effects external to the conduction apparatus proper.
Although the results obtained August 13 and 15 are here given, they
are not included in finding the mean values given at the foot of each
column. At one time it was suspected that an accumulation of dirt
forming near the edge of the upper face of the disk, and checking the
flow of heat there, might account for the discrepancy between the
earlier and the later results. Accordingly, between August 28 and
September 2 the conduction apparatus was dismounted, and the disk
was cleaned and replaced. But the result obtained September 2 was
in close agreement with most of the others.

296

PROCEEDINGS OF THE AMERICAN ACADEMY.

Conductivity, k, of Open-Hearth Steel in C. G. S. Units.

Date.

Temperature.

(Warm above. )

(Warm below.)

C"!'-')

Aug.

13,
17,

1895

[27°.0]
27°.4

[0.1463]
0.1301

[0.1445]
0.1392

[0.1454]
0.1347

il

20,

a

27°.8

0.1333

0.1334

0.1334

u

22,

li

26°.3

0.1334

0.1310

0.1322

((

28,

n

27°.4

0.1306

0.1310

0.1308

Sept

2,

a

27°.0

0.1416

0.1208

0.1312

Mean

, 27°.2

0.1338

0.1311

0.1325

Aug.

24,

1895

44°.2

0.1299

0.1343

0.1321

Aug.

15,
17,

ik

[62°.8]
59°.8

[0.1264]
0.1242

[0.1463]
0.1347

[0.1364]
0.1295

a

22,

((

59°.0

0.1301

0.1287

0.1294

li

26,

11

58°.8

0.1296

0.1327

0.1312

)>0.1347

^0.1316

Mean, 59°.2 0.1280 0.1320

0.1300

Discussion and CRiTicis.^r.

Taking 0.1325 * as the value of k at 27°.2, and 0.1300 as the value
at 59°. 2, we get as the coefficient of change with rise of temperature

* Winkelmann's

Ilandhuch der

Physik, 1895,

gives

: —

Metal.

Temp.

k

Observer.

Eisen

0°

0.1G55

Lorenz.

"

0°

0.1 U88

AngstWim.

"

iiber 0°

0.1587

Berget.

"

ca. 15°

0.1648

Neumann.

"

ca. 15°

0.1133

Wiedemann u. Franz.

'«

0°

0.1509

Mitchell.

K

0°

0.172

Stewart.

Schmiedeeisen

0°

0.2070

Forbes.

"

39°

0.1485

H. F. Weber.

Stahl hart

ca. 15°

0.062

F. Kolilrausoh.

" weich

1-5°

0.111

"

«

15°

0.1104

Wiedemann u. Franz.

Puddelstahl

15°

0.1

375 - 0.1418

Kirchoff u. Hausemann.

Bessemer stalil

15°

0.0946

« «

Mangan stahl

0°

0.03280

Mitchell.

It is probable that open-hearth steel resembles the iron of this table more
nearly than it does the steel.

HALL. — CONDUCTIVITY OF MILD STEEL. 297

0.1300-0.1325 ^^,^,^,

y = 0.1325 X (59.2 - 27.2) = -^-OOOb-.

If we take the values 0.1347 and 0.1316, obtained by including the
bracketed values in finding the means, we get y = — 0.0007+. The
value — 0.0006 is to be iireferred, although it is evident from the pre-
ceding pages that no great confidence can be placed in the digit of
this coetficient. A wider range of temperature, a greater number of
measurements, and evidently a more careful calibration of the copper-
German-silver junctions is needed for a completely satisfactory deter-
mination of y. Moreover, no allowance has here been made for the
change in the specific heat of water between 27° and 59° C. It is to
be observed that the lower temperature is very near the temperature
of minimum specific heat of water. If it should be found that the
specific heat of water at 59° is two per cent greater than at 27°,
which is, to be sure, improbable, the value of y obtained from these
experiments would be 0. If the difference is one third of one per
cent, which seems not unlikely, it will reduce y to — 0.0005. If we
take this as the most probable value, it will, in spite of its uncertainty,
be not devoid of interest, for the tables in Winkelmann give no values
of y in the case of steel, while for iron they give

y — — 0.000038 (Lorenz),
y = — 0.000517 (Angstrom),
y=: -0.0011 (Stewart).

The one measurement of k for an intermediate temperature, 44°. 2,
lies between the mean value for 27°. 2 and the mean for 59°. 2, which
is gratifying evidence, so far as it goes, of the precision of the meas-
urements.

It is a curious fact that every measurement of Jc near 27°, except
the last, gave a smaller value than the preceding measurement. The
reason for this is not evident.

At the low temperature the mean of h^ is a little greater than the
mean of i'j,' At the high temperature the contrary is true. In either
case the difference is to be regarded as mainly accidental, and not
significant, — a fact already mentioned and explained.

"We must now consider more fully the question whether A, as de-
duced from curves like those of Figure 10, represents accurately the
difference between the mean temperatures of the upper and of the
under face of the steel disk. It can hardly be doubted that A, derived

298

PROCEEDINGS OP THE AMERICAN ACADEMY.

"yafues oj r ^

as it is from thirteen well distributed pairs of junctions, represents
with considerable accuracy the mean difference of potential between

the two faces of the steel disk ;
4 but is this mean difference of

potential proportional to the
mean difference of tempera-
ture between these two faces ?
This question would be at
once answered in the affirma-
tive if the disk, instead of being
metallically continuous, were
made up of little columns of
steel, in length equal to the
thickness of the disk, tipped
at both ends with copper, each
column being insulated from
its neighbors. For in such an
aggregation there would be no short-circuiting, and the difference of
potential at the two ends of any steel column would depend merely
upon the difference of temperature at these ends ; and, with the very
small difference of temperature existing, only a fraction of one degree,
the potential difference would be very strictly proportional to the tem-
perature difference.

But the disk is continuous, and there must be short-circuiting cur-
rents within it. IIow will these currents affect the mean difference
of potential of the two faces ? These currents in the steel will be
partly vertical and partly horizontal. "We will consider, first, the

vertical components. Let Figure 11
represent any two equal vertical ele-
ments, S-^ and /So, of the disk, connected
by the copper strips Ci and Co. Let
the e. m. f of element *S'i be E^, directed
upward, and the e. m. f. of S^ be ^j)
also directed upward, and let ^i > E^. If connection between ele-
ments were broken, the difference of potential betweeen the top and
bottom of Si would be E^, and that between top and bottom of tS'2
would be E^. The mean difference of potential would be \ (Ei -f E2).
Let the resistance of *S'i, equal to that of S2, be called r. When
connection exists, as in the figure, we have a short-circuiting current,
C, passing up through Si and down through S^. With this current,
the difference of potential between top and bottom of Si is Ei — rC,

c

c.

Fig. 11.

.5,

HALL. — CONDUCTIVITY OF MILD STEEL.

299

and the difference of potential between the top and bottom of S2 is
E^ -\- rG. The mean of these two differences is \ (E^ + E^), the
same value as if there were no short circuit.

If the element S.^, instead of being of the same size as Si had a
cross-section n times as great, it would have n times as much weight
as Si' in reckoning the mean difference of potential, which would
then be

-4-, (^1 + « ^2)

with open circuit, and
1

« + 1

Ei-rC]-\-n[E,

n

n + ]

(Ei + n e\

with short circuit. This appears to show that the vertical components
of short-circuiting currents within the steel disk do not affect the
mean difference of potential between top and bottom.

iKt^V""-^

^

Aug. 22— ■

\ t

/yf

^

Mean \

Aug. 26--'
Aug. 28-- •

.>S

^

Aug. 17''

.Aug. 13
Sept. 2

. Mean
■-.Aug. 22

Fig. 12.

The non-effect of the horizontal components is shown by the con-
sideration that a horizontal current, maintained in the disk by exter-

300

PROCEEDINGS OF THE AMERICAN ACADEMY.

nal or internal means, would have, of itself, no tendency to establish
any mean difference of potential between the ujjper and under faces
of the disk.

It appears, therefore, to be established that A does represent — that
is, is proportional to — the mean difference of temperature between the
upper and under faces of the disk.*

It may be of interest to compare the difference of potential curves
of the various series of experiments. If in Figure 10 we take the
mean of A and B, we get a new curve, which may be taken as repre-
sentative of the potential conditions, or difference-of-potential coudi-

Aug. 22

tions, prevailing in the experiments at low temperature on September
2. Combining each A curve for low temperatures with its correspond-
ing -B curve, and plotting the resultants, we get Figure 12, except the
heavy line, which is obtained by taking the mean of all the others

* Cf . an article, by Prof. B. O. Peirce, " On the Properties of Batteries
formed of Cells joined up in Multiple Arc." These Proceedings, Vol. XXX-
p. 194.

HALL. — CONDUCTIVITY OF MILD STEEL. 301

except that of August 13. This heavy line is, therefore, the typical
difference-of-potential curve for low temperatures, about 27°. The
heavy line of Figure 13 is similarly the typical curve for high tem-
peratures, near 59°, the curve for August 15 not being used in ob-
taining this composite.

The difference between these two typical curves is noticeable. Each
indicates that the greatest difference of potential is near the edge of
the disk. Each shows a minimum about two thirds as great as the
maximum, but the minimum of the high temperature curve occurs
farther from the centre of the disk. The reason for this dijBference is
not obvious. The rapidity of the flow of water was about the same
for high temperatures as for low, but there was a greater tendency to
development of bubbles in the warmer water, and this may have in-
fluenced to some extent the course of the water over the surfaces of
the disk.

The low temperature curve of September 2 is in marked contrast
with most of the others in its group, showing a relatively great differ-
ence of potential at the edge of the disk. It has already been stated
that the copper coatings of the disk had been cleared of accumulated
dirt shortly before the experiments of vSeptember 2. This dirt had
lodged mostly near the edge of the disk, and its removal apparently
had just the effect that would have been expected upon the flow of
heat in that region. The value of k obtained September 2 accords
well with those obtained shortly before the disk was cleaned.

As to the absolute values found for k, they are very likely in error
to the extent of two or three per cent of their own magnitude, pos-
sibly more. A source of possible error not yet discussed is found at
the ring which encloses the disk (see Fig. 1). The effective diameter
of the disk is taken to be the distance through the ring and disk,
from the groove on one side to the groove on the other. The ring
itself in this grooved part is about 0.08 cm. thick. Its cross-section
is about three per cent of the whole conductive cross-section. The
material of the ring is very similar to that of the disk, if not quite
the same. The uncertainty as to whether the thin part of the ring
can properly be assumed to act like an equal portion of the disk
itself, comes from the dubious nature of the approach to this part of
the ring above and below. The flow into or from it may be partly
lateral, through the edge of the disk proper, but it cannot be wholly
so, for a narrow strip of the thicker portion of the ring is exposed to
the water above and below the disk. It is not likely that the error
here is large, but in future experiments a different method of sup-
porting the disks will probably be used.

302 PROCEEDINGS OF THE AMERICAN ACADEMY.

It may be asked whether the method of investigation which has
been described at such length in this paper will be applicable to the
case of metals in general, and whether it can be extended through
any large range of temperature. It is evident that the thermo-electric
device for determining the difference of temperature between the two
faces of the disk must be used with great caution, and will in some
cases be difficult of operation ; but its use is by no means restricted to
iron. Copper and other metals thermo-electrically near it will cer-
tainly give some trouble, though it is believed that they can be suc-
cessfully dealt with. It is probable that the range of temperature
can be greatly increased by the use of other liquids instead of water ;
but if this is done, the question as to the specific heat of the liquid
will become a difficult one.

The method described is a laborious one in its preparatory stages,
but much of the initial work can be done by a mechanic. The opera-
tion of the apparatus, once prepared, is comparatively easy, and the
calculations are simple. A single experimenter operates the whole
machinery, and makes all the observations. He can make the A and £
observations for a given temperature and calculate the value of k
therefrom in one day, without excessive toil, and with a reasonable
assurance that the result will be in accord with that of the day
before.

This investigation was begun, at the author's suggestion, some years
ago by IMr. A. W.'Slocum, then a graduate student at Harvard, now
of the University of Vermont. Mr. Slocum worked with great zeal
and energy, and made substantial progress, but presently went abroad,
leaving the research in my hands.

I have to thank all members of the Physical Department at Har-
vard, who have given me extremely valuable suggestions in regard to
my work, and have furthered it by every means in their power.

DAVIS. — OUTLINE OF CAPE COD. 303

XVL

THE OUTLINE OF CAPE COD.

By William Morris Davis.

Presented ^larch 11, 1896.

Summary.

This essay attempts to restore the original outline of Cape Cod by reversing
the processes at work on the present outline (p. 308). In order to gain good
understanding of these processes, a review of previous accounts of tlie Cape is
introduced (p. 304), a general consideration of the development of sea-shores is
outlined (pp. 312-317), and the conclusions reached are applied to the problem
iu hand (p. 318). It is thus estimated that the land here once extended at most
two or more miles into the sea on the east, and that perhaps three or four thou-
sand years have been required for the retreat of the shore line to its present
position (p. 326). This period cannot, however, be taken as a full measure of
the time since the glacial deposits of tlie Cape were formed, for there is reason
to believe that the land stood higher than now for an unknown interval between
the building of the Cape and the assumption of the present attitude with respect
to sea level.

The chief interest in the problem here discussed turns on the growth of the
great sand spit of the " Pro vincelands " northwestward from the "mainland"
of the Cape (p. 312), and on the protection thus afforded to the old cliffs of
High head. Brief account is given of the growth and waste of the Provincelands
(p. 323), and of the changes of the western shore line (p. 829). The essay closes
with some practical suggestions regarding the protection of Provincetown harbor
(p. 329), and some speculations concerning the future change of the Cape. The
consumption of the north arm — from the elbow to the hand — will probably
require about eight or ten thousand years (p. 331).

Introduction.

An excursion to Provincetown and the " mainland " of Truro on
Cape Cod with the students of the Harvard summer course in Physi-
cal Geography, in July, 1895, brought to my attention a number of
problems concerning the changes of outline suffered by the Cape.
These problems had taken rough shape on the occasion of a visit to
the peninsula several years ago. Supplementing the observations
made on the ground by a study of the Coast Survey charts and by a
review of what has been written on the subject, the following essay

304 PROCEEDINGS OF THE AMERICAN ACADEMY.

has gradually grown up. Its substance was presented before the
Geological Society of America at the winter meeting in Philadelphia,
December, 1895, and again before the Harvard Geological Confer-
ence in April, 1896.

The end of the Cape is pleasantly reached by a four-hour run in a
steamboat from Boston across Massachusetts bay to Provincetown, in
whose neighb6rhood the most significant of the features here described
are to be found. By driving to High head, the northernmost point of
the "mainland," a general view of the peninsula of Provinceland may
be gained: thence driving or walking to Highland light, one may see
a portion of the long harborless cliff that forms the " back " or eastern
side of the Cape. Walking northwestward along the beach to Peaked
hill life saving station, the action of the surf can be observed at
leisure ; and thence crossing the sandy belt to Provincetown, the
varied forms of the dunes can be studied in detail. A second day
may well include a visit to Race point, the northwestern extremity of
the Cape, and a return southward along the wasting shore to Wood
end, or Long point, whence the town can be regained by boat, pre-
viously arranged for.

Cape Cod is an excellent region for the study of shore forms in the
light of their development from some antecedent outline, and tlieir con-
tinued change towards some future state. Although the " mainland "
of the Cape rises about two hundred feet above sea lev^el, it is built of
uncompacted clays and sands, with occasional boulders, and is there-
fore easily consumed by the waves. Standing far out beyond the
general shore of New England, it receives a violent attack from storm
waves, which alter the shore line so rapidly that the changes are
measurable even in the short time covered by our records.

Extracts from Previous Writings.

The following extracts summarize a number of previous references
to the Cape.

In the Geology of INIassachusetts (I., 1841), E. Hitchcock makes
brief mention of the erosion on the eastern coast and the growth of
the Cape into Massachusetts bay (323), the southward growth of
Nauset beach, a mile in fifty years (324), the dunes of Provincetown
(325), and the "diluvial elevations and depressions " of Truro (367) ;
Provinceland is "alluvial ; that is, washed up by the ocean" (371).

Lieut, (afterwards Admiral) C. H. Davis wrote a " Memoir upon
the geological action of the tidal and other currents of the ocean"
(Jlem. Amer. Acad., Boston, 1849, IV. 117-156), in which he called

DAVIS. — OUTLINE OF CAPE COD. 305

attention to the repeated occurrence along our coast of bars built
northward from coastal bluffs, such as Sandy hook, N. J., and Cape
Cod, and suggested that " a generic term " should be applied to these
forms. He mentioned a place of division of the tidal currents on the
east side of Cape Cod, near Nauset inlet, from which the flood tide
flows north and south.

Thoreau's narrative of his excursions on the Cape in 1849, 1850,
and 1855, tells of various changes in the coast line known to the
people there. A log canoe, buried long before on the inner side of
the bar that forms the eastern wall of the marshy East harbor at the
north end of the mainland, was found many j'ears afterwards on the
Atlantic side of the bar ; that is, the bar had been pushed westward
over the buried canoe as the sea cut away the outer beach. Swamp
peat was sometimes found on the exposed beach, although it was
originally formed undoubtedly on the inside of the bar. Stumps had
been seen off Billingsgate point ; the implication being, not that the
land had been depressed, but that it had been washed away, leaving
the stumps mired in their native soil.* A writer in the " Massachu-
setts Magazine " of the previous century is quoted to the effect that
an island, called Webbs island, formerly existed three leagues off
Chatham, containing twenty acres of land ; the people of Nantucket
carried wood from it ; but in the writer's day a large rock alone
marked the spot, and the water thereabouts was six fathoms deep.
(Cape Cod, in New Riverside edition of Thoreau's works, 1894, pp.
182, 183.)

Freeman's History of Cape Cod (1860) attributes much wasting of
land to reckless cutting of the trees, — a doubtful conclusion as far as
it refers to shore work, although probably applicable to the interior
district of the dunes. He says: "'The work of devastation was too
extensively accomplished ; as is seen on the shores of the Cape since
washed away by tides aided by the force of the winds, so that vast
flats of sand extend in some places a mile from the shore, now, at low
water, dry, or nearly so, and in some instances these flats disclose
large stumps of ancient trees embedded in their native peat " (752).

H. L. Whiting prepared a " Report on the special Study of Prov-
incetown Harbor, Mass." (Rep. U. S. Coast Survey, 1867, pp. 149-
157). He distinguishes Truroland, the mainland of the Cape, " by the

* I have found this explanation of the occurrence of tree stumps on the shoals
off Chatham current among the fishermen of the Cape. See Proc. Bost. Soc.
Nat. Hist., 1893, XXVI. 173.

VOL. XXXI. (n. s. xxin.) 20

306 PROCEEDINGS OF THE AMERICAN ACADEMY.

existence of clay and of boulders, and by the peculiar form of the
' bowl and dome ' drift " ; and Provinceland, " of sand only, — so free
from all earthy matter that it will not even discolor water, — while the
forms which the dunes and ridges here assume are mainly characteris-
tic of wind drift " (155). He concludes that " the outer ridges of the
peninsula of Provincetown were the earliest in date, and that the flats,
marshes, and ponds now existing are subsequent accumulations and
accidents, which have taken place under the shelter and eddy influ-
ences of the outer hooked bar or beach " (155). The narrow outer
bar that connects the cliffs of Highland liglit with the Provincetown
peninsula is described as wasting back with the cliffs, and is said to
be in danger of breaking through at two points,

H. Mitchell wrote a " Report . . . concerning Nausett beach and
the Peninsula of Monomoy " (Rep. U. S. C. S., 1871, pp. 1B4-U3).
Monomoy is described as built of sands derived from the bluff of
Cape Cod during northeast storms ; it grew southward into Nantucket
sound at the rate of 157 feet a year from 1856 to 1868. The changes
in the beach near Chatham are particularly described. The same
author submitted an " Additional report on the changes in the neigh-
borhood of Chatham and Monomoy" (Ibid., 1873, pp. 103-107).

W. Upham published some notes on Cape Cod in the Geology of
New Hampshire (1878, III. 300-305), and a more extended essay on
"The formation of Cape Cod " a year later (Araer. Nat., 1879, pp.
489-502, 552-565). He described the moraine extending eastward from
Sandwich and entering the sea at Orleans (494) ; north of this point,
the Cape consists chiefly of modified drift, rarely containing boulders
(537). When the drift plains were deposited, the land stood some-
what higher than at present (561). Provinceland consists of sea sand,
supplied by erosion on the east side of the Cape (564).

Chamberlin makes a brief reference to Cape Cod in his essay on
the "Terminal moraine of the second glacial epoch." "The great
northward hook of Cape Cod is composed of plains and rolling hills of
sand and gravel, which resemble accumulations that often accompany
the morainic belt on its interior side, and suggest the thought that the
hook may be the modified inner border of the moraine which enters the
sea near Orleans, and may be presumed to curve northward concentric
with the hook, forming thus a loop enclosing the basin of Cape Cod."
(Third Ann. Rep., U. S. G. S., 1883, p. 379.)

H. L. Marindin studied the " Encroachment of the sea upon the
coast of Cape Cod, INIass." (Rep. U. S. Coast Survey, 188^9, pp. 40"-
407, chart 28). From Highland light to Nauset lights, the average

DAVIS. — OUTLINE OF CAPE COD. 307

recession from 1848 to 1888 was 128 feet, or 3.2 feet per annum.
The face of the cliff, whose average height is 50 or 100 feet, has
thus lost a total of 30,231,038 cubic yards, or 755,776 cubic yards
per annum. The bar south of Nauset, enclosing the north side
of Pleasant harbor, extended its length southward some distance in
the same period. The same author has made a detailed report on
the changes in shore line and anchorage areas of Provincetown harbor
in Appendix 8, U. S. Coast Survey report for 1891, with an elaborate
chart.

K. Weule has, in his " Beitriige znr Morphologic der Flachkiisten "
(Kettlers Zeitschr. wiss. Geogr., 1891, VIII. 211-256), discussed Cape
Cod at some length (232-238). The tidal currents are regarded as
the most important factors in its shaping. A misunderstanding of
local conditions is implied when the author asks how " the narrow
mainland of uncompacted materials can remain intact in an exposed
situation, when even so resistant landmasses as rocky Nantucket and
Martha's Vineyard suffer great loss " (232). The present preserva-
tion of the Cape is ascribed to the beach sand, brought from the
shoals on the southeast by the flood tide. Weule follows Whiting in
attributing a greater age to the outer than to the inner side of Prov-
incetown peninsula (234). The existing mainland is regarded as
only a remnant of a great extent of drift land (233) ; this opinion
being taken from a report by A. Agassiz.

A brief article of my own, describing " Facetted pebbles on Cape
Cod" (Proc. Bost. Soc. Nat. Hist., 1893, XXVI. 166-175), argued
from these evidences of teolian action that the plains of gravel and
sand were deposited under the air rather than under the sea.

A " Report of the Trustees of public reservations on the subject of
the Province Lands" (Mass. Legislature, House, Pub. Doc. 339,
Feb., 1893, p. 6) states that "there is evidence that the tides and
waves have built one beach after another, each farther north than the
last, and that the so called Peaked hill bar is a new beach now in
process of formation." The report contains an elaborate map of the
sandy peninsula by J. N. McClintock, on a scale of about five inches
to a mile, with ten-foot contours. The manner in which the outer
beaches overlap the inner ones is very clearly shown. Five photo-
graphic illustrations present characteristic views of the dunes.

A general work on coastal forms — " La geographic littorale " —
by J. Girard (Paris, 1895), briefly compares Sandy hook and the end
of Cape Cod, classifying them with spits formed by littoral currents,
but giving no specific description.

308 PROCEEDINGS OF THE AMERICAN ACADEMY.

Review of Previous Writings.

The structures of the " mainland " of Truro and of the peninsula of
Provinceland are so unlike that their different origins have long been
recognized ; the former being attributed chiefly to diluvial or glacial
and aqueo-glacial agencies, the latter to marine agencies acting on the
former. The general character of existing processes by which the
shores are undergoing change, and the present rate of action of these
processes have been carefully examined by various observers ; but no
systematic attempt has been made to trace the processes and the changes
that they have produced backward to their beginning. This task is
therefore attempted here.

Reconstruction of the Original Outline of the Cape.

The development of the existing outline of Cape Cod must be traced
backward to the original outline. The initial form that it had before
the present cycle of cutting and filling began along its shores may be
roughly reconstructed by reversing the marine processes now at work
and following them until they lead back through earlier and earlier
conditions. The restoration may be regarded as complete, when the
reconstructed forms are everywhere of non-marine origin. Then, re-
versing the order of study, the normal operation of cutting and filling
processes should lead forward again to the existing outline of the Cape,
and should even allow a reasonable prediction of future changes for
some time to come.

Provinceland, the Chatham bars, and Monomoy, and a few small
bars near Wellfleet, must first be removed, as they consist wholly of
sea-carried materials, their arrangement being closely accordant with
action at present sea level. The tidal marshes north of Wellfleet,
along Paraet river, and elsewhere, should be excavated. The '' main-
land," chiefly of glacial and aqueo-glacial deposits, will then stand out
alone, as indicated by the outline NBHQPTC, Figure 1. It descends
to the shore on nearly all sides in steep cliffs of moderate height ; long,
straight, or gently curving beaches running along the base of the cliffs.
Exceptions to this rule are found almost exclusively on the shores
of protected bays, such as those north of Chatham and about Well-
fleet. The cliffed descent of the mainland to the smooth beaches is
manifestly an indication of destructional retreat from a formerly greater
extension seaward, just as the gentle slope of the land to the irregular
shore line of the bays is an indication of small change from construc-
tional form.

DAVIS. — OUTLINE OF CAPE COD. 309

Although no close accuracy is to be expected in restoring the sea-
ward extension of the clifFed mainland, there are nevertheless some
simple principles that will at least serve to guide us towards a not al-
together imaginary reconstruction. First, it must be remembered that
general subaerial denudation has not eflfected significant changes in
glacial topography dui'ing postglacial time. Second, the restored out-
line should possess irregularities of pattern comparable to those in the
protected bays of to-day, advancing from the headlands and retreating
towards the troughs or " valleys " in the high ground. Third, the
amount of land restored should be much less on protected shores than
on exposed shores. Fourth, cliiFs that are now protected by forelands
of marsh and bar must not be built out so far that their recession could
not have been accomplished before the bars began to grow in front of
them.

Possible Changes of Level.

These four guiding principles do not include reference to the effects
of change of level, because, if any change has occurred since the time
of accumulative construction of the mainland, it has been of small
amount, and it has, to my mind, acted on the whole in favor of de-
creasing the land area by submergence, thus co-operating with the
destructive action of the sea. This view is in accord with that ex-
pressed by Upham, who thinks that, when the drift was deposited
hereabouts, the land stood somewhat higher than at present, and that
the numerous small indentations or re-entrants of the shore line, such
as occur along the south side of the Cape, are results of a slight sub-
mergence of trough-like depressions or valleys. The digitate bays of
Martha's Vineyard would seem to lend support to this view ; but they
are otherwise interpreted by my colleague. Professor Shale r, who re-
gards them as having been formed by subglacial streams acting on sea-
floor deposits that had been strewn in front of the ice margin when the
sea stood higher than now, although he suspects also that " at the close
of the glacial period this region was considerably higher than at pres-
ent" (Geol. Martha's Vineyard, 7th Ann. Report U. S. Geol. Survey,
1888, pp. 318, 319, 350). The latter view is further supported by the
small amount of erosion — about three miles — suffered by the low
sandy southern shore of Martha's Vineyard (Ibid., p. 349) since the
present level of the land was assumed.

Without undertaking to determine precisely the original level of
the Cape mainland, the most plausible explanation of the facts seems
to me that the washed gravels and sands correspond to the supermarine

310 PROCEEDINGS OP THE AMERICAN ACADEMY.

sandr of Greenland and Alaska ; that the troughs, by which the plains
of washed sands are trenched, 'result from the channelling by streams
when they carried less waste than while they were previously aggrad-
ing the plains ; and that the indentations of the shore line are the result
of slight depression, whereby the troughs were partly drowned. Tlie
reconstruction of what I have above called the " original outline," will
therefore not necessarily lead us to the shore line that obtained at the
close of the time of accumulative construction, if the land then stood
higher than now ; but only to a contour line drawn on the original
constructional mainland at present sea level. However, between the
actual original shore line and the reconstructed contour line, there
must have been a difference of degree rather than of kind ; the latter
embracing a smaller land area than the former, but the general outline
and disposition of the laud areas probably being of much the same
style in both cases, except for the indentations of drowned valleys
after submergence. For this reason, no further especial attention
will be given to depression in its effect in altering the outline of the
Cape.

A proposed reconstruction of the outline of the Cape has been drawn,
with the four guiding principles above stated in mind. Trifling addi-
tions are made in the bays; none more than 2,000 feet. Significant
additions are made on the west side of the Cape ; some of these meas-
ure 4,000 or 5,000 feet. Two miles or more of land are added on the
east side, or " back," facing the broad Atlantic. The margin of the
restored outline is indented toward the various troughs and valleys that
break the general surface of the mainland. About High head, the
northern point of the cliffed mainland, the fourth of the guiding prin-
ciples comes into play ; and hereabouts the most interesting problem
of the Cape is found. The view of the peninsula of Provinceland
from this commanding point is therefore particularly instructive.

The Problem op High Head.

The cliffed margin of the mainland at High head, H, Figure 1, is
notably even both on the northern and western sides. At present, the
head is protected both on the west and north by forelands of marsh and
bar, the bars springing tangent from cliff fronts farther south or south-
east. The bar, Q R, on the west, is part of a long concave shore line,
TPQR, — the "west concave" shore, — whose excavated curve is mani-
festly dependent on the existence of the peninsula of Provinceland to
the northwest. Before this concave curve was cut, a nearly straight
shore line, CTYQH, — the "west straight" shore, — had been made,

DAVIS.

OUTLINE OF CAPE COD.

311

as indicated by its remnants now seen on the west marginal cliff, QH,
of High head, and again about six miles to the south, TC, on Bound-
brook, Gi-itlitis, and other islands. The cutting qf the west straight
cliff, QM, must have continued until the peninsula of Provincelaud
began to project northwest to High head. Then, as the movement
of the shore currents was somewhat changed by the interference of the
peninsula, the middle of the straight cliff was excavated more rapidly,

Fig. 1.

forming the west concave shore, TPQ, and the northern part of the
straight clilf on High head at the same time came to be protected by
the outspringing concave bar, QR, that now encloses East harbor and
its marshes on the southwest side.

The bar, BJ, on the north of High head, is part of the long eastern
convex shore line, NBJK, whose form is determined by the masterly
Atlantic currents. It is along the outer beach of this bar — or of its
representative in former days — that the sands of the peninsula have
been transported from the southeast ; this being the conclusion of all
observers, unless perhaps of Hitchcock. Now it follows from the re-

312 PROCEEDINGS OF THE AMERICAN ACADEMY,

lation of this northeast bar to the peninsula of Provinceland, and from
the relation of the peninsula to the western bar, that a somewhat
shorter time was allowed for cutting the north cliff of High head than
for cutting its west cliff ; but inasmuch as wave energy was greater on
the north than on the west, time and energy varied inversely, and hence
about the same amount of lost land may be added to each cliff. The
amount of reduction suffered on either side of High head is therefore
roughly proportional to the time before the bar was built in front of
the north cliff.

The north bar, BJ, that for this reason takes our attention, is one
of the class built by marine action, as recognized by Admiral Davis.
It springs tangent to the curve of the long convex cliff and beach, NB,
on the east side or '"back '"' of the Cape. As tiie retreat of the margin
of High head is measured by the time befoi'e the north bar was built,
the question arises whether bars of this kind are built in front of
straight cliffs early or late in the attack made by the sea on the land.
This question may be divided into two ; the first considering the de-
velopment of the cliff; the second considering the stage in the devel-
opment of the cliff when the protecting bar would be likely to grow
out in front of it.

Development of Shore Profiles.

Let the activities of the sea be resolved into two components ; one
acting on and off shore; the other along shore; and let the effects of
the first of these components be now examined alone, postponing con-
sideration of the effects of the second component to the next section.

On some young coasts, the on-and-off-shore movements of the sea
carry out to deep water all of the waste that is abraded from the land
and its submarine slope, leaving the shore line bare.* The rocky
floor seen at low tide on the coast of Brittany illustrates this condi-
tion. Here the sea is able to do more work than it has to do. Its
action is like that of a young river, whose ability to carry load is
greater than the resistance of the load that it has to carry, and whose
valley floor is therefore attacked and deepened. But as the valley is
deepened, the slope, velocity, and carrying power of the river are all
decreased ; at the same time the load, derived chiefly from the val-
ley slopes, is increased : thus ability to do work gradually falls into

* The problem of flat coasts, with shallow off-shore waters, is so different
from the problem here considered that it will be treated independently in a later
section.

DAVIS. — OUTLINE OF CAPE COD. 313

equality with the work to be done. When this happy condition is
reached, the river may be said to have graded its channel. Youth
then passes into adolescence.

A comparable series of changes may be detected in studying the
profile of a seacoast at right angles to the general shore line. As the
sea can at first usually dispose of more waste than it gathers, the coast
is energetically attacked and forced to retreat, and sea cliffs are thus
produced. But in virtue of the changes thus brouglit about, the energy
of on-and-off-shore attack decreases, while the waste coming from the
growing cliffs increases ; thus ability to do work approaches equality
with work to be done, and the sea-floor profile, like that of the valley
floor, may be said to be graded. When a graded profile is attained,
the adolescent stage of shore development is reached.

The amount of retreat necessary before a graded profile is attained
varies with the texture of the coast, and with its exposure to the sea.
A coast of unconsolidated deposits will soon supply a large amount of
waste from its clitfed margin, while the cliffs of a rockbound coast will
shed waste slowly ; hence, on coasts of given exposure, grade will be
assumed with a less amount of cliff"-cutting where the rocks are weak
than where they are strong. This recalls the behavior of rivers in
regions of weak and resistant rocks ; in the latter, they may assume
gentle slopes ; but in the former, rather steep slopes are necessary to
carry off the freely offered waste ; and gentler slopes can be assumed
only as the whole surface is worn down : this general relation having
been pointed out some years ago by Major Powell (Uinta Mountains,
194). Moreover, inasmuch as a greater amount of waste can be
handled on exposed coasts than on i^rotected coasts, a considerable
retreat may develop high cliffs on the former before enough waste is
shed from the cliff face to give the shore waves all the work they can
do ; while on protected coasts a moderate retreat, producing low cliffs,
will supply as much waste as can be handled by the sea.

The under-water form of a graded profile, when first developed, also
depends largely on the violence of the ou-and-off-shore movements of
the sea. On a protected coast, the bottom will be degraded so as to
descend from the shore line by a gentle slope to an eroded platform of
moderate depth ; but on an exposed coast, the bottom will be degraded
so as to descend from the shore line by steeper slope to a platform of
greater depth.

314 PROCEEDINGS OP THE AMERICAN ACADEMY.

Typical Shore Profiles.
A graded profile being ouce attained, its graded condition will be
preserved through all the rest of an undisturbed or normal cycle of
shore development ; shore profiles and river profiles being alike in this
as in so many other respects. Before grade is assumed, the ability of
the sea may be so far in excess of its load that it undercuts the shore
and forms sea caves at tide level, as in profile 1, Figure 2. When
grade is first assumed, the coast is usually cut back to a steep cliff, like
profile 2. Much later, when the sea has cut back the shore so that
the waves must transverse a submarine platform before attacking the
laud, their strength is thereby so much lessened that the cliflf leans
back to a moderate slope, as in profiles 3 and 4, and even then supplies
enough waste to keep the waves at its foot fully occupied.

Fig. 2.

There is something more than analogy in the comparison that may
be drawn between the longitudinal profile of a stream and the trans-
verse profile of a shore. In youth, each usually has its torrent or
upper portion, where ability to carry load is greater than load to be
carried ; but as development progresses, the graded couditiou of mid-
stream extends headward, and after a time reaches all the way to the
headwaters. At the same time, the lower or fioodplain-delta portion
extends seaward, its grade being rather steeper in adolescence, when
nmch material is brought from the headwaters, than later, in maturity
and old age, when the supply of waste is very slow. The critical point,
wliere marine action changes from degrading the near-shore bottom to
aggrading the ofF-shore bottom, migrates seaward, as 1', 2', 3', 4', in
Figure 2. At the same time, the seaward extension of the bottom
deposits increases. Furthermore, the comparison between stream and
sea suggests the need of examining that process on the sea floor, which
corresponds to corrasion in the stream bed. Sea-shore profiles make
it clear that a considerable deepening is accomplished on the floor of
the platform, landward from the critical points, 1', 2', etc. Off the
eastern cliif of Cape Cod, this deepening can hardly have been less
than fifteen or twenty fathoms : off the Chalk cliffs of Normandy, a
similar scouring and deepening of the bottom may be inferred. We

DAVIS. — OUTLINE OF CAPE COD. 315

are accustomed to study transportation and deposition as submarine
processes, but little attention has been given to decomposition, disin-
tegration, corrasion, or any other process by which the sea floor is
degraded. The subject deserves careful investigation.

It is manifest from the preceding paragraplis tliat a graded profile
may be attained much earlier on one part of a shore line than on
another ; for the texture, the original profile, and the exposure of a
coast all vary from place to place. But in a region like Cape Cod,
where the original shore line consisted wholly of uncompacted mate-
rials, this aspect of the problem need not be considered further.

Development of Shore Outlines.

It is not, however, only in on-and-o£E-shore action that a close com-
parison may be drawn between the operations of marine and fluviatile
agencies. The 'long-shore action of the sea also is in many respects
comparable to the down-stream action of rivers. Beginning on an
unevenly deformed land surface in a region of moderate rainfall,
where there are many heights and hollows, the drainage will at first
consist of many small independent systems, each one transporting waste
from the initial divides down the initial slopes into the initial hollows.
Every stream proceeds, by degrading and aggrading its course, to
develop a line of slope on which its ability to do work shall every-
where equal the work that it has to do. As the eminences are worn
down and the hollows are filled up, local systems that were at first
independent become confluent, and the drainage of the higher ones
is discharged to the lower ones. Every change of this kind will call
for rearrangement of the degraded and aggraded slopes in the con-
fluent basins. Ultimately, all the separate systems will, in one combi-
nation or another, find outlet to the sea, and the waste will be carried
a long distance from the main divides to the main river deltas.

It is much the same with the action of the sea. Leaving the on-
and-otf-shore action out of consideration for the moment, let us view
only the 'long-shore action, as determined by the dominant rather than
by the prevailing movements of the littoral waters. The projections
or headlands of the constructional shore line act as so many divides, on
either side of which the 'long-shore currents flow away from the apex,
as in the uppermost outline in Figure 3. The re-entrants or bays are so
many basins into which the 'long-shore currents converge from the ad-
jacent headlands. The headlands are slowly worn back, and the waste
is carried along their sides into the bays, where it forms aggrading

316

PROCEEDINGS OP THE AMERICAN ACADEMY.

pocket beaches or bridging bars, as in the second and later outlines of
Figure 3. The initial irregularity of shore outline is thus replaced
by a graded outline; grade being first attained in the bays, and last
on the headlands, much as was the case with stream action. As the
headlands are cut farther back and beaches are formed at the base
of their cliffs, then the 'long-shore action is moi-e and more thrown
into one direction or the other from the chief headlands, transporta-
tion is carried on past many of the subordinate headlands, and much
of the waste finds its way into the chief re-entrants of the shore line,
as in the lowermost outline of Figure 3. We should expect to find

Fig. 3.

inside the long-sweeping curve of the aggrading shore line of the chief
bays more or less distinct record of the sharp-curved pocket beaches
of an earlier stage.

However irregular the initial shore line was originally, and how-
ever many divisions were then made in the direction of the 'long-shore
currents, the time will come when only a few of the most prominent
and resistant headlands survive, as in the later outlines of Figure 3 ;
elsewhere the 'long-shore action is developed into a continuous move-
ment. Truly the direction of transportation along the graded shore
line is sometimes one way, sometimes the other, according to the
sweep of storm winds ; but if the dominant currents alone are con-
sidered, the movement is essentially constant. The graded condition,

DAVIS. — OUTLINE OF CAPE COD. 317

first reached on the pocket beaches, comes to prevail all along the
shore ; ability to do the work of transportation is everywhere equal to
the work of transportation to be done.

In the river problem, the number of independent river systems that
occupy the originally deformed surface varies with the strength of the
initial relief and with the rainfall. A light rainfall and a strong,
rapid-growing initial relief of resistant rocks produce many indepen-
dent river systems, and a long time must elapse before a general grade
is attained. The early stage of this condition is illustrated in the
lava-block mountains of southern Oregon, so well described by
Russell (4th Ann. Report U. S. Geol. Survey, 435). But a heavy
rainfall and a faint, slow-growing initial relief of weak materials may
allow the immediate development of a single river system, soon attain-
ing grade over the whole ai-ea concerned. So with the sea. Moderate
'long-shore action and strong initial irregularity of resistant rocks break
up the 'long-shore currents into many systems at first; the grading
of the shore line and the union of the many currents can be accom*
plished only after a long time of endeavor. But strong 'long-shore
action and moderate initial irregularity of weak materials may permit
continuous 'long-shore movements for a long distance on well graded
beaches almost from the very first.

Both in valleys and on coasts — in rivers and on shores — the graded
condition will be reached sooner on certain stretches than on others ;
and just as an alternation of rough rapids and smooth-flowing reaches
indicates a youthful stage of river life, so an alternation of ragged head-
lands and smooth-beached bars indicates a youthful stage of shore line
development. But in time even the more resistant parts will be
trimmed off so as to accord with the less resistant, and then down-stream
transportation — or 'long-shore movement — is well developed ; the
adolescent stage is reached. From this time forward, on a shore as in
a river, the grade is normally changed only where and when a change
of load calls for readjustment ; the readjustment necessitating an ag-
gradation or degradation of the valley floor, or an advance or retreat
of the shore line, as the load may increase or decrease.

It should of course be understood that comparisons of this kind are
not formal comparisons in which the condition of one member may be
inferred immediately from those of its analogue. The purpose of the
comparison is not to compel explanation, but chiefly to borrow illus-
tration of the systematic processes of land sculpture from the better
known examples of river action, and apply them to the less studied
examples of shore action ; less studied certainly in this country,

318 PROCEEDINGS OF THE AMERICAN ACADEMY.

where our great interior areas have for some decades past absorbed the
attention of geologists ; more studied than river action in Great Britain,
but not from the point of view here taken.

Under favorable conditions, irregular shore lines may be smoothly
graded early in their cycle of development. This is well illustrated
in the case of Martha's Vineyard. Here an extremely irregular con-
structional shore has been reduced to a remarkably even and well
graded outline in a relatively early stage of the attack of the sea on
the land ; for although a matter of two or three miles of the southern
headlands of the island have probably been cut away by the sea,* a
good part of the original shore line still remains in the branching bays
behind the bridging bars. The straight-cliffed headlands stand perfectly
in line with the bars across the bays. The later stages of outline on
graded shores are considered in the third section below.

Application of the Foregoing to Cape Cod.

The foregoing account of the development of shore lines is perhaps
an overlong preparation for the application of the simple principles
that govern shore changes to the case of Cape Cod ; but the excuse
for the details into which I have entered is the desire to show good
ground for the conclusion which they support ,* namely, that on a coast
as weak as the mainland of Cape Co^l, any originally irregular shore
line would soon be reduced to grade by the action of a sea so ener-
getic as the Atlantic, with its frequent southeast and northeast storms.
Only a moderate time and a moderate recession is therefore necessary
for the production of the even northeast cliff of High head. It does
not. however, follow from this that only a short time actually elapsed
in this work, for as far as has yet been stated, the High head cliif that
we see may have been cut far back from the first position of an even
cliff on this part of the coast line. Whether the time was long or short
can be best determined by examining into the conditions which deter-
mine the development of the bar by which the clilf is now protected,
this being the second problem announced above.

It should be noted that when the northeast cliff of High head
formed the open shore line of this part of the Cape, the outline must
have extended in a sympathetic curve, HBFjA, for some distance
southeast of its present limit ; and from this early form there must
have been a gradual change to the shore line of to-day. At some time
during this change, the protecting bar, BJ, must have been built out

* Shaler, loc. cit., 349.

DAVIS. — OUTLINE OF CAPE COD.

319

to the northwest. The problem is to determine at what stage in the
history of a clifFed shore line such a bar or spit might grow out from
one part of its face and protect another part.

Off-shore Bars.

In order to avoid misapprehension, it is advisable to make careful
distinction between those bars or spits which spring as tangent attach-
ments to a clifFed shore, often extending into comparatively deep
water, and those off-shore bars which are built up from the bottom in
shallow water, not immediately connected with the mainland. Exam-

MAINLAND

CM L MB

SE^

^^^^.

b' s b"

pies of the latter class are common along a great extent of our southern
coast, especially where the tides are weak. Briefly stated, their his-
tory seems to be as follows. When waves roll in upon a shelving
shore, as in Figure 4, much of their energy is expended on the bottom.
Between the line of their first action far off shore and their final ex-
haustion on the coast, C, there must be somewhere a zone of maximum
action. This zone must lie farther seaward when large storm waves
roll in than when the sea is slightly ruffled in fair weather. Let the
zone of maximum action for storm waves be shown by Z in profile.
Here the bottom is deepened; the coarser particles are moved land-
ward, forming a shoal and in time a bar, B', enclosing a lagoon, L; while

320 PROCEEDINGS OP THE AMERICAN ACADEMY.

the finer particles are moved seaward, beyond the limits of Figure 3,
where they are distributed in moderate thickness over a considerable
area. During this process, we may imagine the storm waves to say :
" We cannot to advantage attack a coast where the off-shore water
shoals so gradually ; let us therefore first deepen the off-shore bottom,
so that we may afterwards make better attack on the coast." So say-
ing, a preliminary off-shore bar is built up by the storm waves in posi-
tion B' ; and afterwards, at times of exceptional storms, successive
additions may be made on its outer side, as B". Wind action builds
the bar up with dunes, and carries much sand over into the lagoon.
But a time will come when the bottom farther to seaward has been
deepened enough to enable even the greatest waves to act severely on
the outer slope of the bar, taking from it more than they bring to it ;
then the outward advance of the bar is changed to a landward retreat,
and it is pushed back to such a position as B'". Tliis change in be-
havior may be taken to separate the stages of youth and adolescence
in the development of a shore line of this kind.

Young bars that are advancing or that have advanced seaward may
often be recognized by belts of dunes, B', B", roughly parallel to the
shore, enclosing lines of marsh or " slashes," S, as they are called on
the coast of New Jersey. Adolescent bars, retreating landward like
B"', may be distinguished by the exposure of the dark mud of the
lagoon marsh, M, on their outer slope, as is sufficiently explained by
the diagrams. Many examples of this kind miglit be cited. In time,
the retreat of the bar will carry it back to the mainland ; then, as long
as the marginal cliff is not too high, the dunes, D"", will be heaped
directly on the land slope, and the mature stage of shore development
is reached. In this stage, the depth of water near the shore is much
greater than it was originally ; degradation of the sea floor reaching to
depths much below low tide.

An interesting variation on this type of coastal forms is found on
coasts whose submarine slope varies, so that off-shore bars are formed
in one district, but an immediate attack is made on the land in a
neighboring district. The coast of New Jersey gives a standard ex-
ample of this kind. About Atlantic city the bars are built off shore ;
about Long Branch, the land is cut back in a retreating cliff of moder-
ate height. Although now generally retreating and exposing marsh
mud on their ocean side (Ann. Report N. J. Geol. Survey, 1885, p.
80 et seQ.), the bars frequently possess dune ridges and slashes, as if
they had once advanced seaward. Somewhere in the earlier history
of this coast, there must have been a point or fulcrum of no advance

DAVIS. OUTLINE OF CAPE COD. 321

or retreat between the advancing bars and the retreating cliff. It
should not be overlooked that 'long-shore action has a share, often a
large share, in the development of compound forms of this kind ; but
it is quite conceivable that they might be developed essentially under
the control of on-and-off-shore action alone. A second example of
this kind is perhaps to be found in the combination of the bars from
Chatham to Nauset with the cliffed margin of the Cape mainland
farther north ; but into this problem it is not desirable to enter further
at present. The origin of tangent bars or spits, built out into compar-
atively deep water, may now be taken up.

Tangent Bars or Spits.

In order to understand more clearly the conditions under which
tangent bars would form, it is necessary to return for a few moments
to the problem of the varying outline of a graded shore as dependent
on an .increase of load. It is advisable to enter this phase of the
problem through comparison again with the development of rivers and
valleys.

In the case of adolescent rivers, the increasing dissection of the
drainage basin by growing headwater branches may frequently cause
the load to continue to increase after the first attainment of a graded
slope along the trunk river. As a consequence, the trunk river must
aggrade the valley floor, forming a flood plain, until the load begins to
decrease later on in maturity. Much in the same way, 'long shore ac-
tion of the sea on a coast of graded outline may gather an increasing
load as the cliffs retreat and become longer and higher ; and with this
increase of load, certain parts of an early-graded outline may have to
be built forward into the sea. But on pursuing this comparison a step
further we find here, as in some earlier cases, a contrast replacing the
agreement thus far traced between the river and the 'long-shore action.
Not only the load, but also the volume of a river increases from youth
to maturity by reason of the better development of stream lines all
over the drainage basin ; and this increase of volume tends to prevent
the aggradation asked for by the increase of load. Similarly, the vol-
ume of water involved in the 'long-shore movements becomes greater
as the inequalities of a young shore line are reduced to the smooth
curves of adolescence and maturity ; but here the increase of volume
causes the shore waters to move in curves of larger radius than before,
and this change may require the beaches to grow forward on certain
concave or incurved parts of the shore line. In such case, increase in
the volume of 'long-shore water movements may co-operate with the

VOL. XXXI. (k. s. xxiii.) 21

322

PROCEEDINGS OF THE AMERICAN ACADEMY.

increase of load in tending to build the land out iuto the sea. Here
rivers and 'long-shore currents have unlike behavior.

One of the best examples of this kind that has come to my notice is
found on the coast of Georgia and Florida, where the better adjust-
ment of coastal bars to shore currents and the consequent increase in
volume and strength of the latter seems to have led to the out-building
of the several bars that are involved in the southward migration of
Cape Canaveral.* The accompanying digaram, Figure 5, illustrates the

essential features of the changes here
inferred. The genei-al attack that is at
first made nearly all along the ragged
coast soon comes to be resolved into
two diverse actions ; a persistent attack
on the chief medial headlands, while
the subordinate headlands are protected
by the growth of off-shore bars. Let
the ragged outline of Figure 5 repre-
sent, the original shore line of an un-
compacted land mass. The general
attack by the sea first cuts off all the
headlands, forming cliffs 2, 3, more or
less connected by bars. When longer
and higher cliffs, 4, are developed, they
supply so large an amount of waste and
allow the movement of so large a vol-
ume of water along shore that the less
exposed cliff of earlier intention in the
upper part of the figure is no longer
attacked, but is protected by a spit, 4', that springs out from the main
cliflP, prolonging its curve in one direction or the other, — here, up-
ward, — according as the tides and the on-shore winds determine the
direction of the 'long-shore movement. In this case, on-and-oflP-shore
action and depth of water have little to say. Wherever the dominant
'long-shore movement advances, there the tangent bar must grow,
whether the water is shallow or deep.

Fig. 5.

* This peculiar chancre in the situation of the cuspate foreland known as
Cape Canaveral was briefly stated by the writer in Science, 1895, 1. 606. It has
later been found that Weule had previously noted the fact of migration {loc. cit.,
p. 253), although not mentioning the cause here suggested to account for it.

DAVIS. — OUTLINE OF CAPE COD. 323

Illustration from the Coast of New Jersey.

An example suitable for illustration of this case is found in the rela-
tion of Sandy hook to the Long Brancli cliffs on the New Jersey coast,
as exhibited on the excellent topographical maps of that State. Al-
though now protected by the spit of Sandy hook, both Riirasor neck
and the Highlands of Navesink are truncated by sea cliffs. The
truncation must have been accomplished before the spit was built, and
therefore before the Long Branch cliff had been pushed back to its
present position. Stage 3, Figure 5, essentially represents this rela-
tion. In the change from earlier stages to the present, the 'long-shore
action has increased in consequence of the general smoothing of the
outline, and the direction of 'long-shore movement has been somewhat
changed ; so that now instead of carrying the waste from the Long
Branch cliff directly to the truncated headlands next north, it is carried
along an independent path forming the spit of Sandy hook outside of
the line of truncation. It is interesting to notice that the Long Branch
cliffs were evenly graded, and that the spit was formed rather early in
the general attack of the sea on the land hereabouts, and that a very
slight change in the outline of the chief cliff sufficed to cause the growth
of the spit outside of the subordinate cliffs further north. The various
fluctuations in the growth of the spit and the intermittent destruction
of its slender bar are described in the Annual Report of the New Jersey
Geological Survey for 1885, p. 78.

The Long Branch cliff has for some time been retreating under the
blows of the Atlantic breakers. The farther it retreats the longer the
stretch of cliff becomes ; it is undoubtedly much longer now than for-
merly. It may be fairly inferred that the two great spits, to the south
as well as to the north of the cliff, have always been, as now, essentially
tangent to the cliff front. It follows necessarily that the point of the
attachment of the spits to the mainland has shifted, and that the spits
have also been pushed backward at equal pace with the retreat of the
cliff. With these conclusions in mind, the problem of High head and
the northeast bar may at last be taken up.

Growth of the Provincelands.

There is good reason to think that the analogy between Sandy
hook and the Provincelands pointed out by Admiral Davis may be
carried much further than he suspected. The great convex cliff line
on the back of the Cape corresponds to the slightly convex line of the
Long Branch cliff ; the northeast cliff of High head is the counterpart

324 PROCEEDINGS OF THE AMERICAN ACADEMY.

of the protected cliff of the Navesink highlands ; the slender bar that
springs tangent to the curve of the back of the Cape and runs to the
broad peninsula of" the Provincelauds is essentially a repetition of the
slender bar that springs north from the Long Branch cliff and runs to
the broadened peninsula of Sandy hook. The point where the bar
now springs northwestward from the long convex back of the Cape is
not the point where the bar first began to grow. Its original point of
attachment must have been southeast of the present point ; and in the
change from the original to the present arrangement, both the cliif and
the slender bar must have been forced back, in the very manner already
described for the example in New Jersey. Marindin's report gives pre-
cise data for the retreat of the cliff ; and the story of the buried canoe,
recorded by Thoreau, gives support to the retreat of the bar near its
point of attachment. In both examples, the farther part of the great
spit has grown by addition to its seaward side in order to keep the out-
line in a curve sympathetic with the retreating cliff; the outward or
eastward growth of Sandy hook being described in the Annual Report
of the New Jersey Geological Survey for 1885, p. 77 ; the similar
growth of the Provincelands is more fully stated below. As a result
of the outward growth of the spit while the cliff is retreating, there
must be a neutral point or fulcrum of no change somewhere on the
connecting bar : and with the further straightening of the cliff front,
the position of this fulcrum must generally shift toward the spit, as
shown by F^, Fj, F3, Figure 5.

The original point of attachment of the connecting bar on Cape Cod
must have been at the intersection of two converging lines determined
by the northeast cliff of High head and the innermost or oldest of the
bars in the Provinceland peninsula. The first of these lines is well
defined, HB, Figure 6 ; the second is less distinct, but appears to be
recorded in a sand bar on the line EFo. The form of this bar has
probably been somewhat changed by wind action , yet the trend of its
inner margin along the shore of East harbor is comparatively straight,
as if it had not been much altered from the form given when it was
built. Its trend departs slightly from the direction of the adjacent
Atlantic shore, as if had been determined by conditions now vanished.

The intersection of the two guide lines, HB and EF2, when pro-
longed to the east-southeast, is found at a point Fi, about 4,000 feet
off the present shore, and about a mile and two thirds east-southeast
from the present point of attachment of the springing bar. Judging
by the present rate of retreat of the cliff line, this outer position must
have been occupied about 1,200 years ago. These figures are of neces-

DAVIS. — OUTLINE OF CAPE COD.

325

sity only approximate, but they are believed to give a fair iiulicatiou
of the order of magnitudes involved, both in space and time. We may
then infer that when the general outline of the back of the Cape had
assumed the position of the line AFiBH, the shore was well enough
graded to supply material for the building of a spit ; and that the
curvature of the shore at the point Fi, assigned for the beginning of
the spit, was such that the dbminaut 'long-shore currents, moving from
south to north in flood tide or under southeast storms could no longer

Fig. 6.

follow the shore, but departed from it outwardly by a small angle-
Thus the protecting bar, FiE, began to grow in front of the High
head cliff.

At an earlier stage the 'long-shore currents must have been much
interrupted by the irregularities of the original shore line. No large
and well developed current could at that time follow these irregulari-
ties. But as the headlands were cut back and the bays bridged across,
and the shore assumed the outline ABH, then the resistance to the
development of the current became less and less ; thereby the
current became stronger and stronger, and desired a straighter
and straighter path for its movement. At the same time, a greater
and greater volume of waste was supplied from the growing cliffs.
As long as the back of the Cape pi-ojected farther into the sea than

326 PROCEEDINGS OF THE AMERICAN ACADEMY.

now, the northward shore current may have swung pretty well around
the mainland, as sketched in line ABH. But as the east side of the
Cape was cut away and straightened, and as the shore current grew
stronger and stronger, it became increasingly difficult for the waters to
turn the curve that led to High head ; and at last, when the turning
was impossible, the spit began to form on the line FjE. As the
change progresses, the current swings on a fulcrum, F'., ; the spit
broadens by the external addition of new bars, FgG, as well as by the
formation of sand dunes inside of the curve ; and the fulcrum shifts
along the shore to the northwest, as indicated by the points F^, F3,
Figure 6, in the manner already explained for Figure 5.

The important point to note is tluit here^ just as on the New Jersey
coast, the grading of the initial irregular shore line into a curved cliff
shore, and the straightening of the curved cliff" shoi'e enough to require
the growth of the tangent bar, must have been accomplished early in
the development of so weak a land mass as Cape Cod in face of waves
80 strong as those of the Atlantic.

Dimensions of the Original Cape.

Now inasmuch as no very long time can have been required for the
Atlantic waves to wear back the original shore line of the Cape to a
graded outline, ABH, of which the High head cliff is a part, and
inasmuch as the growth of the springing spit must have been begun
soon after the grading of the shore, it follows that the original con-
structional outline of the land in front of the High head cliff cannot
have extended far into the sea. I have given it an extension of 3,000
feet in Figure 1. A similar original extension of all the mainland of
the Cape may be assumed outside of the graded shore line ABH,
that existed before the springing spit was formed ; and thus the origi-
nal outline of the eastern side of the mainland has been roughly
sketched in. As drawn in Figure 1, the greatest retreat from the
original shore to the present shore is nearly two and a half miles, and
at the present strength of wave action, 3,000 or 4,000 years may be
roughly taken to have sufficed for the accomplishment of this change.
This time is probably too long rather than too short, for the retreat
now must be slower than when the cliff" was lower.

It should be carefully understood that the period here computed
does not measure postglacial time ; for, as already stated, it is believed
that the land hereabouts stood somewhat higher than now during the
accumulation of the stratified sands, and that only after the time of

DAVIS. — OUTLINE OF CAPE COD. 327

accumulation were the valleys and low grounds slightly submerged
by a moderate depression of the laud, and the work whose duration is
here computed begun. The time that passed while the sea was at
work on some lower shore is not measured. There is no indication
of a recent elevation of the land hereabouts, as far as the shore fea-
tures testify : even the protected cliffs of High head are cut down
to present se^ level.

The Nauset bar extends southward from the cliff at the point N.
The earlier positions were prolougations of the lines A, D. The
point of attachment must therefore have migrated to the southwest;
the retreat of the cliff front determining the retreat of the bar that
stands in line with it. How the problematic islands off Chatham
affected the behavior of the bar is not here inquired into.

Inasmuch as the recession of the eastern shore is believed to have
been of moderate measure, the loss on the western shore must have
been still less. This is considered in a later section.

The Origin of Race Point.

Two important consequences follow from the swinging of the shore
current on its movable fulcrum. The first gives explanation of the
overlapping of the newer shore lines outside of the older body of the
peninsula, as stated in the Report of the Laud and Harbor Commis-
sioners, quoted above. This is only a repetition of the process by
which the spit first departed from the beach on the back of the Cape
itself. The outer margin of the Proviuceland peninsula is therefore
its very youngest part, and not its oldest, as supposed by Whiting.
The long bar, FgJK, ending in Race point, is a distinct external
addition to the older body of the Provincelands, and a long narrow
" slash " is included behind it. It has grown out into comparatively
deep water, for the 20-fathom line lies only 1,700 feet offshore to the
northwest. Peaked hill bar may be, as the Commissioners have
plausibly suggested, the embryo of still another external bar.

It may be noted that small spits departing tangentially from curved
beaches are not uncommon. The map accompanying Whiting's report
shows two of them near Wood End, one pointing east, the other north,
from the sharp curve of the bar, as if determined by a strong south-
west storm, whose waves worked eastward and northward from the
apex of the curve at Wood End. A minute spit of this kind is shown
on the chart of Cape Cod bay (Coast chart 110, printed 1890), a little
northeast of Race point; but a later edition of the chart (1892)
carries a smooth curve around the point. Small examples of these

328 PROCEEDINGS OF THE AMERICAN ACADEMY.

forms, trending eastward, were seen on the south shore near Long
point light, at the time of my visit to the Cape last summer.

The Wasting Shore from Race PoixNt to AVood End.

The seconil consequence of the outward deflection of the current
around the peninsula is the rapid consumption of the bar, VW, that
extends south from Race point inlet to Wood End, the long "finger "
at the end of the Cape. This suggests a preliminary digression.
Wonder is often expressed at the ability of sand bars to withstand the
violence of the surf that breaks unceasingly upon them. The sands
are entirely unconsolidated, and their surface layers are moved by
every surge of the waters. Yet the form of the bar changes very
slowly. The reason for this must be found in the continual feeding,
from the cliffs and from the bottom off shore, by which the volume of
the bars is sustained. The bars of our southern Atlantic coast pre-
sumably receive much of their sand from the l)otti)Di. Sandy hook
receives much of its supply from the retreating cliffs at Long Branch.
If the supply be withheld, the bar will be rapidly swept away. It
may not be that llie grains of sand are actually ground to dust, but
that they are brushed along, and when no followers come to take their
place, it is left vacant, and the face of the bar retreats ; its dunes are
cut back, and a low cliff-shore is formed.

As long as the outside of the peninsula formed a continuous curve,
sand was carried along it in plenty from the cliff and the sea floor
on the back of the Cape, and probably also from the shoals where
Webbs island and its vanished mates once stood off Chatham. This
condition is represented in line DGVAV. But as the cliff from Nauset
to Highland was cut fartlier back, and the shore current became unable
to follow its earlier path along the margin of the peninsula, the additional
bar, ending in Race point, was laid out, and tlie long marshy " slash "
was enclosed behind it. From the beginning of this additional bar
until the present time, the supply of sand carried around the western
curve of the peninsula was greatly reduced ; at times it may have
ceased entirely. The supply being thus reduced or cut off, the bar
southward from Race point inlet nearly to Wood End rapidly wasted ;
and the sand taken from it by northwest gales went to supply the cor-
respondingly rapid growth of Long point, WX, into Provincetown
harbor, which Whiting shows to have extended many feet eastward in
the fifty years past. Like Race point, Long point has advanced into
comparatively deep water; the 20-fathom curve lies only 600 feet
off shore ; the same depth is not found for almost three miles off the
cliffed shore of the " back " of the Cape,

DAVIS. — OUTLINE OP CAPE COD. 329

The Western Side of the Cape.

The western side of the Cape offers simpler problems than those of
the eastern side. The first task here attempted by the waves was the
development of the long west straight shore line, HQTC, of which
only the extremities now remain. This does not seem to have required
anywhere a greater recession than 3,000 feet. It must have been
accomplished chiefly by northwest gales and north-to-south shore cur-
rents, by which the waste gathered from the more continuously cliflfed
shore was carried southward to tie together the several islands below
South Truro. If southwest gales and south-to-north shore currents
had been dominant, an acuminate spit should have been formed in
prolongation of High head, where the waste would have been supplied
from both sides of the Cape ; but of this there is no sign.

The modification of the west straight shore line by the excavation
of the present concave shore line, QPT, undoubtedly results, as has
already been stated, from the disturbance of antecedent conditions
that was caused by the growth of the Provincelands to the northwest.
The northwest gales gradually came to have less and less influence ;
for some time past, they must have ceased to be dominant ; the chief
control of shore movements now seems to be in the hand of the
weaker southwest gales ; for both the offsetting spit at the mouth of
Pamet river, P, and the outspringing bar, QR, tliat protects High
head on the west, imply a northward transportation of sands. Some
southward movement, however, still occurs, as might be expected ;
for at the faint angle, T, where the older straight shore line, HQTC,
is now cut by the concave shore line, QPT, a spit projecting to the
southwest seems to have been begun, and its continuation under water
is indicated by a shoal of sympathetic curvature, TU, some five and a
half miles in length. How far this shoal may be a new feature, origi-
nating with the excavation of the concave shore line, or how far it may
be of much greater age, dependent on the extensive Billingsgate shoals,
where outlying islands are thought to have originally stood, is for the
present an undecided question.

Protection of Provincetown Harbor.

A matter of considerable economic importance turns on the changes
experienced by the "wrist" of the Cape, the narrowest part of the
bar that connects the mainland or " forearm " of the Cape with the
peninsula or " hand." The people of Provincetown feel anxiety lest
tlie sea should breach the bar and wash a great amount of sand west-

330 PROCEEDINGS OF THE AMERICAN ACADEMY.

ward past High head into their excellent harbor. The records of
changes in the bar that connects Sandy hook with the Long Branch
cliffs give ground for this anxiety. The point that I wish here to
call attention to is that the only part of the northeast shore that is
liable to be broken through lies on the stretch, BFg, between the
point where the connecting bar springs northwest from the great cliff
and the point where the '' fulcrum " is at present located. Within
this stretch, the bar is generally retreating, being cut on the outer
side, and reconstructed on the inner side.

Two safeguards may be suggested. One would cause the fulcrum
to migrate southeastward, thus diminishing the length of the narrow
and breakable bar, and at the same time increasing its breadth and
strengtli. This would be accomplished by the construction of bulk-
heads along the outside of the narrow bar, or wrist, so as to catch the
drifting sand instead of allowing it to pass by ; thus the bar might be
broadened and strengthened. Judging by the rapidity with which the
body of a wrecked vessel causes an accumulation of tand on its south-
eastern side, a significant addition to the narrow bar might soon be
made in this manner. Manifestly, the greatest economy in the use of
the drifting sand requires that the bulkheads should be continually
built out so as always to project a little beyond the aggrading shore
line. There are indications that this very result is at present being
accomplished by natural process, for the beach in the narrow stretch,
BFg, is now notably broadened in front of its former line at the base
of the surmounting dunes.

A more economical and enduring protection of Provmcetown harbor
than the above plan suggests has been already secured by complet-
ing the extremity of the bar, QR, that some years ago almost en-
closed East harbor ; so that if storm waves should temporarily breach
the narrow connecting bar on tlie ocean side, — the " wrist " of the
" hand " of Provincetowu at the end of the " bended arm of Massa-
chusetts," — all the sand that was carried through the breach would
settle in East harbor, and thereby strengthen the embankment against
further encroachments. A second protecting dike has been built
across the marsh, northeastward from near High head. The fear that,
in case the narrow connecting bar or " wrist " should be breached,
the whole action of the Atlantic 'long-shore currents would thereafter
be directed through the breach into Provincetowu harbor, is ground-
less. The whole history of the growth of the peninsula demonstrates
that the 'long-shore currents must continue to swing in long curves
of large radius in the future, as in the jiast.

DAVIS. — OUTLINE OF CAPE COD. 331

The danger of silting up the Provincetown harbor by drift eonnng
from the west concave shore line along the west protecting bar of
High head does not appear to be imminent, for the processes of trans-
portation are comparatively slow on the iimer side of the Cape ; but
the danger is nevertheless real, and nothing but an extensive and ex-
pensive system of bulkheads from North Truro northward, on the
stretch PQ, appears to be sufficient to avert it.

The destruction of the narrow strip of sand-bar shore, VW, between
Race point and Wood Kud seems to me to threaten Provincetown
harbor with a greater danger than any that it is exposed to from the
east. This shore is now wasting rapidly. Once broken through,* the
currents driven by nortliwest gales, as well as by the rising tide,
would no longer have to swing around Wood End, W, and deliver
their load of drifting sand to Long point, X; they would in all prob-
ability invade the harbor directly, cutting away the low-tide flats that
now expand south of the village, and throwing the detritus thus
gained into the harbor. Attention has been called to this danger by
Marindin in the Coast Survey report for 1891, Appendix 8. While
bulkheads may delay the destruction of the narrow bar, they can
hardly preserve it even through a brief historical period. It has
been proposed to abandon the wasting bar to its fate, and to protect the
harbor by building a dike from the west end of the village across the
flats to Wood End. A partial protection might be gained by building
bulkheads on the northern shore of the peninsula, two or three miles
east of Race point, K. Drifting sand from the east would then be
stopped there. Race point, no longer so well supplied with sand as
now, would be wasted by the northwest storms, and the sands carried
from it would go southward to repair the shore towards Wood End.
The protection of the bar northeast of High head near Fg would, to a
certain extent, work in the same direction by diminishing the supply
of sand for the Race point bar ; but a considerable time might elapse
before any advantageous effect from this cause would be felt.

The Future of the Cape.
The encroachment of the sea on the back of the Cape is undoubtedly
destined to continue until the Truro mainland is all consumed north
of Orleans, the '' elbow " of the bended arm. At the present rate of
recession — 3.2 feet a year — eight or ten thousand years will be re-
quired for this task ; and this without considering the aid given by the

* A small breach lias been marie in this bar during the past wmter, as I
have learned from a recent visit to Provincetown.

332 PROCEEDINGS OF THE AMERICAN ACADEMY.

waves of Cape Cod bay, whose concave sweep along the Truro shore
shows their competeuce to do no insignificant share of the work.

It does not seem at all likely that, while the rest of the Truro main-
land is wearing away, the spit at Race point will of itself curve around
to the south, and thus save from destruction the narrowing bar which
encloses Provincetown harbor on the west. A great volume of trans-
ported sand would be needed to continue the bar in the deep water
through which its present curve would lead. Moreover, the shoal
known as Peaked hill bar may, as has been suggested, mark the begin-
ning of a shore line exterior to that of the present Race point curve.
It is possible that as additional tangent spits are lapped on the outside
of the curve. Race point will be cut back by a current from the north-
west, working opposite to the great current that rounds the peninsula
from the east ; a cuspate or acuminate spit being then formed in the
angle between the two, such as now exists at Great point, Nantucket.
There, the transportation of shore waste is northward on the east
shore and southward on the west shore, according to the memoir by
Admiral Davife ; this being proved by the drift of coal and bricks
from vessels wrecked on the east shore {op. clt., 139). The occur-
rence of these '' cuspate forelands," as Gulliver has called them (Bull.
Gaol. Soc. Amer., 189G, VII.), is not so much of a rarity in nature
as might be imagined from the little that appears about them in books;
their growth being sometimes attributable to accordant currents that
flow towards the point on either side ; sometimes to opposing currents,
one flowinfj inwards, the other outwards. Good reasons iiave been
given by Abbe for believing that Cape Hatteras and the other cuspate
capes of the Carolina coast have been built between opposing currents
(Proc. Bost. Soc. Nat. Hist., 1895, XXVI. 489).

The Provincetown peninsula may be expected to outlast the Truro
mainland ; for as long as the latter exists, the former must receive
contributions from it. But when the mainland is washed away, —
ten thousand years hence, at the present rate of wearing, — then
Provinceland must rapidly disappear. Sable island, a long sand bar
oflPNova Scotia, is perhaps to be regarded as the vanishing remnant of
a destroyed drift island (see Trans. Roy. Soo. Canada, 1894, XII.pt. 2,
pp. 3-48 ; also, note in Science, 1895. 11. 886). It may in this sense
be taken to represent a future stage in the destruction of Cape Cod.
All these changes are rapid, as changes go on the earth's surface. The
Truro mainland will soon be destroj'ed, and the sands of Provinceland
will be swept away as the oceanic curtain falls on this little one-act
geographical drama.

GOTO. — EMBRYOLOGY OF STARFISH. 333

XVII.

CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF

THE MUSEUM OF COMPARATIVE Z(JOLOGY AT HARVARD

COLLEGE, UNDER THE DIRECTION OF E. L. MARK, LXIV.

PRELIMINARY NOTES ON THE EMBRYOLOGY OF
THE STARFISH (ASTERIAS PALLIDA).

By Seitaro Goto.

PreBented by E. L. Mark, April 8, 1896.

In this preliminary notice I propose to give a brief account of the
principal results of the study which I have been carrying on during
the past winter. Since I began my study, two papers in the " Quar-
terly Journal of Microscopical Science," covering the same ground in
other species, have come to ray notice, one by Bury, and the other by
MacBride. But as the results of these observers are at variance on
some important points, it will, I believe, not be without interest to
publish my own, which, it will be seen, are mostly, but not altogether,
confifmations of those of MacBride.

The material was collected last summer in Mr. Agassiz's Laboratory
at Newport, R. I., where I was enabled to work for several weeks
through the courtesy of the owner, to whom my best thanks are due.
It will be observed that I have studied the same species as that on
which Mr, Agassiz himself worked about thirty years ago, — a spe-
cies which in the course of its development passes through typical
Bipinnaria and Brachiolaria stages.

(1) A few words as to the orientation of the body. In the Bra-
chiolaria stage the sagittal plane of the body passes through the dorsal
Brachiolaria arm, and is at right angles to the line joining the tips of the
other two arms. It also passes through the anus and the middle of the
mouth. In later stages, as metamorphosis approaches more and more
its end, the anus shifts its position considerably, and the sagittal plane
can be determined externally only by means of the Brachiolaria arms.
I find the plane determined by this means constant for all practical
purposes ; that is to say, it bisects the stomach as well as the body as
a whole, and passes through both the dorsal pore and the point where

334 PROCEEDINGS OF THE AMERICAN ACADEMY.

the definitive mouth is formed. As the Brachiolaria arms are visible
until a very late stage of metamorphosis, this method of determining
the principal plane of the body is very convenient, and at the same
time tolerably accurate.

The aboral disk arises, as already known, as a thickened patch of
the ectoderm on the right side of the body, and extends also some dis-
tance on to the dorsal side ; but as metamorphosis progresses, it shifts
its position, so that it becomes more and more inclined to the sagittal
plane, and at the same time more nearly parallel to the crossplane,
i. e. a plane perpendicular to the chief axis of the body. The posi-
tion of the dorsal pore remains nearly, although not quite, constant.

From the above it follows that the oral side of the adult is the ante-
rior, the aboral the posterior, the madreporic (interradius) the dorsal,
and the side opposite this the ventral, side of the larva. Right and left
sides are evident from what precedes.

(2) As has been pai'tly suggested in the foregoing lines, both the
definitive mouth and anus are new formations.

(3) Of the body cavity in the larva I distinguish, with Bury, four
portions, an anterior and a posterior on either side of the body. In
the fully developed Bipinnaria, these portions are all directly or indi-
rectly continuous with one another; but in the course of the Brachi-
olaria stage the right posterior portion is entirely cut off from the
remaining parts, and later, when the tentacles of the Bipinnaria have
been largely drawn in, another septum is formed anterior to the
first, while the left posterior portion also becomes constricted off from
the rest ; so that at this stage there are three separate cavities, viz.
(1) the right posterior, (2) the left posterior and the right middle
portions, which still form but one cavity, and (3) the right and left
anterior portions, which freely communicate with each other in the
Brachiolaria arms. The anterior cavity (3) persists in the adult as
the axial sinus and the water-vascular system.

The right posterior cavity shifts its position hand in hand with the
aboral disk, and finally occupies the posterior end of the larva just be-
neath the disk. With the development of the arms of the star, it sends
out diverticula into them, and thus assumes the bibrachiate, o-radiate
star-shape which it has in the adult.

(4) The formation of the water-vascular ring is not a mechanical
result of the breaking through of the adult mouth, for the ring exists
as such some time before the mouth is formed.

(5) With Bury and MacBride I distinguish sharply two structures,
the pore-canal and the stone-canal. There is a stage when the pore-

GOTO. — EMBRYOLOGY OF STARFISH. 335

canal alone is present. I believe that they are also phylogenetically
distinct, the stone-canal being a later formation. In comparing the
Echinoderms with such a group as the Enteropueusta it is, as it
seems to me, the pore-canal alone that is primarily to be taken into
consideration.

(6) The opening of the pore-canal and the stone-canal into the body
cavity persists throughout life. This is true not only of Asterinn gib-
bosa and Asterias pallida, but also of Asterias tenera, Solaster endeca,
and Gribrella sanguinolenta, and inferentially of all starfishes. The
opening of the stone-canal is always on the right side of the sagittal
plane.

(7) In agreement with MacBride, I find that the " dorsal organ " of
Bury, the " Centralblutgeflecht " of Ludwig, forms the perioesophageal
portion of the body cavity of the adult. It arises in the form of a tube
from the left posterior enterocoel just behind the pore-canal. In a
young Brachiolaria it encircles about one fourth of the whole circum-
ference of the cardiac portion of the stomach, and it forms a complete
ring in a young star that has just finished its metamorphosis. The
septum that divides- it from the left posterior enterocoel is subsequently
completely absorbed except in one place, viz. the madreporic inter-
radius, where it appears to persist throughout the life of the animal.
The term "oral coelom," used by MacBride, seems to me unfortunate,
as that term has been applied to another and entirely different cavity
(left coelom) in Crinoids. I therefore prefer to call the structure in
question " perioesophageal enterocoel."

(8) The perihsemal spaces (the inner ring excepted) as well as the
peribranchial spaces are, according to my observation, of true schizoccel
origin.

Cambridge, April 8, 1896.

PROCEEDINGS OF THE AMERICAN ACADEMY.

XVIII.

ON THE GROUP OF REAL LINEAR TRANSFORMA-
TIONS WHOSE INVARIANT IS AN ALTERNATE
BILINEAR FORM.

By Henry Taber.

Presented February 12, 1896.

Let G denote the group of linear automorphic transformations of
the alternate bilinear form

2n 2n
1 1

with cogrediant variables and of non-zero determinant. On page 575
et seq., Volume XLVI. of the Mathematische Annalen, I have shown
that a transformation of group G can be generated by the repetition
of an infinitesimal transformation of group G if, and only if, it is the
second power of a transformation of group G. I now find, if J is
real, that the same theorem holds for the sub-group of real trans-
formations of group G. That is, if J is real, a real transformation
of group G can be generated by the repetition of a real infinitesimal
transformation of group G if, and only if, it is the second power of a
real transformation of tliis group. Furthermore, if JF is real, the sec-
ond power of a real transformation of group G is the (2 ???)th power of
a real transformation of this group for any even exponent 2 m*
If the transformation T'is defined by the system of equations

x\ = a^i Xi + «,-2 a:2 + . . . . -f «r, 2n ^2n (^ = 1 9 2, . . . . 2 n — 1,2 n),
let 7a denote the transformation defined by the equations

X'r= (Orl^l + ^r2^2 + • • • • + f'^r. 2n^2n) Xx^ (f = 1 , 2, . . . 2n — 1, 2n),

A being a root of multiplicity m of the characteristic equation of T.

* For an odd exponent 2 m + 1, any real transformation of group G is the
(2 m -f l)th power of a real transformation of this group.

TABER. — LINEAR TRANSFORMATIONS. 33T

The nullity * of 7\ is then at least one, and the nullity of successive
powers of Tk increases until a power of exponent /x = m is attained
whose nullity is equal to m. The nullity of the (jj. + l)th and higher
powers of T\ is then also m. If we designate respectively by

nil, 1)12, rn^_-i_,m^ = ?n,

the nullities of

rp rpi rpn. — x rp^,.

•^ A.J -^ A' • • • • -^ A J -^ A>

then

mi ^ ^2 — nil = .... = m^j. — m^^-i = 1.

The numbers /^i, [l^i ^tc, may be termed the numbers belonging to the
root A. of the characteristic equation of T.

If now Th the second power of a real transformation of group G,
the numbers belouging to each negative root of the characteristic
equation of T are all even. These conditions are probably not only
necessary but sufficient in order that a real transformation T of group
G may be the second power of a real transformation of this group.

* The nullity of the transformation defined by the system of equations

x'r — ari^i + ar2a-2 + . . . + a^, iv^-jv r = (1, 2, . . . N),

is m if all the(7n — l)th minors (the minor determinants of order N—m-\-V)
are zero, but not all the rath minors (the minor determinants of order N—m) of
the matrix.

3'ii) a^isi • • •

®21> a22> • • •

VOL. XXXI. (N. S. XXIII.) 22

PROCEEDINGS.

Eight hundred and seventy-fifth Meeting.

May 8, 1895. — Annual Meeting.

Vice-President B. A. Gould in the chair.

The chair announced the death of James Dwisrht Dana
and John Newton, Associate Fellows.

The Corresponding Secretary read the following letters:
from Sir F. Pollock, acknowledging his election as Foreign
Honorary Member ; and from the Natural History Society of
Bonn, announcing the death of its Curator, Reinhard Peck.

The Corresponding Secretary read the Report of the
Council.

The Report of the Treasurer was read and accepted. The
following is an abstract.

General Fuxd.
Receipts.

Balance, May 1st, 1894 |2,301.55

Sale of rights, Mass. Cotton Mills 30.50

2,332.05

Assessments $975.00

Sale of publications 48.52

Subscriptions for publications .... 461.00 $1,484.52

Investments 3,847.50

Return of bank tax 51.40

Gift of E. C. Clarke 600.00 5,983.42

$8,315.47
Expenditures.

General expenses $2,630.26

Publishing expenses 1,788.29

Library expenses 1,221.68 $5,640.23

Balance 2,675.24

$8,315.47

340 PROCEEDINGS OF THE AMERICAN ACADEMY

RuMFOKD Fund.

Receipts.

Balance May 1st, 1894 $4,110.21

Investments $1786.00

Return of bank tax 97.55 1,883.55

$5,993.76
Expenditures.

Books and binding $120.06

Kent 10.00 $130.06

Balance 5,863.70

$5,993.76
Warren Fund.
Receipts.

Balance, May 1st, 1894 $334.89

Income 840.00

$1,174.89
Expenditures.

Investigations $800.00

Balance 374.89

$1,174.89

Building Fund.

Receipts.

Balance, May 1st, 1894 $418.68

Income 425.00

843.68
Balance 843.68

On the motion of William R. Livermore, it was
Voted.) That the thanks of the Academy be tendered to
the Treasurer for his timely gift of six hundred dollars for
clerical assistance.

The Librarian made a verbal report on the condition of the
library, by which it appeared that 2,773 books and pamphlets
were added to the library daring the past year, of which
1,866 were obtained by gift and exchange, 645 purchased with
the appropriation from the General Fund, and 262 with the
appropriation from the Rumford Fund. 465 volumes were

OF ARTS AND SCIENCES. 341

bound at an expense of $580.15, $35.80 of this amount being
charged to the Rumford Fund. The total expenses for books,
periodicals, and binding amounted to $1,326.68. There were
242 books borrowed from the library by 32 persons, of whom
23 were Fellows of the Academy.

The following reports were presented : —

Report of the Rumford Committee.

The Academy voted at the annual meeting. May, 1894, to
appropriate the sum of one thousand dollars from the Rumford
Fund to be expended at the discretion of the Rumford Com-
mittee in aid of investigations in Light and heat, payments
from this sum to be made on the order of the chairman of the
Committee. The Committee has made the following appro-
priations from this sum : —

$250 to Professor Nichols of Cornell University for investi-
gations in the radiations from carbon at different temperatures.

$250 to Professor E. H. Hall of Harvard University, in aid
of his investigations on the thermal conductivity of metals.

$250 to Professor Webster of Clark University, in aid of an
investigation upon the relation between the velocity of light
waves and electrical waves.

The Committee has also voted to approve the purchase of
such volumes of the Memoirs of the International Bureau of
Weights and Measures as are needed to complete the set in
the library of the Academy, and has also passed the following
votes : —

Voted., To recommend to the Academy to appropriate $250
to Professor B. O. Peirce in further aid of his investigation
on the thermal conductivities of poor conductors.

Voted., To recommend to the Academy to appropriate
$2,000 to the Rumford Committee to be expended at the
discretion of the Committee in aid of investigations in Light
and heat, payments from this sum to be made on the order
of the chairman of the Committee.

Voted., For the second time to recommend to the Academy
that the Rumford Medals be awarded to Thomas A. Edison,
for his investigations in electric lighting.

John Trowbridge, Chairman.

342 PROCEEDINGS OP THE AMERICAN ACADEMY

Report of the O. 31. Warren Committee.

8 May, 1895.
In behalf of the C. M. Warren Committee, I have to report
that during the past year a grant of six hundred dolhirs ($600)
has been made to Professor C. F. Maber}-, of Cleveland, Ohio,
in aid of his investigation of American and Canadian Petro-
leums, and a grant of two hundred dollars (1200) to Profes-
sor F. C. Phillips, of Allegheny, Pa., in furtherance of his
researches on Natural Gas.

F. H. Storer, Chairman.

A report of the Committee of Finance was read and
accepted.

On the recommendation of the Cpmmittee of Finance,
it was

Voted., To appropriate nineteen hundred dollars (-^1,900)
for general expenses.

Voted., That an appropriation of thirteen hundred dollars
($1,300) be made to the Librarian, and that this appropriation
cover also the petty expenses of the Assistant Librarian.

Voted., That an appropriation of eighteen hundred dol-
lars ($1800) be made for Volume XXX. of the Proceedings,
now in press, provided that a part of this sum be expended
for plates.

Voted, That an appropriation of four hundred dollars ($400)
be made to be used in the preparation of the plates and print-
insT the text of Roland Thaxter's Memoir on the Laboulbeni-
acese.

On the recommendation of the Rumford Committee, it was

Voted., To appropriate the sum of two hundred and fifty
dollars ($250) from the income of the Rumford Fund to B. O.
Peirce, in further aid of his investigation on the thermal con-
ductivities of poor conductors.

Voted., To appropriate from the income of the Rumford
Fund the sum of two thousand dollars ($2,000), to be ex-
pended at the discretion of the Rumford Committee in aid of
investigations in Light and Heat, payments from this sum to
be made on the order of the chairman of the Committee.

OF ARTS AND SCIENCES. 343

Voted, To award the Rumford Premium to Thomas Alva
Edison for his investigations in electric lighting.

On the recommendation of the C. M. Warren Committee,
it was

Voted, To appropriate six hundred dollars ($600) from the
income of the C. M. Warren Fund to C. F. Mabery, of Cleve-
land, Ohio, to aid him in continuing his researches on the
chemistr}'- of petroleum.

On the motion of the Treasurer, it was

Voted, That the assessment for the ensuing year be five
dollars (|5).

On the recommendation of the Committee on amending the
statutes, it was

Voted, To amend chapter 10, section 2, line 8, of the stat-
utes, by changing the word " eight " to " seven."

The following gentlemen were elected members of the
Academy : —

Arthur Gordon Webster, of Worcester, to be a Resident
Fellow in Class I., Section 2 (Physics).

Arthur Michael, of Boston, to be a Resident Fellow in
Class I., Section 3 (Chemistry).

John Fiske, of Cambridge, to be a Resident Fellow in
Class III., Section 3 (Political Economy and History).

William Price Craighill, of Baltimore, to be an Associate
Fellow in Class I., Section 4 (Technology and Engineering),
in place of the late William Helmsley Emory.

The annual election resulted in the choice of the following
officers : —

Alexander Agassiz, President.
Benjamin A. Gould, Vice-President for Class I.
George L. Good ale, Vice-Presiderit for Class II.
Augustus Lowell, Vice-President for Class III.
Charles L. Jackson, Corresponding Secretary.
William Watson, Recording Secretary.
Eliot C. Clarke, Treasurer.
Henry W. Haynes, Librarian.

344 PROCEEDINGS OF THE AMERICAN ACADEMY

Councillors.

Benjamin O. Peirce, \

Henry Mitchell, \ of Class I.

Leonard P. Kinnicutt, )

Benjamin L. Robinson, \

Henry W. Williams, \ of Class II.

Henry P. Bowditch, )

Andrew M. Davis, \

Thomas W. Higginson, S of Class III.

James B. Thayer, )

Member of Committee of Finance.
Augustus Lowell.

Rumford Committee.
John Trowbridge, Edward C. Pickering,

Erasmus D. Leavitt, Charles R. Cross,
Benjamin O. Peirce, Amos E. Dolbear,

Benjamin A. Gould.

C. M. Warren Committee.

Francis H. Storer, Henry B. Hill,

Charles L. Jackson, Leonard P. Kinnicutt,
Samuel Cabot, Arthur M. Comey,

Robert H. Richards.

The chair appointed the following Standing Committees : -

Committee of Publication.

Charles L. Jackson, William G. Farlow,

William W. Goodwin.

Committee on the Library.
Samuel H. Scudder, Amos E. Dolbear,

G. Stanley Hall.

Auditing Committee.
Henry G. Denny, John C. Ropes.

OF ARTS AND SCIENCES. 345

Henry W. Williams read an obituary notice of Hermann
von Helmholtz.

The following papers were presented by title : —

On the Thermal Conductivities of certain poor Conductors.
By B. O. Peirce and R. W. Willson.

On the Simultaneous Partial Differential Equations :

Byk -/<^^^

By B. O. Peirce.

On certain Derivatives of unsymmetrical Tribrombenzol.
By C. L. Jackson and F. B. Gallivan.

Trinitrophenylmalonic Ester, ^y C. L. Jackson and J. I.
Phinney.

On Phenoquinone. By C. L. Jackson and G. Oenslager.

On the Nitrite of Bromdinitrophenylmalonic Ester. By
C. L. Jackson and J. H. Moore.

On Dinitrobromtoluol and some of its Derivatives. By
C. L. Jackson and M. H. Ittner.

Contributions from the Gray Herbarium of Harvard Uni-
versity, New Series, No. 9 : — 1. On the Flora of the Gala-
pagos Islands, as shown by the Collections of Dr. G. Baur.
2. New and noteworthy Plants, chiefly from Oaxaca, collected
by C. G. Pringle, L. C. Smith, and E. W. Nelson. 3. A
Synoptic Revision of the Genus Lamourouxia. By B. L.
Robinson and J. L. Greenman.

Studies in jMorphogenesis. 3. A Preliminary Catalogue of
the Processes concerned in Ontogeny. By C. B. Davenport.

On the motion of the Recording Secretar}^ it was

Voted, That it is expedient to publish in the Memoirs
Roland Thaxter's paper on Laboulbeniaceae.

Eight hundred and seventy-sixth Meeting.

October 9, 1895. — Stated Meeting.

The President in the chair.

The chair announced the death of Edward Samuel Ritchie,
Harold Whiting, and Henry Willard Williams, Resident Fel-

346 PROCEEDINGS OF THE AMERICAN ACADEMY

lows ; Daniel Cady Eaton, William Wetmore Stoiy, Associate
Fellows; and Thomas Henry Huxley, Louis Pasteur, Sven
Ludwig Loven, and Carl Friedrich Wilhelm Ludwig, Foreign
Honorary Members.

The Corresponding Secretary presented the following
letters : from the International Catalogue Committee of the
Royal Society of London ; from the Royal Society of New-
South Wales, announcing the conditions of competition for
its medal ; from the Royal Academy of Sciences of Turin,
announcing the death of its Secretary, Giuseppe Basso ; from
the family of Etienne Leopold Trouvelot, announcing his
death ; and from the Physico-economical Society of Konigs-
berg, announcing the death of its Honorary President, Franz
Ernst Neumann.

The following gentlemen were elected members of the
Academy : —

Paul Sebastian Yendell, of Dorchester, to be a Resident
P^'ellow in Class L, Section 1 (Mathematics and Astronomy).

Hammond Vinton Hayes, of Cambridge, to be a Resident
Fellow in Class I., Section 2 (Physics).

Benjamin Kendall Emerson, of Amherst, to be a Resident
Fellow in Class H., Section 1 (Geology, Mineralogy, and
Physics of the Globe).

Marie Alfred Cornu, of Paris, to be a Foreign Honorary
Member in Class I., Section 2 (Physics), in place of the late
Hermann Ludwig Ferdinand von Helmholtz.

Jacobus Henricus van't Hoff, of Amsterdam, to be a For-
eign Honorary Member in Class I., Section 3 (Chemistry),
in place of the late Jean Charles Galissard de Marignac.

John Trowbridge exhibited and read extracts from a col-
lection of letters written by Count Rumford to Marc Auguste
Pictet, and presented to the Academy by Jules Marcou, who
obtained them from the late Auguste de la Rive. It was
thereupon

Voted, That the thanks of the Academy be tendered to
Jules Marcou for his valuable gift.

Voted, That the letters be referred to the Rumford Com-
mittee and the Committee of Publication.

OP ARTS AND SCIENCES. 347

!Eiglit hundred and seventy-seventh Meeting.

November 13, 1895.

The Corresponding Secretary in the chair.

The chair announced the death of Asahel Clark Kendrick,
an Associate Fellow.

A letter was read from Thomas W. Higginson, tendering
his resignation as Councillor.

The following papers were presented by title : —

Thermo-electric Interpolation Formulae. By Silas W.
Hoi man.

Melting Points of Aluminum, Silver, Gold, Copper, and
Platinum. By S, W. Holmau, with R. R. Lawrence and
L. Barr.

Pyrometry : Calibration of the Le Chatelier Thermo-electric
Pyrometer. By Silas W. Holman.

Calorimetry : Methods of Cooling Correction. By Silas W.
Holman.

On some Points in the Development of ^cidia. By Her-
bert M. Richards.

William E. Story gave an informal talk on the new methods
of representing mathematical surfaces, and exhibited Pla-
teau's apparatus and a variety of models.

Eight hundred and seventy-eighth Meeting.

December 11, 1895.

The Academy met at the house of the President, at
Cambridge.

The President in the chair.

The resignation of Thomas W. Higginson as Councillor
was accepted.

The following papers were read: —

On the Temperature of the Crust of the Earth at great
Depths. By Alexander Agassiz and P. C. F. West.

Palestine in the Fifteenth Century B. C, according to
recent Discoveries. By Crawford H. Toy.

348 PROCEEDINGS OF THE AMERICAN ACADEMY

£iglit hundred and seventy-ninth Meeting.

Januaiy 8, 1896. — Stated Meeting.

The President in the chair.

The Corresponding Secretary read letters from A. Cornu
and J. H. van't Hoff, acknowledging their election as Foreign
Honorary Members ; and one from B. K. Emerson, accepting
Fellowship.

The vacancies in the Council occasioned by the death of
Henry W. Williams and the resignation of Thomas W.
Higginson were filled by the election of

William M. Davis, of Class H.,
Horace E. Scudder, of Class HI.

The following gentlemen were elected members of the
Academy: —

Augustus St. Gaudens, of New York, to be an Associate
Fellow in Class HI., Section 4 (Literature and the Fine Arts),
in place of the late William Wetmore Story.

Hermann Graf zu Solms-Laubach, of Strasbnrg, to be a
Foreign Honorary Member in Class II., Section 2 (Botany),
in place of the late Marquis de Saporta.

Carl Gegenbaur, of Heidelberg, to be a Foreign Honorary
Member in Class II., Section 3 (Zoology and Physiology), in
place of the late Thomas Henry Huxley.

Willy Kiihne, of Heidelberg, to be a Foreign Honorary
Member in Class II., Section 4 (Medicine and Surgery), in
place of the late Charles Edouard Brown-Sdquard.

The President alluded to the fact that the Massachusetts
Historical Society had informally signified to the Academy
its willingness to make a ten years' lease of quarters in its
proposed building on the Back Bay. After a brief discussion,
it was

Voted, That the question of securing new quarters for the
Academy be referred to the Committee of Finance, to report
at the next meeting.

On the motion of the Corresponding Secretary, it was

Voted, To meet, on adjournment, on the second Wednes-
day in February.

OP ARTS AND SCIENCES. 349

Edwin H. Hall presented by title a paper On the Thermal
Conductivity of Mild Steel.

Henry Taber announced that he had proved the" sufficiency
of certain conditions that in Volume XL VI., page 583, of the
Mathematische Annalen, he had shown to be necessary in order
that a transformation of the group whose invariant is a cer-
tain linear complex may be generated by the repetition of an
infinitesimal transformation of the group.

Eight hundred and eightieth Meeting.

February 12, 1896. — Adjourned Stated Meeting.

The President in the chair.

The chair announced the death of Martin Brimmer and
Richard Manning Hodges, Resident Fellows.

The Corresponding Secretary read a letter from Augustus
St. Gaudens, acknowledging his election as Associate Fellow.

The President reported that no change of quarters for the
Academy was feasible at present.

George L. Goodale read a paper on Forestry under New
England Conditions.

John Trowbridge gave an informal account of his experi-
ments with the cathode rays.

No evidence of refraction was detected. Wooden lenses,
both double convex and double concave, were tried, and
apparently did not affect the rays. In Helmholtz's discussion
of the electro-magnetic theory of light, there is a longitudinal
wave which travels with an infinite velocity. Such a wave,
travelling faster than the velocity of light, would not suffer
refraction. The photographs procured by the new rays are
not strictly shadow pictures, such as may have been obtained
before in the electro-static field. They show a specific ab-
sorption which is a new phenomenon. For instance, a disk
of microscopic cover glass, j^-^ of an inch thick, perfectly
transparent to the ordinary rays of light, throws as strong a
shadow as a board one inch thick. The bones of the hand
throw a stronger shadow than the flesh surrounding them.

350 PROCEEDINGS OP THE AMERICAN ACADEMY

The new phenomenon promises to be of great assistance in
studying the extremities of the human body, and in detect-
ing the presence of metallic bodies and fragments of glass in
them. A shot, placed on the back of the hand, can be photo-
graphed through the flesh of the thickest portion of the hand.
Of course, therefore, it could be detected if it were imbedded
in the hand.

Pictures have been taken at the Jefferson Physical Labora-
tory, after an exposure of one minute, which show the bones
of the hand, and there is no doubt that the time of exposure
can be reduced to a few seconds for certain portions of the
extremities. The Crookes tubes can be protected from injury
by immersing their terminals in a suitable oil, — preferably,
boiled linseed oil.

The following paper was read by title : —

On the Group of real Linear Transformations whose Invari-
ant is an Alternate Bilinear Form. By Henry Taber.

Eight hundred and eighty-first Meeting.

March 11, 1896. — Stated Meeting.

Vice-President B. A. Gould in the chair.

The Corresponding Secretary read letters from Carl Gegen-
baur, W. Kuhne, and H. Graf zu Sohns-Laubach, acknowledg-
ing their election as Foreign Honorary Members ; from the
Natural History Societ}^ of Bonn, announcing the death of its
Secretary, Philipp Bertkau ; from the Joint Commission of the
Scientific Societies of Washington, D. C, with regard to the
proposition for a Director in Chief of Scientific Bureaus in
the Department of Agriculture.

On the motion of the Corresponding Secretary, it was

Voted, To meet on adjournment on the 8th of April.

W. M. Davis read a paper on the Outline of Cape Cod.

C. L. Jackson made the following announcement : —

C. L. Jackson and A. M. Comey have found that potassic co-
balticyanide is converted by boiling nitric acid into a red jelly,
having the formula KH.2Co,;(CN)iiH^O, from which a silver salt,
Ag3Co3(CN)uH20, and a barium salt, BaHCo3(CN)xiliH20,

OF ARTS AND SCIENCES. 351

were obtained. All of these substances have been analyzed.
Potassic ferricyanide gives a similar black jelly under the
same conditions.

On the motion of Eliot C. Clarke, the following resolutions
were adopted : —

Whereas, The Committee on Coinage, Weights, and Meas-
ures of the National House of Representatives has reported
a bill requiring the early adoption of the Metric System of
Weights and Measures by all Departments of the Govern-
ment, and its adoption by the whole nation, at a subsequent
fixed date, as the only legal system, —

-Resolved, That the American Academy of Arts and Sci-
ences, renewing its recommendation made in former years,
urges Congress to consider favorably this bill, and thus permit
the United States to join the majority of civilized nations as
regards its system of weights and measures.

Resolved^ That this Academy believes that the universal
adoption of this system will aid researches in physical science,
and also commercial transactions, and will tend to bring about
the fraternity of nations.

Resolved, That copies of these resolutions be sent to the
presiding officers of the two Houses of Congress, to the Sen-
ators from Massachusetts, to the chairman of the Commit-
tee of Finance of the Senate, and to the chairman of the
Committee on Coinage, Weights, and Measures of the House.

Eight hundred and eighty-second Meeting.

April 8, 1896. — Adjourned Stated Meeting.

The Academy met at the Walker Building of the Massa-
chusetts Institute of Technology, Boston.

Vice-President B. A. Gould in the chair.

The following gentlemen were elected members of the
Academy : —

John Stone Stone, of Boston, to be a Resident Fellow in
Class I., Section 2 (Physics).

Robert Wheeler Willson, of Cambridge, to be a Resident
Fellow in Class I., Section 2.

352 PROCEEDINGS OF THE AMERICAN ACADEMY.

Theobald Smith, of Boston, to be a Resident Fellow in
Class II., Section 4 (Medicine and Surgeiy).

Charles Lane Poor, of Baltimore, to be an Associate Fellow
in Class L, Section 1 (Mathematics and Astronomy), in place
of the late James Edward Oliver.

Robert Simpson Woodward, of New York, to be an Asso-
ciate Fellow in Class I., Section 4 (Technology and Engi-
neering), in place of the late John Newton.

Basil Lanneau Gildersleeve, of Baltimore, to be an Associate
Fellow in Class III, Section 2 (Philology and Archaeology), in
place of the late William Dwight Whitney.

Thomas Raynesford Lounsbury, of New Haven, to be an
Associate Fellow in Class III., Section 2, in place of the late
Asahel Clark Kendrick.

Karl Theodor Weierstrass, of Berlin, to be a Foreign Hon-
orary Memberr in Class I., Section 1 (IMathematics and As-
tronomy), in place of the late Arthur Cayley.

Michael Foster, of Cambridge, Eng., to be a Foreign Hon-
orary Member in Class II., Section 3 (Zoulogy and Physiology),
in place of the late Carl Friedrich Wilhelm Ludwig.

Alexander Onufrijevic Kovalevskij, of St. Petersbui-g, to be
a Foreign Honorary Member in Class II., Section 3, in place
of the late Sven Ludwig Loven.

Karl Weinhold, of Berlin, to be a Foreign Honorary Mem-
ber in Class III., Section 2 (Philology and Archaeology), in
place of the late Sir Henry Creswicke Rawlinson, Bart.

Friedrich Hermann Grimm, of Berlin, to be a Foreign
Honorary Member in Class HI., Section 3 (Pliilosophy and
Jurisprudence), in place of the late Sir John Robert Seeley.

On the motion of the chairman of the Rumford Committee,
it was

Voted, To present the Rumford Medal by proxy at the
annual meeting.

David G. Lyon gave an informal talk on recent Assyrian
discoveries, with illustrations.

E. L. Mark presented the following paper by title: —

Preliminary Notes on the Embryology of the Starfish
(^Asterias pallida). By Seitaro Goto.

AMERICAN ACADEMY OF ARTS AND SCIENCES.

Eeport of the Council. — Presented May 13, 1896.

BIOGRAPHICAL NOTICES.

Richard Manning Hodges By David W. Cheeveu.

Harold Whiting John Trowbridge.

Edward Samuel Ritchie Amos E. Dolbear.

Martin Brimmer William Everett.

Henry Wheatland F. W. Putnam.

James Edward Oliver Gustavus Hay.

Viscount Ferdinand de Lesseps Henry Mitchell.

VOL. XXXI. (n. S. XXIII.)

23

REPORT OF THE COUNCIL.

Since the annual meeting of the 8th of May, 1895, the
Academy has lost by death thirteen members : — five Fellows,
Martin Brimmer, Richard Manning Hodges, Edward Samuel
Ritchie, Harold Whiting, and Henry Willard Williams ; four
Associate Fellows, Daniel Cady Eaton, Atticus Greene Hay-
good, Asahel Clark Kendrick, and William Wetmore Story ;
four Foreign Honorary Members, Thomas Henry Huxley,
Sven Ludwig Loven, Carl Friedrich Wilhelm Ludwig, and
Louis Pasteur.

RESIDENT FELLOWS.

RICHARD MANNING HODGES.

Dr. Richard Manning Hodges was born in Bridgewater, Mas-
sachusetts, November 6, 1827. He was fitted for Harvard College at
the Boston Latin School, and was graduated in 1847.

Dr. Hodges took the degree of A. M. in due course, and that of
M.D. in 1850. He was Demonstrator of Anatomy from 1853 until
1861. In 1866 he was appointed Adjunct Professor of Surgery. He
resigned his professorship in 1872.

In 1863 he was chosen one of the visiting surgeons of the Massa-
chusetts General Hospital. In 1885 he resigned his position, and
gave up the practice of surgery. In 1891 he entirely ceased to
practise.

His last illness was short, and he died on February 9, 1896.

Dr. Hodges served twice on the Board of Overseers of Harvard
College. He was a Resident Fellow of the American Academy
of Arts and Sciences.

As a man he was sincere, straightforward, open, just, positive,
punctual, not to say punctilious. Possessed of a strong body, he was

356 HAROLD WHITING.

active, and an untiring worker. He held the confidence of his
patients to a remarkable degree. He was a good diagnostician ; a
logical reasoner ; and possessed great common sense.

He was equally loyal to his profession, and t-o his brother physi-
cians. As an anatomist he was exact and thorough ; expert and
dexterous, his dissections were more than excellent, they were
beautiful.

He contributed largely and wisely to the Warren Anatomical
Museum with the work of his own hands. Many of his preparations
of coarse anatomy are, and will remain, unsurpassed.

From practical anatomy to surgery the step was short. He be-
came a rapid and skilful operator. He was also so well grounded
in Surgical Pathology that he was a thoughtful, level-headed, and
much valued consultant.

As a writer he was concise and clear. His " Dissector " went
through several editions, and was of the utmost value to students.
Unsullied by pictures, it ligliteued the work of the young anatomist by
clear and true descriptions, by accuracy, and by brevity. His essay
on the " Excision of Joints " won the Boylston Prize in 1861.

His observations on " Spiroidal Fractures" and on " Pilo-nidal
Sinus " were original. His latest work was a " History of the Dis-
covery of Anaesthesia," which will endure as a complete and careful
account of that great surgical event.

Dr. Hodges did a large share of public gratuitous service in the Chol-
era Hospital ; in the Boston Dispensary ; at the Massachusetts General
Hospital; at the State House during the Civil War, on the examining
board for surgeons ; and also as a volunteer surgeon sent to the seat
of war.

His quick and buoyant manner, his keen insight, decision of char-
acter, and honesty, would have insured success in any pursuit ; and
they won for him an enduring reputation as a skilled anatomist, a bold
yet conservative surgeon, and a reliable observer and physician.

1896. David W. Cheever.

HAROLD WHITING.

Professor Harold Whiting was born in Roxbury, May 13,
1855. He was fitted for Harvard University at the Roxbury Latin
School, and graduated from the University with the degree of A. B.
in 1877, of A. M. in 1878, and of Ph. D. in 1884. He was Instructor
in Physics in the University from 1883 to 1891. In 1892 he was
appointed Associate Professor of Physics in the University of Cali-

HAROLD WHITING. 357

fornia, and was lost at sea, with his wife and four children, while
returning to Cambridge, on May 27, 1895.

This bare recital of the principal epochs in Lis life is like a mere
pen and ink sketch of a vivid personality, lacking color, and convey-
ing no adequate idea of the man whose career was so suddenly closed.
This personality was so intense that one feels it difficult to realize
that he has left us, and one half expects to meet him on turning some
corner.

He early manifested a remarkable aptitude for scientific subjects.
When little older than six or seven, it is related that he used to sit in
an arm-chair for long periods, his head sunk on his breast, and when
spoken to he would say, " Please don't interrupt me ! I have almost
got the theory." He was always observing, as well as thinking, even
before he could speak plainly, coming home from drives or walks with
such revelations as this : " I have found that an island is a steady
thing."

My attention was first called to "Whiting when he was a Sophomore.
I was hearing a recitation in Physics, and had made some remarks
upon a scientific point. He arose and stoutly denied the truth of my
assertion. The class tried to suppress him by hissing, and by " Wood-
ing up," but he maintained his ground. I found that he was right,
and from that time began to observe him more closely. He was both
morally and physically courageous. While sailing in the harbor of
Plymouth he was capsized, and remained for a long time in a perilous
position in the water shut in by a fog. At length a fisherman hove
in sisht. Whitinsf, immersed in the water, took off his hat and made
the fisherman a low bow. The latter remarked, " Had n't you better
git in? "

Dr. Whiting had a keen sense of humor mingled with a subtle wit.
There was nothing unkind in this wit, for he had too generous a heart
to knowingly wound any one. On looking over the proof of one of
his scientific papers I was puzzled by a certain involved mathematical
expression, and turned to him for an explanation. It was in reality a
very simple formula, and he remarked, in apologizing for its abstruse
form, " I have been so much annoyed by the involved mathematical
expressions of the English school of mathematicians that I determined
to give them a nut to crack."

Dr. Whiting's idiosyncrasies were so strong that no team work was
possible for him. He could not be hitched up with any one. His
mind seemed to play about a subject much as certain forms of electricity
dart hither and thither about a summer cloud, frequently illumining

358 HAEOLD WHITING.

obscure regions in a surprising way. It was highly interesting to see
him make an imperfect piece of apparatus give wonderfully good re-
sults ; and while he was Instructor in Physics in the Jefferson Physi-
cal Laboratory he was of great service to the department, acting like a
skirmisher in the growing subject of Laboratory teaching of Physics.
Those who followed him profited both by his mistakes and his suc-
cesses, and could afford to pardon the mistakes, which were those of a
courageous explorer in a new field. His fertility of mind was remark-
able, and he often said of himself that he was like a codfish which lays
a million eggs, and only one or two perchance hatches. This fertility
and brilliancy were such that many of those with whom he was asso-
ciated often remarked that they would not be surprised if Whiting
should hit upon something remarkable in Science. If he had dis-
covered, for instance, the X-rays, many of us would have said, " It was
just like Whiting to look through his hand at a Crookes tube."

The physical department of Harvard University is indebted to him
for many valuable suggestions, and also for pecuniary contributions.
He did not hesitate to aid it whenever he saw its needs, and by his
will he gave twenty thousand dollars for Fellowships in Physics in
Harvard University.

His scientific work began with an investigation of magnetic waves
on iron and steel rods, which was published in the Proceedings of the
American Academy of Arts and Sciences. His thesis for the degree
of Doctor of Philosophy in Harvard University was on the Tlieory of
Cohesion. While Instructor in the University he published a Sylla-
bus of a Hundred Physical Measurements, for the use of the Jefferson
Physical Laboratory ; and also a valuable treatise on Physical IMeasure-
ments, consisting of three large volumes. While at the University of
California he also published full sets of laboratory notes. When we
examine his life's work we find that it gave a decided stimulus to
the modern laboratory method of teaching Physics by quantitative ex-
periments rather than by qualitative. If he had elaborated many of
his ingenious methods into papers for periodicals, the list of his works
would have been much longer.

Many could come forward and testify to his generous hand, as well
as to his generous mind. With his rich qualities of scientific imagina-
tion, experimental skill, and mathematical ability, joined to the stead-
iness of middle age, much could have been expected of him. He
still lives in his gift to the young Physicists of the University.

1896. John Trowbridge.

EDWARD SAMUEL RITCHIE. 359

EDWARD SAMUEL RITCHIE.

Edward Samuel Ritchie, son of John and Eliza (Eliot) Ritchie,
was born in Dorchester, Massachusetts, August 18, 1814. After liv-
ing some years in Dorchester, his father moved to North Bridgewater.
During the years 1827 and 1828 he attended school at the Friends
Academy in New Bedford. In 1829 he was taught by Rev. John
Goldsbury, in North Bridgewater, studying mornings and working for
a furniture maker in the afternoons, as he had mechanical aptitudes,
and wished to learn the use of tools.

Early in life he showed great interest in art and in science. He
was the only surviving child of a family of six, and his father gave him
every advantage to help him in studies in which he was particularly
interested. His health was extremely delicate in youth, and that added
to a very sensitive nature prevented him from taking a collegiate educa-
tion, which his father was anxious he should have. He had a labora-
tory to work out experimentally what interested him, and was a very
close student. Having great power of concentration, he was entirely
oblivious to everything around him when he was particularly interested
in any subject.

He had also a great love for music, and was a good musician, giving
his services as an organist for several years to the Episcopal Church
in New Bedford, in which he was senior warden.

While living in New Bedford he constructed a telescope for his
own use, which he afterwards sold to the Friends Academy, where he
had formerly been a scholar.

He was much interested in sculpture, and has left very creditable
work in several cameos and a nude figure, two thirds life-size, of a
nymph of his own posing. He made the clay figure, plaster cast, and
cutting in marble, doing all the work from the beginning. He thought
seriously of going to Rome to make that art a life study, but, being a
devoted son, was unwilling to be separated from his aged mother.

In 1850 he entered into partnership with N. B. Chamberlain, a
philosophical instrument maker. His business previous to this had
never been pleasant to him, but this was quite to his taste. After a
short time the partnership was dissolved, and Mr. Ritchie continued
the business alone.

His improvement in the induction coil brought him into public notice.
In 1851 Ruhmkorff of Paris constructed the coil which yet bears his
name. He succeeded in producing sparks about two inches in length.
Ritchie perceived that the defect of the Ruhmkorflf coil was insuflSicient

360 MAETIN BRIMMER.

insulation of the secondary coil. He concluded that, if this were divided
into sections properly insulated from each other, the device would be
more efficient and give a longer spark. On trial, his expectations were
realized. One of these coils was exhibited at a meeting of the British
Association held in Dublin in 1857, and afterwards at the University
of Edinburgh. A description of Ritchie's coil was published in
Silliman's Journal and in the Journal of the Franklin Institute.
M. Ruhmkorff procured one, and, copying it successfully, received a
prize from the French government for it, — a proceeding which
greatly disappointed Mr. Ritchie, who was entitled to it. The
improvement of Mr. Ritchie transformed the coil from being a toy
giving a two-inch spark to an instrument capable of giving a flash
two feet or more in length, and approaching the characteristics of
lightning.

At the time of our Civil War Mr, Ritchie's attention was called to
the need of a better compass for our navy. The English Admiralty
Compass, considered the finest in the world, was in general use at that
time. In order to aid his study in making his improvements in this
instrument, he made a support so constructed as to give the motions
of a vessel at sea.

After much thought and labor he invented the Monitor and Liquid
Compasses. The former did good service during the war, and the latter
was at once adopted by the Navy, and is now in use all over the world.

He also constructed about that time another instrument which was
a great help to the Navy, the Theodolite, fastened to a pendulum hang-
ing in a tank of water, which enabled surveys to be taken of the harbors
on the Atlantic and Gulf coasts. For these inventions of high merit
he will be long remembered by the scientific world.

He was an exceedingly conscientious man, and was ever ready to
help others over difficulties which he had overcome himself, and some-
times such persons received the credit and financial profit which rightly
belonged to him.

He died on June 1, 1895, in his eighty-first year.

1896. A. E. DOLBEAR.

MARTIN BRIMMER.

The various distinguished bodies to which our deceased associate,
Hon. Martin Brimmer, belonged, have already paid him such
varied and appreciative tributes that a detailed biography, in the
ordinary sense of the word, would be quite out of place. Nor was his
life itself so distinguished by striking adventures or significant dis-

MARTIN BRIMMER. 361

coveries that a biographer could find the details, if he were disposed
to use them. As far as his life had a story, it is shortly told.

Martin Brimmer was the son of Hon. Martin and Harriet
[Wadsworth] Brimmer, and was born in Boston, 9 December, 1829.
His father was a well known and most public spirited citizen, twice
Mayor of Boston. For him is named the Brimmer School, and on
the older maps of Boston T Wharf is described as " Brimmer's T."
Our associate entered the Sophomore Class of Harv^ard College at the
age of sixteen, and graduated in 1849. Without being distinguished
as a scholar, he won the very peculiar regard of all who were
associated with him, as instructors or companions, even under circum-
stances where many young men would have made a different impression.
He travelled in Europe soon after graduation, and soon after his
return began a connection with a great number of literary and chari-
table societies as trustee, with one or other of which he was constantly
engaged to the last. He was chosen a Fellow of Harvard College at
an unusually early age, but with universal approval, and, having
resigned this post, was again chosen to it and held it till his death,
sitting for a part of the interval on the Board of Overseers. He was
a member of the Massachusetts House of Representatives in 1859,
1860, and 1861, and of the State Senate in 1865. He was a Presi-
dential Elector in 1876, and a candidate for Congress in 1878. In
1869 he was chosen the first president of the Boston Museum of Fine
Arts, and held the post till his death. He visited Europe and Egypt
more than once after his first journey ; and in the early days of the
struggle for the territory of Kansas, had travelled there to find out
for himself the truth of a situation so passionately discussed by oppos-
ing partisans. A journey in Kansas in 1855 was a more arduous
affair than one to Egypt a generation later. Mr. Brimmer died
January 14, ]8'.)6. This may be said to be his biography, unless
one gave a dptailed list of all the bodies of which he was the devoted,
energetic, intelligent servant, — unless, also, one went into the details
of private life, where Mr. Brimmer indeed shone with an unequalled
light, but one whose lustre was far too tender and sacred for public
exposure.

But if life means not events but character, not what one has
done but what one has been, Mr. Brimmer's is a memory which it
is peculiarly incumbent on us to record and to cherish. In Virgil's
matchless and immortal roll of those who have won eternal happiness,
he ranks with the patriot soldier, the inspired bard, the stainless
priest, and the keen inventor, " those who have made others remember

362 MARTIN BRIMMER.

them by deserving it." No man ever deserved to be remembered
better than our late associate.

He was born and brought up in the midst of all those things which
are commonly held to excuse and incapacitate men from hard work.
He had an ample fortune, so secured that it might be enjoyed and not
dissipated ; he had an assured social position, which exempted him
from all toil or strife to bring his name into prominence ; and a slight
physical infirmity might have been held in his case, as in that of the
historian Prescott, rather to justify idleness than otherwise. Mr.
Brimmer yielded to none of these allurements. He conceived that he
held all his personal and social advantages as in trust for the com-
munity. He positively enjoyed to work for the strengthening of all
that is good, and the suppression of all that is evil, in modern society.
By example and by precept, by personal labor and by contributions of
money, perhaps most of all by the fact that he was known to be
always upon the side of what was high, noble, strong, and lovely,
whether he was actually speaking, giving, or working, he was a living
proof that what are sometimes censured or ridiculed as the showy
fungi of a decaying civilization may be really the healthy flowers of a
new and hopeful republic.

The thoughtful student of our society, its merits and its wants, must
see clearly that one serious danger to our happiness and prosperity
arises from the temper of second thought, — of suspicion and distrust
of ourselves and others. A vast number of our ablest, wisest, and
most virtuous citizens seem unable to execute their highest purpo.ses
without tormenting themselves all the time by some speculation as to
what secondary effect their action may have on themselves or others.
Mr. Brimmer combined with the soundest and most cultivated intelli-
gence an absolute simplicity of character. Open or reserved as the
case demanded, whenever he did speak or act he was perfectly sincere.
He was by no means without honorable ambition ; but it was an am-
bition held in strict subservience to courtesy, to honor, and to conscience.
He was firm in his opinions and distinct in their expression ; but it
could only be a very mean or a very brutal person who could be
offended by his high-minded and polite refusal to agree to what he
thought wrong. The word " culture," so sadly soiled and travestied at
the present day, had in Mr. Brimmer its perfect fulfilment. He stood
to uncultivated men as an apple does to a crab. In this age, which
fancies mere tartness or bitterness constitute flavor, such a presence as
his was a living instance of how much the raciest nature is improved
by the development of sweetness and tenderness.

HENRY WHEATLAND. 363

Of all Mr, Brimmer's public services, if we are to make the invidi-
ous task of selection, the highest place may be given to his work in
the Art Museum. Perhaps other men could have filled his place in
other institutions equally well ; in this he was without a possible
rival. By disposition and training alike, he was fitted to be a perfect
judge and patron of fine art ; and if Boston is ever to keep her head
above the overwhelming gulf of pretension and mediocrity that is pour-
ing over the country in matters of art, she will owe her salvation to
him more than to any single man. This work elicited from him other
work of exquisite power, for which his adaptation had hardly been
suspected. He delivered one or two addresses on the importance of
the fine arts, which were not merely sound, elegant, and manly, but
rose in more than one passage to thrilling and convincing eloquence of
a kind rare indeed in these days.

This Academy, like the community, was the better for his member-
ship, and his place will long be unsupplied.

1896 William Everett.

HENRY WHEATLAND.

Henry Wheatland was elected a Fellow of the Academy in
1845. He was born in Salem, January 11, 1812, and died there,
February 27, 1893. His father was Richard Wheatland, born in
Wareham, Dorset County, England, in 1762, who came to America
in 1783. For several years he sailed from the port of Salem as com-
mander of vessels in the India trade. In 1801 he retired from the
sea and became one of the prosperous India merchants who helped to
make the fame of the old town in the palmy days of its commerce.
In 1796 Captain Wheatland was married to his second wife; and
Henry was the sixth and youngest child of this marriage. As a boy
he was of a delicate constitution, and, being naturally disposed to study,
his parents had him fitted for college in the Salem schools. At the
age of sixteen he entered Harvard, and was graduated in the class of
1832. His taste for natural history was evidently formed in boyhood,
for we find that in the last year of his college course he was active in
the formation of " The Harvard Linnean, " of which college society
he was the Secretary. The Constitution of this society, as he wrote it,
is among his papers. This was probably the immediate precursor of the
present Harvard Natural History Society, which was formed in 1837.
On leaving college he returned to Salem and became an active
worker in the Essex County Natural History Society and the Essex
Historical Society.

364 HENRY WHEATLAND.

In order to carry on his scientific studies he followed the course
which seemed at that time almost essential to a student of natural his-
tory ; he entered the Harvard Medical School, attending lectures in
Boston in the winter and studying with Dr. Abel L. Pierson in Salem
during the remainder of the year. In 1837 he received the degree of
M. D. That it was not his intention to practise medicine, unless
forced by circumstances to earn a living in that way, I know from
frequent conversations with him and from his advice to me, when,
aroused by Agassiz' visit to Salem, in 1856, I wished to accept the
offer made to me to become his student. At that time Dr. Wheat-
land said, "You can go to Cambridge and study under Agassiz,
Wyman, and Gray, and prepare yourself to enter the Medical School
and become a doctor, as I did ; then you can get your living in that
way, if you have to, and study natural history too. That is the way
most naturalists have done. " In my early days, and still more in
his, to follow natural history as a profession and a means of liveli-
hood was hardly to be considered. It is evident that Dr. Wheatland
gave as much attention to the comparative anatomy of animals as he
did to the special anatomy of man, for during this time he prepared
many skulls and skeletons for the collection of the Essex County
Natural History Society, which, with others prepared at a later time,
are still preserved in the Peabody Academy of Science in Salem.

The Doctor was always filled with a quiet enthusiasm for his work,
never demonstrative, and even painfully reserved in his manner in
public ; only those who knew him best and were by their work closely
associated with him found out his true nature, and realized how much
he accomplished in his quiet, j^ersistent way. Many a time I have
seen the face of this reserved and quiet man beam with delight on
obtaining some skull new to the collection, or when bringing up in his
little dredge a seaweed or shell new to him. Often when in a dory
dredging off the shore of AVinter Island, Marblehead, Swampscott,
or Manchester, his favorite localities for a half-day's outing, I have
seen him as enthusiastic and happy over the contents of the little
dredge as any naturalist of to-day could be on seeing for the first time
the animals brought up from great depths by the modern appliances.
I think it can be safely claimed for Dr. Wheatland, that he was the
first to dredge in our New England waters, and I believe he was the
first naturalist in America to adopt this means of collecting animals
and plants living on the ocean bottom at moderate depths.

It was during the most active time of his natural history days that
the Geological, Botanical, and Zoological Survey of the State was car-

HENRY WHEATLAND. S65

ried on, and the Doctor contributed his full share in specimens and
observations during his constant association with Emerson, Storer,
Gould, Harris, and others ; while Stimpson of a later date always
acknowledged that he took his first lessons in dredging of Dr. Wheat-
land. To Agassiz he sent many specimens when the latter began to
make his famous Museum at Cambridge ; and with many conchologists
abroad he carried on active exchanges, which added much to the early
importance of the natural history collections in Salem. For years
after I became intimately associated with him, in my boyhood, in the
work of the Essex Institute, the Doctor continued his preparation of
skulls of such mammals as he could obtain, many heads being brought
home to him from foreign countries by Salem sea captains. These
heads the Doctor soaked in tubs of water kept in the yard at his home,
and bleached on shelves he prepared for the purpose on the roof of the
barn. He daily watched and worked over these specimens for hours
at a time, and finally placed them clean and white in the cases in the
Institute. In those days every collector was obliged to prepare his
own specimens ; and if a rare fish or reptile came to the Society and
there was no money for the purchase of alcohol, which was generally
the case, the Doctor would prepare the skin and "mount" the speci-
men. It was his hands that prepared, over sixty years ago, the large
specimen of horse-mackerel which still hangs upon the walls of the
Peabody Academy of Science, and the enormous lobster, the wonder
of the present day, which is treasured by the Academy.

It was the Doctor's practice of saving in some way every important
specimen which he secured that made the series of '• stuffed " turtles
and their prepared shells and skeletons of such importance as to call
Agassiz to Salem. On this occasion I was first brought in contact
with the great naturalist, which event changed the whole course of
my life ; and it was thus through the training of Doctor Wheatland
that I entered upon my career of scientific pursuits. In acknowledg-
ing him, my life-long friend, as my first instructor in science, I but
give credit due to one who helped many others in a similar manner, —
one whose friendship was always true and lasting, and whose useful-
ness and influence in the community were widespread.

While Dr. Wheatland was a true naturalist and did much to encour-
age the study in others, and unquestionably aided to a considerable
extent the impetus given to its study in Salem, he became in later
years equally interested in local historical and genealogical researches.
As younger men gradually took up his natural history work, he turned
his attention wholly to historical matters and his brain became a

366 HENRY WHEATLAND.

wonderful treasure-house of genealogy and local history. While he
published but little, he was ever helping others to prepare papers.
Many were the hours and days he gave to rendering such assistance,
and to making critical revisions of manuscripts submitted to him. It
was in such work that his kindly nature was tested to the utmost, but
uever did he refuse to give to others for use or publication the results
that he had worked long and diligently to secure. I have often heard
him remark, " It makes no difference who publishes or gets credit for
a fact that I have found, so long as it is made known to the world, or
a mistake is corrected."

The life work of such a man as Henry Wheatland would naturally
culminate in some important result to the community in which he
lived, and the result of his life work can be best expressed by the
words The Essex Institute. This important and remarkable institu-
tion is his memorial. Henry Wheatland is rightfully the acknowledged
founder of the Essex Institute. It was through his efforts that in
1848 two societies were brought together which for a number of years
had their home and principal membership in Salem, — the Essex His-
torical Society and the Essex Institute. To the subjects already fos-
tered by these societies was added the encouragement of art and
horticulture. The formation of a library was also included in the
new organization, and plans were made for the publication of the
Proceedings of the Institute and of scientific and historical papers.

Of this Institute Dr. Wheatland was the Secretary and sustaining
power, giving of his moderate competence to its needs, and working
day and night for its advancement, without compensation or thought of
reward except in the successful growth of the institution and the ac-
knowledgment of its usefulness by the community. From the small be-
ginnings of nearly half a century ago, the Institute has grown as a sturdy
tree of knowledge. It now has a considerable membership of devoted
workers, who appreciate what has come to them and realize its useful-
ness and influence in the community. It has a home of its own and
considerable invested property, which insure its perpetuity. It is a
power for education and culture, and for all that calls forth the higher
aspirations of man. It has set an example which has been followed in
many places, and it has added to the sum of human knowledge by its
numerous publications. Such has been the result of the life and labors
of Henry Wheatland, - that gentle persistent worker whose aim was
ever to help others in their researches ; to save from destruction for
the use of the future student the manuscripts he would require in his
studies ; to furnish to the people the ready means of obtaining a knowl-

JAMES EDWARD OLIVER. 367

edge of the natnral history of the region by forming a perfect col-
lection of the rocks, minerals, plants, and animals of Essex County ; to
practically encourage the cultivation of fruits, flowers, and vegetables ;
to form a scif iitific and historical library for the benefit of all who
wished to study ; to foster research and to aid in the diffusion of
knowledge. All this Doctor Wheatland lived to see carried forward
far beyond his expectations. He died content with his work ; and he
has left a priceless legacy to the city of his birth. With his death the
last of the old school of naturalists has passed away. New methods
and new theories have made rapid advances, and a second genera-
tion, after his active working days, has entered the ever-widening
field of scientific research, until now the times are changed, and in-
stead of its being necessary to become a doctor of medicine in order to
be a naturalist, a physician must be something of a naturalist in order
to hold his position in the medical profession.

1896. F. W. Putnam.*

ASSOCIATE FELLOWS.

JAMES EDWARD OLIVER.

James Edvtard Oliver, who died on March 27, 1895, in the sixty-
sixth year of his age, was born in Portland, Maine, July 27, 1829,
of Quaker parentage. The family subsequently removed to Lynn,
Massachusetts, and there young Oliver fitted for college at the Lynn
Academy. He entered Harvard as a Sophomore, graduated in 1849,
and was the class poet.

One of his classmates writes of him that " he was a modest, diffident,
retiring, self-absorbed person in college, doing work not to be ashamed
of in other branches, but achieving distinction only in mathematics."

* In these brief reminiscences of the career of Dr. Wheatland, and of the
remarkable influence he exerted on the life of many young men and women,
as well as upon the community in which he lived, I have not attempted a sketch
of his life, nor have I alluded to many events of special interest. Some of these,
and a list of the important offices he held, the societies that conferred member-
ship upon him, and the titles of his publications, are to be found in the pamphlet
published by the Essex Institute, containing an account of the meeting of the
Essex Institute held on April 17, 1893, "in memory of its late President";
also in the Memoir by William P. Uphani, printed in the Proceedings of the
Massachusetts Historical Society, 1895; and in memorials of various other
societies.

368 JAMES EDWARD OLIVER.

It may be added that he was one of the ablest pupils of the elder
Peirce.

With regaid to his mathematical ability, Cajori * writes that " in
1849 he had already displayed extraordinary mathematical power,"
and " in the Harvard Catalogues of 1854 and 1855 we find J. E. Oliver
taking advanced courses of mathematics such as were offered at that
time by no other institution in the land."

Shortly after graduation he received an appointment in the Nautical
Almanac office in Cambridge under Professor Peirce, where he met
several men of unusual mathematical ability.

In 1861 he was elected a Fellow of this Academy, and in 1873 an
Associate Fellow. In 1871 he was appointed an Assistant Professor
of Mathematics at Cornell University, and in 1873 he was appointed
Professor, and retained the office during his life.

He was also a Fellow of the American Philosophical Society, of the
American Association for the Advancement of Science, and of the
National Academy. He was also a member of the Council of the
American Mathematical Society.

Professor Oliver's published communications on mathematical sub-
jects may seem fewer than might have been expected, considering his
great ability. He seems to have been actuated less by a regard for repu-
tation than by what he considered as his immediate duty. Mrs. Oliver
writes that '• his chief original work was done before his advanced
students," and that, " when his intellectual curiosity was satisfied, he
begrudged the time necessary to write it out for publication."

Professor Burr (Cornell Daily Sun, April 3, 1895) writes of him
that "his mind was too discursive in its method and too unpractical
in its bent to lead him largely into publication, and it is as a teacher
and a man that he will be longest and most aflfectionately remembered.
He was absent-minded, unmethodical, prone to digression, but his
acuteness of mind, his power of sustained research, his comprehensive-
ness of view, his utter freedom from bias, his unflagging enthusiasm,
made his leadership for those who had the wit and mettle to follow it
a thing of perpetual inspiration."

Besides these peculiarities of his intellectual temperament, if I may
use such an expression, which were without doubt unfavorable to
publication of original results, there was also another difficulty. The
excessive work required of him as mathematical professor at Cornell

* Bureau of Education, Circular of Information No. 3, 1890, p. 178. The
Teacliing and History of Mathematics in the United States, by Florian Cajori,
M. S. (University of Wisconsin), etc. Washington, 1890.

JAMES EDWARD OLIVER. 369

left him little time for the preparation of his material for publication.
He has alluded to this in one of his official reports.

In the Appendix to the Annual Report of the President of Cornell
University for 1886-87, Professor Oliver (see Cajori, he. cit.) writes :
" We are not unmindful of the fact that by publishiug more we could
help to strengthen the university, and that we ought to do so, if it were
possible. Indeed, every one of us five is now preparing work for
publication, or expects to be doing so this summer, but such work
progresses very slowly, because the more immediate duties of each
day leave us so little of that freshness without which good theoretical
work cannot be done. . . . The greatest hindrance to the success of
the department, especially in the higher kinds of work, lies, as we
think, in the excessive amount of teaching required of each teacher, —
commonly from seventeen to twenty or more hours per week."

I am indebted to Mrs. Oliver for the following list of Professor
Oliver's published notes and papers connected with mathematics.

Demonstration of the Pythagorean Proposition. Math. Monthly, Vol. I.,

1858, p. 10.
On Mr. ColUns's Property of Circulates. Math. Monthly, Vol. I., 1859,

p. 345.
Introduction to Treatise on Determinants. Math. Monthly, Vol. III.,

1860, p. 86.
Partial Investigation on the best approximate Representation of all the

Mutual Ratios of k Quantities by those of Simple Integers. Proceed-
ings of American Academy of Arts and Sciences, Vol. VI.
Mathematical Note on Linguistic Resemblances. Trans. Amer. Philos.

Soc, Vol. XIII.
On some Focal Properties of Quadrics. Proceedings American Academy,

Vol. VII.
Note on Query concerning Ball held in Jet of Water. Analyst, I. 29,

1874.
On the Law of Distribution for certain Plant-Numbers. A Method of

finding the Law of Linear Elasticity in a ]\Ietal. Abstract Proceedings

Amer. Assoc. Adv. Sc, Vol. XXXL 1882.
A Projective Relation among Infinitesimal Elements. Annals of Math.,

Vol. I., May, 1SS4.
On the General Linear Differential Equation. Annals of Math., Vol. III.,

August, 1887.
Elementary Notes. I. General and Logico-math. Notation. Aunals of

Math., Vol. IV., December, 1888.
Preliminary Paper on Sun's Rotation. Read before the Spring Meeting

of the National Academy, 1888.

The Soaring of Birds. Science, January 4, 1889.

VOL. XXXI. (n. S. XXIII.) 24

370 VISCOUNT FERDINAND DE LESSEPS.

Some Difficulties in Lasage-Thomson Gravitation Theory. Abstract Pro-
ceedings Amer. Assoc. Adv. Sc, Vol. XLI., 1892.

A Mathematical Review of the Free-will Question. Phil. Review, Vol. I.,
March, 1892.

Review of Mathematical Recreations, by W. W. Rouse Ball. Bulletin
N. Y. Math. Soc, November 1, 1892.

Estimates of Distance. Science, March 11, 1892.

Oliver, Wait, and Jones. Text-books on Mathematics for Colleges.
Algebra, especially chapter on Imaginaries, etc. Trigonometry.

Cornell University. Reports on Courses, Aims, and Methods of Mathe-
matical Teaching at Cornell University.

Papers and Discussions at vai'ious Educational Meetings on Teaching,
■with application to the study and teaching of Mathematics.

The above sketch refers to matters which, being related to his
scientific career, present themselves more easily to our notice.

But this was only a part of his life. Professor Oliver was inter-
ested in much outside of his special duties as teacher of mathematics.
His moral qualities were of a superior order. His personal relations
with his friends and colleagues were such as to gain for him their
respect and affection.

But I feel that any attempt on my part to portray the social and
moral side of his life would be inadequate, and must refer for infor-
mation in this regard to the affectionate notices* of him written by
those who had enjoyed the privilege of intimate companionship with
him, and who regarded him as a man of exceptionally exalted
character.

1895. G. Hat

FOREIGN HONORARY MEMBERS.

VISCOUNT FERDINAND DE LESSEPS.

In the biographical notice that follows we do not expect to make an
adequate exhibit of the work and honors of a life so long and impetu-
ous as that of M. de Lesseps, but hope, by pi-esenting the salient
points in his career, to indicate what manner of man he was, from first
to last, without intruding mei'e opinion.

Of his boyhood we know very little, except that he had every
advantage of refined social life and education. As he reached man-
hood he found himself down at the front where volunteers for the

* See Christian Register, May 2, 1895; Cornell Daily Sun, April 3, 1895;
The New Unity, Chicago, August 1, 1895.

VISCOUNT FERDINAND DE LESSEPS. 371

advance were called, and he went forward. Fortune was inconstant,
but he never looked back. For more than a half century he was the
most conspicuous and interesting figure in the rush of the busy world ;
— he turned its tide, but it overtook him and whelmed over him
when his footsteps faltered in old age. He was the hero of one
generation, and the victim of another.

He was born at Versailles in the year 1805, and educated at the
Lycee Napoleon for the foreign service, to which the family had a
sort of traditional claim. His grandfather, Martin de Lesseps, was
Consul at St. Petersburg before the Revolution, and his father,
Matthew de Lesseps, held the Commissariat-Generalship of Egypt at
the time that Ferdinand was born. Subsequently he was Imperial
Commissary at the Seven Islands (Ionian), where he won tlie good
will of everybody, and in the fantastic diction of the period he was
declared to be "liberal even to fanatical generosity." In 1817 he
was sent on a mission to Morocco, and shortly after ap[»ears as French
Consul at Philadelphia, where he assisted at a Commercial Convention,
and was elected a member of the American Philosophical Society.
He married Mademoiselle de Grivegnir, daughter of a distinguished
jurist in Malaga. It was through tliis Spanish mother that Ferdinand
de Lesseps came to be a kinsman of the Empress of the French, in the
unfolding [)lot of this family's missions.

Tlie rirst professional employment for Ferdinand was offered by his
uncle, Jean Baptiste de Lesseps (best known to scientific men as that
Viscount de Lesseps who, in 1787, crossed Siberia from the Okhotsk
Sea to bring a report — which proved to be the last tidings — from
La Perouse). This uncle in 1825 was French Consul at Lisbon, and
Ferdinand, then twenty years old, was sent by him on a diplomatic
errand. Shortly after this we find him a " student consul " at Alexan-
dria, under his father, Matthew de Lesseps.

In 1833 he was given a sub-consulate at Cairo. It was in the
following year that the great plague broke out, memorable as perhaps
the most fatal visitation of modern times. Young Lesseps was then
left in management of the Consul-Generalship, and he won such golden
opinion that he was decorated with the cross of the Legion of Honor.
In 1838 he went as Consul to Rotterdam, in 1839 to Malaga. It was
as French Consul at Barcelona that he won, during the revolt of 1842,
the admiration of Europe as a humanitarian. His personal courage
and his devotion to suffering people induced four governments to send
him decorations. He was made an officer of the Legion of Honor.
The city of Marseilles awarded him a medal, and the city of Barcelona

372 VISCOUNT FERDINAND DE LESSEPS.

caused his bust to be set up. A few years later he received the
insignia of Chevalier of the Grand Cross of Isabella the Catholic.

lu 1848 he was summoned to Paris by Lamartiue, and ?ent as
Minister Plenipotentiary to Madrid. After about one year he was
withdrawn by Prince Napoleon, and sent in the same capacity to
Rome, at the time of Garibaldi's occupation, — ''a man," said M.
Odilou Barrot, " who enjoys our full confidence, whom we have put to
the test in very trying circumstances, and who has always served the
cause of liberty and humanity."

M. de Lesseps has told the vexations of his Roman mission very
interestingly in his " Recollections," but with singular absence of
personal consciousness even in failure. He often defends himself, and
designs to do so ; but it is always an argument for the merit of the
work upon which he is engaged, rather tlian any declaration of his
own higher motives, except that he attributed all his successes to the
energy of " patriotism," — a word which seemed to embrace about all
the world. He seems to have used tliis word simply as the most
modest admissible expression for public duty, and the obligation to
serve in the great march.

He learned in the foreign service to respect and sympathize with
earnest people everywhere, and formed enthusiastic friendships among
even those of radically different blood and radically different traditions.
To bring people together in some common interest and intent seemed
to him to be the cure for national prejudice, and he saw in foreign
trade this catholicon. In the administration of foreign affairs during
periods mostly peaceful, he acquired a pretty clear impression of those
larger principles of reciprocal trade that escape the merchant's more
short-sighted view. The consulates were for him not only schools of
commercial jurisprudence, but they enabled him to distinguish the
interest of the community from that of tlie individual. An obstruction
or difficulty in the path of trade may be of value to the few wlio know
or can afford the roundabout way ; but it is the mission of public
spirit to equalize opportunities as well as to shorten process. M. de
Lesseps, in his letter to Cobden (1854), advised those statesmen who
opposed the Suez Canal because it would reduce the number of ships
and men by shortening the route to India, to induce shipmasters to
take the Cape Horn route, and thus employ more men and more ships.
This retort is a pretty good illustration of his way of meeting disin-
genuous criticism by the reductio ad absurdum.

In this earlier part of his life, following in the footprints of his
fathers, he often missed complete success, — although always in

VISCOUNT FERDINAND DE LESSEPS. 3^3

earnest. But all the while he was unwittingly fitting himself for his
true mission, and gaining strength, and knowledge, and courage for a
practical interpretation of a sublime thought : " Je veux detruire cette
muraille de sable qui arrete le progres."

Tills muraille de sable, this wall of sand, separates two worlds.
In one, man plays the roll assigned to him in the original setting of
the piece ; in the other, he assists in never-ending creation. In one
he awaits his orders in harness, in the other he is charioteer. But
from whichever point of view we look backwards across the Isthmus,
from Islam or from Christendom, history foreshadows the canal.
And now that it is finished and in full operation, so that its direct
effects upon the commerce of the world have ultimated, one reads the
history of trade by a new light, which discovers prophetic meaning in
many events that anticipated this new dispensation.

If we seek the origin of the idea, we are taken back to a remote
past; and we may follow the thought down through a hundred genera-
tions, which it dimly pervades, till the sifting of the French Revolu-
tion discovers its " fixity, that true sign of the law " ; thenceforth, it is
a pressing obligation hastening to maturity.

In Napoleon's message to the Directory, " Whatever European
power holds Egypt permanently is in the end mistress of India,"
he put the cart before the horse, as we have since learned ; but
the necessity for the canal in the scheme of human progress is re-
flected even in such inverted conceptions, and to the short-lived
Egyptian Institute we owe the first scientific investigation of the
problem of joining the two seas, although the errors of svirvey, which
deferred the project by placing the two seas out of level, reflect little
credit upon the pupils of the Normal School. To make errors that
should prove stepping stones to the truth was, however, in the spirit
of the age ; and the report of Lapere — reasserting "• the wisdom of
the ancieiifs" that had already afflicted and separated the nations
of the earth for thirty-five hundred years — only aroused a new and
defiant generation, that with better observations and in better temper
reconciled the two seas forever.

M. Mimut, "one of the most distinguished diplomats ever in the
service of France," holding the great work of the French expedition
in his hands, gave to M. de Lesseps the first quickening thought, and
from that moment his mind and heart received the ancient hope re-
newed, and he became its champion. It was a religious experience.
"You have," said Renan, addressing him at the Academy, " caused
to blossom once more a flower which seemed faded forever. You

374 VISCOUNT FERDINAND DE LESSEPS.

have given to this sceptical age of ours a striking proof of the efficacy
of faith."

We are satisfied that here is the point of view from which the con-
fidence and the enthusiasm of M. de Lesseps is comprehensible. How
could he possibly have expected to convert such a man as Lord
Palmerston ? The result of their interview was that M. de Lesseps
doubted the sanity of the Premier, while the latter regarded him as an
adventurer, — a soldier of fortune, — employed, perhaps, in the interest
of some French " move '"' in Egypt. In reality, there stood before the
great statesman a simple-minded Da Gama, who had discovered a new
route to India and offered himself as the pilot. It was nobody's in-
terest then to make him a figure-head, — he was at the other end of the
ship ; it was his trick at the wheel.

After the failure of the Egyptian expedition, Napoleon, in 1803,
instructed Matthew de Lesseps, Political Agent in Egypt, to nominate
for election and for the Sultan's approval an officer of ability to serve
as Pasha of Cairo. M. de Lesseps named one who was then in com-
mand of a regiment of Basha Bazouks, a Macedonian, who could
neither read nor write, and who had come to Egypt as a subordinate of
contingents. This man was Mehemet Ali, — the wise and terrible, —
who subsequently made himself master and mortgagee of Egypt. It
was he who built the Mahmoudieh Canal, — the last, and perhaps the
greatest, of non-militant works ever executed by unaided human
hands. He also inaugurated work on the Barage — the dream of
Hassan — over which Egypt liad brooded five hundred years ; and it
was he that discovered the potential energies of young Lesseps, whom
he caused to sit at his feet and listen to the narrative of his slaughter
of the Mamelukes, and the now possible project of a cut through to
Suez, which the Viceroy was ready to undertake, under a grand corvee,
except that he feared the would-be and could-be mistresses of India.

It was at this time (while consular pupil in 1832) that Ferdinand
de Lesseps became the companion, and incidentally the teacher, of
Said Pasha, the son of the Viceroy, on whom was laid the futurition
of the father's dream, — and the dream of all the Pharaohs. The
" memorandum " prepared for this prince by M. de Lesseps, long
after, connects itself in our minds with these boyhood days, when he
says, " The names of the Egyptian sovereigns who erected the Pyra-
mids remain unknown, — the name of the prince who opens the great
maritime canal will be blessed from century to century down to the
most distant posterity," — so well did this "grand .Frenchman"
understand and share the hope of glory upon which he counted.

VISCOUNT FERDINAND DE LESSEPS. 375

Ever since the establishment of British power in India, the best
minds in England had seen the necessity of securing a right of way
across the Delta of the Nile, and a step in this direction seemed to be
made in the treaty of Warren Hastings in 177G ; but nothing came of it
till the vigorous movement of Mehemet Ali to open a transit route via
the Mahmoudieh Canal brought Waghorn to the front in 1829, — the
Indian mail service really opened two years later, — although it was
not till 1840 that a steamship company used the Red Sea route.
Young M. de Lesseps was witness of Waghorn's triumph, and, more
than all the world besides, did him the honor of a just appreciation.
He was an example and an inspiration all his life, and when his own
hour of triumph arrived he raised a statue to Waghorn at Suez.

Professional detractors discover, now, that the Suez Canal never
really presented any physical difficulties or dangers ; but how many of
this class resisted the heresies of the French expedition, or forsook
them at the bidding of a Bourdelau? Many of us are old enough to
remember when in popular belief the two seas were out of level, the
sand storms of the desert buried caravans and armies, and thousands of
dead fellaheen were used each month to raise the banks of the Mah-
moudieh Canal. Children in our Sabbath schools were taught these
things of the land of Mehemet Ali, — that bold bad Napoleon of
Islam. And yet the Revue des Deux Mondes, away back in 1835,
ventured to say, " Mehemet Ali is working for Europe, which will
become his heir."

None of these bugbears troubled the mind of M. de Lesseps, who
was the first to declare that the Isthmus was only a muraiUe de sahle
qui arrete le proyres. This was not the "bluff" of the speculator,
but the faith that moves mountains. Yet, as late as July, 1852, in a
letter to the Consul General of Holland in Egypt, he says, " I confess
that my scheme is still in the clouds, and I do not conceal from myself
that, as long as I am the only person who believes it to be possible, that
is tantamount to saying it is impossible." At this time he had forsaken
Egypt and was setting up a model farm in the Berry district and
restoring the castle of Agnes Sorel. Abbas Pasha was then Viceroy
of Egypt, and this prince knew not Joseph. It was under this
Viceroy that the English Transit Railway was built.

In 1854, when M. de Lesseps was nearly fifty years old. Abbas
Pasha died, and the " sympathetic Mohammed Said " became Viceroy
of Egypt, and immediately sent for his old master to return ; and before
the close of the year, the concession of powers for the formation of the
canal company had been issued.

376 VISCOUNT FERDINAND DE LESSEPS.

A few little touclies of nature in the journal of Lesseps, at this
critical turning point of his career, reveal the entire consecration of
every faculty to the great mission of his life. As he left his tent in
the early morning of that anxious day, when the Viceroy was to hear
his story and decide upon building the canal, he beheld a rainbow in
the sky : " I confess that my heart beat violently, and that I was
obliged to put a rein upon my imagination, which was tempted to see,
in this sign of alliance spoken of in the Scriptures, the presage of the
union between the Western and the Eastern world, and the dawning of
the day for the success of my project." And again, on the same day,
" When I leave the Viceroy to go and get my breakfast, I jump my
horse over the parapet. You will see that this foolhardy act was one
of the reasons that induced the Viceroy's entourage to support my
scheme, — the generals at breakfast telling me as much." On the
same day there was target practice, and a wliole regiment had failed
to hit the mark. M. de Lesseps seized a musket and put a bullet
through the bull's-eye. Again, as he stood there, a bird hovered in
the sky, — he raised his piece and fired, — an eagle fell at his feet.
Even misfortune betokened success, as when on the Nile his cabin
took fire and he was severely burned, he said, "The accident was of
good omen, in that we had acquitted our debt to ill luck." Later, he
found himself accidentally a lodger in the building which had once held
the Institute of Egypt, and " this too was a good sign." All signs
point our way when we are on the right road.

The gods were propitious, the prince was gracious. M. de Lesseps
did not pause to consider whether these gods were of his own creation,
or this prince only his docile pupil. Besides, he was not alone to
carry the burden of this project, even in its initial stage. He had
taken two friends into his counsels long ago. These were Linant
Bey and Mougel Bey, engineers of the Barage and countrymen of
his. These three put their earnings together, and with a capital
of six thousand dollars they formed the nucleus of the " Compagnie
Universelle."

Linant Bey had long before tried to raise a company to build the
canal, and he and Bourdelau had run double lines of levels reconciling
the two seas. He had also traced a line of location, which was very
nearly that adopted in exploitation. Mougel, well known since as the
chief engineer of the Barage, was recognized as the essential third point
of support. In his journal of this initiatory period, M. de Lesseps
notes down his dependence upon the knowledge, skill, and devotion of
these two friends ; and in his days of greatest glory he published

VISCOUNT FERDINAND DE LESSEPS. 377

these notes, that all the world might be just to those who ventured
with him in the forlorn hope.

These three kindred spirits, with only the Bible for their guide-book,
made a reconnaissance of this route. The journal of Lesseps, which
contains over twenty quotations from tlie Old Testament, is full of the
confidence and courage of hope and health. In that desert land the
skies are clear and the north wind full of joyous life and stored up
energies. The whole horizon is in view, and he who has singleness of
purpose may march straight to his destiny, whether he holds the Koran
or the Bible. And here we are minded of that fearful contrast offered
by the Isthmus of Panama on the otiier side of the earth, — with its
weary mountains and dark forests breaking down the trade winds, and
shortening and degrading the vision till high hopes and purposes are
starved out, and man becomes a timid, sickly animal.

We agree that, as it turned out, the Suez Canal did not involve the
solution of any new problems of physics, or very greatly tax the skill
of French engineers. Indeed, it was not the practical and direct diffi-
culties that really made the nations timid. In "this wall of sand,"
separating the Moslem from the Christian world, there lay sealed up
with the seal of Solomon afreet and genii that made nations tremble
to think of; for these, rashly let loose, might disturb the balance
of power and throw out of adjustment trade and industries all over the
world. It proposed radical change, and who could tell what might
happen ? Engineers and laborers stood ever ready to do the work, but
the world waited for the prophet who could forecast a healthy and
happy result. M. de Lesseps filled this office. His training, his
knowledge, and his enthusiasm commanded respect. His promises of
advantage east and west once seemed florid beyond the measure of his
careful computations and great array of statistics, but, as we read his
articles now, we are struck with his acuteness of foresight and his
moderation.

He seems to have been a man of unusual singleness of purpose,
— in something wider than a moral sense, — and to have absorbed
himself absolutely in the work before him without ulterior design.
He was in politics a republican, but he " never even from curiosity
attended a political meeting." He was a partisan of the Prince
President, but he could not follow him in the coup d'etat, and only
submitted to the Empire in the interest of peace, — and the canal.
The Empress was a kinsman of his, and he had rendered her, in her
humbler life, a personal service which she requited in personal good
will. These relations may have procured for him the entree to the

378 VISCOUNT FERDINAND DE LESSEPS.

family circle, and brought the Emperor within the sphere of good in-
fluence. French journals were quick to discover an epigram, — the
expression of good omen : " The marriage of two families, and the
marriage of two seas." And this ran through the newspapers all
around the world: ''There is an Arab proverb (much affected by
M. de Lessens) quite apposite here: 'The dogs bark, — the caravan
passes.' "

M. de Lesseps was not, by early training, an engineer, but a diplo-
mat. To us his real assimilated rank is that of a discoverer. Why
not, as well as Da Gama and Magellan ? These declared that a ship
could reach the Pacific Ocean by sailing around the continents, and
they proved it. De Lesseps declared that a ship could reach the
Pacific Ocean by sailing through the continents, — and he proved it.
" He discovers who proves," said Aristotle. None of these men
originated a new thought, but each of them did a new thing for the
relief of mankind. Vasco de Gama wa^ a prize drawn in a lottery in
answer to prayer. He received his instructions through a great
prince ; and the miracle of his selection and vicarious appointment •
raised him above all fear. M. de Lesseps presents the antithesis. He
was a volunteer who taught princes a good doctrine, and held them
down to it. "Hear me for my cause" was all he asked, and those
who paused to listen fell under the spell of his enthusiasm and received
his testimony.

That was a period of great strain, when Robert Stephenson, one of
his own professional caste, turned upon him in Parliament and spoke
of his project as " one of those chimeras so often formed to induce
English capitalists to part with their money, the end being that these
schemes leave them poorer, though they make others much richer."
The good temper and even the good sense of M. de Lesseps gave way,
and he crossed the Channel to demand explanation. The explanation
was made, and we are constrained to say that, if M. de Lesseps had
not long afterwards, in his old age, published the correspondence, we
should have overrated these contending champions of land and sea.

We can appreciate the causes for anxiety that afflicted intelligent
men of affairs in Great Britain. Any change in the course of trade
involves national risk. In this way the Venetian Indian trade had
dwindled away after the Cape route was opened ; then arose Portugal
and Holland, to be outdone by England only after a tremendous
struggle. Who could say which way the wheel might turn if the
Egyptian Transit Railroad came to be supplanted ?

England hardly attempted to disguise her apprehensions of danger

YISCOUNT FERDINAND DE LESSEES. 379

to her trade from the success of the canal built by a rival nation ; but
when she prevailed upon the Sultan to order the withdrawal of
the fellah labor, our American journals charged her with complacent
hypocrisy, and very naturally, since she had just completed the
Transit Railway under the corvee in its most cruel form. Under this
corvee, Egypt had always draughted men for public works, much in
the same way that other countries procure soldiers in time of war.
In a country whose existence is involved in a system of canals of irri-
gation and dikes, there must be no hesitation at critical moments.

M. de Lesseps had prevailed upon Mohammed Said greatly to mod-
ify the corvee so far as to provide wages and hospitals, but the Viceroy
insisted that the abolition of the system would ruin Egypt. Inducing
the peasant to work for government by offers of reward was an
untried experiment, and one too dangerous to try, since its failure
involved, as alternative, the calling in of foreign labor, — an abomina-
tion to the Egyptian. England's suggestion that, to avoid the foreign
force, the dimensions of tlie canal should be reduced to the capacity of
the native labor, was perfectly logical, and consistent with her alarm
policy. But England's policy, like our own, is the net result of con-
flict among varied interests, theories, and sentiments. The Transit
Railway had been practically a national interest, and in the composition
of forces this interest had been strong enough to determine the direc-
tion of the resultant till the work was finished. But reports from the
scene gradually aroused the humanitarians, who threw their weight
into the scale where the interests of the fellahs seemed to lie. In this
way, Lord Palmerston, wholly misunderstanding the signs of the
times (as events proved), worked more wisely than he knew, and the
spell that had darkened Egypt from her birth broke forever.

M. de Lesseps — always single in his purpose — made the with-
drawal of the corvee another stepping stone to success. He procured
the reference of the question of damages to the Emperor of the
French, who made a generous decision in his favor, — far too gener-
ous, perhaps, but it enabled the company to introduce machinery in
place of hand labor, till the Suez plant excelled that in use in any
other part of the world.

The writer of this notice, who made an inspection under full
authority from M. de Lesseps the year before the canal was opened,
can bear witness that never before or since, in his long experience, has
he seen laborers so kindly cared for and so free under the most abso-
lute discipline. M. de Lesseps said, " I have no difficulty in controlling
my laborers, because I treat thera kindly and make them comprehend

380 VISCOUNT FERDINAND DE LESSEPS.

that they are working for all mankiiKl." His labor then was drawn
from many races, but did not include the " sambos," " bravos," coolies,
and mongrels of Panama.

His sympatliies were always with the fellaheen, even to his own
prejudice. He had lived among them, they had served him, and
when the struggle came for the possession of the canal he did not
ask that it should be French, but that it should be neutral under a
pledge from Araby Pasha, the chief of revolting fellaheen and Arab
troops. One can hardly conceive of a professional diplomat so blind
to his own interests and the interests of the Compagnie Uiiiverselle,
but, like his grandfather of the Seven Islands, "he was generous even
to fanaticism."

It must not be overlooked that the Suez Canal is an extension of
the Mediterranean and an improvement upon the old route of trade
between Europe and Asia, along which ports have been made, ware-
houses established, and political relations adjusted. Its opening was
an easement to all the world. Its construction was sure, because it
was the next legitimate step forward under the pressure of an enor-
mous demand. It was, as it proved, no experiment. Tlie ground
had been profiled and bored, the climate had been tested and found
healthy and clieerfu! ; and laborers were near at hand, not likely to
suffer in temper and spirits from the slight change of scene. In short,
there were no dilficultie.s, except familiar vicissitudes — and the pride of
kings.

The successful opening of the Suez Canal, near the close of 18G9,
induced throughout the world, perhaps for the first time, a conviction of
common interest. The struggle in Egypt had been with tiie common
enemy, and the victory belongs to us all. It was the old allegory
adapted to our age, with Count de Lesseps in the part of St. Michael.
Among the Mohammedans, as among Christians, there are sects that
believe in an internal sense of the word. To these, human life and
history, and all events, both great and small, are allegories, and he
who catches so much as a glimpse of the esoteric meaning of the piece
plays his part like a god. This is the source of enthusiasm.

Some newspaper men are comparing the traffic of the Sault de Ste.
Marie with that of the Suez Canal. — as if these two works were of
the same world-wide interest, or of the same dramatic import. There
may be more tonnage passing through the lock at the " Soo " than
through the desert of Suez, and there may be more yet passing through
a city street, but the gonfalon was borne by the Compagnie Universelle.
I During the construction of the canal, we never heard of a single

VISCOUNT FEEDINAND DE LESSEES. 381

American purchaser of a share. Yet these shares are worth now
many times their original face value, and we, indirectly or through
foreign sliips, are among the best customers. The Report for last
year shows that of the three thousand three hundred and fifty-two
ships that passed through the canal (averaging considerably over two
thousand tons each) England sent two thousand three hundred and
eighty-six and the United States only five.

Since the days of the Conquistadors, the project for a canal through
the American isthmus had been an alternative to the shorter route to
the Indies by way of Suez, to be considered in case physical or politi-
cal difiiculties should intervene. The completion of the Suez Canal
and the guaranty that British control gave it, reduced very much the
commercial demand for the westward route. Nevertheless, the immense
revenue at Suez excited the popular mind, especially in France, with
the hope of a great speculation in the establishment of a rival company,
which, while sharing in some measure the overflowing trade of India
and China, might secure the interchanges between the two coasts of
America, and perhaps the whole of our trade with Japan.

The Panama project was high born and burst into life a full grown
scheme. There were no prime ministers, emperors, and sublime portes
lying in wait to stifle it. With Count de Lesseps for its godfather
holding the lamp of Aladdin, all the world attended its baptism with
complacent expression, — except, perhaps, that the Monroe Doctrine
cast a sinister shadow over the scene, — a very thin shadow, but
enough to depress the market for the securities after the first rush was
over. But what the scheme lacked from first to last was justification
in immediate necessity. It had an illegitimate and premature birth,
and its sponsors limited their risks to broker's charges, — exce^at M.
de Lesseps, who gave all. He gave his past earnings in the best ser-
vice of our age, and he gave his fair fame as endowment enough for
the whole credit of the company — at the start.

Professor Nourse and other clever writers have said that the
American canal would be of greater benefit than the Egyptian, be-
cause it would connect greater oceans, and that commerce demands a
navigable zone around the world. Half in sympathy with these ideas,
we cannot help thinking that in this aphoristic form they lack practical
merit. Man does not inhabit the sea, and the road that traverses
or connects intimately the most inhabited portions of the earth must be
the most valuable. Our system of overland railways to the Pacific is
practically a supplement to the Suez Canal in the all around commerce
of the world, and it was the building of this system with its counec-

382 VISCOUNT FERDINAND DE LESSEPS.

tious that for many years diverted our capital from the oceau, to which
it reluctantly returns.

" Fear only the unforeseen " is a classic proverb, much affected by
the French, but among no other people has it less practical honor.
For the Panama project, the almost unprecedented depth of the cut,
the peculiarly obdurate ledges, the great rainfall interrupting labor and
causing sloughing of the banks, the necessity for turning tlie Chagres
River, were difficulties weighed and discounted at the start, by able
engineers and by a very large and very intelligent company in France.
But the ''bodily slipping of the hills," in the excavations near the
summit level, was a frightful disappointment. It necessitated a post-
ponement of the sea level cut, and the adoption of a scheme of locks,
which involved the ponding of the Chagres River. This change of base
"was fixtal, and the company broke. A new company has been formed
and may keep the project alive till a stronger call from the commercial
world comes to its aid.

The King of Spain, looking from his chamber window, shaded his
eyes and said, " I am looking for tlie walls of Panama, — they have
cost enough to be seen from here."

With regard to the charges of fraud against the financial manage-
ment of the Panama Canal Company, we do not feel competent to
speak, except to call attention to their diminution as investigations
proceed. But with regard to the waste of plant, it does not seem in
undue proportion to the magnitude of the experiment. Charges of
waste were made against the Suez Canal Company, — especially by
our journals and those of other distant people, — and they attend all
great enterprises. The plant that does not prove equal to the new
work must be cast aside, and lost, if far away from the junk market.
The greatest and most successful works that we have visited are
strewn with wreckage, marking the field where the battle was won.

Throughout the whole period of the construction in Egypt, M. de
Lesseps was actual manager in chief. Many thousands worked under
his direction, holding all gradations of rank, but he was the real
master spirit. Flis reports from the beginning are full of acknowl-
edgments of the services and merits of his subordinates. At the
outset he leaned upon the superior engineering of Linant Bey and
Mougel Bey, and for many years preceding the completion of the
canal his reports place Voisin Bey in the foreground, not as a shield
from responsibility, but as the support of an enthusiastic company
whose millions were thus insured. In his old age, the dependence
upon others necessarily became greater and greater, and in the

VISCOUNT FERDINAND DE LESSEES. 383

Panama scheme he no longer verified the statements of the engineers
by adequate inspection and exploitation, but absolutely fell back upon
others. His old attitude of control and command continued, but he
merely indorsed the reports of the chiefs of divisions, — whose figures,
we now know, were correct, — v^ithout discovering their misleading
limitations. His famous promise that the canal should be open to the
passage of vessels in 1889 was based, as he stated, upon the unani-
mous acquiescence of his chiefs of division. Moreover, if we plot on
profile paper the amount of work done from date to date, the curve of
increments, projected, seems to justify the prediction. His own belief
in it is enthusiastically stated in his '' Recollections of Forty Years, "
and he adds : " I am an octogenarian. Old age foresees, and youth
acts." This last mot was lost upon the public, who saw him now only
as the figure-head of a brave ship given over to pirates.

In 1869, the year of the triumphal opening of the Suez Canal,
M. de Lesseps was sixty-four years old. This is the age of compul-
sory retirement for our United States Engiueei's. But M. de Lesseps
was so vigorous that he was for many years later a most valuable man
in council. He aided the ship canal of Corinth, and the grand canal
from tlie Elbe to the Baltic. It was not until he was seventy-six that
the Societe du Canal de Panama was constituted, and he was eighty
when he crossed the ocean to make a personal inspection of the work
in progress. The execution for a year or two, under the full con-
tracts, seemed to realize his predictions, and warrant great expecta-
tions ; but after that everything went wrong. The master's mind
failed before he could have discovered how much he had been betrayed.
He lived to be eighty-nine, and died a poor man. His widow and her
children are now dependent upon a pension from the Suez Canal
Company.

1896. Henry Mitchell.

Note. — The foregoing had been written out for transmission to
the Academy, but was withheld under misgivings as to the adequacy
of our very much foreshortened view of the causes of the failure
at Panama, until Mr. Nathan Appleton (American Agent of the
Panama Canal Company) sent us the last word spoken on the sub-
ject by a competent witness. This was in the form of a biographical
notice by M. Gabriel Gravier, which the Countess de Lesseps sent
to Mr. Appleton with her autograph indorsement. This writer takes
the ground that all was going well, till a senseless panic upset the
market. " The work was marching to a certain success ; the original

384 VISCOUNT FERDINAND DE LESSEPS.

estimate had proved very nearly sufficient. Whence came the cyclone
which swept away the company and its four hundred thousand share-
holders? From Paris, the distracted brain of France!" With some
personal and professional knowledge of the American isthmus, and
from reading the recent reports of Kimball, Rogers, and others, we
distinctly see that the difficulties in the Panama scheme were really
intrinsic, — although not insurmountable, — and we have let our account
stand.

The Academy has received an accession of ten Resident
Fellows, six Associate Fellows, and ten Foreign Honorary
Members.

The Roll of the Academy, corrected to date, includes the
names of 196 Fellows, 96 Associate Fellows, and 73 Foreign
Honorary Members.

May 13, 1896.

LIST

FELLOWS AND FOKEIGN HONOEARY MEMBERS.

(Corrected to August 1, 1896.)

RESIDENT FELLOWS. — 195.

(Number limited to two bundred.)

Class I. — Mathematical and Physical Sciences. — 73.

Section I. — 18.
Mathematics and Astronomy.
Solon I. Bailey, Arequipa, Peru.
Seth C. Chandler, Cambridge.

Alvan G. Clark, Cambridgeport.
J. Rayner Edmands, Cambridge.

Benjamin A. Gould,

Cambridge.

Francis M. Green,

Boston.

Gustavus Hay,

Boston.

Henry Mitchell,

Boston.

Edward C. Pickering

Cambridge.

John Ritchie, Jr.,

Boston.

John D. Runkle,

Brookline.

T. H. Safford,

Williamstown

Edwin F. Sawyer,

Brigliton.

Arthur Searle,

Cambridge.

William E. Story,

Worcester.

Henry Taber,

Worcester.

O. C. Wendell,

Cambi'idge.

P. S. Yendell,

Dorchester.

Section H. — 21.

Physics.
A. Graham Bell, Washington.

Clarence J. Blake, Boston.
Francis Blake, Weston.

VOL. XXXI. (N. S. XXIII.) 25

John H. Blake,
Chai'les R. Cross,
Amos E. Dolbear,
Edwin H. Hall,
Hammond V. Hayes,
Silas W. Holman,
William L. Hooper,
William W. Jacques,
Alonzo S. Kimball,
T. C. Mendenhall,
Benjamin O. Peirce,
A. Lawrence Rotch,
Wallace C. Sabine,
John S. Stone,
Elihu Thomson,
John Trowbi'idge,
A. G. Webster,
Robert W. Willson,

Boston.

Boston.

Somerville.

Cambridge.

Cambridge.

Boston.

Somerville.

Newton.

Worcester.

Worcester.

Cambridge.

Boston.

Cambridge.

Boston.

Lynn.

Cambridge.

Worcester.

Cambridge.

Section III. — 21.

Chemistry.
Samuel Cabot, Boston.

Arthur M. Comey, Cambridge.
Thos.M. Drown, So. Bethlehem, Pa.
Charles W. Eliot, Cambridge.
Thomas GafSeld, Boston.

Henry B. Hill, Cambridge.

386

RESIDENT FELLOWS.

Henry M. Howe, Boston.

Charles L. Jackson, Cambridge.
Leonard P. Kiunicutt, Worcester.
Charles F. Mabery, Cleveland, O.
Arthur Michael, Boston.

George D. Moore, \Vorcester.
Charles E. Muuroe, Washington.
John U. Nef, Chicago.

Robert H. Richards, Boston.
Theodore W. Richards, Cambridge.
Charles R. Sanger, St. Louis.
Stephen P. Sharpies, Cambridge.
Francis H. Storer, Boston.
Charles H. Wing, Ledger, N. C.
Edward S. Wood, Boston.

Sectiox IV. — 13.
Teclmology and Engineering.
EUot C. Clarke, Boston.

Gaetano Lanza, Boston.

E. D. Leavitt, Cambridgeport.

William R. Livermore, Boston.
Hiram F. Mills, Lowell.

Cecil H. Peabody, Boston.
Alfred P. Rockwell, jNIanchester.
Andrew H. Russell, Rock Island, 111.
Peter Schwamb, Arlington.

Charles S. Storrow, Boston.
George F. Swain, Boston.

William Watson, Boston.

Morrill Wyman, Cambridge.

Class II. — Natural and Physiological Sciences. — 63.

Section I. — 13.

Geology, Mineralogy, and Physics of
the Globe.

Milton.

Boston.

Boston.

Cambridge.

Amherst.

Cambridge.

Boston.

Cambridge.

Cambridge.

Boston .

Cambridge.

Cleveland, O.

Cambridge.

H. H. Clayton,
Algernon Coolidge,
William O. Crosby,
William M. Davis,
Beuj^ K. Emerson,
O. W. Huntington,
Robert T. Jackson,
Jules Marcou,
WilHam H. Niles,
John E. Pillsbury,
Nathaniel S. Shaler,
Warren Upham,
John E. Wolff,

Sectiox II.
Botany.
William G. Farlow,
Charles E. Faxon,
George L. Goodale,
H. H. Hunnewell,
B. L. Robinson,
Charles S. Sargent,
Arthur B. Seymour,
Charles J. Sprague,
Roland Thaxter,

— 9.

Cambridge.

Boston.

Cambridge.

Wellesley.

Cambridge.

Brookline.

Cambridge.

Boston.

Cambridge.

Section IH. — 25.

Zoology and Physiology.

Alexander Agassiz, Cambridge.
Robert Amory, Boston.

James M. Barnard, Milton.
Henry P. Bowditch, Boston.
William Brewster, Cambridge.
Louis Cabot, Brookline.

Samuel F. Clarke, Williamstown.
W. T. Councilman, Boston.
Charles B. Davenport, Cambridge.
Harold C. Ernst, Boston.

J. Walter Fewkes, Boston.
Edward G. Gardiner, Boston.
Samuel Henshaw, Cambridge.
Alpheus Hyatt, Cambridge.

John S. Kingsley, Somerville.

Theodore Lyman, Brookline.
Edward L. Mark, Cambridge.
Charles S. Minot, Boston.
Edward S. jNIorse, Salem.
George H. Parker, Cambridge.
James J. Putnam, Boston.
Samuel H. Scudder, Cambridge.
William T. Sedgwick, Boston.
James C. White, Boston.

William ]\l. Woodworth, Cambridge.

RESIDENT FELLOWS.

387

Section IV. — 16.

Medicine and Surgery.

Samuel L. Abbot, Boston.
Edward H. Bradford, Boston.
Arthur T. Cabot, Boston.

David W. Cheever, Boston.
Benjamin E. Cotting, Roxbury.
Frank W. Draper, Boston.

Thomas Dwight,
Reginald H. Fitz,
Charles F. Folsom,
Frederick I. Knight,
Francis Minot,
Samuel J. Mixter,
W. L. Richardson,
Theobald Smith,
Henry P. "VValcott,
John C. Warren,

Boston.

Boston.

Boston.

Boston.

Boston.

Boston.

Boston.

Boston.

Cambridge.

Boston.

Class III. — Moral and Political Sciences. — 59.

Section I. — 10.
Philosophy and Jurisjyrudence.

James B. Ames,
Charles C. Everett,
Horace Gray,
John C. Gray,
G. Stanley Hall,
Nathaniel Holmes,
John E. Hudson,
John Lowell,
Josiah Royce,
James B. Thayer,

Cambi'idge.

Cambridge.

Boston.

Boston.

Worcester.

Cambridge.

Boston.

Newton.

Cambridge.

Cambridge.

Section H. — 21.

Philology and Archceology.

William S. Appleton, Boston.
Charles P. Bowditch, Boston.
Lucien Carr, Cambridge.

Franklin Carter, Williamstown.
Joseph T. Clarke, Boston.
Henry G. Denny, Boston.

Epes S. Dixwell, Cambridge.

William Everett, Quincy.

William W. Goodwin, Cambridge.
Henry W. Haynes, Boston.
David G. Lyon, Cambridge.

Bennett H. Nash, Boston.

Frederick W. Putnam,
Edward Robinson,
F. B. Stephenson,
Joseph H. Thayer,
Crawford H. Toy,
John W. White,
Justin AVinsor,
Jolin H. AVright,
Edward J. Young;,

Cambridge.

Boston.

Boston.

Cambridge.

Cambridge.

Cambridge.

Cambridge.

Cambridge.

Waltham.

Section HI. — 16.

Political Economy and History.

Charles F. Adams,
Edward Atkinson,
Mellen Chamberlain,
John Cummings,
Andrew M. Davis,
Charles F. Dunbar,
Samuel Eliot,
John Fiske,
A. C. Goodell, Jr.,
Henry C. Lodge,
Augustus Lowell,
John C. Ropes,
Denman W. Ross,
Charles C. Smith,
F. W. Taussig,
Francis A. Walker,

Lincoln.

Boston.

Chelsea.

Woburn.

Cambridge.

Cambridge.

Boston.

Cambridge.

Salem.

Nahant.

Boston.

Boston.

Cambridge.

Boston.

Cambridge.

Boston.

388

RESIDENT FELLOWS.

Section IV. — 12.
Literature and the Fine Arts.

Francis Bartlett, Boston.

John Bartlett, Cambridge.

George S. Boutwell, Groton.

J. Elliot Cabot, Brookline.

Francis J. Child,
T. W. Higginson,
S. R. Koehler,
Charles G. Loring,
Pei"cival Lowell,
Charles Eliot Norton,
Horace E. Scudder,
Barrett Wendell,

Cambridge.

Cambridge.

Boston.

Boston.

Brookline.

Cambridge.

Cambridge.

Boston.

ASSOCIATE FELLOWS.

389

ASSOCIATE FELLOWS. — 96.

(Number limited to one hundred. Elected as vacancies occur.)

Class L — Mathematical and Physical Sciences. — 38.

Section I. — 16.

Mathematics and Astronomy.

Edward E. Barnard,
S. W. Burnham,
George Davidson,
Fabian Franklin,
Asaph Hall,
George W. Hill,
E. S. Holden,
James E. Keeler,
Emory McClintock,
Simon Newcomb,
H. A. Newton,
Charles L. Poor,
William A. Rogers,
George M. Searle,
J. N. Stockwell,
Chas. A. Young,

Chicago.
Chicago.
San Francisco.
Baltimore.
Washington.
Washington.
San Jose, Cal.
Allegany, Pa.
New York.
Washington.
New Haven.
Baltimore.
Waterville, Me.
Washington.
Cleveland, O.
Princeton, N.J.

Section II. — 7.

Physics.

Carl Barus, Washington.

J. Willard Gibbs, New Haven.
S. P. Langley, Washington.

A. M. Mayer, Hoboken, N. J.

A. A. Michelson, Chicago.

Ogden N. Rood, New York.

H. A. Rowland, Baltimore.

Section III. — 8.

Chemistry.
Wolcott Gibbs, Newport.
Frank A. Gooch, New Haven.
S. W. Johnson, New Haven.
M. Carey Lea,
J. W. Mallet,
E. W. Morley,
J. M. Ordway,
Ira Remsen,

Philadelphia.
Charlottesville, Va.
Cleveland, O.
New Orleans.
Baltimore.

Section IV. — 7.
Technology and Engineering.
Henry L. Abbot, New York.
Cyrus B. Comstock, Washington.
W. P. Craighill, Washington.
F. R. Hutton, New York.

George S. Morison, Chicago.
William Sellers, Philadelphia.
Robt. S. Woodward, New York.

Class IL — Natural and Physiological Sciences. — 31.

Section I. — 15.

Geology., Mineralogy., and Physics of
the Globe.

Cleveland Abbe,
George J. Brush,
Edward S. Dana,
Walter G. Davis,
Sir J. W. Dawson,
G. K. Gilbert,

Washington.
New Haven.
New Haven.
Cordova, Arg.
]\lontreal.
Washington.

James Hall,
Clarence King,
Joseph LeConte,
J. Peter Lesley,
J. W. Powell,
S. L. Penfield,
R. Pumpelly,
A. R. C. Selwyn,
G. C. Swallow,

Albany, N.Y.
New York.
Berkeley, Cal.
Philadelphia.
Washington.
New Haven.
Newport, R.I.
Ottawa.
Columbia, Mo.

390

ASSOCIATE FELLOWS.

Section II. — 3.

Botany.

A. W. Chapman, Apalachicola, Fla.
W. Trelease, St. Louis.
John D. Smith, Baltimore.

Section III. — 8.

Zoology and Physiology.

Joel A. Allen, New York.
W. K. Brooks, Baltimore.
George B. Goode, Washington.

O. C. Marsh,

H. N. Martin,
S. Weir Mitchell,
A. S. Packard,
A. E. Verrill,

New Haven.
Baltimore.
Philadelphia.
Providence.
New Haven.

Section IV. — 5.

Medicine and Surgery.

John S. Billings, Washington.
Jacob M. Da Costa, Philadelphia.
A. Hammond, New York.

W

Alfred Stille,
H. C. Wood,

Philadelphia.
Philadelphia.

Class III. — Moral and Political Sciences. — 27.

Section I. — 5.
Philosophy and Jurisprudence.

T. M. Cooley, Ann Arbor.Mich.

D. R. Goodwin, Philadelphia.

Charles S. Peirce, New York.

T. R. Pynchon, Hartford, Conn.

Jeremiah Smith, Cambridge.

Section II. — 7.
Philology and Archceology.

A. N. Arnold, Pawtuxet, R.I.
Timothy Dwight, New Haven.

B. L. Gildersleeve, Baltimore.

D. C. Gilman, Baltimore.
T. R. Lounsbury, New Haven.

E. E. Salisbury, New Haven.
A. D. White, Ithaca, N.Y.

Section HI. — 9.
Political Economy and History.
Henry Adams, Washington.

G. P. Fisher, New Haven.

M. F. Force, Cincinnati.

H. E. von Hoist, Chicago.
Henry C. Lea, Philadelphia.

Edward J. Phelps, Burlington, Vt.
W. G. Sumner, New Haven.
J.H.Trumbull, Hartford, Conn.
David A. Wells, Norwich, Conn.

Section IV. — 6.

Literature and the Fine Arts.
Jame.s B. Angell, Ann Arbor, Mich.
L. P. di Cesnola, New York.
F. E. Church, New York.

R. S. Greenough, Florence.
Augustus St. Gaudens, New York.
W. R. Ware, New York.

FOREIGN HONORARY MEMBERS.

391

FOREIGN HONORARY MEMBERS. — 69.

(Number limited to seventy-five. Elected as vacancies occur.)

Class I. — Mathematical and Physical Sciences. — 25.

Sectiox I.
Mathematics and
Arthur Auwers,
Francesco Brioschi,
J. H. W. Dollen,
H. A. E. A. Faye,
Hugo Gylden,
Charles Hermite,
William Iluggius,
Otto Struve,
J. J. Sylvester,
H. C. Vogel,
Karl Weierstrass,

— 11.

Astronomy.
Berlin.
Milan.
Dorpat.
Paris.
Stockholm.
Paris.
London.
Karlsruhe.
Oxford.
Potsdam.
Berlin.

Sectiox II. — 3.

Physics.
A. Cornu, Paris.

Lord Rayleigh, Witham.

SirG. G. Stokes, Bart., Cambridge.

Section III. — 8.

Chemist

Adolf Baeyer,
Marcellin Berthelot,
Robert Bunsen,
J. H. van't Hoff,
Mendeleeff,
Victor Meyer,
Sir H. E. Roscoe,
Julius Thomseu,

ry.

Munich.

Paris.

Heidelberg.
Amsterdam.
St. Petersburg.
Heidelberg.
London.
Copenhagen.

Section IV. — 3.
Technology and Engineering.
Sir Henry Bessemer, London.

Lord Kelvin,
Maurice Levy,

Glasgow.
Paris.

Class II. — Natural and Physiological Sciences. — 25.

Section I. — 5.

Geology, Mineralogy, and Physics of
the Globe.

Alfred Des Cloizeaux, Paris.
A. E. Nordenskiold, Stockholm.
C. F. Rammelsberg, Berlin.
Henry C Sorby, Sheffield.

Heinrich Wild, Zurich.

Section II. — 7.

Botany.

J. G. Agardh, Lund.

E. Bornet, Paris.

Sir Joseph D. Hooker, Sunningdale.

Baron von Mueller, Melbourne.

Julius Sachs, Wiirzburg.

Solms-Laubach, Strasburg.

Eduard Strasburger, Bonn.

392

FOREIGN HONORARY MEMBERS.

Section III. — 9.
Zoology and Physiology.

Du Bois-Reymond,
Michael Foster,
Carl Gegenbauer,
Ludimar Hermann,
Albrecht KolUker,
A. Kovalevskij,
Lacaze-Duthiers,

Berlin.
Cambridge.
Heidelberg.
Konigsberg.
Wiirzburg.
St. Petersburg.
Paris.

Rudolph Leuckart, Leipsic.

J. J. S. Steenstrup, Copenhagen.

Section IV. — 4.

Medicine and Surgery,

W. Kiihne, Heidelberg.

Sir Joseph Lister, Bart., London.
Sir James Paget, Bart., London.
Rudolph Virchow, Berlin.

Class IIL — Moral and Political Sciences. — 19.

Section I. — 3.
Philosophy and Jurisprudence.

James Martineau, London.

Sir Frederick Pollock, Oxford.
Henry Sidgwick, Cambridge.

Section II. — 7.

Philology and Archoeology.

Ingram By water, Oxford.

Sir John Evans, Hemel Hempstead.
Pascual de Gayangos, Madrid.
J. W. A. Kirchhoff, Berlin.
G. C. C. Maspero, Paris.
Max Miiller, Oxford.

Karl Weinhold, Berlin.

Section III

. — 6.

Political Economy and History

Due de Broglie,

Paris.

James Bryce,

Oxford.

W. E. Gladstone,

Ilawarden.

Hermann Grimm,

Berlin.

Theodor Mommsen,

Berlin.

William Stubbs,

Oxford.

Section IV

. — 3.

Literature and the Fine Arts.

Jean Leon Gerome, Paris.
John Ruskiu, Coniston.

Leslie Stephen, London.

STATUTES AND STANDING VOTES.

STATUTES.

Adopted May 30, 1854 : amended September 8, 1857, November 12, 1862,
May 24, 1864, November 9, 1870, May 27, 1873, January 26, 1876,
June 16, 1886, October 8, 1890, January 11 and May 10, 1893, April
11, Mrt?/ 9, a7id October 10, 1894, anJ March 13, ^jor«7 10, a/i^Z May
8, 1895.

CHAPTER I.

Of Fellows and Fokeign Honorary Members.

1. The Academy consists of Fellows and Foreign Hoiiorary Mem-
bers. They are arranged in three Classes, according to the Arts
and Sciences in which they are severally proficient, viz. : Class I.
The Mathematical and Physical Sciences ; — Class II. The Nat-
ural and Physiological Sciences ; — Class III. The Moral and
Political Sciences. Each Class is divided into four Sections, viz. :
Class I., Section 1. Mathematics and Astronomy ; — Section 2.
Physics ; — Section 3. Chemistry ; — Section 4. Technology and
Engineering. Class II., Section 1. Geology, Mineralogy, and
Physics of the Globe ; — Section 2. Botany ; — Section 3. Zoology
and Physiology ; — Section 4. Medicine and Surgery. Class III.,
Section 1. Philosophy and Jurisprudence ; — Section 2. Philol-
ogy and Archseology ; — Section 3. Political Economy and
History ; — Section 4. Literature and the Fine Arts.

2. Fellows, resident in the State of Massachusetts, only, may
vote at the meetings of the Academy.* Each Eesident Fellow
shall pay an admission fee of ten dollars and such antiual assess-
ment, not exceeding ten dollars, as shall be voted by the Academy
at each Annual Meeting.

* The number of Resident Fellows is limited by the Charter to 200.

VOL. XXXI. (n. S. XXIII.)

394 STATUTES OF THE AMERICAN ACADEMY

3. Fellows residing out of the State of Massachusetts shall be
known and distinguished as Associate Fellows. They shall not
be liable to the payment of any fees or annual dues, but on remov-
ing within the State shall be admitted to the privileges,* and be
subject to the obligations, of Resident Fellows. The number of
Associate Fellows shall not exceed 07ie hundred, of whom there
shall not be more than forty in either of the three Classes of the
Academy.

4. The number of Foreign Honorary Members shall not
exceed seventy-five ; and they shall be chosen from among per-
sons most eminent in foreign countries for their discoveries and
attainments in either of the three departments of knowledge
above enumerated. And there shall not be more than thirty
Foreign Members in either of these departments.

CHAPTEK II.

Of Officers.

1. There shall be a President, three Vice-Presidents, one for
each Class, a Corresponding Secretary, a Recording Secretary, a
Treasurer, and a Librarian, which officers shall be annually elected,
by ballot, at the Annual Meeting, on the second Wednesday in
May.

2. At the same time, and in the same manner, nine Councillors
shall be elected, tliree from each Class of the Academy, but the
same Fellows shall not be eligible on more than three successive
years. These nine Councillors, with the President, the three
Vice-Presidents, the two Secretaries, the Treasurer, and the
Librarian, shall constitute the Council. It shall be the duty of
this Council to exercise a discreet supervision over all nomina-
tions and elections. With the consent of the Fellow interested,
they shall have power to make transfers between the several
Sections of the same Class, reporting their action to the
Academy.

3. If any office shall become vacant during the year, the
vacancy shall be filled by a new election, and at the next stated
meeting, or at a meeting called for this purpose.

* Associate Fellows may attend, but cannot vote, at meetings of the Academy.
See Chapter I. 2.

OP ARTS AND SCIENCES, 395

CHAPTER III.

Of Nominations of Officers.

1. At the stated meeting in March, the President shall appoint
from the next retiring Councillors a Nominating Committee of
three Fellows, one for each class.

2. It shall be the duty of this Nominating Committee to
prepare a list of candidates for the offices of President,
Vice-Presidents, Corresponding Secretary, Recording Secretary,
Treasurer, Librarian, Councillors, and the Standing Committees
which are chosen bj^ ballot ; and to cause this list to be sent by
mail to all the Resident Fellows of the Academy not later than
four weeks before the Annual Meeting.

3. Independent nominations for any office, signed by at least
five Resident Fellows and received by the Recording Secretary
not less than ten days before the Annual Meeting, shall be in-
serted in the call for the Annual Meeting, which shall then be
issued not later than one week before that meeting.

4. The Recording Secretary- shall prepare for use, in voting
at the Annual Meeting, a ballot containing the names of all
persons nominated for office under the conditions given above.

5. When an office is to be filled at any other time than at the
Annual Meeting, the President shall appoint a Nominating Com-
mittee, in accordance with the provisions of Section 1, which shall
announce its nomination in the manner prescribed in Section 2
at least two weeks before the time of election. Independent
nominations, signed by at least five Resident Fellows and received
by the Recording Secretary not later than one week before the
meeting for election, shall be inserted in the call for that
meeting.

CHAPTER IV.

Of the President.

1. It shall be the duty of the President, and, in his absence,
of the senior Vice-President present, or next officer in order as
above enumerated, to preside at the meetings of the Academy ;
to summon extraordinary meetings, upon any urgent occasion ;
and to execute or see to the execution of the Statutes of the

396 STATUTES OF THE AMERICAN ACADEMY

Academy. Length of continuous membership in the Academy
shall determine the seniority of the Vice-Presidents.

2. The President, or, in his absence, the next officer as above
enumerated, is empowered to draw upon the Treasurer for such
sums of money as the Academy shall direct. Bills presented on
account of the Library, or the Publications of the Academy, must
be previously approved by the respective committees on these
departments.

3. The President, or, in his absence, the next officer as above
enumerated, shall nominate members to serve on the different
committees of the Academy which are not chosen by ballot.

4. Any deed or writing to which the common seal is to be
affixed shall be signed and sealed by the President, when thereto
authorized by the Academy.

CHAPTER V.

Of Standing Committees.

1. At the Annual Meeting thej-e shall be chosen the following
Standing Committees, to serve for the year ensuing, viz. : —

2. The Committee of Finance, to consist of the President,
Treasurer, and one Fellow chosen by ballot, who shall have
charge of the investment and management of the funds and trusts
of the Academy. The general appropriations for the expendi-
tures of the Academy shall be moved by this Committee at the
Annual Meeting, and all special appropriations from the general
and publication funds shall be referred to or proposed by this
Committee.

3. The Eumford Committee, of seven Fellows, to be chosen
by ballot, who shall consider and report on all applications and
claims for the Kumford Premium, also on all appropriations from
the income of the Eumford Fund, and generally see to the due
and proper execution of this trust.

4. The C. M. Warren Committee, of seven Fellows, to be
chosen by ballot, who shall consider and report on all applica-
tions for appropriations from the income of the C. M. Warren
Fund, and generally see to the due and proper execution of this
trust.

5. The Committee of Publication, of three Fellows, to whom
all memoirs submitted to the Academy shall be referred, and to

OF ARTS AND SCIENCES. 897

wh.om the printiug of memoirs accepted for publication shall be
intrusted.

6. The Committee on the Library, of three Fellows, who
shall examine the Library, and make an annual report on its
condition and management.

7. An Auditing Committee, of two Fellows, for auditing the
accounts of the Treasurer.

CHAPTER VI.

Of the Secretaries.

1. The Corresponding Secretary shall conduct the correspond-
ence of the Academy, recording or making an entry of all letters
written in its name, and preserving on file all letters which are
received ; and at each meeting he shall present the letters which
have been addressed to the Academy since the last meeting.
With the advice and consent of the President, he may effect
exchanges with other scientific associations, and also distribute
copies of the publications of the Academy among the Associate
Fellows and Foreign Honorary Members, as shall be deemed
expedient ; making a report of his proceedings at the Annual
Meeting. Under the direction of the Council for Nomination,
he shall keep a list of the Fellows, Associate Fellows, and Foreign
Honorary Members, arranged in their Classes and in Sections in
respect to the special sciences in which they are severally profi-
cient ; and he shall act as secretary to the Council.

2. The Eecording Secretary shall have charge of the Charter
and Statute-book, journals, and all literary papers belonging to
the Academy. He shall record the proceedings of the Academy
at its meetings ; and after each meeting is duly opened, he shall
read the record of the preceding meeting. He shall notify the
meetings of the Acadeni}', and apprise committees of their ap-
pointment. He shall post up in the Hall a list of the persons
nominated for election into the Academy ; and when any indi-
vidual is chosen, he shall insert in the record the names of the
Fellows by whom he was nominated.

3. The two Secretaries, with the Chairman of the Committee
of Publication, shall have authority to publish such of the pro-
ceedings of the Academy as as may seem to them calculated to
promote the interests of science.

398 STATUTES OF THE AMERICAN ACADEMY

CHAPTER VII.

Of the Teeasurer.

1. The Treasurer shall give such security for the trust reposed
in him as the Academy shall require.

2. He shall receive officially all moneys due or payable, aud
all bequests or donations made to the Academy, and by order of
the President or presiding officer shall pay such sums as the
Academy may direct. He shall keep an account of all receipts
and expenditures ; shall su.bmit his accounts to the Auditing
Committee ; and shall report the same at the expiration of his
term of office.

3. The Treasurer shall keep a separate account of the income
and appropriation of the Eumford Fund, and report the same
annually. >

4. All moneys which there shall not be present occasion to
expend shall be invested by the Treasurer, under the direction of
the Finance Committee, on such securities as the Academy shall
direct.

CHAPTEPt VIII.

Of THE Librarian and Library.

1. It shall be the duty of the Librarian to take charge of the
books, to keep a correct catalogue of same, and to provide for
the delivery of books from the Library. He shall also have the
custody of the publications of the Academy.

2. The Librarian, in conjunction with the Committee on the
Library, shall have authority to expend, as they may deem ex-
pedient, such sums as may be appropriated, either from the Rum-
ford or the General Fund of the Academy, for the purchase of
books, and for defraying other necessary expenses connected with
the Library. They shall have authority to propose rules and
regulations concerning the circulation, return, and safe-keeping of
books ; and to appoint such agents for these purposes as they
may think necessary.

3. To all books in the Library procured from the income of
the Rumford Fund, the Librarian shall cause a stamp or label to
be affixed, expressing the fact that they were so procured.

OP ARTS AND SCIENCES. 399

4. Every person who takes a book from the Library shall give
a receipt for the same to the Librarian or his assistant.

5. Every book shall be returned in good order, regard being
had to the necessary wear of the book with good usage. And if
any book shall be lost or injured, the person to whom it stands
charged shall replace it by a new volume or set, if it belongs
to a set, or pay the current price of the volume or set to the
Librarian ; and thereupon the remainder of the set, if the volume
belonged to a set, shall be delivered to the person so paying for
the same.

6. All books shall be returned to the Library for examination
at least one week before the Annual Meeting.

CHAPTEE IX.

Of Meetings.

1. There shall be annually four stated meetings of the Acad-
emy ; namely, on the second Wednesday in May (the Annual
Meeting), on the second Wednesday in October, on the second
Wednesday in January, and on the second Wednesday in March.
At these meetings only, or at meetings adjourned from these
and regularly notified, shall appropriations of money be made,
or alterations of the statutes or standing votes of the Academy
be effected.

2. Fifteen Fellows shall constitute a quorum for the transac-
tion of business at a stated meeting. Seven Fellows shall be
sufficient to constitute a meeting for scientific communications
and discussions.

3. The Recording Secretary shall notify the meetings of the
Academy to each Fellow residing in Boston and the vicinity; and
he may caiise the meetings to be advertised, whenever he deems
such further notice to be needful.

400 STATUTES OF THE AMERICAN ACADEMY

CHAPTER X.

Of the Election of Fellows and Honorary Membees.

1. Elections shall be made by ballot, and only at stated
meetings.

2. Candidates for election as Resident Fellows must be pro-
posed by two or more Resident Fellows, in a recommendation
signed by them, specifying the Section to which the nomination
is made, which recommendation shall be transmitted to the
Corresponding Secretary, and by him referred to the Council for
Nomination. No person recommended shall be reported by the
Council as a candidate for election, unless he shall have received
a written approval, signed at a meeting of the Council by at least
seven of its members. All nominations thus approved shall be
read to the Academy at a stated meeting, and shall then stand on
the nomination list during the interval between two stated meet-
ings, and until the balloting. No person shall be elected a Resident
Fellow, unless he shall have been resident in this Commonwealth
one year next preceding his election. If any person elected a
Resident Fellow shall neglect for one year to pay his admission
fee, his election shall be void ; and if any Resident Fellow shall
neglect to pay his annual assessments for two years, provided
that his attention shall have been called to this article, he shall be
deemed to have abandoned his Fellowship; but it shall be in the
power of the Treasurer, with the consent of the Council, to dis-
pense (sub silentio) with the payment both of the admission fee
and of the assessments, whenever in any special instance he shall
think it advisable so to do.

3. The nomination of Associate Fellows shall take place in
the manner prescribed in reference to Resident Fellows ; and
after such nomination shall have been publicly read at a stated
meeting previous to that when the balloting takes place, it shall be
referred to the Council for Nomination ; and a written approval,
authorized and signed at a meeting of said Council by at least
seven of its members, shall be requisite to entitle the candidate
to be balloted for. The Council may m like manner originate
nominations of Associate Fellows, which must be read at a stated
meeting previous to the election, and be exposed on the nomina-
tion list during the interval.

OF ARTS AND SCIENCES. 401

4. Foreign Honorary Members shall be chosen only after a
nomination made at a meeting of the Council, signed at the time
by at least seven of its members, and read at a stated meeting
previous to that on which the balloting takes place.

5. Three fourths of the ballots cast must be affirmative, and
the number of affirmative ballots must amount to eleven to effect
an election of Fellows or Foreign Honorary Members.

6. Each Section of the Academy is empowered to present lists
of persons deemed best qualified to fill vacancies occurring in the
number of Foreign Honorary Members or Associate Fellows
allotted to it ; and such lists, after being read at a stated meeting,
shall be referred to the Council for Nomination.

7. If, in the opinion of a majority of the entire Council, any
Fellow — Resident or Associate — shall have rendered himself
unworthy of a place in the Academy, the Council shall recommend
to the Academy the termination of his Fellowship ; and provided
that a majority of two thirds of the Fellows at a stated meeting,
consisting of not less than fifty Fellows, shall adopt this recom-
mendation, his name shall be stricken off the roll of Fellows.

CHAPTER XI.

Of Amendments of the Statutes.

1. All proposed alterations of the Statutes, or additions to
them, shall be referred to a committee, and, on their report at a
subsequent meeting, shall require for enactment a majority of
two thirds of the members present, and at least eighteen affirma-
tive votes.

2. Standing Votes may be passed, amended, or rescinded, at
any stated meeting, by a majority of two thirds of the members
present. They may be suspended by a unanimous vote.

CHAPTER XII.

Of Liteeary Performances.

1. The Academy will not express its judgment on literary or

scientific memoirs or performances submitted to it, or included in

its publications.

26

402 STATUTES OP THE AMERICAN ACADEMY

STANDING VOTES.

1. Communications of which notice had been given to the
Secretary shall take precedence of those not so notified.

2. Kesident Fellows who have paid all fees and dues charge-
able to them are entitled to receive one copy of each volume or
article printed by the Academy, on application to the Librarian
personally or by written order, within two years from the date
of publication. And the current issues of the Proceedings shall
be supplied, when ready for publication, free of charge, to all the
Fellows and members of the Academy who desire to receive them.

3. The Committee of Publication shall fix from time to time
the price at which the publications of the Academy may be sold.
But members may be supplied at half this price with volumes
which they are not entitled to receive free, and which are needed
to complete their sets. ,

4. Two hundred extra copies of each paper accepted for publi-
cation in the Memoirs or Proceedings of the Academy shall be
placed at the disposal of the author, free of charge.

5. Resident Fellows may borrow and have out from the
Library six volumes at any one time, and may retain the same
for three months, and no longer.

6. Upon special application, and for adequate reasons assigned,
the Librarian may permit a larger number of volumes, not exceed-
ing twelve, to be drawn from the Library for a limited period.

7. Works published in numbers, when unbound, shall not be
taken from the Hall of the Academy, except by special leave of
the Librarian.

8. Books, publications, or apparatus shall be procured from the
income of the Rumford Fund only on the certificate of the Rum-
ford Committee that they, in their opinion, will best facilitate
and encourage the making of discoveries and improvements which
may merit the Rumford Premium.

9. The Annual Meeting and the other stated meetings shall be
holden at eight o'clock, P. M.

10. A meeting for receiving and discussing scientific commu-
nications may be held on the second Wednesday of each month
not appointed for stated meetings, excepting July, August, and
September.

OF ARTS AND SCIENCES. 403

RUMFORD PREMIUM.

In conformity with the terms of the gift of Benjamin, Count
Rumford, granting a certain fund to the American Academy of
Arts and Sciences, and with a decree of the Supreme Judicial
Court for carrying into effect the general charitable intent and
purpose of Count Rumford, as expressed in his letter of gift, the
Academy is empowered to make from the income of said fund, as
it now exists, at any Annual Meeting, an award of a gold and
silver medal, being together of the intrinsic value of three
hundred dollars, as a premium to the author of any important
discovery or useful improvement in light or in heat, which shall
have been made and published by printing, or in any way made
known to the public, in any part of the continent of America, or
any of the American islands ; preference being always given to
such discoveries as shall, in the opinion of the Academy, tend
most to promote the good of mankind ; and to add to such
medals, as a further premium for such discovery and improve-
ment, if the Academy see fit so to do, a sum of money not
exceeding three hundred dollars.

INDEX.

JEcidia, development of, 255.
Agassiz, A., and West, P. C. F.,

Oil the Temperature of the

Crust of the Earth at great

Depths, 347.
Alaminum, melthig point of, 218.
Aiiimon-cupriammonium aceto-ox-

ide, 79.
Ammonic dicupric acetate, 84.
Anemone, ^cidiuni on, 262.
Aromatic hydrocarbons, 34, 58.
Asterias pallida, embryology of, 333.
Atomic Weight of Zinc, 158.

B.

Bancroft, W. D., The Chemical
Potential of the Metals, 96-
122.
Baric Sulphate, Occlusion of Baric

Chloride by, 67.
Barr, L. See Holman, S. W., Law-
rence, R. R., and Barr, L.
Bars, oflf-shore, 319.

tanofeut, 321.
Benzol, 34, .58.

derivatives of, 123.
Biographical Notices, list of, 353.
Martin Brimmer, 360.
Richard Manning Hodges, 355.
Viscount Ferdinand de Le ^seps,

370.
James Edward Oliver, 367.
Edward Samuel Ritchie, 359.
Henry Wheatland, 363.
Harold Whiting, 356.

c.

Calorimetry, 245.

Calvert, Sidney. See Jackson, C. L.,
and Calvert, Sidney.

Canadian petroleums, 43.

Cape Cod, future of, 331.
outline of, 303.

Cathode rays, 349.

Chemical Laboratory of the Case
School of Applied Science,
Contributions from, 1.

Chemical Laboratory of Harvard
College, Contributions from,
67, 78, 87, 96, 123, 136, 158.

Chemical Potential of the Metals,
96.

Conductivity of Mild Steel, 271.

Copper, melting point of, 218.

Council, Report of, 353

Cryptogamic Laboratory of Har-
vard University, Contribu-
tion from, 255.

Cupriammonium acetate, 8.5.

Cupriammouium double salts, 78.

Cuprianiline acetobroraide, 89.

Cuprianiline bromide, 88.

Cuprianiline salts, 87.

D.

Davis, W. M., The Outline of Cape

Cod, 303-332.
Dibrommetaphenylene dicetamide,

149.
Dicuprianiline acetomonobromide,

92.

406

INDEX.

E.

Edison, T. A., award of Rumford

medal to, 3il, 343.
Embryology of the Starfish, 333.

F.

Fellows, Associate, deceased, —

James Dwight Dana, 339.

Daniel Cady Eaton, 346.

Asahel Clark Kendrick, 347.

John Newton, 339.

William Wetmore Story, 346.
Fellows, Associate, elected, —

William Price Craighill, 343.

Basil Lanneau Gildersleeve,
352.

Thomas Raynesford Louns-
bury, 352.

Charles Lane Poor, 352.

Augustus St. Gaudens, 348.

Robert Simpson Woodward,
352.
Fellows, Associate, list of, 389.
Fellows, Resident, deceased, —

IMartin Brimnaer, 349.

Richard Manning Ilodges, 349.

Edward Sanmel Ritchie, 345.

Harold Whiting, 345.

Henry Wi Hard' Williams, 345.
Fellows, Resident, elected, —

Benjamin Kendall Emerson,
346.

John Fiske. 343.

Hammond Vinton Hayes, 346.

Arthur IVIichael, 343.

Theobald Smith, 352.

John Stone Stone, 351.

Arthur Gordon Webster, 343.

Robert Wheeler WiJlson, 351.

Paul Sebastian Yendell, 346.
Fellows, Resident, list of, 385.
Foreign Honorary Members, de-
ceased, —

Thomas Henry Huxley, 346.

Sven Ludwig Loven, 346.

Carl Friedrich Wilhelm Lud-
wig, 346.

Louis Pasteur, 346.
Foreign Honorary Members,
elected, —

Marie Alfred Cornu, 346.

Jacobus Henricus van't Hoff,
346.

Foreign Honorary Members, list of,
391.

G.

Gallivan, F. B. 5ee Jackson, C. I*.,

and Gallivan, F. B.
Gold, melting point of, 218.
Goodale, G. L., Forestry under New

England Conditions, 349.
Goto, S., Preliminary Notes on the

Embryology of the Starfish

(Asterias pallida), 333-335.
Greenman, J. L. See Robinson, B.

L., and Greenman, J. L.

H.

Hall, E. H., On the Thermal Con-
ductivity of Mild Steel, 271-
302.

High Head, 310.

Holman, S. W., Calorimetry: Meth-
ods of Cooling Correction,
245-254.
Pyrometry : Calibration of the
Le Chatelier Thermo-electric
Pyrometer, 234-244.
Thermo-electric Interpolation
Formula}, 193-217.

Holman, S. W., Lawrence, R. R.,
and Barr, L., Melting Points
of Aluminum, Silver, Gold,
Copper, and Platinum, 218-
233.

Houstonia caerulea, -3icidium on,
260.

Hydrocarbons C,,H2„ + 2> 22.

Ittner, M. H. See Jackson, C. L.,
and Ittner, M. H.

Jackson, C. L., on potassic cobalticy-
anide, 350.

Jackson C. L., and Calvert, S., On
the Behavior of certain Deriv-
atives of Benzol containing
Halogens, 123-135.
Bromine Derivatives of Meta-
phenylene Diamine, 136-157.

INDEX.

407

Jackson, C. L., and Gallivan, F. B.,
On certain Derivatives of un-
syininetical Tribrombenzol,

Jackson, C. L., and Tttner, M. H.,
On Dinitrobromtoluol and
some of its Derivatives, 345.

Jackson, C. L., and Moore, J. H.,
On the Nitrite of Brom-
diiiitropheuylmalonic Ester,
345.

Jackson, C. L., and Oenslager, G.,
On Phenoquinone, 345.

Jackson, C. L., and I^hinney, J. I.,
Trinitrophenylmalonic Ester,
345.

L.

Lawrence, R. R. See Ilolman, S.
W., Lawrence, R. R., and
Barr, L.

Le Chatelier pyrometer, 234.

Librarian, Report of, 340.

Linear transformations, 336.
of a bilinear form, 181.

Lyon, D. G., Recent Assyrian dis-
coveries, 352.

M.

Mabery, C. F., On the Composition
of the Ohio and Canadian
Sulphur Petroleums, 1-66.

Marcou, J., gift of, 346.

Melting points of certain metals,
218.

Metals, chemical potential of, 96.

Metaphenylene diamine, 136.

Metric system of weights and
measures, 351.

Monobromphenylene diamene, bro-
mide of, 156.

Moore, J. H. See Jackson, C. L.,
and jNIoore, J. H.

Moulton, F. C. See Richards, T. W.,
and Moulton, F. C.

N.
New Jersey, coast of, 323.

0.

Octocupriammonium mono-iodide
acetate, 81.

Octocuprianiline acetomonobro-
mide, 93.
monoacetobromide, 91.
triacetobromide, 91.

Oenslager, G. See Jackson, C. L.,

and Oenslager, G. ; also Rich-
ards, T. W., and Oenslager, G.

Officers elected, 343, 348.

Off-shore bars, 319.

Ohio petroleums, 12.

Parker, H. G. See Richards, T. W.,
and Parker, H. G.

Peirce, B. O., On Simultaneous par-
tial Differential Equations,
345.

Pierce, B. O., and Willson, R. W.,
On the Thermal Conductivi-
ties of certain poor Conduc-
tors, 345.

Peltandra, ^Ecidium on, 256.

Petroleum, origin of, 64.

Phinney, J. I. See Jackson, C. L.,
and Phinney, J. L

Physical Laboratory of the Massa-
chusetts Institute of Tech-
nology, Contributions from,
193, 218. 234, 245

Platinum, melting point of, 218.

Proceedings of JNIeetings, 339.

Provinceland, growth of, 323.

Provincetown harbor, 329.

Pyrometry, 234.

E.

Race Point, origin of, 327.
Ranunculus septentrionalis, ^cidi-

nm on, 261.
Richards, H. jM., On some Points

in the Development of ^ci-

dia, 25.5-270.
Richards, T. W., and Moulton, F. C,

On the Cuprianile Acetobro-

mides, 87-95.
Richards, T. W., and Oenslager, G.,

On the Cupriamraonium Dou-
ble Salts, 78-86.

408

INDEX.

Richards, T. W., and Parker, H. G.,
On the Occlusion of Baric
Chloride by Baric Sulphate,
67-77.

Richards, T. W., and Rogers, E. F.,
A Revision of the Atomic
Weight of Zinc: Analysis
of Zincic Bromide, 158-180.

Robinson, B. L., and Greenman, J.
L., Contributions from the
Gray Herbarium of Harvard
University, New Series, No.
9, 345.

Rogers, E. F. 5ee Richards, T. W. ,
and Rogers, E. F.

Rumford Committee, Report of,
341.

Rumford, Count, letters of, 346.

Rumford Fund, papers published
by aid of, 193, 218, 234, 245,
271.

Romford Medal, award of, 341.

S.

Sambucus, ^cidium on, 264.
Shore outlines, 315.

profiles, 312.
Silver, melting point of, 218.
Starfish, embryology of, 333.
Statutes and Standing Votes, 389.
Story, W. E., On New Methods

of representing Mathematical

Surfaces, 346.
Sulphur Petroleums, Composition

of, 1.

Taber, H., Note on the Automorphic
Linear Transformation of a
Bilinear Form, 181-192.
On the Group of Real Linear
Transformations whose In-
variant is an Alternate Bilin-
ear Form, 336, 337.
Transformation of the Group
whose Invariant is a certain
linear Complex, 349.

Tangent bars, 341.

Tetrabrombenzol, symmetrical, 133.
unsymmetrical, 132.

Tetrabromdinitrobenzol, action of
sodic ethylate on, 134.

Tetrabromphenylene diamine, 156.

Tetrammon-tricupriammouium io-
dide, 83.

Thaxter, R., On Laboulbeniacese,
345.

Thermo-electric Interpolation For-
mulaB, 193.

Toluol, 34, 59.

Toy, C. H., Palestine in the Fif-
teenth Century B. C, 347.

Treasurer, Report of, 339.

Tribrombenzol, symmetrical, 134.

Tribromchlorbenzol, unsymmetri-
cal, 133.

Tribronidinitrobenzol, reduction of,
144.

Tribromiodbenzol, action of fuming
nitric acid on, 131.
behavior of, 128.

Tribrommetaphenylene diamine,
140.

Trowbridge, J., experiments with
the cathode rays, 349, 350.

w.

Warren (C. M.) Committee, Report

of, 342.
Fund, aid from, 1.
West, P. C. F. See Agassiz, A., and

West, P. C. F.
Willson, R. W. See Peirce, B. O.,

and Willson, R. W.

Xylols, 35, 59.

X.

z.

Zincic Bromide. Analysis of, 158.
Specific gravity of, 162.

Zoological Laboratory of the Mu-
seum of Comparative Zoology
at Harvard College, Contri-
bution from, 333.

MBL/WHOI LIBRARY

UH lAflE $

^5^H

Vi.il\ti