sien | seeker

meses ifs3 : renee ‘
eee reese ete ” $ resins mac «2
fea Wey is ne Best ie cee F eee ‘ ie ; . P pare Re Galt: WEY ; NEAT: ogitt | Caco Ne
send Sedicdhy dea te td : ps one » eel
wr cea Race el Me roar
ee ees ti
ihonefs

eve Fra ‘ : ; Padtadearh a ; vite d va +4: en eet
» Se lth ele NANG at : Rete eae BH
pi het “he es ve " flap i ; eo 2 ? i‘ ~ ‘ 7” om Sa

1 hots

A adr AE: i ; wince: : Se ee me a pa I Hh “tis et He “ses : , 4 a.
:: a 24 see - vi Far tents seg fo : een st eects ety lp Nes . 4 AN hte
5 ~ . Sf Shh ean bis ea od wes > z

ieee re Pf y er ene tha ito ieee
7 de teiaite eth? a "ihe wits ol * ‘ ‘be: en its . oe wise. datestae 2 ABB Cte satan
hg en Vinh ois he ss 4 bone fv as “a te Renter wie ete irene: ae H: eee: He iy AAG: ~ . “ap hip . i, S Praia , ¢ Al nd i la Y ewe ; : By te

A crap nate ven ey ee ee ee , wots penn owes ris Syerhetal citi beh i ae Ae nin ates heal . , ppt tie tee Psi ie atone: Me ve stoi yi :

ate emetdiet os ath a A Si prog De ociael silent “is i Hate. “ee age , eto Aiea mR AE Feat f > a ind ie Hs . ae Aine — neti apne ene jose is
wrt hee= hie 4 Cy eae ee 4 Ke At AE iy" Qeinck ies oe atin aie | nig Ge ghee xfer 5 4 , oe ae re ph a FR ire Ps RET maths chp, eee es ur hat sh, chee cteateit | (Vaiteyooge tert.
we Evhakon : . rae 5%, . seb oe . ; So i oe aoe ii: ry 4a a eee ; cand rete hes § ete eis iz j fe a dbo iw Hea. vipa Mond . . ‘ wiebiatolinine ts jibatenibad
yg : - ss ‘: Ste ny Moong " or . © ‘ - : 4 . t

wiyiut ste % path eee Ton Bo aero Reoemeatnw thre oe fou a
2 Spa Be siratirs nah phat sett Lao-ayinat Tea ish st
*h wr ag PO y
3 fy Ten fan te Ses BANA

ap Abb Sage Pew eatin A Ronin te
Speed } We ‘ea iat ee MO nb
te te ee * ‘ ane : F

ews Me oes

: LIBRAR¥-QFo( CONGRE 88.4

[SsuuTHSONTAN DEPOSIT]
Chap 4 Uy

HS $0 -72.¢ S
a a
UNITED STATES OF Ad Rica.’

PDP PY BD PB BH, BOOS SSoe caf

x deme eg ae

JSESTSCSESSESES*

THE

LONDON, EDINBURGH, anp DUBLIN

PHILOSOPHICAL MAGAZINE

AND

JOURNAL OF SCIENCE.

CONDUCTED BY

SIR ROBERT KANE, LL.D. F.BR.S. M.R.LA. F.C.S.
SIR WILLIAM THOMSON, Kyr. LL.D. F.R.S. &e.

AND

WILLIAM FRANCIS, Pu.D. F.L.S. F.R.A.S. F.C.S.

‘¢ Nec aranearum sane textus ideo melior quia ex se fila gignunt, nec noster
vilior quia ex alienis libamus ut apes.” Just. Lies. Poli. lib.i. cap. 1. Not.

’

VOL. XX.—FIFTH SERIES.
JULY—DECEMBER 18835.

, $ \= = * ‘ tS >
f Vv hes 75
—
I A rh S
v . e o. ~~
4 Bagh 2 ary. < Fiz :
‘ ae a. < .
hea kes nee ae
~* ‘ mo eD - §
_ J; aI ~ a
>» gig +
LONDON: “Se ny,
hi ’/ aR pS -<
x wn, tA ree T Pd

ating ‘oa ee
TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET.

SOLD BY LONGMANS, GREEN, AND CO.; KENT AND CO. ; SIMPKIN, MARSHALL, AND CO. ,;
AND WHITTAKER AND CO.;—AND BY ADAM AND CHARLES BLACK, AND
T. AND T. CLARK, EDINBURGH; SMITH AND SON, GLASGOW 5-—
HODGES, FOSTER, AND CO., DUBLIN ;——-PUTNAM, FEW YORK ;—
VEUVE J. BOYVEAU, PARIS ;—AND ASHER AND CO., BERLIN.

“‘Meditationis est perscrutari occulta; contemplationis est admirari
perspicua..... Admiratio generat queestionem, queestio Investigationem,
investigatio inventionem.”—Hugo de S. Victore.

—‘ Cur spirent venti, cur terra dehiscat,

Cur mare turgescat, pelago cur tantus amaror,
Cur caput obscura Phoebus ferrugine condat,
Quid toties diros cogat flagrare cometas,

Quid pariat nubes, veniant cur fulmina ceelo,
Quo micet igne Iris, superos quis conciat orbes

Tam vario motu.”
J. B. Pinelli ad Mazonium.

CONTENTS OF VOL. XX.

(FIFTH SERIES).

NUMBER CXXII.—JULY 1885.

Pa
Dr. L. B. Fletcher on a Determination of the B.A. Unit in

Terms of the Mechanical Equivalent of Heat............
Mr. J. Edmondson on Calculating-Machines..............
J. A. Groshans on a New Law, analogous to those known

under the names Law of Avogadro and Law of Dulong and

CER ay Una s iy se ee « ou sare UN
Mr. H. I. Newall on Colliding Water-jets................
MM. Elster and Geitel’s Observations on the Electrical Pro-

Rpeemeeee OUT CIOS os ce evi ee we ts St me
_ Assistant-Prof. J. W. Clark on certain Cases of Electrolytic

9 DL CELL anal etl aoe RT i ap eee esata iene tee
Assistant-Prof. J. W. Clark on the Determination of the Heat-

eee a Thermometer... ... 2+... ccs n ch ee ewes
Mr. A. M. Worthington on the Error involved in Professor

Quincke’s Method of Calculating Surface-Tensions from the

Mimensions of fiat Drops and Bubbles ................
Prof. Oliver Lodge on the Stream-Lines of Moving Vortex-

rings. (Plates IJ., III., IV.)
Notices respecting New Books :—
Dr. E. Hoppe’s Geschichte der Elektrizitat ..........
Proceedings of the Geological Society :—
Mr. T. Mellard Reade on the Evidence of the Action of
Land-Ice at Great Crosby, Lancashire..............
Mr. D. C. Davies on the North-Wales and Shrewsbury
Ee eer d ts cree hs oak kg eee alt 5 oe &
On the Depth to which Daylight penetrates in Sea-water, by
eee Heh iid Mil soarasi oye e a 5s os cs ce ses Se
On the Electrical Conductivity of Solid Mercury and of Pure

Metals at Low Temperatures, by MM. Cailletet and Bouty.
Optical Method for the Absolute Measurement of Small

Meneths, by M. Mace de Lepinay .... 2.0.00... ee.

e 2. SO 8) we Ae AP Be Se.) OS) C.. Bae) 0) (ye

NUMBER CXXTII—AUGUST.

Prof. L. Henry on the Polymerization of the Metallic Oxides.
Mr. J. Buchanan on the Thermoelectric Position of Carbon.

ae UNM Rites nee) as ais by thd a a etaiels ie ee ee ees
Dr. J. A. Fleming on the Use of Daniell’s Cell as a Standard

of Electromotive Force. (Plate V.) ......5.50 06. ends

74

iv CONTENTS OF VOL. XX.—FIFTH SERIES.
Page
Dr. J. A. Fleming on Molecular Shadows in Incandescence
TAMPS ok. Gi ALE. cons be ve ee ee . 141
Dr.J.D. Dana on the Origin of Coral Reefs and Islands. (PlateI.) 144
Dr. J. H. Gladstone on the Specific Refraction and Dispersion

of Light by the Alums ......... 503 5s.).42. 5 rrr 162
Mr. C. Tomlinson on the Bleaching of Iodide of Starch by

means of Heat ... 1... 0. SEES Dias cme aoe 2 168
Capt. W. de W. Abney on the Production of Monochromatic

Light, or a Mixture of Colours, on the Screen .......... 172
Mr. W. Sutherland on the Mechanical Integration of the

Product of two Functions . ......0..+0+:..00 90 175
Mr. S. Bidwell on the Sensitiveness of Selenium to Light, and

the Development of a similar Property in Sulphur ...... 178

J. A. Groshans on a New Law, analogous to those known
under the names Law of Avogadro and Law of Dulong and
OG ng oes ee ee nee be be vem eer 191

Notices respecting New Books :—

Lieut. Lefroy’s Diary of a Magnetic Survey of a Portion
of the Dominion of Canada, chiefly in the North-western
Territories, executed in the years 1842-44.......... 204
Proceedings of the Geological Society :— :
Prof. T. G. Bonney on the so-called Diorite of Little
Knott (Cumberland), with further Remarks on the
Oceurrence of Picrites in Wales...........: 73 205
Mr. W. H. Penning’s Sketches of South-African Geology.
—No. 2. ASketch of the Gold-Fields of the Transvaal,
South Arica 6... 6... a es we > ee 206
Dr. Charles Ricketts on some Erratics in the Boulder-clay
of Cheshire, &c., and the Conditions of Climate they

GETIOEC | gals kee sw mee oe Oe ane ec 207
Messrs. J. E. Marr and T. Roberts on the Lower Paleozoic
Rocks of the Neighbourhood of Haverfordwest...... 208

Mr. W.S. Gresley on certain Fossiliferous Nodules and
Fragments of Hematite (sometimes Magnetite) from
the (so-called) Permian Breccias of Leicestershire and

South Derbyshire .. 2.0.0.0 505 een eas 209
Prof. J. W. Judd and Mr. C. Homersham’s Supplementary
Notes on the Deep Boring at Richmond, Surre 210

of the Breidden Hills in East Montgomeryshire and

West Shropshire 1.0... ee ae ee 211
Captain I’. W. Hutton’s Correlations of the Curiosity-

Shop Beds, Canterbury, New Zealand.............. 212

On Hygrometry, by M. Jamin: 2.35 -.: 2. 4. $4. 21

On the Deformation of the Luminous Wave-surface in the

Magnetic Field, by E. von Fleisch] ...........2......7.) 216
On a new Method for Determining the Mechanical Equivalent

of ' Meat, by AG: Webster... ce. bids Vins ce 217

New Form of Hygrometer, by M. Bourbouze. ............ 220

EE LE A
Pi ; ,
:

CONTENTS OF VOL. XX.-——FIFTH SERIES. Vv

Page

On the ae of Heat in the Swelling and Solution
of Colloids, by E. Wiedemann and Ch. Ludeking ........ 220
On a Mercury-Galvanometer, by G. Lippmann ............ 220

NUMBER CXXIV.—SEPTEMBER.

Dr. J. A. Fleming on the Distribution of Electric Currents
in Networks of Conductors treated by the Method of
Meee (Plater VE. VILL). oo Se 221
Dr. T. Carnelley on the Periodic Law, as Illustrated by certain
Physical Properties of Organic Compounds.—Part I. The
Alkyl Compounds of the Beeb pies gt le 259
Dr. J. D. Dana on the Origin of Coral Reefs and Islands .. 269
Dr. J. Hopkinson on an unnoticed Danger in certain Appa-
eee ror Histribution of Electricity .........6.......- 292
Dr. W. W. J. Nicol on Supersaturation of Salt-Solutions.... 295
Notices respecting New Books :—
Messrs. Glazebrook and Shaw’s Textbook of Practical
Rpmmrrmesceneree yee es FER ee IIIS PO Sd |, 300
On the Electrical Resistance of Alcohol, by M. G. Foussereau. 301
The Volcanic Nature of a Pacific Island not an Argument for

litle or no Subsidence, by J. D. Dana .......2...0.0.-. 303
On the Quantity a Electrical Elementary Particles, by E.
Mere nae oe ele ee OL PPE PR ee 303

NUMBER CXXV.—OCTOBER.
Mr. F. Stenger on the Electric Conductivity of Gases. (Plate

I reheat! 2. ARR held GULLY. ORs 305
Mr. R. H. M. Bosanquet on Electromagnets.—LV. Cast Iron,
Charcoal Iron, and Malleable Cast Iron ................ 318
Dr. J. J. Hood on the Influence of Heat on the Rate of Che-
EEE I es A a ee oes EU, IN 323 -
Mr. 8. Bidwell on the \Gehention of Electric Currents by
(LE SL SS a oe ce eee a Se 328
Dr. J. Hopkinson on the Seat of the Electromotive Forces in
ie Orr ifr oA ed eects Suns Smee che bee ots oa 336

Mr. R. T. Glazebrook on a Comparison of the Standard Resist-
ance-coils of the British Association with Mercury Standards
constructed by M. J. R. Benoit of Paris and Herr Strecker
Seem re FPN ot, SOU CUE ihe 343

Lord Rayleigh on the Accuracy of Focus necessary for Sensibly
Renee Poe nL eI ALS OLY no Bk I oe, 354

Lord Rayleigh on an Improved Apparatus for Christiansen’s _
imeem Me Ac cee Pe I oS tad 398

Lord Rayleigh’s Optical Comparison of Methods for observing
Seemal, SEQUINS... Priam ep eee we oe EM 360

vil CONTENTS OF VOL. XX.—FIFTH SERIES.

: Page
Lord Rayleigh on the Thermodynamic Efficiency of the Ther-.

SOMO... Sh etd al) Deis eid aie Ree 361
Dr. J. Kerr on Electro-optic Action of a Charged Franklin’s
PIR Ss Pe Se vee eer 363°

Prof. O. Lodge on the Seat of the Electromotive Forces in a
Voltaic Cell. Theories of Wiedemann and of Helmholtz... 372
On a Differential Resistance-Thermometer, by T. C. Men-
denhall ..00. 05605 waa ie GE Oh ae 384
On the Determination ef the Coefficients of Tuboeual Friction,
by Walter Konig .....:....08 04.01. i 387
Note on the Transmission of Light by Wire-Gauze Screens,

by 5. P. Langley... 050 seth ass. ssn een 387

NUMBER CXXVI.—NOVEMBER.

Prof. De Volson Wood on the Luminiferous Atther......... 389

Prof. F. Himstedt on a Determination of the Ohm ........ 417

Mr. J. Larmor on the Molecular Theory of Galvanic Polariza- ;
PINS ives reeke Le he a Dh shew hale Bibs ese 422

Assist.-Prof. J. W. Clark on the Influence of Pressure on
certain cases of Electrical Conduction and Decomposition. 4385
Prof. H. Hennessy on the Winters of Great Britain and

Ireland, as influenced by the Gulf-Stream .............. 439
Prof. H. Hennessy on the Comparative Temperature of the

Northern and Southern Hemispheres of the Earth ...... 442
Mr. J.J. Hood on Retardation of Chemical Change ........ 444
Prof. C. Michie Smith on Atmospheric Electricity.......... 456

Notices respecting New Books :—
Prof. C. Winkler’s Handbook of Technical Gas-analysis. 461
On the Separation of Liquid Atmospheric Air into two diffe-

rent Liquids, by 8. Wroblewski .........:¢<..3aeern 463
On the Source of the Hydrogen occluded by Zine Dust, by

Greville Williams .. 2.52... eee ea ee ho 464
On two new Types of Condensing Hygrometers, by M. G.

ee a ECM 468

NUMBER CXXVII.—DECEMBER. .

Prof. O. Reynolds on the Dilatancy of Media composed of

Rigid,.Particles in.Contact.. (Plate X.) ).. ..:./15 ae gee 469
Mr. G. Gladstone on the Refraction of Fluorine ... ...... 481

Mr. T. Gray on Measurements of the Intensity of the Hori-
zontal Component of the Earth’s Magnetic Field made in
the Physical Laboratory of the Unive ersity of Glasgow.
hg DS he ee core Meme 484

CONTENTS OF VOL. XX.—FIFTH SERIES. vil

Page
Prof. T. Carnelley on the Periodic Law, as illustrated by cer-
tain Physical Properties of Organic Compounds.——Part I.
The Melting- and Boiling-points of the Halogen and Alkyl

Compounds of the Hydrocarbon Radicals .-............ 497
Drs. W. Ramsay and 8. Young on some Thermodynamical
(8 TLE LSS 2 EON Cre ore en eee 515

Mr. H. Wilde on the Velocity with which Air rushes into a
Vacuum, and on some Phenomena attending the Discharge of
Atmospheres of Higher into Atmospheres of Lower Density 531

Notices respecting New Books :—

Prof. O. J. Lodge’s Elementary Mechanics, including

Peyerasiiiec and !ReuMAlics. 2... ss ke ge et ee 545

Mert. Cates Vratici Tables tor 1886 .. 2.2.2... 54. 547
Visible Representation of the Focus of Ultra-red Rays by Phos-

femerecnee by Hy LOUMNC! se lee eh ee a ee ee wk 547

On the Tones produced in a Plate or a Column by frequent
Discharges of an Electrical Machine, by M. E. Semmola .. 548

ERRATUM.

Page 163, line 4 from bottom, for '2478 read :2784.

PLATES.

I. Illustrative of Dr. J. D. Dana’s Paper on the Origin of Coral
Reefs and Islands.
IL, Iil., & IV. Lllustrative of Prof. Oliver Lodge’s Paper on the
Stream-Lines of Moving Vortex-Rings.
V. Illustrative of Dr. J. A. Fleming’s Paper on the Use of Daniell’s
Cell as a Standard of Electromotive Force.
VI. & VII. Illustrative of Dr. J. A. Fleming’s Paper on the Jistribution
of Electric Currents in Networks of Conductors.
VIII. TWlustrative of Mr. J. Buchanan’s Paper on the Thermoelectric
Position of Carbon. :
IX. Illustrative of Mr. F. Stenger’s Paper on the Electric Conductivity ©
of Gases.
X, Illustrative of Prof. O. Reynolds’s Paper on the Dilatancy of Media
composed of Rigid Particles in Contact.

XI. Illustrative of Mr. T.Gray’s Paper on Measurements of the Intensity
of the Horizontal Component of the Earth’s Magnetic Field.

THE
LONDON, EDINBURGH, ann DUBLIN

PHILOSOPHICAL MAGAZINE
| AND

JOURNAL OF SCIENCE.

[FIFTH SERIES.]

JULY 1885.

I. A Determination of the B.A. Unit in Terms of the Mecha-
nical Equivalent of Heat. By LAwRENcE B. FLETCHER,
2.J).*

. experimental work of the following investigation was

completed in 1881, and forms the subject of a thesis
submitted to the Johns Hopkins University in that year. In
the present paper a more accurate method of calculating the
currents from the deflection-curves is used, and some of the
other calculations have been revised. The results of the two
papers are substantially the same.

The experiment consisted of simultaneous thermal and
electrical measurement of the energy expended by a current
in a coil of wire immersed in a calorimeter. The result de-
pends upon the values of the mechanical equivalent and the
unit of resistance, and gives a determination of either in terms
of an assumed value of the other.

The old determinations of Quintus Icilius and Lens have
no value, as the resistance is uncertain, as pointed out by
Rowland and H. FI. Weber.

Joule}, in 1867, made a determination of the mechanical
equivalent by this method, assuming the B.A. unit as deter-
mined by the Committee in 1863-4 to be equal to 10° C.G.S.
units. The value of the mechanical equivalent thus obtained
is more than 1 per cent. greater than Joule’s water-friction

* From an advance proof from Silliman’s American Journal, commu-
nicated by the Author.

+t Report of British Association Committee on electrical standards, 1873.

t Phil. Mag. S. 5. vol. v. pp. 30, 127, 189. .

Phil. Mag. 8.5. Vol. 20. No. 122. July 1885. B

2 Dr. Fletcher on the Determination of the B.A. Unit

value. H. F. Webert, in 1878, used a similar method, em-
ploying the Siemens unit, the value of which he also measured
in C.G.S. units. Weber’s value of the mechanical equivalent
is about one part in two hundred greater than Joule’s water-
friction value, and one part in four hundred greater than
Rowland’s water-friction value.

In both Joule’s and Weber’s experiments a possible source
of error seems to have been ignored. The wire was assumed
to be at the temperature of the water in which it was immersed,
and its resistance was calculated on this assumption. It is
evident, however, that the wire was hotter than the water,
inasmuch as it was giving heat to the water. The error due
to this cause is of uncertain amount. If corrected for this
error, the values of the equivalent would be increased and
their excess over the water-friction values would become
greater than before. To avoid this source of error, the research
described below was planned. The suggestion and general
plan of the research I owe to Professor Rowland.

The theory of the method is as follows:—A current ¢, flow-
ing through a wire of resistance R, for a time t, generates an

2
amount of heat represented by h= beak where J is the me-

chanical equivalent of heat. The wire being immersed in a
calorimeter and put in a circuit with a galvanometer, h, c, and
t can be measured. Then if R is measured in B.A. units, the
experiment will give a relation between the value of that unit
and the mechanical equivalent. In this research R was
measured during the actual experiment by connecting its
terminals with those of a large resistance R/ and measuring
the current c’, which flowed through the latter. With this
/

rae ie cc’ R’t
arrangement cR=c'R’, or R=—R’. Hence J= in
c iia:

which R does not appear, and the uncertainty attaching to its
temperature has no effect.

The calorimeter was a cylindrical cup of sheet copper hold-
ing about 800 cubic centim. On the bottom of the cup lay a
sheet-copper frame which supported three vertical glass rods.
Around these the wire R was coiled, forming a helix. The
ends of the wire were soldered to stout copper wires, which
inswated by short vulcanite tubes, passed through the wall oF
the calorimeter and turned down so that they could be placed
in mercury-cups. The cover of the calorimeter rested in con-
tact with the water to secure uniformity of temperature. The
cover had an expansion-tube and a smaller central tube. which
formed one bearing for the stirring-apparatus, another bearing
being given by a brass socket on the bottom of the calori-

in Terms of the Mechanical Equivalent of Heat. 3

meter. The stirrer consisted of a spiral blade of sheet copper
supported on a brass frame, the upper part of which was
tubular, and passed through the central tube of the cover.
The stirrer was kept in motion during the experiment by a
silk thread, which passed over a vulcanite wheel at the top of
the stirrer and ran to a driving-clock. The stirrer formed
the escapement of the clock, which ran very uniformly with
this arrangement. I estimated the heat generated by the
stirrer as two thirds of the whole work of the weights. This
is about one thousandth part of the heat generated by the
current, and only a rough determination of the correction is
needed. The thermometer passed through the tubular upper
part of the stirrer, and was clamped to a shelf above in such
a manner that its bulb was in the centre of the calorimeter
and surrounded by the stirring-blade, which, in turn, was
surrounded by the wire which carried the current. The wire
was composed of an alloy of platinum and iridium, and was
varnished to prevent conduction to the water. Its resistance
was about 1°8 ohm. The calorimeter was supported on legs
of vulcanite within a copper vessel with double walls, the
space between which was filled with water. This water-jacket
was provided with a hollow cover, also filled with water, and
its inner surface and the outer surface of the calorimeter were
nickel-plated and polished. Thus the calorimeter was nearly
surrounded by an envelope of fairly constant temperature,
the thermometer, stirrer-thread, and connecting wires passing
through openings in the jacket. 7

From the mercury-cups, in which the electrodes of the
calorimeter dipped, the wires of the main circuit ran to the
battery and galvanometer. These wires were 2°5 millim. in
dizmeter, cotton-covered, carefully paraffined, and twisted
together to eliminate direct action on the needle. The battery
consisted usually of 24 one-gallon bichromate cells, arranged
4 in series and 6 abreast, and gave a very steady current. In
one experiment only 20 cells were used, 4 in series and 5
abreast. The galvanometer-coil for the main current was a
single turn of stout wire laid in a groove on a wooden circle
of about 80 centim. diameter. A sine-galvanometer was so
placed that its needle was in the axis of the single-wire coil
and about 1 centim. distant from its plane. This excentricity
was rendered necessary by the length of the suspending fibre.
The coil of the sine galvanometer was connected with the
calorimeter-eJectrodes by a second circuit, in which a resist-
ance-coil of 30,000 ohms was included. The wires of this
circuit were kept apart, as the current was too small to exert
an appreciable direct action, and as great irregularity in some

4 Dr. Fletcher on the Determination of the B.A. Unit

preliminary experiments in which the wires were twisted
together was finally traced to leakage, although the wires had
a double covering of silk. Both circuits were provided with
commutators. The sine-galvanometer had a horizontal bar
parallel to the axis of the coil. To one end of this was
attached a telescope, beneath which was a short scale which
was seen by reflection in the mirror of the needle, and allowed
avery accurate setting to be made without bringing the needle
to rest. The needle consisted of two thin strips of steel 1:2
centim. in length, separated by a piece of wood ‘6 centim. in
thickness. The circle of the galvanometer was graduated to
half-degrees, and read by verniers to one minute,

The needle was acted upon by both currents simultaneously,
and by means of the commutators the actions were caused to
be in the same and in opposite directions alternately. The
current through the sine-galvanometer is c’ in the formula
5 oe cc’ R’t

Oe
is c+c’, and was assumed equal to ¢, as c’ was less than :00007e.
Let G denote the constant of the fixed coil, G/ that of the
sine-galyanometer, H the horizontal magnetic force, 6 and 0
the deflections when the actions are in the same and in opposite
directions respectively. Then

Gecos@ +G/c'=Hsin 8,
Ge cos ’—G'c'=H sin 6.

The current through the coil on the wooden circle

Hence

tai EH : : ,__ H sin 3(0—6’)

c= Gg tan3(+0), « ~ @ cosk(O+0)
Let / denote the length of the wire in the fixed coil, and } the
distance of the needle from its plane. Then

Aq?

GS 7
(1+ sake

F)

Hence the equation for J becomes

J= EZe ‘aps ]
ig H% tan (6+ 0’) sin 1(0—@/)
Amr? (G’ h cos 3(0 + 0’)
I shall discuss in order the quantities contained in this
expression.
R’, the resistance of the secondary circuit, is the sum of the
resistances of the 80,000-ohm coil, the sine-galvanometer, and

m Terms of the Mechanical Equivalent of Heat. 5

connecting wires. The whole was measured by connecting
the terminals of the circuit with a Jenkin-bridge and com-
paring with other coils, using a high-resistance Thomson-
galvanometer. The provisional standard was a 10-ohm coil,
A, made by Warden, Muirhead, and Clarke. From this coil
the resistance of a 100-ohm standard, B, was obtained by means
of a comparator, C, of ten coils, each nearly equal to A, each
coil of C being compared with A, and C in series then com-
pared with B. Then A, B, and C were arranged to form a
bridge with D, a 1000-ohm standard, whose resistance was
thus fixed. HH and F, two 1000-ohm coils of a resistance-box,
were then compared with D. Finally, a bridge was formed
with A, B, D+H+F, and R’, the secondary circuit, giving
R’ in terms of A. LE lliott’s coils were used in making the
adjustments, which were always very small, and the tempera-
tures were carefully observed. The result is

f= S0012-4 at 19°33 C,

PR’ consisted principally of the 30,000-ohm coil, and the varia-
tion of this only need be considered. Its temperature varied
from 19° to 24° when in use, the mean temperature being
22°°3 C. At this temperature R’=30052, which value was
used throughout. | |

The length of the wire in the fixed coil was determined by
measuring with a steel tape the distance between two threads
fastened on the wire before it was placed on the circle. When
the wire was in position, the interval of a few centimetres
between the threads was measured. The tape had been com-
pared with standards. Care was taken to avoid difference of
tension in the two positions of the wire. ‘The result is

l=264°49 centim.

The quantity b (the excentricity of the needle) was esti-
mated by holding the tape horizontally over the top of the
circle, and reading the positions of the centre of the wire and
the galvanometer-fibre. For most experiments b=1°2 centim.
It was frequently remeasured, and a correction applied when
it varied. The method of measurement is not very accurate ;
but an error of 10 per cent. in b, which could hardly occur,
would only involve an error of 1 part in 3000 in J.

G’ had been determined by Professor Rowland* by mea-
surement during the construction of the coil, and also b
comparison with another coil. The values are 1832:24 by

* Silliman’s American Journal, 1878.

6 Dr..Fletcher on the Determination of the B.A. Unit

measurement and 1833°67 by comparison. The mean, giving
the second value twice the weight of the first, is 1833°19.

Hence the constant term = 10996 +10".

G’ has recently been remeasured and found to be 1832°53.
My final result is corrected to this value.

H was measured in the following manner :—The circle
bearing the fixed coil carried four smaller wires, which could
be connected with the battery and an electrodynamometer of
the form described in Maxwell’s treatise. These four wires
with the needle formed a tangent-galvanometer, the other
coils being open. Hight pairs of simultaneous readings of
the galvanometer and electrodynamometer were taken, com-
prising all possible combinations of signs of the currents in
the galvanometer and the two electrodynamometer-coils. I
am greatly indebted to Professor 8. H. Freeman, then Fellow
of the University, for assistance in these readings.

The expression for H is

27/2
nC V1 62) pee

vs oe Ann v1 ’ l? J} Wsinae .
e Ttangd ~

where C is a function of the dimensions of the electrodyna-
mometer-coils, | the moment of inertia of the suspended coil,
nm the number of turns of wire in the galvanometer, l’ the
mean length of one turn, 0’ the mean distance of their planes
from the needle, T the time of vibration of the small coil,
a and ¢ the mean deflections of electrodynamometer and gal-
vanometer. 4

C was known from measurements during the construction —
of the instrument, and I had been determined by observing
times of vibration with and without the addition to the sus-
pended coil of bars of known moment of inertia. These values
of C and I had been verified by Dr. E. H. Hall and myself, in
connection with a previous research, by comparing the values
of H obtained by this method and by the magnetic method,
the arrangement of the experiment being such as to make the
two results obtain for the same point and time. The value of
Ov T is 018567.

The measurements of l’ and b’ were made in the same
manner as those of / and b.

The results are l/=263°91 centim., b’=2-07centim., for most
experiments. A correction was applied when 0’ varied. Hence

2 7 eehre
dor s v +(1- ome ) ="11069.

Kach of the angles, « and q, is the mean of eight readings

in Terms of the Mechanical Equivalent of Heat. 7

taken to 1’. The former was about 13°, the latter 6°. Twas
obtained by observing ten transits with a seconds-clock, allow-
ing the coil to vibrate for several minutes, and then taking ten
more transits. The difference between the mean times of the
two series divided by the number of vibrations gives T very
exactly. The difference between the values before and after
the experiment never exceeded 1 part in 8000. The mean
value is about 2°42 seconds. H was determined before and
after the main experiment. |

The quantities in the formula for J remaining to be dis-
cussed aret, h, and the deflections. To treat these intelligibly,
I proceed to describe the method of experiment exactly.

First, a determination of H was made. The calorimeter was
then weighed, filled with distilled water at a temperature
usually 2° or 3° below that of the air, carefully wiped with a
towel to remove moisture, again weighed and placed in the
water-jacket. Its amalgamated electrodes were placed in the
mercury-cups with the terminals of the two circuits, the main
circuit being broken at the commutator. The water-jacket was
kept permanently filled, and stood in a room of fairly constant
temperature, so that its temperature changed little during the
experiment. ‘The thermometer was placed in position and the
stirrer started. During a few minutes readings were taken of
the thermometer and of three auxiliary thermometers, giving
the temperatures of the jacket, the 30,000-ohm coil, and the
air near the stem of the principal thermometer, the time of
each reading being noted by a seconds-clock. The circuit was
then closed and a galvanometer-reading taken, one of the com-
mutators was reversed and another reading taken, the time of
each reading being noted. The time of passage of the mercury
of the thermometer over several successive scale-divisions was
then taken, also readings of the other thermometers. Two more
commutator-reversals and galvanometer-readings followed,
then another set of thermometer-readings; and this alternation
was continued for about 40 minutes, during which time the
thermometer rose about 12° C. Usually sixteen galvano-
meter-readings were taken and seven groups of thermumeter-
readings comprising 35 or 40 readings of the principal ther-
mometer. Then the circuit was broken and the calorimeter
allowed to cool for two or three hours, during which time
groups of readings were taken as before, the stirrer being kept
in motion. While this radiation-experiment was in progress
another determination of H was made. ['inally, the thermo-
meter was removed and the calorimeter taken out and weighed.

The mean of each group of thermometer-readings, corrected
for stem error, gives very exactly the temperature of the ther-

8 Dr. Fletcher on the Determination of the B.A. Unat

mometer for the mean time of that group. The difference
between any two of these mean temperatures, corrected for
radiation, gives by multiplication into the capacity of the
calorimeter and contents the heat generated in the interval.
Hence any two groups give a determination of J when com-
bined with the proper values of @ and @. :

I have combined groups taken 18 to 25 minutes apart, the
rise of temperature being 6° to 8°.

In this calculation the differences of temperature of coil,

water, and thermometer are assumed to be constant for this
interval. The water is cooler than the coil, and the thermo-
meter cooler than the water. Both differences depend upon
the rate of generation of heat, and may be put approximately
proportional to the square of the current. The rise of the
thermometer after breaking the circuit is due to these dif-
ferences, and was found to be less than 0°05. The variation
of this quantity during the interval in question would be
about 3 per cent., as the current changed 1°5 per cent. Hence
the variation is 0°:°0015; and as the rise of the thermometer is
6° or 8°, the error is negligible.
- Two thermometers were used, designated as Baudin 6165,
and Baudin 7320. ‘The former is graduated in millimetres, of
which about 12 equal 1° C. It had been used by Professor
Rowland in his determination of the mechanical equivalent,
and compared several times with the air-thermometer. Baudin
7320 is graduated to 0°-1 C., one degree occupying about a
centimetre. It had been compared with standard thermo-
meters, its errors plotted, and the error for each degree obtained
from the curve. The following tables give the reduction to
the absolute scale :—

BavupiIn 6165.

Temperature Temperature
oe on absolute reer on s claie
‘ Iscale from O° OC. ’ Iseale from O° C,
fe) fe)

35 0:0 320 24:54:7
50 1°368 330 25°365
100 5°839 340 26:174
150 10°183 350 26981
200 J4:450 360 27°782
250 18°709 370 28°584
260 19°557 380 29-376
270 20°401 390 30°170
280 21°242 400 30°965
290 22:076 410 31°786
300 22:907 420 32°581

310 23°731

in Terms of the Mechanical Equivalent of Heat. 9

| BaupDIn 7320.

Temperature Temperature
Reading.| on absolute || Reading.) on absolute
scale from 0° C. scale from 0° C.
(2) oO Oo 2)
0 0-122 23 23-108
+) 5-092 24 24:122
10 10°110 25 25°137
15 15°090 26 26°152
16 16-093 27 27°166
fain 4 17-094 28 28°179
18 18-091 29 29°192
19 19-086 30 30°205
20 20°081 31 31:217
21 21-085 32 32230
22 22095

The table for 6165 is condensed from Professor Rowland’s
paper* on the mechanical equivalent. Change in the zero
point has no effect on the differences of temperature used, but
the zero points were determined occasionally in order to get
the mean absolute temperature.

The correction for radiation was made in the following
manner :—The groups of thermometer-readings taken after
breaking the circuit were reduced to mean readings at mean
times. Any two of these mean readings gave the radiation
for the intervening time. If ¢/ and?” are the temperatures
at the beginning and end of an interval of T minutes, and +
is the mean temperature of the jacket during the interval, then

{— i" =el| £( + 0") —T],
where c is the coefficient of radiation. In the calculation of
ce, stem-corrections were applied and a correction made for the

heat generated by the stirrer. Hence in the main experiment
the temperature-correction for an interval T’ is
A=cl’[3(s' +s")—7'|+K,

where s‘ and s” are the observed temperatures corrected for
stem-error, 7’ is the mean temperature of the jacket, and Kx is
the stirrer-correction. The sum of the corrections A from the
beginning of the experiment added to the stem-corrected ob-
served temperature at any point, gives the temperature which
would have been reached in the absence of radiation. The
difference between any two of these theoretical temperatures
multiplied by the heat-capacity, gives the heat generated in
the interval.

* Proceedings of American Academy of Arts and Sciences, 1880.

10 + Dr. Fletcher on the Deterinination of the B.A. Unit

The coefficients of radiation were found to decrease with
decreasing difference of temperature between calorimeter and
jacket. When this decrease was regular the corresponding
value of ¢ was used for each small interval of the main ex-
periment. When the decrease was small and irregular the
mean value of ¢ for that day was used throughout. In the
revision of the calculations, stem- and stirrer-corrections were
neglected in the calculation of both ¢ and A, it being obvious
that, both being small and quite regular, they are eliminated
in this way; and the value of c¢ corresponding to the difference |
between calorimeter and jacket for each small interval of the
main experiment was used in all cases. The mean results of
the two methods differ about 1 part in 1000, and the figures
in the table of results below are the means of both calcula- |
tions. The mean values of ¢ for the different experiments |
vary between 0°:0035 and 0°-0046, the general mean being

0°:0040.

The mean radiation-correction is about 5 per cent., and is
the most important source of variable error in the experiment,
as the temperature-differences are small, and errors of reading
have a large effect. But a 10-per-cent. error in the radiation
would only involve an error of 1 part in 200 in J; and as the
errors are irregular they are, in a great measure, eliminated
from the final result. |

The calorimeter was composed of 246 gr. copper, 45 gr.
brass, and 6 gr. solder. The specific heat of a mixture of
these proportions was measured with Regnault’s apparatus.
Six determinations gave the value ‘(0898 +°0005 for the mean
specific heat between 24° and 100°. Reduced by Bede’s law
for copper to the mean temperature of my experiments, it
becomes ‘0877. The capacities of the coil and glass rods
were calculated from published tables. ‘The whole capacity
is as follows :— |

Calorimeter . . . . 3802:1x:0877=26°49
Qo... Aw 4. 88D R082 eee
GHass roue :' 5.0 8's. 4% 9Ox “177= 1:59
Thermometer estimated at . . . . . 1°25

ce ee

Total capacity) . °° 4 = 60e

The values of the deflections were obtained by a graphical
method. The galvanometer-readings fell into 4 groups, lying
about 26° and 8° 45’ on each side of the zero-point. The
readings of each group were plotted separately as functions of
the time. From each curve the theoretical mean readings for
each interval between two temperatures used in the calcula-

in Terms of the Mechanical Equivalent of Heat. 11

tion of J were obtained by measuring a large number of
equidistant ordinates, calculating the area of the curve, and
dividing by the base-line. If a, 6,c, and d are the mean
readings thus obtained, 20=a—b and 26’=c—d as the galva-
nometer was graduated from 0° to 360°. Thus the zero-
point was not used, though observed before and after each
experiment.

Below are the results of one experiment in detail. Hach
value of J is calculated from the two temperatures found in
the same horizontal line.

Series of December 9th.

or.
Weight of calorimeter and water before experiment . 1157°2
Weight of calorimeter and water after experiment . 1157°0

11D 7-1
Sue calorimeter... 1.) 6 we eo BART

| 813°4
Seemeeer or calorimeter’. 60 ep Beare ce eo 804

; 843°8
Capacity reduced to weight in vacuo. . . . . . 844'8

Horizontal maguetic force before experiment . . . ‘1960
Horizontal magnetic force after experiment . . . ‘1963

es

1S CORE SS OMRAGISG a MG Ty Bla ae le Big WS 1 24

Temperature of jacket. . . . 21:5 to 21°6
Temperature of air near stem . 22°8 to 23:4
Temperature of 30,000-ohm coil. 21'5 to 22°3
Thermometer in calorimeter. . Baudin 7320

| . | .
. Reading . Reading
Time. 7390. Stem. =A. Time. 7390. Stem. =A.

‘
840 | 19-45 | —-010 0 25 14 | 25-40 | +-009 | +-056
fae. 1 9145 | —-007 | —-018 || 31. 9 | 97-40. | 021 | 167
90 38 | 23:30 | -000 | +005 || 35 58 | 29898 | -099 | -299
oie | 95490 | +009 | -056 || 49 54. | 31-15 | ~-039 | Hep

ee eee re a

20 29! J

51 841 7 35° || 41810000
50 57°5 | 7 28-5 || 41640000
50 48°5 | 7 24-4 | 41640000
50 40-4 | 7 21-4 | 41840000

12. Dr. Fletcher on the Determination of the B.A. Unit

Time. | Galvanometer. Time. Galvanometer.
m..") 43: ees, ma,” 1s ieee
5 50 237 14 28 «+5 237 «5
f0 259 12 29 15 258 46
10 20 229 34 32 45 229 42
11 30 207 42 34 0 208 O
15 40 237 11 —689: «COO ie aaa
Ty ab 258 56 4 Q 258 40
22 O 229 40 44 15 229 44.
23 25 207 58 45 25 208 4

Following is the general Table of results:—

ae th Nov. 24.| Dec. 9. | Dec.-14. | Dec. 20. | Dec. 22. | dan: See anne

Thermometer ...| 7320 7320 6165 6165

Temp. of water...| 24°°3 25°°4 2420 26°°6 26°°3 aoe 26°:

es

—

Pomp 01 3! °.....- 23°°0 22°"4 23°°8 21 23°°4 19° 0 20°:

4229 4164 4217 4206 4232 4205 4174515

aga Pea: |e 4204 | 4218 | 4193
J BBG Aha eS ANE OS 4182
T0000 4217
4904
4225
4925
4994

(ees a ee ee ee | ee | ee ss |

Meat is... .«. 4230 | 4173 | 4206 | 4199 | 4216.| 4196 | 4200 |)

The greater number of results on Nov. 24 is due to the
fact that single thermometer-readings were taken instead of
groups. ‘The two experiments with 7320 show the greatest
variation from the mean; but the mean of these two agrees
closely with the general mean. Hxperiments made on Noy.
29 and Dec. 6 were rejected on account of leakage of 3 and
6 grams respectively. The results, however, are 4220 and
4227, falling within the limits of the series. The duration of
the experiment was less than one fifth of the interval between
the two weighings of the calorimeter; and probably the loss
during the experiment was the same fraction of the whole loss.
Furthermore, the leakage during the radiation-experiment
would affect the radiation-coefficient in such a manner as to
approximately compensate for the effect of leakage during
the main experiment. or these reasons I have retained the
experiments of Jan. 26 and Feb. 16, which showed a leakage
of 1 gram and 1°5 gram respectively, but have given the results

6165 | 6165 | 6165.

4240 4181 4180 4207 4198 4194 4207 |)
4248 4164 4222 4195 4216 4192 4219 |)

es

!
ql
\

i

1
“|
uy

|
wh
Ml
it
| |

ill
i

2 OOP nage Cm a

POPE ieee WO Qe ee I a , diieat
~ rte

Der 3a
i ee
a re e

m Terms of the Mechanical Equivalent of Heat. 13

half the weight of the other. The remaining experiments are
satisfactory in this respect, the loss being a few tenths of a
gram, due principally to the removal of the thermometer.

The result of the experiment is J =42,039,000 O, where O
is the value of one tenth of the 10-ohm coil in earth-quadrants
per second. Reduced to the new value for the constant of the
sine-galvanometer, it becomes J = 42,055,000 O.

I have also calculated the experiment from the formula
<i “3 where R is the resistance of the calorimeter-coil as
measured in the ordinary manner, corrected to the mean tem-
perature of the water, and further corrected for superheating.
I estimated the superheating from observations of the main
and derived currents when the strength of the former was

/}Q/
varied. The expression se should give the true resistance
of the coil at any instant. When the currents are smaller, the
“fl P4
superheating is less, and the comparison of the value of a

for a zero-current, obtained by graphical extrapolation, with
its value for the full current, should give the superheating.
The method is not very accurate, as the observations with
the smaller currents are uncertain. I find the increase of
resistance due to superheating to be about 1 part in 700, cor-
responding to a difference of temperature of 2° C. When
this correction is applied, the second method of calculation
gives J =42,156,000 O.

This result depends upon the square of the main current ;

-and as the temperature of the coil changed 6° or 8° during the

experiment, its mean resistance is somewhat uncertain. Hence
this result has not the weight of the former.

The discovery of this discrepancy has greatly delayed the
publication of this paper. It may be due to conduction
to the water, which was guarded against by varnishing the
wire and using distilled water, but was not proved to exist.
For let E be the difference of potential of the ends of the
coil, e the E.M.F. of polarization, R and ,r the resistances
of coil and water respectively. Then the current in the coil is
— and the current in the water is =—— .

The energy converted into heat is
i? R ae
2 ee ak Euler gre
CR+er= z[14+=(1 ao cell

In the first method of calculation above the energy is com-

14 On the Determination of the B.A. Unit.

puted as
i? R €
WC+e)= y[1+ (1 uli 12
In the second method it is computed as
2

(C+c)’R= ws E -- ““(1- nD + smaller terms.
Hfswas over 6 volts, ¢ is about 1:5 volts. Hence the second
result is larger than the first, which agrees with the facts,
and the error of the first is less than one fourth of the difter-
ence between the two. The discussion shows that the first
method of calculation is to be preferred; and I therefore take
J =42,055,000 O as the result.

Since the completion of my experiments, a 10-ohm Elliott
standard in the possession of the University has been com-
pared with the Cambridge standards and found correct at
20°°9 C. My standard has been compared with this, with the
following result:—

W.M. & C.’s coil ak
Tiaikaies Ee 00168 in 1878,

Mi » ~=1:00170,, 1882,
ri , =1:00173 ,, 1883.

In these comparisons the Elliott coil was taken at 16°3 C.,
as marked. Also we have

Elliott coil at 20°9 —1-0014.
Elliott coil at 16°°3

Hence O= ++! =1-0003 B.A. units and J=42,068,000 x

value of B.A. unit in earth-quadrants per second.

Rowland” has discussed Joule’s values, and reduced them
to the air-thermometer and the latitude of Baltimore. The
mean of the best results from the friction of water, in 1850 and
1878, thus becomes 426°55 kilogram-meters or 41,805,000
C.G.S8. at 14°:1 C. This, according to Rowland’s results for the
temperature-variation, corresponds to 41,608,000 at 26°, the
mean temperature of my experiments. Rowland’s value at 26°

is 41,720,000. Combining the mean of these, 41,664,000,

with my result, I find 1 B.A. unit= minder ales = 9904 earth-
42,068,000

quadrant per second.

This research cannot compare in weight with the elaborate
determinations of the ohm by direct methods which have been

* Proceedings of American Academy of Arts and Sciences, 1880.

Mr. J. Edmondson on Calculating Machines. 15

made in England and this country since the conclusion of
my experiments ; but as few results by this method are at hand,
I publish it as a slight contribution to the history of this vexed
subject.

Marlborough, N. Y., April 15, 1885.

Il. Summary of Lecture on Calculating Machines, delivered

before the Physical Society of London, March 28, 1885, by
JosEPH Hpmonpson, of Halifax*.

ALCULATING machines are of two classes, the automatic
and the semi-automatic. The former were invented by

Mr. Charles Babbage between 1820 and 1834, and were de-
signed mainly for the computation of tables. When the primary
factors of a table are placed on such a machine, it will calcu-
late the table and impress it line by line upon a stereo-mould

with great celerity and absolute correctness. Mr. Babbage’s

machine, so far as it was completed, is in the South Kensing-
ton Museum; and another machine by M. Scheutz (for the
principles of which he confessed his indebtedness to Mr. Bab-
bage) is in the office of the Registrar General at Somerset
House. These machines proceed by the method of Differences,
and are therefore termed ‘ Difference-Hingines.”” Their work
is limited to addition, but they indirectly perform subtraction
by adding the complement of the subtrahend.

If it be required to construct a table (of cubes for instance),
the law of its growth must be found by reckoning mentally
the first few terms of the table. Placing them in a line, the
difference between each term and its successor is placed under

the former. This process is repeated with the differences

until at last a line is reached in which all the differences are
cyphers. Thus :—

Meee ee ea ee ek Of B25" S16

fr onerence (“bs Te 19 “370-61 91

2nd _ Sete bes Sas a)

pe re ee

Ath bs Br yO

The machine may now be set to work by placing the first
column of figures in their respective places upon it. During
the first half of the first revolution of the machine, the wheels
containing the terms A and C have the terms B and D added
on to them and become 8 and 18 respectively. During the
second half-revolution, B and D have the terms C and E added
on to them, becoming 19 and 6 respectively, and the second
column is complete. Hach succeeding revolution of the

* Communicated by the Physica Society.

16 Mr. J. Edmondson on Calculating Machines.

machine similarly produces each succeeding column, the upper
line being the term required for the table.

The apparatus consists of identical parts, equal in number
to the digits required for the last column of the tables it is to
compute. It is therefore very complex and very costly, and

will not serve the purpose of computers in general, who must

have recourse to the semi-automatic class of instruments.
These are portable, of moderate cost, allow of very rapid
working, and require no special mathematical skill.

In 1663 Sir Samuel Moreland produced an instrument by
which additions and subtractions could be worked, digit by
digit ; but it took more time than the ordinary mental opera-
tion. It was left to Viscount Mahon (afterwards Hari of
Stanhope) to produce the first really practical instrument.
Besides a machine dated 1780, which was a great advance on

that of Sir Samuel Moreland, though on the same lines, he ©

invented three machines. Those of 1775 and 1777 were on
the table for inspection after the lecture, and abound in beau-
tiful and effective contrivances. The third machine the lecturer
had not seen, and it has never been described. In that of
1775 is found the “ Stepped Reckoner,’’ the basis of the only
instruments that have come into extensive use.

The reckoner of the modern machines, patented in 1851 by
M. Thomas de Colmar, consists of a cylinder divided into
10 sections, on which there are respectively U0, 1, 2, 3, 4,5, 6,
7, 8, and 9 teeth. ‘The teeth of one section being coincident
with an equal number of those of the next section, the whole
presents a stepped appearance. Hach revolution of this
reckoner moves a pinion of 10 teeth, in gear with one of its
sections, as many teeth as there are upon that section. The
motion of the pinion is communicated to a dial with the digits
0 to 9 in orderly succession upon it. Thus, if the figure seen
through an aperture in the covering of the dial were 0, and
the pinion were in gear with the section having 3 teeth, the
first revolution of the reckoner would move the dial to 3; the
second to 6; the third to 9; this result being the multiple of
8 (teeth on the reckoner) by 3 (revolutions). A series of,
say, 8 reckoners, pinions and dials, each pinion being set to
the section having 3 teeth, would give in 3 revolutions the
product 99,999,999. The next revolution would bring in a
new feature—the carrying of the tens. Here lies the great
difficulty of Calculating Machines ; but the difficulty has been
overcome, though not without leaving room for improvement.
The 1 that will have to be carried from each dial to that im-
mediately on its left cannot be added while the latter is being
operated upon by its reckoner. ‘The machine must therefore

Ce ee

4 a
>
e

|
|
)

4

Mr. J. Edmondson on Calculating Machines. 17

make an independent note of it until it can be dealt with.
Hach dial has a tooth (the primary carrying-tooth) which
makes this note at the moment that the dial passes from 9 to
0, by pushing along the axis of the reckoner belonging to the
next higher digit a sliding-piece which revolves with the
reckoner and carries a single tooth (the secondary carrying-
tooth). Above this sliding-piece, and upon the same axis as
the pinion in gear with the reckoner, is a second pinion of
10 teeth. Before the sliding-piece was moved by the primary
carrying-tooth, the secondary carrying-tooth passed on the
side of this second pinion without affecting it ; but being now
brought into line with it, the pinion will be moved 1 tooth,
and will therefore add 1, as soon as the reckoner has added its
own proper number ; so that the fourth revolution, giving the
product of 33,333,333 by 4, brings the dials to 183,333,332.
When the secondary tooth has done its work, an incline on
its side comes in contact with a projection on the frame of the
machine, which restores it to its original inoperative position.
Each reckoner is set to operate a little later than its neighbour
to the right, so that the carrying on the latter may be com-
pleted before that on the former is commenced. ‘The before-
named blank space on the reckoner is to allow time for the
carrying; and for a “‘dead-point”’ in the machine, at which

_the first-named pinions, which are movable longitudinally on

their axes, may be set to their required sections of the reck-
oners, and other changes may be made in the setting of the
machine. The position of the pinions is shown by indicators
on the face of the machine, each of which can be set to any
required figure.

Hach pinion communicates its motion to its dial by bevel-
gear in the well-known combination of three wheels, which
permit the action to be reversed at will, so that the change
from + to — or from x to + is made instantaneously.

In multiplying by more than a single digit, the dials require
to change their position in relation to the multiplicand, to
imitate the “stepping’”’ of the lines in long multiplication.
They are therefore placed in a slide which can be moved, step
by step, from right to left or vice versa.

Up to a recent date the machine of M. Thomas de Colmar
was the only one in use. A few years ago Mr. George B.
Grant, of Boston, Mass., brought out an instrument somewhat
on the lines of the Stanhope machine of 1777. It is very
compact, is beautifully and substantially made; but as sub-
traction is performed by adding the complement of the
subtrahend, reversing is so tedious as to be fatal to the
general adoption of the machine. Its range is also very

Phil. Mag. 8. 5. Vol. 20. No. 122. July 1885. C

18 Mr. J. Edmondson on Calculating Machines.

limited, and there are other defects inherent in the method of
its construction.

Still more recently Mr. Tate, of London, has introduced a
machine on the lines of M. Thomas, in the internal mechanism
of which there are several improvements. The workmanship,
too, is very superior to that of the French instrument.

Mr. Edmondson (the lecturer) having, in practice, found
certain inconvenient limits to the utility of these machines,
conceived the idea of obviating them by throwing the instru-
ment into the circular form, and thus obtaining an endless
instead of a limited slide. This has produced a machine of
greatly extended powers ; for it can deat with a multiplier or

dividend of any number of figures, and can carry a quotient

to any number of decimal places. But its chief utility does
not consist in dealing with such large numbers. It frequently
happens that quotients produced on a machine require to be
further operated upon ; but as in the straight machines they
are recorded on a special set of dials, which are not in the
general working line, they must be transferred by hand before
they can be dealt with in any other way than as multipliers.
In the circular machine they come up in the working line
ready for further operations. The circular form, in fact,
gives a certain elasticity in the applications of the machine
which could not be described within the limits of the lecture,
and which can only be appreciated by actual use of the in-
strument. Although the special set of dials for recording
multipliers and quotients are dispensed with as non-essential
in the circular machine, they could be easily and conveniently
added, and would still further extend the powers of the
instrument.

The lecturer expressed his indebtedness to the following
gentlemen for the loan of valuable instruments on the table,
which, along with Mr. Tate’s machine and several belonging
to the lecturer himself, attracted much attention after the
lecture :—

General Babbage, Bromley, Kent. A piece of the Differ-
ence-ingine of the late Charles Babbage, Hsq.; Stanhope
Machines of 1775 and 1777 ; Sir Samuel Moreland’s Machine
of 1663,

The Rev. Professor Harley, F.R.S., Huddersfield. Stanhope
Machine of 1780,

Theodore B. Jones, Esq., Harrogate. The American
Machine of George B. Grant, Esq.

I’, H. Bowman, Hsq., D.Se., Halifax. A Machine of Sir
Samuel Moreland’s, constructed on the principle of ‘‘ Napier’s
Bones.”

ee

+42 3H gers bat
III. Ona New Law, analogous to those known under the names,

Law of Avogadro and Law of Dulong and Petit. ByJ.A.

GROSHANST. 7
HEN the boiling-points and densities of substances are

compared with their atomic composition, we find that
there is a law regulating this relationship which is analogous
to the laws mentioned in the title of this paper.

I have called this law, The Law of “ Density Nuinbers.”
The density numbers (Densitdtzahlen) form a new class of
constants, which present themselves as attributes of the ele-
ments, and are of the following nature.

They are whole numbers. Hach element possesses only
one, which may be easily determined; but two or more
elements may possess the same number.

Table I. contains a list of these new constants so far as they
have yet been determined. An asterisk affixed to a number
denotes that this is the number which appears to me probably
the correct one, but that verification is necessary. I venture
to hope that, when the scientific public realize the interesting
nature of this new law, the constants given in the Table will
be rigorously verified, and the gaps at present so numerous
will then be filled up.

TABLE I.

Density-numbers of the Elements already determined.

re Bag BS FE 4
1? See | Be 2 Rese Na: «<4
oe Se ae iw vee Al 4
Si 4
Cl 4
Mg 5 Ca 7 v* 8 Cr 9
As 8 Mn 9
Se 8 Fe 9
Br 9
Ni lt Sr 13 ZTr* 14
Go. ‘il Nb* 13 Sn 14
Cajtii Sb .13 'Pe,,. 14
Zn Ii I 14
Rb* 11
Pd* 16 Mo* 17 Ba 19
Ag 16 Cs* 17
Cd 16
Ta* 23 Hg 26 W* 29

Bi* 23 Ti* 26 Pt* Zs

+ Communicated by the Author. Translated by W. W. J. Nicol, M.A.,
D.Sc., F.R.S.E.
C2

20 M. J. A. Groshans on the

The density-numbers of the elements increase with their
atomic weights, but are not proportional to these. _

The exact nature of these numbers is not yet distinctly
made out; but the following provisional hypothesis, which,
I believe, may ultimately prove to be correct, will help in
forming a conception of their nature.

According to this hypothesis, carbon, hydrogen, and oxygen
are simple bodies ; but the other elements are compounds of
other simple substances, the number of atoms of which is
shown by the density-number of each element.

The law itself may be enunciated in the following simple
form :—

_ “The densities of substances are proportional to the density-
numbers.”’

For example, in the case of two compounds of carbon,
hydrogen, and oxygen (comparable with one another), and
with the formule C, H,O, and C, H,O,, if m and n’ equal the
sum of p+q+r and p’+49'+7’ respectively, and the densities
are 6 and 0’,

rs) n n n
7 =F OF 5 Gipuon constant = &.

~The law of Avogadro leads in the same way to a constant ;
for if the vapour-density of a substance at 0° C. and 760 mil-
lim. equals D, and its molecular weight is a, then

~ = whence D = constant.

This constant is unique, and is true of all bodies which can
be converted into vapour. Such is not the case, however,
with the law of molecular specific heats (a xc). This constant
differs in the various groups of compounds; each group
possesses a single constant which is peculiar to it. The law
of density-numbers resembles in this respect that of Dulong
and Petit. |

This new law is applicable to substances in every state of
aggregation ; but it is necessary that the substances should
be compared under similar conditions, such as :—

(1) as gases at the boiling-point ;

(2) as liquids at the boiling-point ;

(3) as solids (crystals hydrated or anhydrous);

(4) in very dilute solutions.

When dealing with substances which contain only carbon,
hydrogen, and oxygen, I shall continue to use the letter » for
the sum of the numbers of density ; but when other elements
enter into the composition of a body, I shall employ B to
indicate the above sum,

a

> ——————

Law of Density- Numbers. 21

In the following pages I shall endeavour to give a fairly
complete account of the numerous applications of this law,
the study of which has occupied my leisure hours for many
years, ever since a happy chance disclosed to me a special
case, an account of which was published in Poggendorft’s
Annalen, Ixxviii. (1849) p. 112.

When the law is applied to hydrated crystalline salts con-
taining the same number of molecules of water, a hitherto
concealed relationship is discovered.

If B is the sum of the numbers of density of the salt, and
6 its density, it is found that i: =k for all salts with similar
formule and containing the same amount of water of crystal-
lization, as will be seen from Table II., which contains the
data relating to six salts of the general formula MR, . 6H, O.
B in this case is equal to the sum of the density-numbers for
the anhydrous salt plus 18, which is the value of B or n for
H,O. ‘Thus, a reference to Table I. shows that CaCl, 6 H,O
has B=7 + 2(4) + 6(3) =33.

TABLE II.

Salt. B. 0. k.
SrCl,, Gag... 21418—39 1-964 19:86
Beer 8). 15-+18=33 1-654 19-95
wee nl... 19+18—37 1:84 20°11
oo) 184+18=31 1-562 19:85
Ni(NO,), 6aq....| 23+18=41 2-05 20-00
Zn(NO,), 5, +-.| 28-+18=41 2-065 19°85

The data for this table and for others are taken from the
collections made by Landolt and Bornstein, Clarke’s ‘Constants
of Nature,’ and the experiments of Bodeker, Playfair and
Joule, Schréder, Topsoé, and others on the densities of com-
pounds.

The table shows clearly enough that the volumes & are
practically equal, and by means of such groups it is possible

_ to determine the density-numbers for the various elements.

All the density-numbers in Table I. have been determined by
employing the data given by hydrated salts and (where pos-
sible) the boiling-points, and the densities of aqueous solutions
as well as all other available analogous data.

At this point I shall return to Table I., in order to bring
forward some views of a more general nature regarding the
density-numbers.

22 M. J. A. Groshans on the

Triads.

The name of Triads has been given to groups of elements
such as chlorine, bromine, and iodine, or phosphorus, arsenic,
and antimony ; ‘and I shall retain the name, notwithstanding
the fact that each group contains at least four elements. Of
these four elements three are characteristic, while the fourth
may be regarded as a precursor, which, although participating
in the nature of the three principal elements, yet is distinct,
and possesses properties more or less different.

The following table contains twenty elements arranged in
five triads. I have chosen those which appear to be complete,
and have added to the table the atomic weights and the
density-numbers with their differences.

The density-numbers of the three principal elements in each
triad have a constant difference, which is in two of the groups
five units, in the other three six units.

TaBLE III,—Triads.
No. 1. No. 2. No. 3. No. 4. No. 5.

eLlaw.lp.|a. eLlAw.Bla./ EL AW. B. a. eLlaw.lB.|a.l| EL Aw. (Bla.

a SS sg en — | | | ee | |

Aik} Slo 36 2 =e

4| ||Na] 23 | 4| [Mg] 24 | 5
0 1

0 1
Pi oe S | 32 | 2 Cl} 35°5) 4 i) po-i3 Ca} 40 | 7
9) 6 6

5
As| 75/8) ||Se/ 79/8] |/Br| 80 | 9| | Rbi 85 [11] |/Sr| 87-5118

5 6 5 6 6
Sb [120 |13} || Te|125 |14} | I |127 |14) |] Cs /133 |17| |] Bal137 |19

El = nee AW = atomic weight. B = density-number.
= difference of density- numbers,

The differences in the atomic weights of the three principal
elements all lie between 44 and 49:5. The equality of this
difference in the same triad is only approximate; and the
opinion, formerly held, of their exact equality has been ren-
dered untenable by the experiments of Stas and Marignac, as
shown in Table IV.

Tase LY.

. Lt. Hit. TY, | \\«

P31 S 32 Cl 35:5 K, 30 te ean
44 47 4G:

44°5
As 75 Se 79 Br 80 Rb 85'2 Sr 87°5
45 46 47 47°
Sb 120 Te 125 1. ror Os: 132°7 Ba . 13z.

ae te

Law of Density-Numbers — 23

The differences in the atomic weights between the precursors
and the first member of the triad proper varies from 16 to 17,
and the difference in density-numbers from 0 to 2, as in
Table V.

TABLE V.

Atomic Weights.

J. ql. ETT, iV. ¥:

14 (7 IG F 19 Na 23 Mg 24
17 16 16°5 16 16
P31 S 32 Cl 355 Kk 33 Ca 40

Density-numbers.

a tees: 5 | KF =4 Na=4 Mg=5
) 1 0 1 2
P=3 S=2 | Cl=4 K =) Ca =7

Among the elements in Table I. besides those included in
the above triads, are to be found groups of two or more
elements which present the same pronerties as the triads.

Thus there are three pairs with the difference 16.

TaBie VI.
| |
Atomic | Density- Atomic | Density- Atomie | Density-
weight. | number. weight. | number. weight. | number.
ied a i | aeioees Ss

li 7 2 resid 3 CO 12 :

16 2 16 1 16 3
Na 23 + Al 27 4 Si 28 4

The difference of 16 in the atomic weights does not corre-
spond to a constant difference in the density-numbers; but a
difference of 45 to 48 corresponds to 5, asis seen in Table VII.

In the last pair (Ag, Tl) the differences are doubled.

TasuE VII.

Atomic Density- | Atomic Density- Atomie Density-
weight. | number. weight. | number. || weight. | number.

—— ———— --

Zn 65°5 il Ag 108 | 16
44:5 5 | 46: 5 96 10

5
eo 108 16 “4 112 16 | Tl. 204 26

“te

24 M, J. A. Groshans on the

Again, comparing the pairs Hg and Zn, which are diad in
compounds such as Zn(C, H;)., and Pb and Sn, which are

TABLE VIII.

Atomic Density- Atomic Density-
weight. number, weight. number.
Zn 65 11 || sds es
135 Dn 89 15
He 200 26 Pb 207 29

tetrad in corresponding compounds, we find that in the first

pair, where the difference in the atomic weights 1385 =3 x 45,
that in the density-numbers = 15 = 3x5; but that this is
not the case with lead and tin; for while the difference in
atomic weights (=89=2 x45) is double, the difference in
density-numbers is treble. In time, I trust, as the gaps in
Table I. are filled up, it will be possible to extend and
generalize considerations of the above nature.

After this digression I shall return to the application of the
law to hydrated crystals, and shall apply it to the results of
the well-known experiments of Schiff on the density of alums
and isomorphous sulphates.

TABLE IX.
Experiments of Schiff.

Salts. B. aU. ve.
(a) Alums:—
OU NGE (WF Bis beds I 6 Be eae 57 1°722 — 3831
PUB, 595 Grad? 56 1°641 34°]
Oe E62 ae a ne 62 1°845 33°6
(6) Sulphates :—
M8, TEL SO is ids dacs shes 32 1:685 19:0
EO isl cath in ere neaes 38 1:953 oe
hy aaa, iat teat ater ae 38 1-951 19°7
ROT Ey. o> sieeph sss anna 38 1-924 19:8
Rg kh tap chawaaess vals 36 1°884 19:1
(c) Double Sulphates :—
1 OR te 6 45 1-995 22°4
A TE Na ity) | x) fowiv'es 51 2°153 23°7
i ABO) yy! ewaives 51 2°123 24:0
LIOMER AFH Nat hisgguatct hy... 51 2°154 23°7
Wem, 1. 24, jaded 49 2°189 22°4
CdK, (SO,), 3) ween 56 2-438 23-0
OG ae gs fenss, 51 2'137 23°9

Se er eS

.
;
om

Law of Density-Numbers. 25

I have omitted from the table the data for the compounds
in which potassium is replaced by ammonium. The value of
k is here also constant, but slightly greater than that for the
potassium compounds. The salts here considered have ap-
proximately equal molecular volumes*. ‘Thus the alums are
thus related :-— 3

Atomic 3 Molecular

weight. : volume.
meee (50),), 12H.0 .....- 474 1-722 275°3
AlNa(SO,), 12,0 ...... 458 1-641 279'1
CrK (SO,), 12H,0 ...... 499 1:845 270°5

Application of the Law to Bodwes in Vapour at the
Boiling-point.

If the boiling-point be s° Centigrade, and Ds be the density
of the vapour at that temperature, the pressure being 760
millim., then T=273+s°; and the laws of Avogadro and Gay-
Lussac may be expressed thus :—

rola
ear e e

On the other hand, the law of density-numbers gives
Der nub :
Dist = BE

Hence it follows that

ate = constant ;

or, for bodies containing only carbon, hydrogen, and oxygen,

Txn

= constant (see above).

* The equality (approximate) of the molecular volumes of isomorphous
bodies, which was long ago recognized, may be regarded as to some extent |
an anticipation of the law of density-numbers, and may be explained as
follows :—

The constants k= B are equal according to the law. In certain cases
(such as the salts in the table) the relation (5) between the atomic

weight and the density-numbers is nearly the same for different members

of the same group; hence it follows that the values of = (that is, the

molecular volumes) are approximately equal.

26 M. J. A. Groshans on the

The value of this constant for water (density-number =3) is”

2734+100x3
13
In Table X. are given the data relating to 13 substances

which contain C, H,0O,, the value of g being in the first group
10, in the second 8. |

== 62°2,

TaBLE X.
Constant Tn/a=62°2.
|

Name. Formula. | a. N. SS. TT. Daya
Hthylic ether......,......1, ony Oe (C Sa &5) 34 | 308 | 62-4
Ethyl allyl ether ......... C;H,,0 86 | 16 64 | 337 | 62:7
Ethyl propionate ......... O35 4 (802 4 oie 100 | 373 | 622
Hithyl carbonate ......... C.F gi: 11S 4 re 126 | 399 | 60°9
el yhe-ebher sce ietidcc).: Cf, 0 08 rr 86 | 359 | 62:3:
Allylic propionate ...... C AL,,O8.4 14 18 123 | 396 | 625
Hthylie oxalate ............ CHO) | 446.4. 20 186 | 459 | 629
mulyl oxalate ........2<:.5s. O,H,,0, | 170 | 22 | 206 | 479 | 62:0
PM MBOG Es. aatds ab dehy, C,H, 92 15 108 | 381 | 62-1
| ep EOS ee as C,H,O 103 16 150 | 423 | 626
Ethyl pyromuconate ...| C,H,O, 140 18 ; 209 | 452 | 620
Methyl benzoate ......... C,.H.O, 136 18 198 | 471 | 62°3
Methy! salicylate ......... CjH,0,21°. 42 19 223 | 496 | 62:0

Table XI, comprises substances with fourteen atoms of
hydrogen. The constant haa is different from thatin Table x

but is here the same for all the bodies.

TasBLe XI.
Names. ec] OR. Oe | ae: M.S ow Autho- Tr [a
rities.

ARG. Yale ude agen eye vos eh dg Os 6 | 14]... | 86) 20] 45] Go.* | 73-9
Hihyl propylate | ici... 0. ieee 6 | 14.) 1, |102) 21 | 85) CRS ae
Dieting! elyaol, 4) i505 anaes 3- <0 6 | 14.) 2 F118} 22 | 123) We) Fae
MHL VOICPAGG vi... ines es esses see 7 | 14] 2./ 180] 23 | 148) Ta. 736
ABMs Falerate © vivysiekessicavdsvene 8 | 14] 2 | 142) 24 |165| Wu. | 74:0
Ethyl ethylacetocarbonate ...... 8 | 14] 3 |158| 25 | 195) Wo. | 74:0
PAV BUCRLIGLG gyi .cy ceajeas aad v90b 8 | 14} 4 |174| 26 |217| Kp. | 732
Diethylene diacetate ............ 8 | 14| 5 |190| 27 | 248| Wo. | 740
Allyl acetoacetate...............+8. 9 | 14] 38 [170] 26 | 206| Wo, | Jag
thy] itaconate....s..20..-:se00r60 9 114) 4/186) 27-| 227) ie. eee

Mean value of Tn =e ly Gs
a

* Go. means Goriainow; Ch. = Chancel: W6. = Dictionary Fehling;
Li. = Linnemann; Wu. = Dictionary Wirtz; Kp. = Kopp; Ke. =
Kékulé, Lehrbuch.

Law of Density- Numbers. 27

Connection between the different Values of the Constant

In an homologous series one observes generally certain
properties common to all the members ; e. g.:—

(1) A regular increase in the boiling-points of the succes-
sive members of the same series.

2) The amount of this increase varies slightly, but is from
16° to 20° for every addition of the group CH). It was for-
merly supposed that the differences were equal ; but, so far
as experiment has gone, we are able to point to only approxi-
mate equality.

In order to apply the law of density-numbers to homologous
series, it will be convenient to number the members of a series
1, 2, 3, §e., according to the place in the series occupied by
each. I shall call these numbers m: thus benzol has m=4;
it is therefore the fourth member of its series according to the
new Law, though at present the first one known*.

Table XI. contains the substances each containing fourteen
atoms of hydrogen ; each is the seventh member of its respec-
tive series, for m here equals 7. Such substances I shall call
“corresponding ’ compounds. They have, as a rule, similar
properties, and have frequently the same constants Tn/a.

I may now sum up the results of observations on the con-

stants = as follows :—

(1) These constants, of which we have already met two
(62°2 and 73:6), form an algebraic series of numbers which
are the same for all bodies, whether metals or non-metals, and
their compounds.

(2) In an homologous series the constants increase with the
successive additions of CH,, 7. e. with the value of m.

(3) There are numerous homologous series the correspond-

ing members of which possess the same constant, e. g. :—

ig. Mehers, . >) C,, He, Oz,
Bf oY Se Be sth:
Y- 9 <i} = 62 i: ae O.
a. Denzel he. 300 Hg,

* M.-Groshans has not made it quite clear how he obtains the num-

bers m otherwise than by experiment. They are found in compounds of

carbon, hydrogen, and oxygen by dividing the carbon and hydrogen by
the group CH, and adding one to the number thus obtained; the residue
left, after deducting all the CH, groups, being the first body: thus ether,
C,H,, 0, is the fifth, for it = 4(CH,)+H,O, and H,O is the first of the
series C,, Hon+2O0.— Translator. —

28 M. J. A. Groshans on the

(4) A comparison of the constants of these four series and
of many others has shown that it is possible to express the

series of constants zt by the formula

tn 97-8, Va,
a
where z=1, 2, 3, 4, &e.
(5) In the four series above,

LSM 5
but in other series,
L=M+Y,

y being a small number ; thus, in the ethylic alcohol series,
2=m+os.

The constant an in Table X. corresponds to r=5. Itis

also the constant for water, and it is for this reason that I
have adopted the value 27°8. For with water we have

a i

On this basis are calculated the constants for various values of
z given in Table XII.

Taste XII.
£, Tn/a. x, Tn /a. i. Tn/a.
1, 27°80 4, 55°60 7. 73°56
2. 39°32 5. 62°17 8. 78°63
3. 48°15 6. 68°10 9. 83°41

In the following table are given the boiling-points, observed
and calculated, for the first sixteen ethers of the series C,, Hoy, Ov.
The formula used in the calculation is

_ T=(273 +8) = = 97-8 Vm.

*
Se

Law of Density-Numbers. 29

TaBLE XIII.
Boiling-points (s), observed and calculated, for the Hthers

n on Qe

se Se Differ-| Autho-

~ penal. - observed.|calculated.| ences. | rities.
moor HO.) ....-: 46 Spe eaten Ree —17-2
2, G, Be Onc 60 | 8 Bechet Bris Moeaot LS
fee er, Of”... 74 | Ii 53 51:0 25-6 | El
meeeee HO, . ....-- 88 14 it 76°6 93-4 | Li
rr ©. Os 102 17 100 100 99-9 | Pu.
er eH ©), 65: 5. 116.5, 20 122 122-0 90:7 Li.
mere e. 1 O,* ...... 130 | 23 143 142°7 19: | Lie
epee dO. nn. aes 144 | 26 162 162°5 19-0 | Fe:
= 0 oes AS 9 See 15S P29 176 181°5 181 Wo.
nee Og FES Oy ac cc.- 172 | 32 200 199-6 17-4 Sch,
Pipe i.O./) »s.--- 186 | 35 217 217:0 16-9 | Ca
emp te O) ......| 200 f' 38° | 7... 233°9 16-4
_ 5 SSO 0 ee 214 | 41 250 250°3 15:8 | Zi
Pe eth Os) i <5... 228 | 44 269 266:1 15-4 | 2
15. | (C,;H,,9,) ea TU ae ne 281°5 15-2
eee, ...... 256 | 50 298 296°7 Zi
TasLe XIV.
Boiling-points, calculated and observed, of Benzol and it
Homologues. ternary.
Differ-
ae Soa Autho-
epee. (Formula, 4.) observed. jcalculated. Weer rities
ae C, Bodied cee deg 60:7
ee Se Ss ees 1 ee eee 54-7 15:1
sl ee ES Cee te ONE yb oe 69°8 137
4. | Benzol Cp Hyo )) 78% 12 85 88°5 19.0 | Henry.
5. | Toluol C, H, 92] 15 108 108-4 19:7 Noad.
6. | Xylol ...... CoH, 106, 18 129 128°1 19-4 | Cahours.
fe) cumel -..| C, Hi. | 120; 21 152 147°5 18-7 | Gerhardt.
ee cymol ..., C,H, | 134) 24 171 166-2 18:0 | Noad.
9. | Laurol ...| C,,H,. | 148} 27 188 184°2 Fittig.

In the columns headed “ differences ”’ in the last two tables
I have given the increase of calculated boiling-point for each
successive member of the series, and a similar relationship is
to be found in the case of other series. Thus, in :—

* Kp. stands for Kopp; Pi. = Pierre; Li. = Linnemann; Pu. =
Puchot; Fe. = Fehling; W6. = Worterbuch (dictionary) Fehling
Sch. = Schorlemmer; Ca. = Cahours; Zi. = Zincke,

30 On the Law of Density-Numbers.
(1) The Ethers C,, Ha,+2 0, beginning with OC; H,0, where

w is equal to m=4, the differences are:—
27°68) B68) 205; a8 BOe ¢ 00% eee
(2) The series beginning with Methyl Benzoate, C, " O2,

where c=m=5,
16-7, 16:2, 15°8, 15:4, 15:0.
(3) The Alcohol series C, Ho.+2 O, beginning with CH, O,

where c=m+3=5,
16°6;). 19-6," 20°25 (19°38, 19°4) ee eee
The first three substances in Table XIV. do not exist, and

it is very commonly found that the first members of homo-—

logous series, those corresponding to m=(1), (2), (3), are
wanting. There are, however, cases where similar bodies are
to be met with: thus the vobied Cy, Hon+2 gives rise by sub-
stitution to C, Ho,4,Cl and C, H.,Cl. And all bid" corre-
sponding members of the three series possess the same
constants: T'/a (or T B/a) and Vsn/ja (or Vs B/a).

The first members of the substituted series are chlorine and
hydrochlorie acid, which both have e=1*.

As a general rule; the value of x is a whole number, but it
sometimes happens that this is not the case. Thus, e=84 for
acetic acid, CyH,O,; acetic anhydride, O, He Os; an a
similar substance eat from the anhydride in possessing
one atom of oxygen more, C, H, O,, methyl oxalate. All the
three are well known to a anomalously. Thus the
vapour-density of acetic acid-is abnormal. All three are solid
at temperatures below which their higher homologues are still
liquid ; and, finally, the aqueous solution of acetic acid is
remarkable for possessing a maximum of density at 80 per

cent.
[To be continued. |

* With regard to gases such as oxygen, and some others such as CO,
CO,, SO, CS,, &c., we find that the value of 2 is a fraction much less
than unity.

Just as experiment has shown that there are cases to which the laws

of Avogadro and those analogous to it are not applicable, so we find
similar cases in which the law of density-numbers no longer holds good.

aoee

IV. On Colliding Water-jets.
By H. Frangx. Newatu, .A.*

gaa RAYLEIGH has recently (Proc. R. 8. vols. xxviii.,

xxix. 1879, and vol. xxxiv. 1882) investigated many
of the phenomena of liquid jets. I will recall one form of
experiment, and record an observation and a few experiments
bearing on it. |

Two horizontal jets of water, issuing from similar glass
nozzles and fed from two glass bottles, are made to collide at
a small angle. When certain precautions are taken, such as
using clean and tolerably dust-free water, the jets rebound
from one another ; but they are made to unite if each bottle
is connected with the pole of a cell (sach as Grove’s or
Leclancheé’s). It is convenient to introduce a key.

I have lately, in repeating this experiment, observed that
the colours of thin plates (Newton’s Rings) are formed with
remarkable brilliancy between the colliding jets. The jets
are of circular section, but on impact they become flattened
against one another, so that the surfaces of separation are
more or less plane,and vertical. Between these surfaces, which
Lord Rayleigh has shown to be electrically insulated from
one another, there is a very thin film (of air, I presume) in
which the colours are visible. These afford a mode of obser-
vation by which one may possibly gain information as to the
nature of the action of electricity in determining the coales-
cence of the jets.

I will first describe the appearance of the colours. The
ease | have examined most carefully is that when the jets
start in the same horizontal plane and the surfaces of sepa-
ration of the colliding jets are vertical. There is some dif-
ficulty in getting a good view of the colours, on account of
the corrugations produced in the jets at collision. J have
found it best to take the plane of incidence of the light hori-
zontal, and to let the light fall against the direction of the
water in the jets, and very nearly at normal incidence ; and

LIuse a magnifying-glass to observe with.

The figure shows the appearance of the jets about the place
of collision, the bold lines denoting outline and marks on the
near surface ; the dotted lines show the isochromatic curves:
those to the right are those which seem to be least distorted
by refraction. The bold lines at R show the small stationary
ripples on the cylindrical surface of the jet just before the

* Communicated by the Author.

32 Mr. H. Frank. Newall on

point of collision. The jets come from the right, as shown
by the arrow.

The form of the isochromatic curves is more or less con-

stant in character, though the colouring of the figures varies
with the velocity of the jets and the angle at which they
collide ; increase in either of these conditions lowers the order
of the colours produced, as is to be expected. ‘There is a
certain symmetry in the figure, more marked on the right
than elsewhere. I describe a particular case, which may be
taken as a sample of the general appearances.

Starting, then, from the right, close up to the ripples R,
there appears green of the fitth order in a straight line per-
pendicular to the axis of the jet, as shown at 1. The colours
of higher order I have not made out, on account of the ripples.
The line 2 is red of fourth order; line 3 green of fourth

order ; lines 4 are red of the third order. Here the symmetry
in colour ceases. Lines 5 show yellow of the third order ;
and line 6 blue of the third order. The curves appear to be
drawn in to a point on the left; but this I take to be an effect
due to refraction through the uneven thickness of the flattened
jet.. These broader bands of colour are bordered at top and
bottom by the colours of higher order succeeding very closely
on one another, as suggested in the figure by the lines 7;
and towards the left at the top, 8, there appear rings, which Ll
peace are probably seen reflected from within the nearer
jet.

It is necessary to see that the tubes leading the water from
the bottles to the nozzles are kept quite still ; otherwise there
are irregularities in the velocity of the jets which give rise
to flickering in the colours. And in order to keep the con-
ditions of collision as nearly the same as possible throughout
one set of observations, it will also be found convenient to
separate the jets, if they coalesce, by holding an electrified
stick of sealing-wax close to either of the jets for a moment,

Having thus found means of avoiding changes in the colours,
I proceeded to test the effect of gradually increasing from
zero an electromotive force across the film of air between the
rebounding jets, by connecting the water in the bottles eiec-

ee I ce

Colliding Water-jets. | 33

trically with different points on a fine wire, which carries the
current between the poles of a bichromate cell. I have ex-
amined the cases when the colours are of tolerably low order,

_that is when the separating film is thin ; and I have not been
able to detect any change of colours, pointing to the gradual

approach of the jets. The colours remain constant, until the
electromotive force is raised to such a point that coalescence
takes place ; and then they disappear quite suddenly. I had
determined to give the matter one more trial, wishing to
observe the effects when the jets collide at an extremely small
angle. In this case it may be arranged that the colours pro-
duced are not of lower order than the fourth in the thinnest
part of the film. This would be the most delicate means of
testing. But I have been completely foiled in my attempts ;

for the jets persistently refuse to rebound from one another

for longer periods than a small fraction of a second. [I attri-
bute this to the large amount of pollen in the air, blown from
fir-trees near the house ; for I find quantities of golden powder
collected on my table and elsewhere in the house.

My observations then, so far as they have gone, eliminate
conclusively the explanation, which Lord Rayleigh has re-
ferred to as possible though not probable, namely that the
action of electricity in promoting union may be ascribed to
the ‘‘ additional pressure called into play by electrical attrac-
tion of the opposed water-surfaces, acting as plates of a con-
denser” (Proc. Roy. Soc. vol. xxxiv. p. 145).

The extraordinary capriciousness of union—the jets some-
times coalescing with very much smaller electromotive forces
than at other times—made me for a time think that dust
might still be at work ; that the dust, which in the ordinary
state of the jets is not able to cause coalescence, might, under
the possibly directive influence of the electromotive force
(much as in an experiment described by Faraday), be turned
across the separating surfaces of the jets, and so rupture them.
If this were the case, one would expect that a larger electro-
motive force would be required to bring about coalescence,
when the water supplying the jets had been allowed to settle.
I left the water for five days to settle and found no appreciable
difference between the electromotive forces required to pro-
duce union, when this ‘settled’? water was used, and when
dusty water was used immediately afterwards. Nor did
diminution in the velocity of the jets affect the magnitude of
the requisite electromotive force.

Lord Rayleigh regards the union as most probably due to
perforation of the separating skin, brought about by disrup-
tive discharge. Let us consider the case when the colour of.

Phil. Mag. 8. 5. Vol. 20. No. 122. July 1885. =D

34 MM. Julius Elster and Hans Geitel on the

lowest order produced in the film between the jets is green
of the third order at normal incidence: the thickness of the
film is, roughly speaking, 24/2 x 6000 x 10-8 centims., that is
0:000075 centims. It appears that the highest minimum
electromotive force to produce union is about 0°75 volt. This
gives a difference of potential in volts per centimetre of 10,000.
Now Sir W. Thomson’s experiments show that, to produce a
spark between brass plates nearly 100 times further apart
than the jets in the case we are considering, an electromotive
force at the rate of about 80,000 volts per centimetre is re-

quired. De La Rue and Miiller have lately shown that the
substitution of saturated aqueous vapour for dry air between

the plates does not make any great difference. It is, however,
conceivable that there is diminished pressure between the jets.
Crowthorne, Wokingham, June 1885. |

V. Observations on the Electrical Processes in Thunder-
clouds. By Jutius Ester and Hans GeiTE.*,.

7 an investigation, with which we have been some time

engaged, on the production of electricity by the friction
of finely divided liquids against solids of various temperatures,
we had occasion to observe the exceeding sensitiveness to
electrical induction of disintegrating jets of liquid as well as
of all solid and liquid particles suspended in a current of air.
This action is the chief source of error to be contended against
in such experiments ; and, as we shall have occasion to show
in a subsequent paper, it makes it very difficult to settle the
question whether electricity is produced when water-spray
rubs against cold bodies. The phenomenon has long been
known for jets of liquid, and has already had a practical appli-
cation in Thomson’s drop-collector and the water-induction
machine. It was an obvious idea to consider that inductive
actions of a similar kind are also at work in the processes
which take place in a thunder-cloud—that is, just to regard
the latter as a self-acting duplicator. |

The principle of this idea will be best seen from an experi-
ment which can be made by the simplest means.

A cylindrical metal tube, A, open at both ends (about
50 centim. long and 8 centim. in diameter), is fixed vertically
to a lateral insulated holder. A small vessel, B, which is also
cylindrical, can be inserted from the top, being held by an
insulated handle. This latter vessel is closed below, and is

* Translated from a separate impression from Wiedemann’s Annalen,
No. 5 (1885), communicated by the Authors.

~ oop *™\ —"

Electrical Processes in Thunder-clouds. ae

provided with an efflux-tube directed downwards and ending
in a narrow aperture. If now a small charge of + electricity
is imparted to the tube A, while at the same time the insu-
lated vessel B filled with water is held over it, the drops from

B become negatively electrified, but fall through the tube A

without coming in contact with the sides. By this means B
itself is positively charged, and in an extremely short time
acquires a potential which is not much lower than that of A,

when the difference between B and A is not too great. If
now B, while still held by the insulating handle, is rapidly

inserted in A, and touches it momentarily, the electricity of
B passes almost completely over to A, according to known
electrostatic laws. If B be withdrawn to its original position,
the same operation may be repeated. It must be remem-
bered that A has now a greater charge than before, and that
accordingly B is also more strongly influenced. It follows
accordingly that, by repeating the operation, the charge of A
increases in geometrical progression. In practice the limit is
soon reached. With sucha duplicator it is easy to show a
considerable increase in the strength of the charge; it is only
necessary to connect A during the experiment with a gold-leaf
electroscope and to be somewhat quick in moving Bb. We
observed almost always a self-excitation of the apparatus, so
that after fortyfold oscillation of the vessel B a spark could be
taken from the tube A.

The process in a thunder-cloud may be considered to be
analogous. Suppose that a given charge of + electricity is
imparted to one place in the lower layer. As long as the cloud
is not dissolved in rain this electricity remains in nearly the
same place, or only slowly spreads with decreasing potential
over the entire cloud according to the conductivity of the
mass of vapour. The case is different when the cloud dissolves
in rain, where it may be assumed that the formation of drops
must take place in the upper, that is the colder, layers of air.
As far as the cloud rains it will acquire positive electricity ;
and if the formation of rain does not begin at too great a height
above the electrical layer, it is of not much lower potential
than the latter. The negatively electrified drops fall to the
earth through the lower inducing layer. 3

But now, as has often been pointed out (so far as we know,
first by Hermann J. Klein), the contraction of surface, which
accompanies the formation of rain, must produce an increase
of potential. As the entire mass of cloud which was pre-
viously charged with positive electricity by induction aggregates
together, and the individual particles of vapour coalesce to
form larger drops, the electricity is confined with increasing

D2

36 On the Electrical Processes in Thunder-clouds.

potential to a smallor space, and must again act by induction
and with increasing power on the newly formed masses of
clouds. As soon as the formation of rain begins in these the

same process is repeated; by a further aggregation of the:

clouds a fresh increase of potential is produced, and so forth.
It can easily be conceived that in this way the electrical po-
tential of a rain-cloud may go on increasing until discharge
takes place.

In the experiment just described the vessel A represents
the lower inducing layer of cloud ; B the upper rainy part of
the same. The contraction of surface, although corresponding
but little to the actual process in the thunder-cloud, is repro-
duced by the introduction of the vessel B into A.

The question of the origin of the initial charge of the lower
layer of cloud will not be so easy to answer. In any casea
very small potential will be sufficient, with a heavy rainfall, to
be increased to the greatest amount. It is perhaps the ordi-
nary atmospheric electricity which, passing to the clouds,
produces a greater local charge. Without being kept up
by the fall of rain, the strongest charge would not last. To
look upon the condensation of aqueous vapour in itself as
a source of electricity appears to us inadmissible, as it is not
clear in what way the separation of the electricities can take
place in this process. [Friction can only explain the initial
charge, and is quite inadequate to account for the enormous
disengagement of electricity which takes place in nature.
We might, moreover, think of a friction of fine aqueous vapour
against large drops or against hailstones. It is not impro-
bable that the difference between the capillary superficial
tension of the smallest and largest drops of water may in
friction produce a separation of the electricities ; but in any
case such an excitation has not been ascertained with certainty.
Supposing it, however, to be so, is the vis viva lost in this
friction sufficient to produce the mechanical equivalent of even
one flash?

It might be urged against the above view, that the diminu-

tion in volume of a cloud consists essentially in the coalescence .

of particles of vapour to form larger drops ; that therefore
the electricity must accumulate on the latter, and ultimately
be carried away with it as it falls. This is undoubtedly par-
tially the case. We must, however, remember that for a
certain time the drops which form in the lower layer of cloud
must be supported by the ascending current of air, the force
of which must decrease with the height; while the drops
which come from the upper layers must reach the lower layers
with greater velocity, and consequently must more easily

ora a

———E—

Se

On certain Cases of Electrolytic Decomposition. 37

overcome the resistance of the ascending air. Part of the
charge is thus lost, but the part is smaller the more powerful
is the ascending current. There is, moreover, a continual
replacement, owing to uninterrupted condensation.

With very-fine rain, which at once falls to the earth, power-
ful charges will scarcely be expected. Much depends, in any
case, on the rapidity of the formation of cloudsand rain. The
rain which falls from a cloud must have the opposite electricity
to the cloud, and can of course act by induction on a second
cloud which it traverses, and which also rains.

Moreover, after each fiash of lightning the remainder of the
electricity will soon attain a maximum if the rainfall is suf-
ficient, and the contraction in volume which is inseparable
from it.

The essence of this theory is that the electricity of thunder-
clouds acts by induction, and that the thunder-cloud is a self-
acting duplicator. The rain which falls from it plays the
part of the jet of water in Thomson’s drop-accumulator, while
the increase of potential is occasioned by the enormous con-
traction of volume and surface.

This theory may claim over preceding ones the following
advantages :—

(1) It enables us to regard the cioud as an aggregate of
discrete drops of water. Hence a charge imparted to a cloud
does not spread by conduction, but as soon as the cloud begins
to rain spreads from point to point by induction.

(2) It does not require electrification by friction. Itis not
impossible that friction may give the start ; but this becomes
unimportant in the further progress of the phenomenon.

If it should come out, anda number of experiments we have
tried do not allow us to settle the point, that a production of
electricity by the friction of water-spray or rain against water
or ice cannot be experimentally proved, the presence of
atmospheric electricity is sufficient to start the phenomenon.

(3) The theory finds the equivalent of the work expended,
in the establishment of a difference of potential, in the vis viva
of the falling drops of water.

VI. On certain Cases of Electrolytic Decomposition. By

J. W. CxarK, Assistant Professor of Physics in University
College, Liverpool*.

HE atomic or molecular conditions which determine me-
tallic or electrolytic conduction are of great interest, but
seem as yet too obscure to allow of any definite general conclu-

* Communicated by the Physical Society: read May 23, 1885,

38 Assistant-Prof. J. W. Clark on certain Cases

sions respecting their nature, beyond regarding a free motion
of the particles resulting either from fusion or from solution
as necessary for electrolytic conduction. Hven this is not
without noteworthy exceptions; for Faraday has described
some binary compound liquids (SnCl,, AsCl;, &e.*) which
neither conduct nor decompose; whilst, on the other hand,
some compound solids are known which conduct metallically
(Cu,Set, Ag,SeT, SnS., CuSf, to which perhaps must be added
PbO,, MnO,, and Ag,O), and some bodies which are solid
and yet conduct electrolytically (Cu,8f, Ag.S§). Further,
zincic iodide§ neither conducts nor is decomposed when ren-
dered fluid by heat; whilst others (Hg1,§, HgCl,§, PbFl,||)
have been considered as conducting without decomposition
under the same circumstances. The nature of the conduction
of the metallic sulphides is very imperfectly known].

Whilst thinking over these facts about a year and a half
ago, it seemed to me very probable that a careful study of
these exceptions to the general laws of electrolytic decompo-
sition might result in more definite conjectures respecting the
condition of internal or molecular structure required for con-
duction and decomposition. To make such a study complete
requires the determination and comparison of a number of
physical constants (e. g. colour, conductivity, expansion, spe-
cific and latent heat, refractive index, specific inductive
capacity, &c.) for substances which are normal and abnormal
in their electrolytic behaviour. This cannot be completely
done at present for want of data; but before passing to the
consideration of those bodies to which my own investigation
refers, I wish to briefly refer to a few previous papers relating
to substances of remarkable electrolytic behaviour from this
point of view. 3

Hittorf states that sulphide of silver fuses at a clear red
heat, but at a temperature of 180° C. it is sufficiently soft to
adapt its shape slowly to that of the surface upon which it
rests, and at the ordinary temperature it is malleable: a cast
stick of it can be slightly bent without fracture, and it can be
cut with a knife or turned in a lathe.. This substance has a
very low resistance, and even at the ordinary temperatures
is electrolytically decomposable by a feeble current. Hittorf
(/. c.) has concluded, from a long and careful series of expe-
riments upon it, that it conducts electrolytically only, and that

* Faraday, ‘ Experimental Researches,’ vol. i.

+ Hittorf, Poge. Ann, Bd. lxxxiv.1851. —{ Hittorf(.¢.), § Faraday.

|| Faraday. Beetz (Pogg. Ann. Bd. xcii. 1854) has since shown that
conduction takes place in a normal electrolytic way. Faraday discovered
that the solid Plumbic Fluoride began to conduct below a red heat.

4] See Faraday, ‘ Experimental Researches’; and Hittorf (7. ¢.).

of Electrolytic Decomposition. 39

its apparent metallic conduction results when a fine thread of

metallic silver has formed between the terminals of the sul-
hide of silver bar.

Hittorf (/. c.) has discovered that cuprous sulphide also con-

ducts electrolytically “und besitzt entweder gar keine oder

ganz geringe metallische Leitung.’’ Unlike sulphide of silver
this body has a high resistance, but, like it, the conductivity
increases with increased temperature. Cuprous sulphide melts
at a white heat, and may be cast in the form of a rod which
at the ordinary temperature is very brittle, although at a
“higher”? temperature it may be bent.

Hittorf (/. c.) points out the difficulty of proving exper-
mentally that the conduction of Cu,S and Ag,S is entirely
electrolytic, because the formation of copper or silver by the
action of the current in the substance of the bars places
their ends in true metallic communication with the battery-
terminals.

Cuprous selenide and argentic selenide closely resemble
their corresponding sulphur compounds to which reference
has just been made, and are described as being “soft’’ and
“slightly malleable’”’ respectively (Watts’s ‘ Dictionary of
Chemistry’). Selenides ordinarily closely resemble sulphides
in their physical and chemical relations, and are consequently
regarded as being possessed of similar molecular constitutions.
Hittorf, however, says of cuprous and of argentic selenide,
that “beide sind gute metallische Leiter.””? Their electrical
behaviour is therefore of an exactly opposite nature to that of

_ their corresponding sulphur compounds.

Faraday has noticed that conduction commences in heated
electrolytes at very different degrees of liquefaction and
softening.

Plumbie chloride * conducts very appreciably at a tempera-
ture far below that at which it fuses, and at which it is not
noticeably soft. Beetz (/. c.) has shown that glass begins to
conduct between 200° C. and 220° C., and Dr. Lodge reminds
me that this has recently been shown to take place at 100° C,

It appears from the behaviour of these bodies that—

(i.) In some solid electrolytes there is a sufficient mobility of
the molecules at the ordinary temperature to enable electrolytic
conduction to take place; in others it is conferred by a rise of
temperature which is insufficient to render the solid liquid or even

* E. Wiedemann, Ber. d. Kgl. Sachs. Gesellschaft der Wissenschaften,
1874. I have unfortunately been unable to refer to this paper; but since
the above was written the author has kindly referred me to ‘Die Elektri-
citatslehrer’ (G. Wiedemann), Bd. i. 8. 558, wherein it is stated that the
iodide and bromide of lead behave similarly.

40 Assistant-Prof. J. W. Clark on certain Cases

soft. A rise of temperature which is insufficient to render a
solid liquid, or so soft as to change its shape, is also sufficient
for the equalization of strain as is shown in the annealing of
glass. Both electrolytic conduction and the annealing of glass
take place more readily at a high than at a low temperature.

(i1.) Substances of apparently similar constitution may exhibit
opposite forms of electrical conduction (Cu,S8e, Ag,Se, and Ag,S,

U.S).

ih bodies to which my investigation refers are mercuric
chloride and mercuric iodide, which Faraday believed to
conduct metallically in the fused condition.

Beetz, as I have recently found, states in a paper, “* Ueber
die Leitungsfihigkeit fiir Elektricitat welche Isolatoren durch
Temperaturerhéhung annehmen”’ (J. c.), that he has obtained
evidence of the electrolytic decomposition of fused mercuric
iodide; and he attributes its apparent conduction without
decomposition to recombination of the products. So far as I
am aware, no attempt has been made to examine the nature
of the conduction of fused mercuric chloride since Faraday
concluded that it probably conducted metallically.

Mercuric Iodide.

Mercuric iodide is dimorphic, and at the ordinary tempera-
ture forms a scarlet powder which at 110° C. becomes yellow,
and at that temperature acquires a very slight electrical con-
ductivity (Beetz). Mercurie iodide melts at 247° C. and
boils at 842° C. It is an interesting substance on account
of the ease with which it volatilizes at temperatures much
below its melting-point ; and it is not unlikely that the ease
with which the molecules are thus shown to be leaving the
solid substance may be connected with a high diffusive rate
when it is fused, and this may partially explain the readiness
with which the products of its electrolytic decomposition often

_ mix and recombine, thus simulating conduction without de-

composition. The mercuric iodide which I have used in the
following experiments was prepared either by the precipitation
of recrystallized and sublined mercuric chloride with pure
potassic iodide and sublimation of the product, or-by the
sublimation of the commercially pure substance.

It may not perhaps be out of place to point out, in the first
instance, the effect of heat upon mercuric iodide; as it is
sometimes stated in text-books of Chemistry that, when heated,

it undergoes partial dissociation with liberation of iodine, where |

the edge of the liquid is in contact with the hot glass vessel.
I have, however, convinced myself that this statement applies
only to the commercially pure substance, which on sublima-
tion leaves a little impure oxide of iron, and which, on the

Oe a SE el a Sas Pa

of Electrolytic Decomposition. Al

edges of the glass vessel in which fusion and sublimation are
effected, may decompose, yielding free iodine vapour. Pure
mercuric iodide may be fused and sublimed without under-
going any such change ; nor have I any reason to believe that,

when strongly heated in a sealed glass tube, iodine is ever set

free, a slight darkening in the colour of the fused substance

being the only apparent alteration which it then undergoes.

Here, too, perhaps it may be convenient to describe the
preparation of the graphite electrodes, which for some years
past I have found very convenient for the decomposition of
such substances as act upon platinum. ‘These electrodes are
best made from the “leads’’ of Rowney’s HH cedar pencils,
which may be easily removed after a few hours’ soaking in
water has softened the glue sufficiently to allow of the pencil
being split in half. The “lead” is then removed with a knife,

-and only requires heating to bright redness in a Bunsen-

flame (to get rid of the shellac (?) which it contains) to render
it fit for use. A platinum wire twisted or bound round one
end makes a good connection for the battery-terminals.

When pure mercuric iodide in a sealed glass tube is kept
in a state of fusion over a gas-flame, and electrolyzed by
means of two platinum-wire electrodes passing through its
ends, evidence of its decomposition may be obtained from the
iodine set free about the + pole, although no mercury is dis-
coverable at the negative. The quantity of iodine thus set
free is, however, small, and does not usually seem to increase
with the length of time that the current is allowed to pass ;
whilst the decomposition, judging from the liberation of iodine
vapour about the positive or upper electrode in the tube,
appears to take place at temperatures very little above the
solidifying-point of the liquid mercuric iodide.

Hlectrolysis of this substance kept fused in narrow V- and
W-shaped tubes over a gas-flame yields much the same results;
but in such tubes the resistance is very high, and it is more-
over difficult to ascertain precisely what is taking place within
them. I therefore adopted a simple V-shaped glass tube of 3ths
of an inch diameter and bent at an angle of about 30°, into
which was placed a sufficient quantity of mercuric iodide for
the experiment. The substance was then kept fused over a
gas-flame, and the two graphite electrodes introduced, one at
each end of the tube. In such a tube the behaviour of the
substance is easily observed ; and by blowing air dried over
chloride of calcium into the end, the issuing vapours are readily
tested for iodine with starch-paper. Under these conditions,

I found that when the substance was at a temperature near

the melting-point iodine could usually be detected with ease,
although at a higher temperature none could be shown to

42 Assistant-Prof. J. W. Clark on certain Cases

exist in the free state. A control experiment with fused
chloride of lead, using a current of the same strength and the
same distance between the electrodes, gave evidence of chlorine
without difficulty. :

Hence it appeared possible that this substance conducted
(as AgS had been supposed to do) in two ways—(a) electro-
lytically, at a temperature at which it just became liquid; and
(8) metallically, at higher temperatures. But during some
observations upon the expansion of fused mercuric iodide ina
thermometer-like tube, I noticed that the liquid underwent
great contraction during cooling from a temperature a little
above the melting-point, accompanied by a distinct loss of
fluidity. ‘To the latter change I at once attributed the evi-
dence of electrolytic decomposition, which I had observed at
a corresponding temperature, to be probably due, as it would
be less favourable to the mixing and recombination of the
products of electrolytic decomposition; and I therefore turned
my attention to the construction of an apparatus in which
they should be so separated as to render this less easy.

The form of apparatus which I finally adopted is shown in
section in fig. 1. The tube (abc)

containing two small porous bat- :
tery-pots (g/), the graphite elec-
trode, and the mercuric iodide to d e

be electrolyzed, is U-shaped, and
the two branches are connected
by a constricted portion (¢), which
further materially hinders the a A
mixing of the fused products of
decomposition in the two branches
of the tube. ‘This apparatus
passes through a hole in the tin
cover (jz) which su>ports it, and
dips into the oil in the beaker,
which is heated to the desired
temperature by means of a gas-
flame or sand-bath, at which it is
kept constant by placing the
bulb of the air-thermometer (/)
in communication with a gas-
regulator. When mercuric iodide
is electrolyzed in such an appa-
ratus witha current ofabout 0°02 — jwetark vex

ampere, iodine is liberated at |
temperatures far above the melting-point of the substance.
If the current was too strong the circuit was usually

~

of Electrolytic Decomposition. 43

broken, apparently in consequence of the heat generated
at the + electrode volatilizing, or otherwise causing the
mercuric iodide to disappear from the porous pot. After
passing the current for some hours the gas was extinguished,

and when the apparatus had become cool the two branches of

the U-tube were cut asunder and broken open for examination.
In the + branch, both within and without the porous pot,
mercuric iodide and black feathery streaks of iodine were
found which gave the starch reaction, and the upper part of

the glass tube was coated with volatilized iodine.

Iodine dissolves freely in fused mercuric iodide, producing
but little change in the colour of the latter ; but just as soli-
dification commences, more or less complete separation seems
to take place with the formation of these black patches rich
in iodine (and Hg,I,?) and an evolution of iodine vapour,
which in the previous experiments also helped to lead to the
conclusion that electrolytic decomposition occurred only near
the melting-point of mercuric iodide.

The contents of the negative branch and porous pot were found
to be of a slightly altered colour; but no free mercury was
discoverable. Repeated exhaustion of its pulverized contents
with absolute alcohol revealed the presence of mercuroso- mer-
euric iodide (Hg,J, or 2HgI,, Hg,I,), which might perhaps
be inferred to be the first product of the electrolytic decom-
position of HglI,, since it is formed when metallic mercury
and mercuric iodide are titurated together in the proper pro-
portions. Mercuroso-mercuric iodide is ordinarily regarded
as a distinct compound, and I suppose rightly so; for it can

exist without decomposition at a temperature at which mer-

curous iodide undergoes decomposition with liberation of free
mercury; but the stability of this body seems to be greatly
increased by the presence of a slight excess of mercuric iodide.

‘Mercurous iodide (Hg,I,) dissolves readily in fused mercuric

iodide in the proportion to form mercuroso-mercuric iodide.
The formation of mercuroso-mercuric iodide by the action of

the current upon mercuric iodide may therefore be represented

by
(i) 2Hel,—HeI,+h,

(ii) 2Hgl,+ Hg.l,=Heg,1,;
or, considering it to take place in one step,
4Hol, — He, lI, -- r..

The action being slow | repeated the experiment, replacing
the mercuric iodide in the apparatus just described by some
mercuroso-mercuric iodide precipitated nearly according to the

44 Assistant-Prof. J. W. Clark on certain Cases

equation
AKI + I, ale 2N.0,Hg,0, — Hg, I, + 4NO,KO.

The precipitate was then washed, dried, fused, and finally
sublimed from an evaporating dish on a sand-bath into a clock-
glass which covered it. The substance contained a little mer-
curic iodide as impurity.

The products of the electrolytic decomposition of this sub-
stance are mercury and iodine, of which the former is depo-
sited in the metallic condition, apparently according to the
equation |

Hg,J,=4H¢ +31;

whilst the iodine is absorbed by the Hg,I, in the positive
branch of the tube, forming mercuric iodide ; thus,

31, + 3H¢,1,=12H gl).

These results may be summarized as follows:—

The conduction of fused mercuric todide is electrolytic; but
decomposition and recombination may take place so rapidly
as to give rise to an apparent metallic conduction; but my
investigation gives no grounds for supposing that uw does not
quantitatively conform to Faraday’s Laws.

The causes rendering the proof of its electrolytic decom-
position dificult may be summed up as follows :—

(1) Lodine is soluble in fused inercuric iodide, and so 1s mer-
curous todide, in the latter case with the formation of Hg,lg.

(2) Mercuric iodide is volatile, and the presence of its vapour

renders the detection of free iodine difficult by the ordinary test,
and also promotes mixture in the electrolytic apparatus by dis-
tillation.

(3) Lt also seems possible that fused mercuric iodide possesses
a high diffusive rate, which would further facilitate the mixture
and recombination of the products of its electrolytic decomposition.

(4) Lhe electrical resistance of fused HgI, is high.

When pure mercuric iodide is fused over a gas-flame in a
straight glass tube, of about 6 centim. in length and 0°5 centim.
in breadth, and electrolyzed between platinum wire or graphite
electrodes with a current of about 0°20 ampere, the resistance
shows some remarkable changes. ‘Thus:—In an experiment
which I copy from my laboratory journal, the resistance of the
mercuric iodide in a sealed tube decreased as the temperature
rose, until the needle of the tangent-galvanometer which was
included in the circuit stood at 20°; and on allowing the
temperature to rise still higher, it fell to 9°. During
cooling the inverse change occurred; for on extinguishing
the gas and allowing the tube to cool, the needle advanced

vee

of Electrolytic Decomposition. 45

from 9° to 20° or 21°, and then fell gradually to 0° as
conduction ceased. The cause of this change in the resist-
ance is not very clear. I.have assured myself that it is
not due to any impurity in the mercuric iodide employed,

nor is there a sufficient change in polarization of the electrodes

at the different’ temperatures to account for it. With a very
feeble current these resistance-changes are not marked, and
indeed may escape observation in an experiment such as that

_ which I have described. I believe this effect to be due, first

to the formation, and then to the dissociation or other altera-
tion at the higher temperature of the mercuroso-mercuric
iodide produced ; in support of which it may be stated that
mercurous iodide undergoes rapid decomposition with separa-
tion of mercury at about the same temperature at which this
change of resistance takes place. Perhaps not entirely uncon-
nected with increase of resistance is the heat generated by the
current, in consequence of the transition-resistance at the com-
mon surface of the + electrode and fused mercuric iodide,
which may occasion the formation of vapour on the surface
of the electrode. I have observed this give rise to a crepi-
tating noise, and to the formation of a wave-motion spread-
ing from the + electrode over the surface of the fused
substance, and, under favourable circumstances, becoming
so marked as to throw the whole tube in which the decompo-
sition was being effected into violent oscillation.

Mercuric Chloride.

Mercuric chloride far exceeds mercuric iodide in the ease with
which it volatilizes at temperatures below its melting-point. It
melts at 265° C. and boils at 295° C., and is more difficult to de-
compose with the current than mercuric iodide. This is to some
extent due to its higher electrical resistance, which prevents
the use of any complex apparatus designed to represent the
products of decomposition and prevent their recombination.
The mercuric chloride which was used was prepared by the
sublimation of the repeatedly recrystallized pure commercial
salt.

The apparatus with which I have succeeded in effecting the
electrolytic decomposition of fused mercuric chloride is shown
in section in fig. 2, and consists of a glass tube of about 3ths of
an inch in diameter bent at an angle of 30°. The porous pots (ab)
and graphite electrodes (¢ d) project at each end, and the requi-
site temperature was supplied by the hot sand of a sand-bath.
With potassic iodide and starch test-paper the evolution of chlo-
rine at the + pole was readily detected, even when a very feeble
current was employed. After the current had been passed

46 Assistant-Prof. J. W. Clark on certain Cases

through the fused mercuric chloride for some time, the appa-
ratus was allowed to cool, and its contents were subsequently
Fig. 2, qi

J.W.,CLARK DEL

pulverized and exhausted with water, which left an insoluble
residue of mercurous chloride (Hg,Cl,), which was not entirely
confined to the inside of the — porous pot.

Mercurous chloride dissolves in fused mercuric chloride,
probably giving rise to a mercuroso-mercuric chloride of
analogous composition to some of the well-known double
chlorides which mercuric chloride forms. The want of time
consequent upon the completion of an investigation on the
influence of pressure on electrolytic conduction, upon which I
have long been engaged, has prevented my examining this point.

The conduction of nercuric chloride is electrolytic, gwing rise
to chlorine and mercurous chloride; and there seems no reason
to doubt that it conforms quantitatiwely to Faraday’s Law;
but the volatility of this substance, as also in the case of
mercuric iodide, and of the products of their decomposition,
would render its further proof difficult.

The causes rendering its electrolytic decomposition difficult —
may be summed up as follows :—

(1) The volatility of mercurous chloride and its solubility in
mercuric chloride.

(2) The near melting- (265° C.) and boiling- (295° C.)
points and great volatility of mercuric chloride facilitate the
mizing and recombination of the products of its decomposition,
and the vapour renders the detection of chlorine by the ordinary
test difficult.

(3) As previously stated in reference to mercuric iodide,
it is possible that the volatility of these substances is con-
nected with a high diffusive rate, when fused, which would
facilitate the recombination of the products of its electrolytic
decomposition. ;

(4) The electrical resistance of the fused substance ts much
higher than that of the mercuric todide.

Before concluding this paper, I wish briefly to refer to the
properties of fused mercuric iodide and chloride with refer-
ence to the porous battery-pots in which the electrodes were

= z
es

of Electrolytic Decomposition. 47

_ placed, and which seems sufficiently important to merit a few

words of description. ‘These porous pots were 2 inches long,
1 inch in diameter, and varied from 74 inch to 4 thick in the
walls. My attention was first attracted to their behaviour by
noticing that when they were partially dipped into fused
mercuric iodide, that liquid rapidly made its appearance in
the pot; and the subsequent analysis of the substance sur-
rounding the porous pots after an experiment, showed the
presence of small quantities of the products of the decompo-
sition effected by the current, such as iodine, mercuroso-
mercuric iodide, and mercurous chloride. Control experi-
ments with water and with fused plumbic chloride showed
that these liquids were unable to penetrate the walls of the
porous pots.

The explanation of the facility with which fused mercuric
iodide penetrates the walls of a porous pot and rises within it
seems to be of acomplex nature. The imbibition of this fused
substance in the porous walls of the pot is due to capillary

action, and does not account for the liquid filling the pot; for

since it is of a capillary nature, this action must cease as soon
as the inner surface or wall becomes wetted. In the case of a
volatile liquid such as fused mercuric iodide this action may be
somewhat prolonged by its volatilization from the inner surface
of the pot-wall, and by the direct formation of crystals from the
vapour. ‘The subsequent fusion of these crystals will account
for the presence of some liquid mercuric iodide within the
porous pot, but then this action must cease. Moreover this
explanation apparently requires a difference of tempera-
ture within and without the porous pot, which, from some
special experiments made upon the subject, can scarcely be
assumed to exist; and I therefore think that the explanation
of the penetration of the liquid through the walls of the porous
pot must be mainly sought in an easy transpiration of the fused
mercuric todide through its pores in consequence of the small
initial difference of level (“head”) of the liquid.

I have already stated that analysis showed the presence of
some of the products of decomposition formed within the
porous pots in the undecomposed substance in which the latter
were partially immersed; and this seems attributable to diffu-

sion through the pot-walls. Little seems known respecting

the rates of diffusion of fused substances through porous dia-
phragms; but the particular difficulties in the way of their
determination for such volatile substances (which may so
readily mix by distillation) seems of itself suggestive of a
molecular activity not unconnected with a long free mean
molecular path and of rapid diffusion.

Pad

VIT. On the Determination of the Heat- Capacity of a Ther-
mometer. ByJ.W.CiARK, Assistant Professor of Phystes
in University College, Liverpool*.

I, thi the determination of specific heats a correction should

be made for that part of the thermometer which is
immersed in the water of the calorimeter, as its specific heat
is not the same as that of an equal volume of water. Very
often this correction is reduced to a mere estimation on
account of the unknown weights of glass and mercury which
constitute the immersed portion of the thermometer.

When two metals of known specific gravity are fused

together, and the volume of the resulting alloy is the sum of
the volumes of its two constituents, it is only necessary to
know the specific gravity and volume of the piece of alloy
to calculate the exact volumes of the metals comprising it.
Similarly, by determining the specific gravity and volume of
the thermometer, the volume of mercury which it contains can
be at once determined from
_ V(S—%:)_
Vi= S835 0. (1)

where V,isthe required volume of mercury in the thermometer,

V the volume of the thermometer,
S the specific gravity of the thermometer,
So 4! e », thermometer-glass,

Sy 9? 70 of Hg.

The mean value of several very closely agreeing determi-
nations of the specific gravity of different specimens of ther-
mometer-glass is 3°199 for lead-glass and 2°512 for soda-glass,
Should it not be known of which sort of glass the thermo-
meter consists, 1t may be readily ascertained by slowly intro-
ducing the upper extremity of the instrument into the re-
ducing-flame of a blowpipe :—Soda-glass yields a yellow
flame, but lead-glass blackens, from the reduction of the oxide
of lead which it contains. It may be assumed that thermo-
meters made on the continent consist of soda-glass ; those
made in England are usually constructed of lead-glass.

The total volume of the thermometer (V) is obtained
from its weight in air and in water.

The volume of mercury contained in the thermometer haying

been found by (1), the volume of that part of the thermometer
which is immersed in the water of the calorimeter has next to

* Communicated by the Physical Society: read April 25, 1885.

Determination of the Heat-Capacity of a Thermometer. 49

be determined. This latter volume, less the contained volume
of mercury, is the immersed volume of glass.

The total immersed volume (V3) of the thermometer is best
found from the weight (W) of the thermometer in air, and its

weight (W,) when dipped in water to the same depth as it

dipped into the water of the calorimeter ; then
ay ae YY ey ee eee ey S64)

But, if the bulb be cylindrical and of the same diameter as the
stem, this volume may be calculated from the measured length
and diameter of the immersed portion; for if of an irregular
shape it may be found by plunging it into a burette graduated
in cubic centimetres and partly filled with water ; but in both
cases with a less satisfactory result than is given ‘by (2).

The volume of the glass of the thermometer (V,) immersed

in the water of the calorimeter is then

V.=V;— Vj.

These volumes of glass and mercury are converted into
their corresponding weights, using the mean specific gravities
of lead- and soda-glass already given. The sum of the pro-
ducts of the weight of the immersed glass and its specific
heat, and of the weight of the mercury and its specific heat,
is the required water-value of the part of the thermometer
dipping into the water in the calorimeter.

Regnault has given the specific heat of thermometer-glass
as 0°2; but for greater accuracy it would be desirable that for
this value the mean specific heat of lead or of soda thermo-
meter-glass should be substituted according to circumstances ™.

The following is an illustration of the application of the
method :—

ay ia aa No. 2. Soda-glass.

Weight of thermometer in air . . . =W 33:97 grm.
3 % water . 4 23°01 grm.

Volume of thermometer 2 gS ARES Ve OGiewe;

Specific gravity of thermometer . . =S 3:098

Mean sp. gr. of serach ti Sigh Dolirssep ea S42

me erivot Hye: 2 sms 3) bree hope eb

Volume of mercury in thermometer=V,=~ 0-5
=0°9799 C.c. ge

* It may perhaps be serviceable to call attention to the Phystkalisch-
Chemische Tabellen of Landolt and Bornstein as a work which contains a
most useful collection of data.

Phil. Mag. 8. 5. Vol. 20. No. 122. July 1885. AD

90 Determination of the Heat-Capacity of a Thermometer.

Weight of thermometer inair. . . . . 83°97 grm.

Weight of thermometer with lower end im-) © |
mersed in water to the same depth as i 2078
was immersed in water of calorimeter

Loss of weight=volume of immersed part 4.94 oo.
of thermometer: . y). =. .acay ee

Subtract contained volume of Ho=V, . =0:5797

to get immersed volume of thermometer-glass =3'661 c.c.

Weight of this volume of thermometer-glass | 2
BOGLX2512:. 5. as oe ee ee
Weight of contained Hg 0°5795x136 . =7°881

Water-value of immersed glass 9°324x0°2 =1'840
Water-value of immersed Hg 7°881 x 0:038385=0°2639

|

|

| :

ha water-value of the immersed part of \ 21039

e thermometer . cc. 9? aa

| To test the reliability of the method, I sacrificed the ther-

| mometer by cutting off the stem at the level at which it was
immersed in the water of the calorimeter, and weighed the
quantities of glass and mercury: there were9°63and 7°64 grms.
respectively, corresponding to a water-value of 2°182, which
ascribes to the above method an error of 3°7 per cent. in the
required correction. ‘This negligible error is due to the volume

of the thermometer—consisting, not only of glass and mercury,

but also of the unfilled bore of the tube—and to slight deviation
of the specific gravity of the thermometer-glass from the mean
specific gravity used in the calculation. It may be just worth
pointing out that when a fragment of the thermometer-tube
is obtained from the maker with the instrument, these errors
may be avoided and the true water-value obtained. Probably
a greater error than the above is introduced into ordinary
specific-heat determinations by the evaporation of the water
in the calorimeter. .

II. A second method for the determination of the water-
value of the immersed part of a thermometer may be employed,
but it requires that the diameter of the bore of the stem be
known. To determine this a fragment of the same thermo-
meter-tube may be obtained from the maker of the instrument,
or a short piece of the upper part of the thermometer may be
very easily drawn-off before the blowpipe-flame, the closed
end of the fragment cut off, and the diameter of the bore
measured with a microscope furnished with a micrometer

On Calculating Surface- Tensions. 51

eye-piece and staye-micrometer. Then the weight of mercury
(w), corresponding to an increase of temperature from 0° to
100° C,, is given by
ee dae ce ee wots SE SR
where a is the area of the tube and / the length in centimetres
of 100° (¢°) on the scale of the thermometer.
The weight (W) of mercury in the bulb is then

Ww
Wr te, . ° “ . e : (2)

where a is the coefficient of apparent expansion of mercury
in glass. For lead-glass «=0°000155, and for soda-glass
0:0001586. Having thus determined the quantity of mercury,
the quantity of glass may be found as in the first method.
With a thermometer containing a known weight of mercury,
this method gave a water-value for the immersed part of the
instrument which was 3:2 per cent. in error. I place much
less reliance upon this than upon the former method, as the
apparent coefficient of expansion of glass varies more than its
specific gravity. The first method is quick and simple, and
the results so excellent, that this second method may be almost
regarded as unnecessary; but should any case arise in which it
is found to possess advantages over the first, it might be possible
to determine the coefficient by observing the increase in length
of the thermometer when heated in steam.
University College, Liverpool.

VIII. On the Error involved in Professor Quincke’s Method
of Calculating Surface-Tensions from the Dimensions of Flat
Drops and Bubbles. By A. M. Wortuineton, JA.,
Clifton, Bristol.

‘N one of a series of well-known papers (Pogg. Annal.
vol. exxxix. part 1; and Phil. Mag. April 1871) Prof.
Quincke has recorded a large number of measures of flat
drops and bubbles, from which he has deduced the value of
the tensions, not only at the free surface of liquids, but also
at the common surface of two liquids in contact.

The numerical results obtained exceed very appreciably
the values of the surface-tensions deduced from observations
with capillary tubes, and Prof. Quincke attributes the diffe-
rence partly to the exposure of the surface of the meniscus
in capillary tubes to atmospheric impurities, but chiefly to
the fact that with capillary tubes the edge-angle is not zero,
and that the quantity measured is not the surface-tension of

* Communicated by the Physical Society : read June 13th, 1885.
Seas | y-

—~

52 Mr. A. M. Worthington on Calculating Surface- Tensions

the liquid, but the surface-tension multiplied by the cosine of
the edge-angle. Thus a special significance has been attached
to the high values obtained by the method of flat drops or
bubbles, and these values have been widely copied and made
the basis of numerical calculations.

The calculation of the results involves the integration of
the equation to the liquid surface |

1 eae:
Ty + a} = Dz.

Prof. Quincke starts by writing Sail 1. € by assuming

that the drop or bubble may be treated not only as flat at the
vertex, but also as having an infinite diameter. It is true that
in § 5 of the paper referred to he himself comments on the
fact that these assumptions are only approximately correct ;
but he does not attempt to show that the error entailed by
these two assumptions is insignificant. It is the object of
this paper to show that the error is very considerable, amount-
ing in most cases to as much as 10 per cent. of the whole
value, and that when duly corrected the values obtained by
this method do not appreciably exceed those obtained with
capillary tubes.

Prof. Quincke’s process consists in measuring by means of
a cathetometer-microscope the dimensions of a large drop or
bubble, such as is represented in section in figs. 1 and 2, that
has been placed on, or below, a horizontal glass plate.

Fig. 1. Fig. 2.

ie

The dimensions measured are

(1) The distance K of the vertex from the plate.

(2) The distance & of the section of maximum radius from
the plate.

(3) The maximum diameter AB.

The first two measures afford the value K—& of the dis-
tance between the vertex and the section of maximum radius,
and the last affords the value L of the maximum radius.

The physical meaning of the assumptions made in the
calculation may be explained in the following manner.

Imagine a central slab to be cut out of the drop between
two parallel vertical planes at small unit distance apart, and
that the slab is again cut in half at right angles to its length,

from the Dimensions of Flat Drops and Bubbles. 53

so that we realize the portion ABCDE'E of the diagram
(fig. 3). Fig. 3.

If the drop were of infinite radius, and therefore flat at the
top, the slab thus obtained would be equivalent to a similar
slab cut out of a mass of liquid, shaped as in fig. 4. If we

Fig, 4.

now, on this supposition, consider the equilibrium of the

_ mass represented in fig. 3, with reference to horizontal forces

parallel to its length, we can equate the hydrostatic pressure
on the rectangular end to the sum of the tensions exerted
along the two edges AD and BC. ‘Thus writing, with
Quincke, AB=K, and writing D for the difference between
the density of the drop or bubble and that of the surrounding
medium, K2D

T+T COs oa 3

whence, when 0=0°, ; KD

| 4 °
Or, if we consider the equilibrium of that portion only of the
slab which lies above the horizontal section of greatest area
(see fig. 5), we may equate T to the hydrostatic pressure on
the rectangular area of unit breadth and depth (K—A),
Fig. 5.

whence oS pe a se OE

ee pen ng re es

ES A SG el -- —— mutes Pag Cm: ie
$A LA TE

_ mentary depth dz is

54 Mr. A. M. Worthington on Calculating Surface- Tensions

It is from this equation that Prof. Quincke calculates the
value of T. It is evident that in neglecting the curvature of
the vertex we are neglecting the pressure due to this curva-
ture transmitted to the whole area K—k. Thus, if 0 be the
radius of curvature in question, the pressure disregarded is

=(K-2). The surface-tension T has to balance this as well

as the hydrostatic pressure due to the weight of the liquid,
and neglect of this term will lead to too small a result.

Again, in neglecting the curvature in the plane at right
angles to the plane of the diagram, we evidently leave out of
account the tension exerted along each edge AH, DE’, of the

slab, which produces a pressure Ry on each unit area of the
surface.

Since the surface is one of revolution, RY is the length of
the normal intercepted by the axis, and writing ¢ for the
inclination of the normal to the axis measured on the side of
the vertex, z.¢. for the edge-angle of the drop- or bubble-
forming fluid at any horizontal section, we have - =
where x is the horizontal radius of the section ; and the
pressure on a horizontal strip of the rectangular end of ele-
Tsing dz _. 7
wae and the total action omitted,

K-k |
{ T sin ¢ dz .
0 M 3
so that the complete equation is

ARR) T Ag
spe aes sid eae { sin ¢ dz __ <a

v

p(k -kyD
= oe

The value of the integral of the last term to a first approx-
imation is shown by Laplace (Mée. Céleste, livre x. 2° Suppl.
p. 483), or by Mathieu (Théorie de la Capillarité, p. 137) to be

1—cos? 2
da Sint

a Bi LL 2
where pare

When 2 is equal to the maximum radius L, then ¢=90°, and

from the Dimensions of Flat Drops and Bubbles. 55

the term becomes

at, (3)
We may use for (a) in this corrective term the value given
by the approximate equation (1),
T ene (K S k)’D .
9 9
whence

Thus the term in question reduces to

K—&

21 S988L °

and the complete equation (2; becomes

(K—k)? 2T(K—k) .
2

== 3-989 ”

D+2(K-2)—

or
. ay : Meee
Be +5 ee ae ae ahear)

Or, writing —C for the value of the factor
the corrective term,

G Be ) in
6 3282 L
q — (K—k)? ; (3)
D 3(1+2K—&C) ;
To find the value of C we must know that of F This is

shown by Laplace (loc. cit. p. 485) or Mathieu (doe. cit. p. 140)
to be equal to

x iP
By Avg pees pee
2/207? Wore tan uy a +5

t
which, when e=L and 6=90°, reduces to
L
i oak. x 4142136 x @ a

b

K—k
in which we may use for a the mean value of 5 found by

Prof. Quincke.
Before giving the numerical results, it is necessary to

56 Mr. A. M. Worthington on Calculating Surface-Tensions
z=K—k
observe that the value of the integral ( —— is calculated

2=0

on the assumption that the fraction = is small; and, again, the
value of : is calculated on the assumption that = is small,
even when ~ is large, i. e. that the drop is very flat, even far
from the vertex. Now the drop or bubble used by Prof.
Quincke was often far from satisfying these conditions. In
some cases the diameter was so small that the fraction 1
amounted to as much as 4, and the curvature at the vertex
was so considerable that the value of : cannot be satisfactorily
calculated in the way described. Thus in the first measure
of an air-bubble in water (Quincke, loc. cit. § 3, table IIL.)

“ —-319 ; while in the case of drops of petroleum in water it

L
is as much as *3768.

I have therefore selected for correction only those measures
for which the value of r was smallest, though the mean value
of the observed quantity K—£ used in calculating was

generally taken from the whole of the measures given by Prof.
(Juincke.

Air-bubbles in Hyposulphite of Soda.

D=1°1248.
: Grams per cm.
1
No.| L. |K—é& BYE 7
3282 L b S. tension S. tension,
(Quincke). corrected.
3. | 1:41 |°3738 ‘07856
‘2161 0492 06853
4, » | 0062 ‘07541
Mean (of 4 measures) ............... ‘07903
By capillary tubes (Quincke) ...... ‘07636

(The mean of K—Z used to calculate 2 was ‘37485.)

from the Dimensions of Flat Drops and Bubbles. 57

Air-bubbles in Distilled Water.
D=1. temp, 25°. C.

Grams per cm.

1 1
No.|} L. |K—4.| — =
3-282 L e S. tension S. tension,.
(Quincke). corrected.
1.| ‘9 |-4112]| -33885 | -3104 ‘08455 08263 Rejected.
2. | 1°38 |-4069 | -2208 | ‘0723 ‘08280 ‘07386
&. | 1:05 | 3972 | -2902 | 1989 ‘07905 07355
4. | 1:095| 4043 | 2783 | 1736 ‘08170 07535
5. | 100 |-4011 | 3047 | -2310 08040 ‘07595
6. | 1:04 | 4225 -2930 | -2049 ‘08920 — 08307
7. | 1:535|-4000 | -1985 | -0443 ‘08000 ‘07122
Ly 5) re 08253 ‘07550
By capillary tubes (Quincke) ...—...... ‘07235 at temp. 16°-2.
If we reject also Nos. (3), ©), and (6) on account of
the largeness of the fraction a the mean result is... °07348
The exps. of M. Wolf (see Terquem’s Capillarits
IIE Ric c52- 5% 03 crees akc ostvncs ne cecedssesecaes tse ‘07345 at temp. 25° O.

Air-bubbles in Bisulphide of Carbon.
D=1-2687. Temp. 25° C.

- Grams per cm,

1 1
Wo; £. |K
3282 L - S. tension S. tension,
(Quincke). | corrected.
1. | 1:25 | -2180 2438 0093 03015 027351
2. | 1°35 | -2230 ‘2226 0046 ‘03157 ‘0287
3 | 175 | °2399 “1741 0 ‘03651 03369
Ris ota niah« Sie. samara damn sadenducsord<- 03274 0299
By capillary tubes (temp. 18°) (Quincke)...} —...... 03343

58 Mr. A. M. Worthington on Calculating Surface- Tensions ; |

Air-bubbles in Olive-oil.
D=-9136.: “Temp. Bos C;

Grams per cm.

1 |
No.| L K—+k c )
8-282 L b S. tension S. tension,
(Quincke). | corrected.
cm. cm. 7 Ny
a “3001 2081 "0095 04113 "03678
me) Lf "2858 "1792 003877 =| )~— 08735 "03387
Se ey | "2885 5 a ‘03804 03453
4. 11:56 | -2905 "1953 00719 03850 °03470
5. | 1:56 | ‘2814 - p C3617 03271
6. | 1:485 | 2751 "2052 701041 ‘03457 03122
7. | 1-485 | -2861 r ss 03741 ‘03364
DROIT och iT eey cc nO Sones acne emaneE Oe hee ‘03760 03392

By capillary tubes (temp. 22°) (Quincke). $4000 ‘03271

Air-bubbles in Oil of Turpentine.
D="8861, Temp 207 1,

Grams per cm.

8. tension S. tension,
(Quincke). corrected.

ee | Se

cm. | cm.
3. | 1:34 |°2665

‘03149
"2274 01199 02818
4. »» | 2659 ‘03134
Mean (of 8 measures) ........csisccccscsonsseede 03033 02818
By capillary tubes (temp. 21°°7) (Quincke) .| —_...... 02765

(The mean value of K—Z used in the calculation of : is ‘2615 cm.)

from the Dimensions of Flat Drops and Bubbles. 59

Air-bubbles in Petroleum.
D=0:7977. | Temp. 24°-2 C.

Grams per cm.

1 1
No.| L. |K—4é : =.
3282 1 b S. tension 8. tension,
(Quincke). corrected.
em.
1. | 1°475 9838 ‘2066 01033 03212 02890
2 |
ie | 2031 0090 03260 02934
3 | , | °2858
Mean (of 3 measures) .........ees0e0: deskay osy bi 03244 02912

By capillary tubes (temp. 22°°3)...sc0s--..0000| eeeeee ‘02566

Air-bubbles in Absolute Alcohol.
D="(906 “Temp. 26°38 GC.

Grams per cm.

1 1
a 6 eee Ae
3282 Li b S. tension S. tension,
(Quincke). corrected.
em, cm
J. | L57 | 2582 || 1959 00326 025383 || 99835
|, | 2560 02591 «| f
8. | 1-415 |-2611 |- 02695
a |. [apg |} 2158 0074 | eile } 02417
5. |1-41 | -2539 | 02548 .
Gia .| 2560 |, > " { 02590 } U2324
7. |1-44 |-2573 17, 02616
loro |} “216 0063 { pee \ 02383
02599 02367
By capillary tubes Mi ee Oe CCOUINCKE) cote osl ak cae 02237

a s eee NG ges Ce ehh er 02365

60 Mr. A. M. Worthington on Calculating Surface- Tensions

Mercury in Air.
D=13:'5432. Temp. 20° C.

Grams per cm.

1 df

= ages iepali 3-282 L’ b° SS. tension | S. tension,
(Quincke). corrected.
6. | 1-625] 2839] 1875 00534 5456 4946
fe 17 ‘2861 "1792 00377 5546 5037
8. | 165 | 2822} (C calculated by propor- "5624 5102
tional parts.) ;
Mean of 8 measures ... 5503 5028

(The mean of K—£ used to calculate : was ‘2863.) -

Bisulphide of Carbon in Water.
D=1:2687—1.

Grams per cm.

1 1 eee
: NV 2
oe ee 3°282 L b° S. tension S. tension,
(Quincke). corrected.
1. | 1275} -5510 ( ‘04069
2. * ‘5478 "2389 "2082 ‘04021 ‘03872
a sao || -03893
4. | 1°3855 | °5830 f AYR 04555 é
| ueelepsagty 2240 1753 {] h4sog |} 04802
Mean of 6 measures...| °04256 04086

(The mean of K—A used to calculate ; was °5583.)
Chloroform in Water.

Grams per cm.
No.| L ea i : 3
, 3282 L b S. tension | S. tension,
(Quincke). corrected.
em. | cm.

4.|} 18 | 3450 02902

1693 ‘0089 ‘026296
5 ‘ 3472 02941 |
Mean of 5 measures ... "03010

(The mean of K—£ used to calculate : was 35076.)

from the Dimensions of Flat Drops and Bubbles. 61
Olive-oil in Water.

D=1—0:9136.
Grams per cm.
N L. |K—& ! :
yar OS =k. 5.9R9 7," 7
3°282 L Y S. tension S. tension,
(Quincke). corrected.
em. em.
6. | 2°37 | 6941 | 02082
"1286 04185 701818
7. 2 °6785 701988
Mean of 7 measures ... "02096

(The mean of K— used to calculate : was °69648. )

Turpentine in Water.

D=1—0°8867.
Grams per em..,
N L. |K-& : :
Oo. , —k. ~ =.
3-282 L b S. tension S. tension,
(Quincke). corrected.
cm. | cm. }
9. | 1:15 | -4610 2649 "1915 01200 01127
Mean of 9 measures ... 01177

(The mean of K—£ used to calculate ; was *4562.)

Mercury in Aqueous Solution of Hyposulphite of Soda.
D=13°543 —1:1248.

Grams per cm.

7 1 1
No L. K—k. a.9000 T°’ FZ
3282 L b S. tension | S. tension,
(Quincke). corrected.
em. | cm.
4. | 1365} -2836 4994
5 5 2730 2232 01405 4628 "4208
6 og 2680 “4459
Mean of 6 measures ... “45107

(The mean of K—£ used to calculate 2 was °2749. )

62 Mr. A. M. Worthington on Calculating Surface-Tensions :

Mercury in Water.

D=138'543—1.

Grams per cm.

1 ft:
No.| L. | K—Aé. GY-O NN ie F
3°282 L b S. tension S. tension,
(Quincke). corrected.
em. | om.
B.D | G0 | ‘4110
4. . ‘2605 A ‘ , “4256 i
7 7 "2527 "4005
Mean of 7 measures...| °4258

(The mean of K—£ used to caleulate ; was ‘26.)

Mercury in Olive-oil.

D=13°543—0°9136.

Grains per cm.

1 1
No L. K—z Wa ye ae
3282 L b S. tension S. tension,
(Quincke), corrected.
cm. cm.
&. | 1:45 | :2862 05238 ,
"21014 ‘00806 31066
4. 2 *2282
Mean of 6 measures ... 3419

(The mean of K~—£ used to calculate ; was ‘2322. )

from the Dimensions of Flat Drops and Bubbles. 63

Mercury in Oil of Turpentine.

D=13°543 —0°8867.

Grams per cm.

S. tension S. tension,
(Quincke). || corrected.

"2449
"2669 2415

2642

1°815| -1970 ‘20313 "2456 "2304

"2554 236

(The mean of K—Z used to calculate : was *2009.)

| Mercury in Petroleum.

D=18°543—0°7977.

Grams per cm.

Za
°
cg
A
!
= :
oe
S
GO
No
ee
ole

ea
. S. tension | S. tension,
(Quincke). corrected.
cm, | cm. :
19 | -2010 3 "2574 ;
il sess \ -1604 0 { cag } 23896
15 | +2220 , ‘ , 8142 ;
: ara 20813 002 {| Bus |} germ
Mean ... ‘2861 ‘2653

(The mean of K—& used to caleulate ; was ‘21155.)

64 Mr. A. M. Worthington on Calculating Surface-Tensions ;

Mercury in Alcohol.
D=13°543—0°7906.

Grams per cm.

No.| L. |K—A. 1 =
3°282 L S. tension | S. tension,
| (Quincke). corrected.
cm. | cm |

1. | 1:5 | :2443] 4 : 3805
2| , | -2574/! -20818 0037 4294 ‘3569.

a ,.. | pee 3740

Mean of 6 measures ... 4025

(The mean of K—£ used to calculate : was ‘248.)

Olive-oil in Alcohol.
D=:9136—:7906.

Grams per cm.

7 :
No.| L. |K-4&.| 3-989T," 5° S. tension | S. tension,
(Quincke). corrected.
33-4 4:405 151975 , ‘00240
21686 0 00210
ac) ge eye” 00215
Mean of 4 measures...!... .. (00226 |

(The mean of K —£ used to calculate : was ‘1917.)

The corrected results with hyposulphite of soda, bisulphide
of carbon, and petroleum do not agree satisfactorily with the
measures made with capillary tubes. |

The general effect of the correction is to reduce the results
by about 10 per cent., but in some cases, especially of drops
of one liquid in another, the value of L is so small that the
accuracy of the correction may be questioned. I hope soon
to be able to lay before the Society fresh determinations of
the surface-tension in these cases. When L reaches or

from the Dimensions of Flat Drops and Sap bhes: 65

approaches 2 centims. the term : is generally insignificant,

though the corrective factor C, which now reduces to 3-989 1)’
is still important. In fact a drop or bubble may be con-
sidered flat before the radius can be regarded as indefinitely
reat.
P In those cases where the corrected value still exceeds the
mean value obtained by the method of capillary tubes, the dif-
ference which was previously considerable is now for the most
part insignificant ; witness water, olive-oil, turpentine, and
alcohol. It must also be remarked that the measures of flat
drops or bubbles agree among themselves far less satisfactorily
than the measures of capillary elevations. ‘Thus the numbers
quoted as obtained with capillary tubes are themselves the
mean of several observations, not differing from each other
as a rule by more than 2 per cent., though made with tubes
of various diameters; while the observations of flat drops or
bubbles differ often by as much as 15 per cent., or even
more.

This very variation is indeed a matter of interest, and not
easily accounted for*. For though M. van der Mensbriigghe
has well pointed out + that the bubble when first blown must
exhibit a higher surface-tension, owing to the absorption of
heat from, and consequent lowering of temperature of, the
surface-layers, yet it is not easy to believe that this deviation
from the normal value would be of long persistence ; nor
indeed do Prof. Quincke’s measures always show a diminution
of tension with the time.

I think, however, that we may draw the conclusion that
the method of capillary tubes, when care is taken thoroughly
to wet the walls above the meniscus, leads to values which
are not discredited, as we had been led to think, by the results
of the method of flat drops or bubbles. ‘This is a satisfactory
conclusion to come to, since the method of elevation in capil-
lary tubes is that which has been most frequently employed
for measuring the surface-tension. Prof. Quincke’s value
(0825 gr.=810C.G.S8. units) of the surface-tension of water

* Mr. Bashforth (see ‘An Attempt to test the Theories of Capillary
Action,’ p. 10) remarks on the difficulty of making any accurate direct
measurement of the height K—z, and Prof. Guthrie has somewhere
noticed the same difficulty. But Prof. Quincke’s measures of K—& do
not show a percentage-variation much greater than that of the quantity
K¥ 2, in determining which the same difficulties do not occur.

+ ‘‘ Etudes sur les variations d’énergie potentielle des surfaces liquides,”
Mém. de l Acad. de Belgique, t. xliii. 1878.

Phil. Mag. 8. 5. Vol. 20. No. 122. July 1885. Fr

66 On Calculating Surface- Tensions.

has been employed by Lord Rayleigh * to calculate the wave- |

length in an oscillating water-jet, and leads, as he himself

points out, to too low a result. Had he made use of the cor- _
rected value (T='0735 gr.=72'1 C.G.S8. units), the results of
p- 82, loc. cit., would have been in very close agreement with
theory.

In conclusion, I would remark that the great value of this
particular paper of Prof. Quincke’s does not, as it seems to
me, depend so much on the absolute accuracy of the results
presented as on the emphasis with which, at the time it was
written, it represented the capillary constant to be measured
as a surface-tension, and on the justice of the general con-
siderations which are put forward, and which remain un-
impeached. |

A. M. WortTHINGTON.
Clifton, Bristol, June 2, 1885.

P.S. For the benefit of readers of Maxwell’s ‘ Theory
of Heat,’ in which Prof. Quincke’s results are given in a
synoptic table, I here reproduce the table with the corrected
values. ‘Those in brackets J have been unable to check. It
must be borne in mind that the numbers in the second column
of tensions are still somewhat uncertain.

The values are given in grammes weight per linear metre.

Superficial Tensions at about 20° C.

Liquid. Sp c cif - Tension separating the liquid from
otal Air. Water. | Mercury.
WY GRE Mia cusee ten teh vies chicas 1:0 7°35 0 23-44
MGrOur ys .c5 .0 sania veenssens 13548 50°28 38°44 0
Bisulphide of Carbon ...... 1:2687 2:99 4:09 [37-97]
POhlorotornt | ..5<.++0.+.s60+ 00. 1:4878 [3°12] 2:63 [40-71]
RMN ancy sansenin rane 07906 [2°36] Aer 35-69
Oly e-O1L. icc cesecceseecesvows 0:9136 3°39 1:82 31:07
Turpentine ...s..s.seseeeree. 0:8867 2°82 1:13 23°60
PV OIOGIN 65 dy sencnians's ones G0 0°7977 2-91 [2:83] 26°53
Hydrochloric acid oa i ag [7°15] ey - [88°41]
| Solution of Hyposulp ve} orcad babe «| .
Me dilns soa. Mhsowses T1246) 686 a 42°08

* “On the Capillary Phenomena of Jets,” Proc. Roy. Soc. no. 196
(1879).

preven

IX. The Stream-lines of Moving Vortex-rings. By OLIVER
Lopez D,S¢e., Professor of Physics in University College,

4 *
Liverpool™. [Plates IL., IIL., IV.]

6 gd object of the present communication is to publish

drawings of vortex stream-lines, some of which I made
originally for my own edification. ‘Taking the lines of a
stationary vortex, as given by Sir W. Thomson in his memoir
on Vortex Motion (Trans. Roy. Soc. Edinb. vol. xxv.), or as
copied into Maxwell’s ‘ Hlectricity ’ (plate 18, vol. i1.), I merely
superpose uniform motion upon them, in the shape of a series
of parallel lines, and join up the corners of the quadrangles
so formed.

Another way of expressing the matter is to say that you
draw the lines of magnetic induction due to a circular ring
conveying a current, placed in a uniform magnetic field with
its lines exactly opposed to those inside the ring.

I choose two strengths of uniform field for the sake of
illustration; one distinctly stronger, the other distinctly weaker,
than the central intensity due to the coil alone. ‘The relative
intensities at. centre of ring due to field and coil respectively
are about as 1 to 5 in fig. 1, and as 64 to 5 in fig. 2 (Plate II.).
Or, taking the curves as representing stream-lines: in fig. 1
the velocity of vortex-motion is equal to the translation-
velocity of the whole ring at a certain circle in its plane
concentric with its core and of 3°3 times the diameter of the
core, and also at two points on the axis; while in fig. 2 the
vortex-velocity and the translation-velocity are equal at a place
1°5 core-radii distant from the centre of the ring outside, and »
at another circle, say two fifths the core dimensions, inside,
the ring. ; ,

In fig. 1 the ring is moving so fast that the translational
flow back of fluid through its centre overpowers the forward
vortex-motion there. In fig. 2 the vortex-motion predo-
minates as far as a point on the axis which I reckon as 1°38
core-radii distant from centre of ring, a point indicated by the
crossing of the partially dotted stream-line. It will be under-
stood that though they look so different, the two Plates repre-
sent the same ring moving at different speeds. The size of
the core or circular axis is the same in both diagrams.

It will be observed that in fig. 2 the portion of fluid
permanently partitioned off from the rest by reason of its
vorticity is truly ring-shaped, and would become thinner or
more wiry if its forward motion were greater—the lines near
the core of the ring being prolate towards the axis; while
in fig. 1 the rotational portion of fiuid, which is being bodily

* Communicated by the sige: Society : read June 27, 1880,
2

68 Prof, Oliver Lodge on the Stream-lines

translated through the rest, forms an ovoid mass with dimples
before and behind-—the dimples, however, becoming less and —
disappearing when the translatory motion is made still slower. —
The lines near the core are in this case rather displaced away —

from the axis. The dottedness of the portion of the line which
crosses the axis of the ring is purely subjective, and only
indicates uncertainty on my part as to its exact course, from
want of knowledge. It is probable that the same defect ex-
hibits itself in my terminology, which is probably incorrect,
or at least unusual. Thus I cannot help calling the actual
circular axis of the ring its “‘core,’’ instead of the whole of
the rotational portion, as is usual in dealing with rings of very
small cross section in proportion to area of ring itself. The
rings drawn are not of small cross section, and so one wants a
name for their innermost axis or core.

We can try to apply Sir William Thomson’s rule* for the
velocity of translation of very thin or high-speed rings, to the
case of fig. 2; though this ring is not nearly thin enough for
the formula to be properly applicable. |

Using the symbol X for the ratio of radius of ring itself to
its cross-section radius, the rule may be written:—

27 X velocity of translation
vortex velocity at centre of ring
‘ia 2r x velocity of translation

vortex velocity at surface of rotational portion

In fig. 2 the value of X is about 38; and accordingly each
of the above terms is about 3 also; or the two vortex velo-
cities specified in the formula are nearly equal, and about
double that of the translational velocity.

This does not agree with what I said before, about the ratio
of uniform field to ring-field at centre being as 64:5; hence
there is something wrong, but I don’t know what. The lines

of uniform velocity in fig. 2 were taken 8 times as close

together as in fig. 1; and this surely represents a velocity
64 times as great. I can only suppose that the ring is much
too fat for the formula. :

Plate ILI. is an attempt to represent a vortex-ring advan-
cing In a very imperfect or viscous fluid, showing its gradual
increase in size, and decrease in forward velocity. It is easily
drawn by superposing a diverging equiangular pencil on the
stationary vortex which forms the basis of all three diagrams ;
but that it really represents the effect of viscosity does not
seem very probable. No slip, due to inertia of displaced fluid,
is shown in any of the diagrams. This figure better repre-
sents a ring moving towards a large distant obstacle.- As

* Phil. Mag. June 1867, xxxiii. p. 511,

=log (8\—4).

of Moving Vortex-rings. 69

drawn, the vortex-velocity at centre is only a trifle greater
than the translational velocity. This plate also represents
the lines of magnetic force due to a circular current with a
repellent pole on its axis, at a point 2°518 diameters away

from the plane of the circle. The dots on the curves indicate

the distribution of the crossing-points which guide the drawing.

Plate IV. shows the attempt of a ring to advance in an
oblique direction, not normal to its plane. It is supposed to
have been knocked out of a hole by a slant impulse. There
is evidently a good deal of vibration, both of the ring as a
whole and of its cross section; and it looks as though a very

little would suffice to break it up altogether. The resultant

velocity at the centre of the ring happens, in the particular
case here chosen, to be about zero.

In cases of oblique progression a tendency to a bodily shifting
of the uniform flow-lines, parallel to themselves, as they pass
from before to behind the ring, is noticeable, and is exhibited
in fig. 1. Perhaps this means a heaving or sinuous path of
motion for the ring. The right mode of joining up the
guiding-points is however in this case by no means obvious ;
and fig. 2 (Plate IV.) is just as likely to be correct as fig. 1.
In fig. 2 no shifting of distant stream-lines occurs, but then
it hardly seems a real case of vortex-motion: at least it looks
only like a ring shaking itself to pieces; while fig. 1 suggests
an attempt of the same ring to pull itself together.

I have.anumber of other diagrams drawn in the rough, in-
dicating various features of the clash or chase of vortex-rings.

The direct clash of two equal opposite rings, or the impact
of one against a looking-glass, is of course very easy. The
clash of two rings of different strength is more complex—one
appears to be opened out over the other.

The chase of two unegual rings, and the penetration of the
front one by its pursuer, are well shown ; but if the rings are
of equal strength they refuse to penetrate, and seem to amal-
gamate or pair, no matter at what different speeds they may
be going.

The deflection of one ring by another whose path is inclined
to it, as calculated by Prof. J. J. Thomson in his ‘ Adams
Essay,’* can also be illustrated, together with what I think
corresponds to vibrations of the core about the circular form.

But all these diagrams I propose to publish in a more
complete form later. This “ experimental” method of investi-
gation, by diagrams based on simple superposition of velocities,
seems capable of great extension, because one is limited by no
approximations or conditions: the only difficulty is the inter-
pretation of results.

* See also Phil. Trans, ii, 1882.

70 Notices respecting New Books.

Last year I examined air vortex-rings produced in the
well-known manner described by Prof. Tait*,in the light of a
powerful intermittent induction-coil Leyden-jar discharge.

‘The motion is, however, of too continuous a nature to exhibit

the advantages of this mode of illumination ; and though the
crispations of vibrating rings are well shown, there is no
obvious. peculiarity noticeable which does not show itself
equally well in a steady illumination.

This paper is only to be regarded as a preliminary note,
and, as Prof. Carey Foster has kindly reminded me, the uni-
form field as I have drawn itis not quite correct. This indeed
will account for the discrepancy between theory and expe-
riment mentioned above.

The appearance of jets of water illuminated intermittently
is, as is well known, very striking; and I have long imagined
that a waterfall illuminated in this way would be a striking
spectacle. The spark is scarcely bright enough for large-scale
illumination, though there is nothing to beat it for instanta-
neousness. A revolving slit-disk would, however, prove a more
manageable and less noisy method, and by a judicious arrange-
ment of special lenses it can be made to give plenty of light.
But the speed of the disk must be high, and its slits narrow,
or the drops will be blurred and their characteristic statical
beauty lost.

X. Notices respecting New Books.

Geschichte der Elektrizitat. Von Dr. Epm. Hoppn.
Leipzig: J. A. Barth (pp. 620).
HE work before us may be characterized as encyclopedic. It
is one of those valuable contributions to scientific literature
of which we owe so many to the laborious research of our German

friends, but which, for some reason, are found in English, for the

most part, only in the form of translations. But Dr. Hoppe’s
book is not only a valuable collection and résumé of all that has
been done in Electricity from the earliest times down to the pre-
sent date, but it is also a veritable romance, in which the story of
discovery in this particular science is told in a most fascinating
manner. We are nowadays so familiar with the achievements
of science, and so accustomed to see the giant of Electricity tamed
and made to serve the purposes of every-day life, that we are
apt to overlook the difficulties with which the earlier investigators
had to contend; and it is just here that the charm of such a
history as the one before us lies,—that, being familiar with the
results of which the original discoverers were in search, we are
able to survey their labours, to trace where, having hit upon the
right track, they have pursued the truth, till their labours have
been crowned with success, and to admire the skill with which
they have overcome the difficulties in their way.

* “Recent Advances, p. 292.

| Ae
+ «
yar t,
a THATS

Notices respecting New Books. 71

Dr. Hoppe divides his work into six Books, which deal with
different epochs in the history of Electricity, and which are, of
course, of very unequal bulk. The first gives us the history of
discovery from the earliest times to Franklin, and occupies 28 pages.
The second embraces the times of Franklin and Coulomb, 1747 to

-1789,.70 pages. Book III. gives us the history from Galvani’s

discovery to the year 1819, treating chiefly of galvanic electricity
and the discoveries of Volta, Davy, Oerstedt, Zamboni, and
others, occupying 73 pages. Book IV. treats of the connection
between Electricity and Magnetism, and of the contributions to

electrical science of Oerstedt, Poggendorff, Schweigger, Ampére,

Faraday, and Nobili, in the years 1820-26, and covers 60 pages.
Book VY. extends from Ohm to the recognition of the law of conser-
vation of energy (1827-47, 260 pages), including, amongst others,
the researches of Ohm, Pouillet, Kohlrausch, Faraday, Schonbein,
Daniell, Grove, Bunsen, Poggendorff, Planté, Faure, Joule, Lenz,
Peltier, Kirchhoff, Gauss, Wheatstone, Weber; and Book VI.
treats of the technical applications of Electricity—of the electric
light, dynamo machines, the electric telegraph, and telephone, &c. :
this occupies 93 pages.

Jt is not practicable, within the limits of a brief notice, to do
more than mention one or two points which may be of interest.
Dr. Hoppe discusses the question of the discovery of the Leyden
jar at some length, and, apparently, gives his verdict in favour of
Kleist as the actual discoverer; at least he seems to adopt the
name ‘ Kleist’s jar” instead of the usual term. Von Kleist, Bishop
of Kammin, in Pomerania, on the 11th October, 1745, placed a
nail in a medicine-glass and held it to the conductor of an electrical
machine; on touching the nail with the other hand he received a
shock, especially if the glass contained mercury; and this appears
to be the first time the experiment was ever made. This result
was communicated by letter to at least three people, in Berlin,
Halle, and Danzig respectively, in November and December of the
Same year. . |

Jn January 1746 a similar observation was made, accidentally,
by Cunzeus, in Leyden, and repeated by Musschenbroek, Professor
of Mathematics and Physics at the University of Leyden, and
Gallamand, Professor of Philosophy at Leyden, who communicated
the discovery to the famous Abbé Nollet, in Paris, who, knowing
nothing of Von Kleist, accredited the discovery to the Leyden pro-
fessors. It would appear, then, to be due to the accident of the
general ignorance of the German language, in comparison to the
widespread use of the French tongue, that this most important
discovery has been connected with Leyden and not with the name

of Von Kleist ; and there is some foundation for Dr. Hoppe’s com-

plaint, that whereas every schoolboy knows what a Leyden jar
means, perhaps not one ever heard the name of Von Kleist.

Some of our common appliances or observations seem to date
back further than is usually supposed. We note that the date of
the first lightning-rod is given by Dr. Hoppe as 1754, and that

72 Geological Society :—

it was erected by Procopius Divisch, although the suggestion was __
made by Franklin in 1749. The first observation of chemical de-
composition by means of the galvanic current is ascribed to Dr. |
Asch, of Oxford, in 1795. Humboldt, repeating the experiment,
first collected hydrogen from the galvanic decomposition of water; |
and Nicholson, in 1800, making the necessary substitution of
platinum electrodes for those of brass or copper, first collected
oxygen from this source. The first electro-plating was effected by
Ritter in 1800, who deposited copper by means of the galvanic
current. The first construction of a secondary battery is assigned
to Gautherot, in 1802; and the first construction of a dry pile,
with which the name of Zamboni is usually associated, is shown to
rest with Behrens, who constructed such a pile in 1803, seven
years earlier than Zamboni.

Respecting the rediscovery of Ohm’s Law by Pouillet, Dr.
Hoppe writes :—‘‘ Pouillet does not say whether or not he was pre-
viously acquainted with Ohm’s law; the attempt has since been
made to claim the priority of the discovery for Pouillet, which is
the more absurd since it has not yet been determined that Pouillet
was not acquainted with the researches of Ohm and Fechner.
For, as I have already had occasion to remark, when Englishmen ~
or Frenchmen do not cite a German work, that is no reason for
supposing that they are not acquainted with it. Since Pouillet had
been for a long time in scientific correspondence with Poggendorff,
as, indeed, appears from the abstracts of his work, which he him-
self prepared for Poggendorff’s Annalen, it appears to me very
probable that Pouillet was acquainted with Ohm’s work, and that
his excellent experiments therefore, as, indeed, Poggendorff re-
marks in a note, were only a confirmation of Ohm’s theory.”

The fifth and sixth books, whick, of course, occupy the most
space, are particularly full and interesting, but space forbids de-
tailed reference. We merely mention the account given of the
introduction of the idea of ‘“ potential” into electrical theory, and
the history of the various forms of incandescent lamp. Dr. Hoppe
does not seem disposed to credit Edison with quite so much origi-
nality as he is generally supposed to have.

Dr. Hoppe has produced a most valuable and interesting work,
which will be cordially welcomed by those interested in the subject
who are acquainted with the German language. It is to be hoped
that an English version may be produced.

XI. Proceedings of Learned Societies.

GEOLOGICAL SOCIETY. |
[Continued from vol. xix. p. 515.] |
May 13, 1885.—Prof. T. G. Bonney, D.Sc., LL.D., F.R.S.,
President, in the Chair.
f iit following communications were read :— |
1. “On the Ostracoda of the Purbeck Formation; with Notes

on the Wealden Species.” By Prof. T. Rupert Jones, F.R.S., F.G.S.

North- Wales and Shrewsbury Coal-Fields. 73

2. “ Evidence of the Action of Land-ice at Great Crosby, Lanca-
shire.” By T. Mellard Reade, Esq., F.R.S.

The Author pointed out that the Triassic rocks under the Low-
level Boulder-clay in the neighbourhood of Liverpool, where they
are not smoothed and striated, are usually broken up into rubble
and red sand, forming a bed of variable thickness occasionally con-
solidated into a breccia. This deposit he had in former papers
attributed tothe action of land-ice. At Mowbrey brick-and-tile works,
Great Crosby, is a section of Keuper marls, the only one existing
for many miles around. ‘The marls are overlain by Low-level
Boulder-clay of the usual type, and between it and the marl is a
deposit from 3 to 4 feet thick, which at first sight is not readily
distinguishable from the marls, but which a careful examination
of the excavations from time to time as they progressed, showed
to be a distinct bed. In this bed, lying at all angles, were
found large blocks of sandstone, some of which were grooved and
striated in an unmistakable manner. ‘The matrix in which they
were imbedded was of the same constitution as the marl, and evi-
dently formed out of it, showing in places strong evidences of
contortion and kneading up. The sandstone blocks belonged to the
Keuper formation, and some of them were very similar to bands
intercalated in the marls near the bottom of the excavation. No
erratic pebbles or boulders of any sort were found in this kneaded-
up marl, whereas the Low-level Boulder-clay is full of them.

The Author considered that the only feasible explanation of the
phenomenon was that the marl had been worked up into a grey
clay by the passage over it of land-ice, which had broken off the
sandstone-bands at their outcrops, forcing the blocks into the dis-
turbed or worked-up marl. These outcrops, concealed by a mantle
of Low-level Boulder-clay, must be to the northward, and therefore
the blocks have travelled in the same direction approximately as
the track of the striations on the neighbouring rocks.

In conclusion, he contended that all the evidence points to the
fact before insisted upon, that the intensest period of cold preceded
the deposition of the Low-level Boulder-Clay, which is clearly a
marine deposit.

3. “The North-Wales and Shrewsbury Coal-fields.” By D.C.
Davies, Esq., F.G.S.

After discussing the origin of Coal-beds, and the causes of their
variation in structure and quality, the Author proceeded to describe
the North Wales and Shrewsbury Coal-field, which consists of three
parts :—(1) The Shrewsbury field south of the Severn, exclusively
composed of Upper Coal-measures; (2) the tracts north of the
Severn, extending from near Oswestry to north of Wrexham; and
(3) the Flintshire Coal-field. ‘The first and second are separated
from each other by the alluvial plain of the Severn and Vyrnwy,
and the second and third by the Great Bala and Yule faults.

Some remarks on the scenery of the Welsh border-land followed,
and then a general section of the Carboniferous system, as developed

74 Intelligence and Miscellaneous Articles.

in the country described, was given, the Permian beds being in- MH i

cluded, as the Author considered them the upper portion of one
great division of Palzozoic time. The section was as follows, aii,
the maximum thickness of each subdivision :—

Thickness in yards.
bark red Santetone...) 0.22... i62:.cydvi search 210
2. Ifton or St. Martin’s Coal-measures ......... 75
3. Red marls with calcareous matter ............ igo ¢ Pemae 580 yarste.
4. Green rocks and Conglomerates ............ 125
> Upper Coal-measures 1 .:...7s...5 daei dee esses 80 )
B.-ieim reek to Geln:doal.. sccps sve awesgemes ser 100 |
7. Cefn coal to Lower yard-coal...........-...006 - 270 | Cone
8. Lower yard-coal to Chwarcle coal ............ 80 | shi,
9. Chwarcle coal to Millstone Grit ..........0 135 )

1255 yards.

A detailed description of the strata was next given, beginning
with the lowest, together with details of each coal-seam as worked
in various parts of the field. After describing the beds from the
Millstone Grit to the Cefn rock in the North-Wales coal-field, the
Author proceeded to notice the Upper Coal-measures and Permian
strata in the Shrewsbury area, andshowed that no break exists between
the two, the former passing gradually into the latter. He then
discussed the probability of Lower Coal-measures existing beneath
the upper beds near Shrewsbury, and showed from sections that the
existence of the lower measures might be anticipated. A similar
inquiry as to the presence of the Coal-measures beneath the New
Red Sandstone of the Vale of Clwyd should also, in the Author’s
opinion, be answered in the affirmative.

The organic remains found in the different beds were briefly
noticed, and then the faults of the district were discussed at some
length. The principal faults run north and south, with an upthrow
to the east, but are crossed by lines of fracture running east and
west.

In conclusion, the correlation of the strata in the North Wales
and Shrewsbury coal-fields, and especially of the coal-seams, with
the beds found in other parts of Great Britain, was discussed, and
a section was given to show the representation of the different
measures in various coal-basins. The Author was disposed to adopt
four subdivisions rather than three only, as usually accepted, and

pointed out some of the characteristics of each subdivision.

XII. Intelligence and Miscellaneous Articles.

ON THE DEPTH TO WHICH DAYLIGHT PENTRATES IN SEA-
WATER. BY MM. H. FOL AND ED. SARASIN,

ys a previous paper * we gave an account of some experiments

which we had made in the Lake of Geneva, with a view to de-
termine the limit of penetration of sunlight in water, and we
announced our intention of making similar experiments in the sea.

* Phil. Mag. January 1885, p. 70.

Bh = — a

Intelligence and Miscellaneous Articles. (6)

Thanks to the kind assistance of Dr. J. Barrois, director of the
zoological station of Villefranche-sur-mer, the ‘ Albatros,’ despatch

-boat of the French Navy, was placed at our disposal for several

days. The intelligent and ready cooperation of the lieutenant of

the ‘ Aboiville,’ the commander of this vessel, and of all the officers,

greatly contributed to the success of these delicate experiments.

Our mode of proceeding was the same as in the experiments on the
Lake. A Monckhoven gelatincbromide-of-silver plate was immersed
to a given depth, in the apparatus, which remained open during a
fixed time. On this occasion we were obliged to preserve the
sensitive layer from the chemical action of the sea-water by covering
it with a thick coating of varnish. The light acted from the back
of the plate and through the thickness of the glass. Repeated
washings with spirits of turpentine and absolute alcohol sufficed to
remove the varnish before proceeding to develop. As on the former
occasion, we employed oxalate of iron as the developing agent,
which in every case was allowed to act for ten minutes.

The experiments were made on the 25thand 26th of last March,
and were favoured by bright and calm weather. We found the
depths we wanted, that is to say from 400 to 600 metres, off Cape
Ferrat, which protects the entrance to the Bay of Villefranche.
Leaving unnoticed the plates of minor interest, we will quote the
following, which appeared to us sufficient to solve the question.

a. Between 10h 30m and 10h 40m a plate exposed at a depth of 200
metres to-begin with ; the boat drifting away from the shore we
were obliged to let out 60 metres more rope in order to prevent
the premature closing of the apparatus.

b. From 12.45 to 12 50 at a depth of 280 m.

e. Between 11.50 and 11.40 at a depth of from 345 to 350 m.

d. Between 10.55 and 11.5 at a depth of 360 m.

ée. From 10.15 to 10.25 at a depth of 380 m.

This experiment was made under especially favourable circum-
stances; there was neither wind nor swell, the boat remained
absolutely motionless, the line perfectly vertical, so that we were
not obliged to let out the line throughout the time of exposure.
jf. From 1.20 to 1.80, under a cloudy sky, but still luminous enough,
at a depth of 405 to 420 m.
All these plates, with the exception of plate f, were exposed while
the sun was shining brightly.

On developing, the plates a and 6 proved to be greatly over-

exposed. On the plates c, d, and e, the strength of the impression
went on diminishing in a ratio corresponding very regularly with the

increase of depth. On the plate e the strength of the impression

was notably inferior to that produced by an exposure during the
same length of time to the air on a clear night with no moon. It

‘is comparable to that produced by an exposure of half the time,

say of five minutes only, under these latter conditions.

Lastly, the plate f did not bear the least trace of any impression
whatever. It is no doubt to be regretted that this last experiment
did not take place, like the others, under a perfectly clear sky.
But the impression on plate e, at 380 m., was already so faint that

76 Intelligence and Miscellaneous Articles.

we may conclude from it, with sufficient certainty, that the extreme
limit could not have been more than 20 metres lower. Moreover,
the experiments made in the Lake of Geneva have shown that
the dispersion of sunlight by a light layer of clouds does not cause
any considerable diminution in the depth to which it can penetrate
in water.

We believe we are justified in concluding from our experiments
that, in the month of March, in the middle of the day and in bright
sunlight, the last glimmers of daylight are extinguished at 400
metres from the surface, in the Mediterranean.

After these results, those of the experiments which we have still
continued in the Lake of Geneva, since the publication of our previous
notice, have scarcely more than a localinterest. To the absorption
peculiar to the water is in this case added that resulting from
particles in suspension, more or less abundant according to the
level. We hope, however, to determine an interesting point relative
to the influence exercised by the seasons on the degree of trans-
parency of these waters.

We know that the experiments of M. Forel have shown that
albumenized silver paper is blackened, in winter, at a depth of 100
metres, whilst in summer it undergoes no alteration at a depth of
45 metres. It would be interesting to ascertain whether this
variation of transparency with the season is peculiar to the super-
ficial layers, or if the same law also holds good at lower levels.

On the 18th of March of this year we repaired to the middle of
the Lake in the ‘Sachem,’ a steam-yacht belonging to M. E. Reverdin,
which her owner had kindly placed at our disposal. As in the
former experiments on the lake, M. I’. A. Forel kindly volunteered
to accompany and assist us. The weather was fairly bright; a
light layer of clouds dispersed the light without completely stopping
the direct rays of the sun. The exposure of the following plates
was made in the manner described in our former paper :—

Plate 10: from 9h 20m to 9.30, at a depth of 158 metres.

Plate 11: from 10.0 to 10.10, at a depth of 192 metres,

Plate 12: from 10.30 to 10.40, at a depth of 255 metres.

Plate 13: from 11.10 to 11.20, at from 240 to 245 metres.

Plate 14: from 11.48 to 12.25, at from 280 to 500 metres.

The time of exposure was therefore uniformly ten minutes in
each case excepting the last, which remained uncovered, at 280
m., during thirty-five minutes. Nevertheless not the slightest trace
of an impression is visible, either upon this plate or upon plates
13 and 12. Plate 11 shows avery faint impression, somewhat like
that of plate ¢ at 380 m. in the sea. Lastly, plate 10 at 158 m. igs
acted on to about the same extent as plate ¢. We place the ex-
treme limit of penetration of daylight in the Lake of Geneva in
winter at about 200 metres. |

It follows from a comparison between this series of experiments
and the preceding, that light does not penetrate in March more
than 20 or 30 m. lower than in September ; in the month of August
the difference is perhaps a little greater. Accordingly the strata of

Intelligence and Miscellaneous Articles. 17

water below 100 metres do not obey the law of variation of trans-

parency established by M. Forel for the more superficial layers.
Compared with the series of plates exposed in the Lake, the

series which we have brought back from the Mediterranean show

amore slow and regular gradation, which leads us to think that,

whilst in the Lake the light would be quickly intercepted by the

more or less turbid deep layers, in the Mediterranean the absorption

due to the pure water alone would be the principal, if not the only

factor in the arrest of the luminous rays.—-Comptes Rendus, April

13, 1885. coke Geena

ON THE ELECTRICAL CONDUCTIVITY OF SOLID MERCURY AND OF
PURE METALS AT LOW TEMPERATURES. BY MM. CAILLETET
AND BOUTY.

The electrical resistance of pure metals increases with the tem-
perature. From the experiments of Matthiessen * and those of
M. Benoitt, the mean coefficient of imcrease in the resistance for
one degree between 0° and 100° differs little in various metals, and
is only slightly removed from 54,5, that is, the coefficient of the
expansion of gases. If the same law held at low temperatures, the
resistance of a metal, varying like the pressure of a perfect gas
under constant pressure, would furnish a measure of the absolute
temperature, and would cease to exist at absolute zero.

Our experiments have been made with mercury and various
other pure metals. The mercury was contained ina spiral capillary
glass tube terminating in two large reservoirs in which dipped two
thick electrodes of amalgamated copper. The reservoir of a
hydrogen-thermometert was in the interior of the spiral; and the
whole being immersed in ice, or in a bath of methylic chloride or
of ethylene cooled by a current of air, according to the method
which one of us has given. To work with another metal, copper
for instance, the wire is coiled in a spiral form on an ebonite tube
in which are long slits, so as to be certain that the liquid was pro-
perly mixed, and that the bath and the resistance were measuring
at the same temperature.

We have only made relative measurements. The resistance in-
vestigated was compared with that of a column of mercury at 0°,
by means of a Wheatstone’s bridge and a very sensitive reflecting-
galvanometer. The following are the results which we have
obtained :— |

1. Mercury.—The empirical formula given by MM. Mascart,
De Nerville, and Benoit, for the apparent resistance of mercury
in glass above 0°, also holds for the freezing-point. When it
solidifies, its conductivity suddenly increases in a ratio, which at
—40° is equal to 4:08. The resistance of solid mercury decreases,

* Proceedings of the Royal Society, vol. xi. p. 516.

Tt Comptes Rendus, vol. lxxvi. p. 342 (1873).

t Hydrogen-thermometer of constant volume, in which the pressure at
0° was 509°3 millim.

78 _ Intelligence and Miscellaneous Articles.

then, regularly as the temperature sinks; between —80° and
—92°'13 it is represented by the formula

nea 1 +aié
ee ay oie 7

in which ¢ is the temperature on the centigrade-scale with
a=0°00407.

This coefficient of variation a, which is almost five times that for
liquid mercury, is very near that of other pure liquids in the solid
state.

2. Silver, Aluminium, Magnesium, Tin.-—For these various metals
the resistance is represented by the formula

re=r(l+az);
and the values of a, deduced from numerous experiments made at
various temperatures, are the following :—

Metal. a. Range of temperature.
Oo

ehh e ere Dare ae 0:00385 +29-97 to —101°75

Aluminium.... 0°00388 +2777 to = S0ay

Magnesium.... 0°00390 0 to — 8831

UNG ad tight Lacdy ids 0:00424 0 to — 85°06

These values of a, which are very near those for the same bodies
near 0°, according to Matthiessen’s experiments, are almost identical
for the first three metals; the value of afor solid mercury is
between the value common to magnesium, aluminium, and silver,
and. that for tin.

3. Copper.—The most complete experiments are those which we
have made with copper. ‘They gave the following values for a
deduced from a series of thirty measurements, which have been
divided into three groups :—

ob. Range of temperature.
0:00418 0 to — 58°22

Venere oer. s 0:00426 — 68°65 to —101°30
0:00424 —113:08 to —122-82

These values are a little greater than those which follow from
the formulas of Matthiessen and of M. Benoit for temperatures
near 0°*. The variation of resistance is almost absolutely regular,
and would enable us, in case of need, to dispense with a hydrogen-
thermometer in measuring temperatures between —20° and —123°.
No appreciable variation of a is observed at this latter temperature,
which would seem to indicate that the agreement, at any rate ap-
proximate, of the hydrogen-thermometer and the copper spiral
thermometer might be pursued still further.

4. Iron, Platinum.—Both these metals differ greatly from the
others in the variation of their resistance above zero; they diverge
from it in the same direction at low temperatures. The formula

r,=7,(1+at)

* @=0°00367 (Matthiessen) ; 0:008637 (Benoit).

Intelligence and Miscellaneous Articles. To

holds for iron from 0° to —92°, with a=0-0049, but it does not
hold for platinum. The value of a deduced from the formula (2),
which near 0° would be about 0°0030, increases as the temperature
sinks, and becomes 0°00342 for a lower limit equal to —94°57;

hence platinum approaches the other pure metals as the temperature

sinks.. ,

In conclusion, our experiments prove that the electrical resistance
of most metals decreases regularly when the temperature sinks
from 0° to —123°, and that the coefficient of variation is appreci-
ably the same for all. It seems probable that this resistance would
become extremely small, and therefore the conductivity very great,
at temperatures below —200°, although our experiments do not
enable us to form any precise idea of what would take place in
those conditions.— Comptes Rendus, May 11, 1885.

OPTICAL METHOD FOR THE ABSOLUTE MEASUREMENT OF SMALL
LENGTHS. BY M. MACE DE LEPINAY.

M. Mouton* gave an ingenious method for measuring, in wave-
lengths, the thickness of a quartz plate cut parallel to the axis,
which depends on a determination, by means of a known grating,
of the wave-lengths of Fizeau and Foucault’s dark bands. Un-
fortunately it depends on our really very imperfect knowledge of
the values n'—n of the two indices of quartz for different radia-
tions, and cannot therefore give numbers which are exact to within
suov: On the other hand, the corrections for temperature are
considerable: ;jhy,> for a degree Centigrade.

I. The method which I have used, which is analogous in prin-
ciple to the preceding, depends on an observation of Talbot’s bandst,
obtained by intercepting half the pencil of sunlight which falls on
a known grating by a quartz plate with true parallel faces. The
spectra deflected on the side of the plate are then channelled with
lines, in general extremely fine, but which for that very reason
tend to a point as exact as that of the spectrum. By using the
third spectrum of a grating of =}, millim., I have been able to
apply the method directly to a quartz plate parallel to the axis, about
4 millim. in thickness. It is known that the wave-lengths corre-
sponding to the centre of each of these dark bands are related to
the thickness of the plate ¢, and the index n, by the formula

gral
aN

€=P,

being an even whole number, which measures the number of the
order of the fringe observed, and which increases by two units in

_ passing from one band to the other, counting from red to violet.

We thus obtain as many values very near the desired thickness as
are observed, and the mean is taken.
In the preceding formula the index is directly given to within
* Journal de Physique, vol. viil. p. 893 (1879).
+ Journal de Physique, vol. 1. p. 177 (1872).

80 Intelligence and Miscellaneous Articles.

one or two units near the fifth decimal place, by either of the formula*

(n —1:52642)(A—1°5182 x 1075) =7°7733 x 1077,

(n' —1°53519)(A—1°5504 x 10-5) =7-8594 x 10-7;
calculated by taking the mean of the very concordant measurements
of the indices made by Rudberg, Mascart, and Van der Willigen,
taking for the waye-lengths the means of the measurements made

by Mascart, Ditscheiner, Van der Willigen, and Angstrom ; the
latter numbers were transformed so as to correspond to the wave-
length Ap, =5°888 x 10-5 adopted by M. Mascart. The constant
of the grating having been determined each time so as to satisfy
this condition, we see that the thicknesses are provisionally
measured as a function of what M. Mouton calls Fraunhofer’s
millumetre.

II. A double correction for temperature must be introduced in
the thicknesses thus calculated. If by 6 we denote the tempera-
ture to which correspond the indices given by the formule above,
by ¢ that at which the measurements were made, the thickness ¢,,
at 0° of the plate, should be calculated by the formula

2e(1+kt)(n+r'0—ri—1)=pr;
k, v, and v' being coefficients determined, the first by M. Benoit,
the two others by M. Dufety. It is more simple and just as exact
to correct the thickness ¢,, originally calculated by one or other of
the two simplified formule,
e, = 0°999741(1 — 0:0000019 ¢)e ordinary rays,
or e, = 0°999709(1 —0:0000001 t)e extraordinary rays.

These two formulet apply to the case in which the plate is cut
parallel to the axis.

III. I applied this method to two quartz plates cut parallel to
the axis made by Hoffmann. The thinnest of them measured
directly (123 bands observed, ordinary rays) gave

€,= 0°402958 cm. +0:000001.

On the other hand, we measured the difference in thickness of two
plates (123 bands observed, extraordinary rays). From this was
deduced for the thickest plate,

é, = 0°602316 cm. + 0°000008.

1V. We may add that this method may lead to a new determi-
nation of the absolute value of the wave-length of the ray D,. The
thickness of the second plate, which M. Benoit kindly measured
with the apparatus of the Bureau international des Poids et
Mesures, was found equal to 0°60236 to within 1/ or 2, From
this we get for the wave-length of the line D,, the number
A=5'8884 x 1075,

12)
which is exact to within 5,)5,;, and is very near that of Angstrom
(5°8889).—Comptes Rendus, June 2, 1885.
* Cornu, Annales de I Ecole Normale supérieure, vol. ix. p.42 (1880).
+ Journal de Physique, 2nd series, vol. cxi. p. 252 (1884).
t It has been assumed that 6=20°. As this temperature is scarcely
known to within 4° or 5°, there is an uncertainty of about 5,1,,5.

THE
LONDON, EDINBURGH, ann DUBLIN

PHILOSOPHICAL MAGAZINE
AND

JOURNAL OF SCIENCE.

[FIFTH SERIES.]

AUGUST 1885:

XIII. The Polymerization of the Metallic Oxides. By Lovts
Henry, Professor of Chemistry in the Catholic University of
Louvain*.

Introduction.

()X# of the principal objects of Physical Science, in fact its

principal object, is to determine the natural relation of
things. Itis by comparisons logically established between the
things themselves that these relationships are disclosed. An
essential condition to be observed in researches of this kind
is to compare only those objects which are really comparable.
Stated in the terms of chemical science, this general condition
of all work undertaken with a truly scientific object consists
in establishing a comparison only between the chemical mole-
cules themselves. Molecules are, in fact, for the chemist the
veritable chemical individuals—they are the bodies themselves
in their simplest, and at the same time their most complete,

expression. Molecules find a graphic representation simple

and concise, and sometimes even sufficiently perfect, and in
every case susceptible of a great degree of precision, in the
chemical formule at present in use.

Formule, however, only maintain their full and complete
utility so long as they are the exact and complete representa-
tion of the molecules of bodies,—of bodies between which che-
mical actions occur, of bodies which we possess in reality, and

which are the veritable objects of our comparisons. This re-

presentation becomes exact and complete only when the for-

* Communicated by Prof. Carnelley.

Phil. Mag. 8. 5. Vol. 20. No. 123. August 1885. G

82 Prof. Louis Henry on the Polymerization

mula indicates the three facts fundamental to the being of the —

molecule, viz. (1) the nature of its constituent elements, (2)
the actual number of atoms of each present, and (3) ). their 7}
combining proportion. i

We are still far from this degree of completeness. Our
knowledge of bodies from the molecular point of view is very
incomplete, and especially so as regards Mineral Chemistry.
Not only is the number of compounds, of which the constitu-
tion and structure have been successfully determined, very
limited, but there are a great number, I might even say the
greater number, to which we are still unable to attribute a
formula expressing in a satisfactory manner the true size and
real weight of the molecule. The formule used in Mineral
Chemistry are in the majority of cases only the empirical for-
mule, indicating merely the nature.and relative weights of
the atoms of the elements in the compound, and are wholly~
silent as to the absolute quantity of matter, or the actual
number of atoms present in the molecule. They do not give
therefore any idea either of its size, or of its absolute weight,
or of its internal structure. The properties of compounds do
not depend on chance ; they depend chiefly no doubt on the
nature of the elements constituting the molecule, but they are
also conditioned by the internal structure of the latter, and by
its size and weight.

It is to Berzelius that we owe the idea of csomerism, without
which a rational study of the innumerable compounds of
carbon would be absolutely unattainable. However it may
be, each of these three great factors in the diversity of com- |
pound bodies exercises a special and preponderating influence |
on certain classes of properties. If, on the one hand, the |
chemical activity of a given compound depends especially on |
the essential nature of the atoms constituting the molecule, —
and on the constitution of the latter, on the other, an increase |
in its weight powerfully affects the physical and mechanical i
properties. “|

Whatever may be the specialization of this dependence, the |
properties of a body, considered individually or relatively, the |
one in relation to the other, their analogies and their differ- |
ences, can only be well understood when we know the size of ||
the molecule. |

Just as the adoption of rational and true formule for the |
various compounds may render great service in a systematic |
study, so in the same degree will the adoption of imperfect
formule, artificial and false, or even incomplete and expressing
only simple relations, lead to results which are untrue. Such
formule mislead by obscuring, or even by totally concealing,

of the Metallic Oxides. 83

the real relations of bodies. They create imaginary relation-
ships, removing apart those bodies which ought to be brought
together, and bringing those together which ought to be kept
apart. Our formule should be like glasses of a perfect trans-
parency, through which the bodies should be seen as they are,
and not like screens to hide, or disguises to disfigure.

Most formule now in use refer to the perfect gaseous state.
These formule are undoubtedly excellent for bodies of which
the normal state is gaseous ; but their exclusive use is insuf-
ficient, and creates a void to be regretted in the comparative
study of bodies, both from the individual and general point of
view. The number of bodies naturally gaseous is in fact very
limited ; and if the number of those capable of becoming so
is very considerable, still there are many which are not, and
with an absolute or relative fixity resist the highest tempera-
tures, or are decomposed.

Our usual molecular formule are therefore necessarily
insufficient, being inapplicable to numerous compounds, and
to the various states which the same body is capable of
assuming. ‘These formule, referring to the perfect gaseous
state, do not tell us anything of the molecular nature of these
same bodies when in the ordinary condition. Thus it is, to
appeal to classical examples, with the formule of sulphur, and
of the fatty acids, or at least with the first terms of the latter
series.

It is in all respects desirable that the molecular formule of
bodies should refer to their natural state, and that the several
formule should be determined for the various physical states
which the bodies can either naturally or artificially assume.
That is the only way to avoid misconceptions, artificial
difficulties, false analogies between unlike bodies, and to
retain for various compounds their own physiognomy,

There is no teacher of Chemistry who has not frequently
noticed, both in mineral and in organic chemistry, differences
in the properties of bodies, which are often most profound,
and totally beyond theoretical prevision, notwithstanding that
the bodies are apparently related as regards their formule
and the analogy of their composition. It is easy to say that
there are exceptions and anomalies. But in spite of these
words, the difficulties still exist in all their entirety, embar-
rassing the mind of both teacher and taught. The greatest
caution should be used in qualifying exceptional and anoma-
lous facts; otherwise confusion is created in the mind, and
place is given to chance in the occurrence of natural facts :
it is the ruin of law and general principles. It is well to
remember the old adage, eee re in the physical as well as

2

84 Prof. Louis Henry on the Polymerization —

in the natural world, that exceptions prove the rule. True
exceptions, true anomalies, have no real existence; these

apparent deviations in the effects of a general known cause ©

are due to the influence of other causes, acting simultaneously,
the existence of which remains momentarily hidden from view.

Among the difficulties which are met with in the compara-
tive study of bodies, there are some which are peculiarly
embarrassing, and which it would be chimerical to try to
solve under present circumstances. But there are others,
and I believe a very considerable number, which are purely
artificial, without real foundation ; difficulties which would
vanish of themselves if we only had a less imperfect know-
ledge of the bodies with which they occur—if, in fact, the
chemical signification of these bodies molecularly in relation
to others were only clearly and surely defined.

This point of view, from which bodies should be studied, is
not absolutely new. Sulphur and the fatty acids, especially
acetic acid, are well-known examples of substances which
have been ‘clearly distinguished in their various physical
states. Moreover, the determination of the relative molecules
of bodies in their various states also touches on a question of
still greater interest, viz. that of the chemical relationships of
the various physical states in which they occur.

The molecular formule at present in use refer to the gaseous
state, and the principal method of determining them depends
on the vapour-density. This process, simple and expeditious,
is undoubtedly excellent in all cases in which it is applicable.
Others, however, are necessary for the determination of mole-
cules in a state other than that of a perfect gas. It is in the

complete study of bodies in their various conditions that we ~

should find indications sufficient for information as to their
molecular nature. The several circumstances which charac-
terize the chemical molecule, both chemically and physically,
are in intimate connection, and to determine this molecule we
can appeal to all connected therewith, either in its past, pre-
sent, or future: in its past, to examine its mode of formation
and the various circumstances beari ing on its chemical origin 5
in its present, to carefully investigate the whole of its pro-
perties—physical, mechanical, and chemical ; in its future, to
study the reactions in which it may take part, the products of
its decomposition, its metamorphoses, Xe.

This complete study of bodies, from both the statical and
dynamical point of view, always gives, though it may be
imperfect, valuable information for solving our present pro-
blem. It is, however, necessary to say, that when I speak of
actual molecules and of actual molecular formule, I do not

a
; wi
a" ‘

| .
ty

of the Metallic Oxides. 85

forget that these molecules, as well as the formule which
express them, are really only relative. We do not know, and
we shall never know, that which is truly and absolutely a
molecule, for we cannot and never shall be able to obtain one
singly and alone. We possess, and we shall always possess,
only groups of molecules. The character of reality which we
wish to give to our molecules and to our formule is essentially,
then, a relative one. What we have to determine are real
relationships between real bodies. These relationships of
analogy and difference remain practically the same if, in the
impossibility of verifying them between the bodies them-
selves, we seek for them between the same bodies ina modified
condition ; modified simultaneously, however, and in a similar
manner. It results from these considerations that it is of

_ great importance to examine chemical phenomena from the

molecular point of view. To establish real molecular formule
for bodies (7. e. formule truly and universally comparable) is
therefore an object eminently worthy of attention. It is this
idea which has induced me to undertake a lengthened series
of theoretical and experimental studies on compounds in both
mineral and organic chemistry, and more particularly as
regards the metallic oxides.

The Peculiar Character of the Metallic Oxides.

A comparative examination of the oxides and chlorides
leads to a general fact. Chlorides are fusible compounds,
and, with few exceptions, volatile. It has been long observed
that chlorine gives wings to the metals, as one may say.
Volatility is especially noticeable in the chlorides of the fixed
elements, and particularly of the metals. The corresponding
oxides, on the contrary, are distinguished by their fixity and
infusibility, either absolutely or relatively. The following
table shows that this difference is general :—

Mole- :
‘cular | Poysical Eusibility. Volatility.
weight. j

Bek, 3 f«c30,- LEY Ee a Oe: Se Boils 17°.

{ meee 70 | Solid. Fuses 577°. Fixed.

| ee By. iat apie ie: Boils 59°.
ae 60 | Solid. { jose a} Fixed.

rogen flame.

{ ee ee Wy PRS ice 47 goa ee Boils 135°.
ee 82 Solid. Infusible. Fixed.
2 ee 231°5| Solid. Fusible. Sublimable.
‘(0 re 121°5| Solid. Infusible. Fixed.

Bethe. wu... e's": ee ee eee Boils 115°°4.
{ . a 150 Solid. Infusible. Fixed.
{ peed. 3. 189 Solid. 249°3. 617°-628°.
BEE Poni nnan> 134 Solid. Infusible. Fixed.

86 Prof. Louis Henry on the Polymerization
Table (continued).

_ | Physical

tt. Fusibility. Volatility.
NbCl, Solid. Fuses 194°. Boils 240°°5.

{ Nb,O; Solid. Infusible. Fixed.
TaCl, Solid. Fuses 211°°3. Boils 241°°6.

{ Ta,O, Solid. Infusible. Fixed.

Al, Cl, Solid. Very fusible. 180°.

| : Fusible in oxyhy- :
Al,O, Solid. { drogen flame. Fixed.
Al, Br, Solid. Fuses 93°. Boils 260°.
Al.I, Solid. Fuses 125°. Boils 350°.
FeCl, Solid. Fuses 3069-3079. Easily volatile.

{ Fe,0, Solid. Infusible. Fixed.
FeCl, cc as Sm pages | ere Volatile.
FeO uns. 5 Solid. Infusible. Fixed.
CrCl Solids derit | besthhee, Volatile.

{ Cr,0, Solid. Infusible. Fixed.
CrO,Cl, ... 44 Tagaididl: ied: (Anis Boils 118°.
CrO, Solid. Fuses about 300°. Decomposes,
WwCl, Solid. Fuses 248°. Boils 275°°6.
wel, Solid. Fuses 275°, Boils 346°°7.
WC1,0 Solid. Fuses 210°°4. Boils 127°°5.
wo, Solid. | Fuses at forge-heat. Fixed.
Vol, Liquid. | {P05 ;POt gousHy |) Boils 154°.
{8 Ree ae CS ee ear Oi cuba ane
Vocl Liquid. |{ Does not solidity 1} Boils 1260-7.

. : Easily volatile,
SeCl, Solid. 1 erin ier \ ae
SeOCl....... Tiguid.) .wiiihtie Boils 179°°5.
Volatile at 300°
6 Spee Solid. Infusible. without melting.

TeCl, Solid. Fuses 224°, Boils 414°.

' Less volatile than
TeO, Solid. i ei | elietaue
AsCl, lagaihidd i Cee Boils 134°.

{ As,O, Solid. |Volatile without fusion. Boils 200°,
Sbel, Solid. Fuses 73° 2. Boils 225°.

{ Sb,O, Solid. Fuses red-heat. —_| Volatile below 1560°.
SbCl, Liquid. | Solid at 0°. | j YntL2 tito BPE,

{ Sb,0, Solid. Nonfusible. Nonvolatile.

f BiCl, Solid. Fuses 227°. Boils 427°-439°.

| Bi,O, Solid. Fusible. Fixed.
HegCl, Solid. Fuses 265°. Boils 295°.

{ HgO Solid. Infusible. Decomposes.

Sublimes 400°-500°

Hg,Cl,...... Solid. | { without melting. } ip
Hg,O Solid. Infusible. Decom poses.
Cu,Cl, Solid. Fuses 434°. Boils 954°-1032°.

| 6x20 Solid. Fuses at red-heat. Fixed.
Cul, Solid. Fuses 498°. Decomposes.

Fuses at bright
OuQ) iearsan ‘5| Solid. red-heat, with de-}| =...
composition.

of the Metallic Oxides. 87
Table (continued).
Mole- ;
cular pe Fusibility. Volatility.
2 weight.) “'°

_) ae 278 Solid. Fuses 498°. Boils 861°-954°.
Ae a oe yas) Solid. |Fuses about a red-heat. { Nagey bs Sy
| Ag,Ol,...... 287 Solid. Fuses 4519 3 ef AA+ Vnjicnevss
{ ae 232 Solid. Infusible. Decomposes.

Sete, ...... 111 Solid. Mees ca af lk ewetn cess

as... 56 Solid. Infusible. Fixed

prtek, <..... 1585} Solid. Besea aS ab ie foe lege

See 103°5| Solid. Infusible. Fixed
a atte 208 Solid. Fuses above 860°, | sn sec eeaee

Fusible in oxyhy-

ae 153 | Solid. { same ti } Fixed

Mact. .....: 95 Solid. 708°. Volatile.
i 40 Solid. Infusible. Fixed.

2 ae 136 Solid. Fuses 262°. Boils 676°-683°.

an 81 Solid. Infusible. Fixed.
| (oe 183 Solid. Fuses 541°. Boils 861°-954°.
oeaey.. 128 Solid. Infusible. Fixed.

ree 191 Sabret ty SOT i eal) Volatile.

Ie cs 3.3 136 Solid. Infusible. Fixed.

; a Sublimes without

MICE, £3: . .2: 129 0) Solid. Fie reaty Bi[ot ter Te aakkd. sas

0) Seas 4°75) Solid. Infusible. Fixed.

[2 5) Sere 80°25) Solid. Fuses 585°-617°. | Volatile below 520°.

ot ae 25°25, Solid. Infusible. Fixed.
lj) 376 lt. Of Reelin A oe eee Sublimes.
PaLO......... 266 | Solid. Infusible. Fixed.

As already stated, this difference in properties is observed
especially in the oxides and chlorides of positive elements, and
in general in elements having a metallic brightness. It may
be noted by the way that it appears to be the greater the higher
the atomicity of the metal.

Whatever may be its meaning, this physical difference is,
by its generality, one of the most remarkable facts in statical
chemistry, strange and, at first sight, quite exceptional.

The properties of compound bodies do not appear by chance;
they are only the final result of the properties of the consti-
tuent elements—a result more or less modified in consequence
of the physical phenomena of combination. Chemical com-
pounds are therefore really mixtures of a special kind. It
cannot be denied that the properties of compounds are closely
related to those of their principal constituents. The differences
between certain elements occur again with greater or less
fidelity in their corresponding and comparable compounds.
That being so, the behaviour of the metallic oxides is especially

88 Prof. Louis Henry on the Polymerization

surprising when compared with the chlorides of the same
elements.

The difference between the volatility of oxygen and chlorine
is enormous. ‘The former, until recently, was considered as a
permanent gas; the other, on the contrary, is easily condensed,
and only becomes gaseous at — 33°°6 (Regnault). An atom
of oxygen weighs 16, and that of chlorine 35°5. The mole-
cular weights of the chlorides are consequently much greater
than those of the corresponding oxides. Now, of mechanical
or physical properties, there is not one which is more directly
dependent on the molecular weight than fusibility and vola-
tility. :

Bearing these facts in mind, we are justified in formulating
a priori the following propositions :—(1) Oxygen and chlorine
are both of them elements which are endowed with great vola-
tility as compared with other elements. The oxides and the
chlorides ought, therefore, to be volatile compounds. (2)
Volatility exists toa greater degree in oxygen than in chlorine.
The oxides ought, therefore, to be more volatile than the cor-
responding chlorides. The oxides ought frequently to be very
volatile, or even gaseous, at the ordinary temperature.

These theoretical deductions are, in fact, realized ina good
number of oxygenated and chlorinated compounds, particularly
in the case of elements or radicals which are negative or

metalloidal, e.g. As, Os, Ru, &e.

Mole-
cular Physical state. Volatility.
weight
Perfect gas; formerl

lex ee 02 considered Sehr
i. 87 Gas, easily condensed.

Seer 64 Gas. Boils —10°.

{ BOCK. aime 119 Liquid. Boils 82°.
8 at ah 80 Solid. Boils 46°.

{ BOO... 135 Liquid. Boils 82°.
oe ingens De pe 44 Gas. Boils —789.
OGG, ise oh 99 Liquid. Boils +8°.
2 Re 154 Liquid. Boils +76°.
ILE 14550 237 Solid. Boils 182°.

1 CU.OLO ™ ,..1, 182 Liquid. Boils 118°.
C0 Sagas 263 Solid. Volatilizes about 100°.
BaQ;, ....:.- 168 Solid. Volatilizes 100°.
As,Os ...... 198 Solid. Volatilizes 200°.

{ 8 ee 208°5 Solid. Boils 148° tf.
FORGO... 153°5 Liquid. Boils 110°.

{ WOl, ...... 395 Solid. Boils 346°.

WCI,0 ...| 342 Solid. Boils 227°,

* CCl, . COC. + Decomposing into PCl; + Cl.,.

of the Metallic Oxides. 89

Relationships of a similar kind, as regards volatility, are
observed between the oxides and the chlorides of the bivalent

_ radicals C,H»,, which are the most analogous to the metals,

for example:—

Molecular

weight, Volatility.
CP ATGCH VOC oS viwccceivcns” 44 Boils ai

0,H,0.....-| | Bthylene oxide ‘i » 135
CG HCl Eanes chloride.... 99 J. un
2 4~"2-**"!| | Ethylene chloride ...... a ee
Propylene oxide......... 58 < o
O6H.O... ... Aidehyde’ #.)..2.d..0060- i a
POG OMG «oc iscideoeicdiee's ij oo
Propylene chloride.....} 1138 ae
C,H,Cl, . | Bropsidee chloride.. ‘3 » oD
Acetone chloride ...... ii se

Let us now attempt, by ales aid of this general idea, to
determine, @ priori, what ought to be the properties of certain
particularly interesting oxides.

To CO;, which is gaseous at the ordinary temperature and
boils at —78°, corresponds OCl,, which is a liquid boiling
at +76°. The difference in volatility between the two
bodies is 154°. SiCl,, TiCl,, and SnCl, boil respectively at
58°, 136°, and 115°. It might therefore be reasonably sup-
posed that the corresponding oxides should be eminently
volatile. SiO, ought certainly to be gaseous at the ordinary
temperature, or at least as gaseous as SO, to which it is
analogous, and the molecular weight of which is but little
different.

Molecular Boiling-

weight. point,
CO, ee sitenpeliokay bie —78
PEO. filo 2 6A —10
pene oto es OD * nonvolatile.

BCl;, with molecular weight 117°5, is a liquid boiling at
18°; B,O;, with molecular weight 70, ought therefore to bea
gas at the ordinary temperature.

Al,Cl, is a-solid, which volatilizes about 100°, and ought
therefore to have a ‘corresponding oxide Al,Q3, if not gaseous,
at least very volatile, and the same should hold good as regards
Fe,QOs.

HgCl, (boiling-point 295°) should have a corresponding
oxide, HgO (containing a so-called permanent gas in combi-
nation with a volatile metal), which, if not gaseous, should at
least be much more volatile than HeCl,, or metallic mercury.

90 ~ Prof, Louis Henry on the Polymerization

The chromates are analogous to the sulphates both in their
general composition and crystallographical properties. One
would expect, therefore, that the difference between CrO,O
and CrO,Cl, would be similar to that between SO,O and
SO,Cl,. Now SOs; boils at 46° and SO,Cl, at 77°, whilst
CrO,Cl, is a liquid boiling at 118°; and from analogy we
should expect CrQOs to be also.a liquid still more volatile, and
boiling even below 100°, whereas CrQ; is a solid, fusible
only at about 300°. The oxides, in fact, instead of being
gaseous at the ordinary temperature, or at least very volatile
liquids, as analogy would have led us to expect, are generally
solid, and frequently almost infusible. What is the expla-
nation of this fact, so strange in its nature and so important
from its generality ? J regard no fact, however strange, as
abnormal or exceptional. In my opinion anomalies and ex-
ceptions have no real existence. They depend most probably,
as already stated, on our ignorance of the consequences, under
certain conditions, of a general known cause.

The Oxides are not Molecularly Comparable to the Chlorides.

Polymerization of the Oaides.—The chlorides and oxides are
almost universally indicated by formule which represent these
bodies as molecularly comparable. But are the oxides really
comparable with the chlorides of the same elements? Is there
a real analogy between them, or is this analogy merely arti-
ficial, and dependent solely on the notation employed? These
are questions which require examination. The chlorides being
compounds which are frequently volatile without decomposi-
tion, the formulse assigned to them have usually been deduced
from the determination of their vapour-density; so that these
formule really represent, both in weight and size, the molecules
of these bodies, at least in the gaseous state. ‘This is not the
case, however, with the metallic oxides. Being fixed, or
volatile only at the very highest temperatures, or else decom-
posing under the action of heat, their vapour-densities are
unknown, and cannot be determined. The formule univer-
sally attributed to the metallic oxides depend solely on ana-
lytical determinations. These formule are therefore simply
empirical, and merely represent the relations between the
weights of the constituent elements, but are perfectly silent
as regards the absolute quantities.

The formule of the oxides have thus their origin, and rest
on an entirely different basis from the formule of the chlorides.
There is no authority for attributing to both the same value
and signification. The relation between the oxides and chlo-
rides, as represented by the usual formule, is purely artificial.

of the Metallic Oxides. 91

The true oxides RO,, which are really comparable with the
known chlorides RCl,,, are for the most part quite unknown.
We only have the polymeric oxides n(RO,), in which n is a
large number. To make this clear, we may take an example
from organic chemistry in illustration, viz. the chlorides of
ethylene and ethylidene:—

CH; p.-p. 60°, rae
CHCI, CH,Cl

Not only are the corresponding oxides, acetaldehyde and
ethylene oxide,

b.-p. 84°.

CHs bp. 21°, VHN0, b=p. 18°,
CHO CH,

known, but also their respective polymers paraldehyde and
metaldehyde, as well as the dioxide and polyoxide of ethylene.

Paraldehyde . . (C,H,O)s, liquid, boiling-point 124°.
Metaldehyde . .° (C,H,O),, solid, nonvolatile as such.
Ethylene dioxide (C,H,O),, solid, boiling-point 102°.
Ethylene polyoxide (C,H,O),, solid, volatile with difficulty.

This interpretation of oxides being admitted, it is no longer
a hypothesis, but a fact, that the physical state and thermal
properties, notably the fusibility and volatility, are a direct
function of the molecular weight, and every abnormal relation,
from the physical point of view, between the chlorides and
oxides ceases immediately. The infusibility and fixity of the
latter, so strange at first sight, return into the natural order
of things.

We have, therefore, to prove the following proposition:—

The known oxides, and notably the metallic oxides, are
polymers n(RO,) of the true oxides RO;, corresponding to the
chlorides RCI...

Proofs of the Polymerization of the Oxides.

Their Additive Power.—There are no compounds which
are more ready to enter into combination than the oxides.
In fact, it would not be far from the truth to say that there is
no class of compounds with which they are incapable of enter-
_ ing into combination. In mineral chemistry they combine
by addition, and often with considerable energy, not only with
oxides of a contrary sign—acid oxides with basic oxides,—
but with all chemical compounds indiscriminately, whatever
may be their special function.

92 Prof. Louis Henry on the Polymerization

Let us limit ourselves more especially to metallic oxides.
They combine not only with the oxides of negative radicals,
but also with the most diverse binary compounds—chlorides,
bromides, iodides, sulphides, selenides, &c.; with oxysalts in
general—hydrates, sulphates, nitrates, carbonates, &c., forming
the large series of basic salts of the polyvalent metals. These
additive compounds are formed in the most varied proportions ;
so that the series of basic salts is illimitable. Many of them are
marked by a definite composition and by a regular crystalline
character; some even occur naturally in the mineral kingdom.

This additive power is observed also, though in a less degree,
with the oxides of organic radicals, and notably in the case of
the glycollic oxides, C,H»,O, which are so similar in other
respects to the metallic oxides RO. The oxide of ethylene
combines with its hydrate, glycol C,H,(OH)., with acetic
anhydride, with acetyl! chloride, with the halogen hydracids, &e.
The aldehydes C,H2n41.CHO behave in a similar manner.

But let us return to the metallic oxides. Metallic oxides
enter into combinations with other metallic oxides which differ
in composition, both in the nature of the metal and in the
quantity of oxygen. Thus, MgO combines with Al,O; to form
spinelle, MgO. AI,O3. Still more remarkable is the combi-
nation of two oxides of the same metal, as in magnetic oxide of‘
iron, FeO. Fe,03;; red lead, 2PbO.PbO,; hausmannite,
MnO. Mn,0,; SnO.SnO,; Ke.

Organic chemistry also furnishes examples of this, which
are still more curious in so far as we have cases where two
oxides, equally oxygenated, combine, as in the case of the
double oxide of ethylene and ethylidene discovered by Wiirtz

; CHO
(Bull. Soc. Chim. iv. p. 16), ( Ce Oa ) Here we

have polymerization properly so called.

In the absence of foreign molecular systems on which they can
exercise their additive power, oxides, both organic and inorganic,
combine with themselves ; and in this combination of a body
with itself lies the gist of polymerization in its largest sense.
This fact may be illustrated by examples from organic chemistry.

CH
Kthylene oxide, RR is a colourless liquid, boiling at 13°.
2

Two distinct polymers of this compound are known :—(a) The
CH,.O.CH, | nies eae
| 1 i

bH,.O. OH,’ 1s obtained directly

from the monoxide and is a crystalline solid, melting-point 9°,
boiling-point 102°. (6) A polymer, (C,H,O)", obtained by

so-called diethylene dioxide,

of the Metallic Oxides. 93

Wiirtz (Compt. Rend. Ixxxiii. p. 1141), is formed sponta-
neously as a white solid crystalline mass, melting-point 56°,

and volatile with difficulty.

In connection with this question there is no body more

Interesting than methylene oxide, CH,: O. The corresponding

chloride, CH;Cl,, is a liquid, boiling-point 40°. The oxide itself
should be naturally a gas, but only assumes this state under
the action of heat ; so soon as the heat is withdrawn, it poly-
merizes, and is converted into a white solid mass which is
insoluble in water, and sublimes at 100°, again changing at
higher temperatures into the gaseous CH,O, as indicated by
its vapour-density. The aldehydes C,H,.CHO are especially
characterized by their tendency to polymerization, giving rise
to products differing from the original aldehyde in all their
physical properties—fusibility, volatility, solubility, &e. Acet-
aldehyde may be taken as a type of thisclass. Under ordinary
circumstances it is a liquid, boiling-point 21°. There are two
distinct polymers known:—

(a) Paraldehyde is a liquid, boiling-point 124°, and in the
state of gas is represented by (C,H,O)3, as proved by its vapour-
density.

(6) Metaldehyde, (C2H,O),, is a solid, crystallizing in
beautiful needles, which sublime at 100°, and volatilize without
melting at 112-115°, being transformed again into ordinary
aldehyde as indicated by its vapour-density.

Trichloraldehyde, or chloral, CCl;. CHO, polymerizes still
more easily. When recently distilled, it is a liquid, boiling
at 94—95°; after remaining at rest for some time, it is spon-

taneously, and sometimes rapidly, converted into what is

called insoluble chloral, which is a white, opaque, porcellaneous,
hard mass, and, when heated to 180°, re-forms ordinary
chloral. | :

The aldol, CH; . CH(OH). CH,. CHO, of Wiirtz also poly-
merizes readily, forming paraldol, which is a beautiful crystal-
line solid, melting at 90° (Comptes Rendus, Ixxxiii. p. 255).
Some aldehydes are only known in the polymeric state, as
oxalic aldehyde or glyoxal*, CHO. CHO.

From analogy, this body ought to be gaseous, or at least
a very volatile liquid at the ordinary temperature :—

* Given the difference in volatility between the chlorides of the nega-
tive radicals and the corresponding aldehydes,

CH; .COCI, b.-p. 55°, CH, .COH, b.-p. 21°,
we should conclude that ordinary succinic aldehyde was a polymer :—
COC1. CH, .CH, .COCI, b.-p. 190°,
COH.CH,.CH,.COH, b.-p. 201°-203°.

94 Prof. Louis Henry on the Polymerization
CHO CHCl
A b.-p. 21°, |) Denps 60°,
H, cHo CHs

|
CHO
CHCl, 402, CHO

I
CHCl, HCl,

In reality, however, glyoxal is only known as a non-
volatile viscous mass. It is unnecessary to extend these —
examples, for the same facts are illustrated in various degrees
by all the members of this group (see Bruylants, ‘On the
Polymerization of the Aldehydes, Jnaug. Diss., Louvain,
1875). The self-additive power of methylene oxide, which is
so evident in its mono-derivatives, the aldehydes, is less marked
in its di-derivatives, the acetones. Though true polymers
of acetone itself are unknown, yet those of some of its chloro-
derivatives have been obtained by Grabowsky (Ber, viil.
p. 1488). |

If we compare the boiling-points of the glycollic, aldehydic,
and acetonic chlorides with those of the corresponding oxides,
thus:—

CH,.CH,.0 :.. b-p.13 | CH,Cl..CH:Cl ss sbemeee

ee ni
OH, OHO Wavy ral ORy CHOLLI » 60
CH,.CH.CH,.O,, 35 | CH;.CHCI.CH,Cl, 96

bie ces aul 0 1g
OH,.0H,.CHO. ,,. 47; |, CH, CH) CHCl ssuamame
CH, . CO . Sin ee ” 56 CH; ° CCl, ° CH; 9 70

and if we further bear in mind the general fact that in the
isomeric substituted derivatives of the paraffins, the volatility
is the greater when the substitution has taken place in a chain
or chains which are least hydrogenated, then we must con-
clude that the aldehydes and acetones are only known to us
in the state of polymers, which depolymerize completely at
the boiling-point. ‘Thus, acetic aldehyde should be gaseous
and more volatile than the oxide of ethylene; propionic alde-
hyde more volatile than oxide of propylene; and acetone more
volatile than the corresponding aldehyde C;H,O. The real
relations are, however, the reverse of these.

b.-p. 87°.

Actual Polymerization of the Oxides themselves.

The actual polymerization of the oxides is readily indicated
by all those characters which have been long observed in the
case of the special variety of oxides called by Chevreul “ oxydes
cuits.” Certain anhydrous and pulverulent metallic oxides,

of the Metallic Oxides. 95

obtained by dehydration of their hydroxides at as low a tem-
perature as possible, suddenly enter into vivid incandescence
when exposed to a higher temperature, verging on redness.
Their properties are thus modified both physically and che-
_mically; they become denser, harder, and, when coloured,
their colour is intensified. ‘They show a greater resistance to
the action of chemical agents than in their former condition,
some even entirely resisting the action of acids and alkalies.
Fe,O3, Cr,O3, ZrO, &e., are examples. The fact of this vivid
incandescence, which is not accompanied by any alteration in
the composition of the product, is the manifest proof of an
energetic combination of the oxide withitself. When hydro-
carbons, such as amylene, terebenthene, &c., are polymerized
under the action of certain agents, such as sulphuric acid,
and especially boron fluoride, a development of heat takes
place which is frequently considerable. Finally, arsenic tri-
oxide is a direct proof of accumulation in a single molecule,
even in the gaseous state, of several normal molecules corre-
sponding to the chloride. Arsenic trichloride being AsCl,,
the corresponding oxide ought to be As,O3. This is in fact
the arsenical oxygenated compound which is present in the
arsenites of methyl and ethyl (Bull. Soc. Chim. xiv. p. 101),
compounds which are volatile without decomposition, and
whose formulz, as deduced from vapour-density determina-
tions, are (CH;),;AsO; and (C,H;),AsO3. The molecule of
arsenious oxide in the gaseous state corresponds not to As,Qs,
but to (As,O3). or As,O,, as proved by its vapour-density. It
volatilizes at about 200° without fusion; and there can be no
doubt that its molecule in the solid state is some multiple of
its gaseous molecule As,O,. There is thus established between
the methylenic compounds and the arsenious ethers a complete
parallel. To this we shall refer again.

Various Methods of Producing the Oxides.
By Dehydration of the Hydroxides.—The various methods of

obtaining the oxides from other compounds containing them
as radicals also tend to show that they are polymers. Among
these methods, one of the most important, on account of its
generality, and also the most interesting from our particular
point of view, is the dehydration of hydroxides. Under the
action of heat the hydroxides, with but few exceptions, lose
water, either wholly or partially. But, however this may be,
the dehydration is the more complete the higher the tem-
perature. This temperature varies according to the nature
of the hydroxide. Some of the latter have only an ephemeral
existence at the ordinary temperature, and are therefore more

oh

96 Prof. Louis Henry on the Polymerization

usually represented by their anhydrides. Inaddition to these _
general facts, it is important to observe not only the nature of
the anhydrides thus obtained, but also the connection which
underlies the two distinct chemical phenomena, viz. dehy-
dration and molecular condensation. Wiirtz developed this
principle in one of his lectures before the Chemical Society of
Paris in 1863. The elimination of water usually depends not
on one but on several molecules of the hydroxide, the residues
of which become soldered together, as it were, by the atoms of
oxygen ; we therefore get an accumulation of the radicals of
these hydroxides in the products formed—a veritable molecular
condensation proportional to this accumulation. This principle
applies to all hydroxides indiscriminately, not only to the
normal hydroxides R*(OH),, but to the oxyhydroxides or
incomplete anhydrides R*O,(OH),-2n. It therefore follows
that the compounds formed under these conditions become
more and more complex, in proportion as this dehydration
itself is more or less complete.

These general facts apply to all hydroxides, both organic

and mineral. The following are illustrations from organic
chemistry :—
_ The acid alcohols are the most interesting in this connection.
The most simple, and therefore the most conclusive cases are
furnished by the glycollic and lactic acids. When heated,
these compounds lose successively a half and then a whole
molecule of water, forming finally glycollide and lactide. I
have shown (Bull. Acad. Belg. xxvii. p. 409) that lactide in
the state of vapour ought to be represented by the formula
[(C3;H,0)O]., and not by the usual formula (C;H,0)O. The
same thing applies also to glycollide, which is nonvolatile.
If we represent the radicals glycollyl (C,H,O) and lactyl
(C3;H,O) respectively by Gl and La, the action of heat on
glycollic and lactic acids will be expressed thus* :—

OH GI—OH Gl
l Z
° on’ °\G1-OH? Gr

OH La—OH La
fs
ba OH? Oa ~O? Oa” if

* The following considerations prove that glycollide cannot be C,H,O,,
but is really a polymer (C,H,O,)n. The minimum molecule of glycollic
acid is represented by CH,OH. COOH, and contains therefore the tetra-
valent radical (=C.CH,—). The corresponding chloride CCl,.CH,Cl is
a liquid, b.-p. 102°. Consequently the oxide, or glycoll, O=C.CH,. #,

ought to be still more volatile, hp as the oxide of methylene is gaseous,
whilst the chloride CH,Cl, is a liquid, b.-p. 40°. In reality, however, this
is not the case, for glycollide is a fixed solid.

of the Metallic Oxides. Fi

Hthylenic hydroxide, CH; . C(OH)3; (or acetic acid,
C,H,O,+ H,O), when distilled, is dehydrated, and forms the
corresponding oxyhydroxide (¢. e. the common acetic acid),
which is represented in the state of vapour, not by the formula

C,H,0, or CH;.CO.OH, but by (C,H,0,2), or
: CH;.C(OH).O.C(OH) . CHs3.
o—

Again, the transformation on prolonged distillation at the
ordinary pressure of certain polyatomic hydroxides, notably
glycerol, C;H;(OH)3, and its monochlorhydrines, C;H;(OH),Cl,
into condensation-products by dehydration are further examples
of this fact.

Returning now to the dehydration of mineral hydroxides,
we have examples in the case of the acid hydroxides.

1. Phosphoric Acid, H,PO, or (HO3)PO, when heated to
200°, is converted into pyrophosphoric acid, H,P,0, or
(HO),PO.0.PO(OH),. Note also, in passing, the various
modifications of condensed metaphosphoric acid, discovered by
Fleitmann and Henneberg.

2. Boric Acid, when dry and crystalline, is represented by
H;BO3. On heating, it dehydrates progressively as follows:—

Dried at ordinary temperature . H;BO3.
Se Pye Pa. + ets. OF, Ea ba).
RST, Ee ee a a te Re ae.
MU ee Ne. ig >, Eo Pg og.

At higher temperatures there is total dehydration and for-
mation of fused anhydride (B,O3),.

3. Silicie Acid.—Normal silicic acid, H,SiO,, is scarcely
known as such. It no doubt forms the gelatinous silica when
freshly precipitated. When heated, it exhibits progressive
dehydration :—

Dried at ordinary temperature . H,Si0; or H,Si,O,.

mee a RO.
pe lgoriey er pd athe 0330 BBEOe
ee merger ates lo) FE SEO,
3 20? 270" . . . ° ° H,S8i,0,,.

After which, total dehydration with formation of (SiQ.),.
4. Titanic Acid.
Dried at ordinary temperature . H,TiQ,.

ee OM Ne ieee atresia kee BAOs Or. TTI,
Beebe scl o aes). iat} xiii athe bl, Oy.
SOE SE: | ee eee ome © 5 bo re

Also similarly with molybdiec and stannic hydrates.

Phil. Mag. 8. 5. Vol. 20. No. 123, August 1885. H

98 Prof. Louis Henry on the Polymerization

It is well to state here that I do not attach an absolute —
value to the formule attributed to a good number of these
incomplete hydrates of mineral acids. It would be difficult
to regard all of them as definite compounds. The fact which
is in evidence, and with which alone we are at present con-
cerned, is that with rise of temperature the dehydration
goes on, and the molecular composition becomes more and
more complicated by the accumulation of radicals in the oxy-
hydroxyl residue. Though these hydrates themselves may
not always have the character of well-defined compounds,
they are often represented by corresponding derivatives
having all the characteristics of true chemical substances.

It is sufficient to mention Troost and Hautefeuille’s oxy-
chlorides of silicon, the numerous polysilicates, so varied in
composition and often so well defined, the condensed meta-
phosphates, the molybdates, &c. Certain anhydrides, which
are totally wanting in oxyhydroxides, are represented by well-
defined salts, as in the case of chromic acid, of which the bi-
and tri-chromates, R,Cr,Q; and R,Cr3;0,9, are well known.
Metallic hydroxides also behave like the acid hydroxides,
thus :-— |

1. Normal Plumbous Hydroxide should be Pb(OH),, but
as ordinarily produced it is already partially dehydrated.
Dried at a low temperature, it has the formula H,Pb,O,
(Tiinnermann, Schaffner, &c.) or H,Pb;0,, in the form of
minute crystalline grains (Payen). At 100° it loses all its
water. | |

2. Hydrates of Copper.—Cupric hydrate, Ou(OH)s, is the
bluish precipitate obtained on adding very dilute alkalies to
solutions of cupric salts. Heated in water, this hydroxide
dehydrates partially and becomes black. It then has the
following composition :—

Dried over sulphuric acid . . H,Cu;0, (Harms).
5» eb100°’..'. . . . “Hea

Normal cuprous hydrate would have the minimum formula
Cu,(OH),. On partial dehydration it gives H,(Cug),0,
(Millon and Commaille), and H,(Cu,),0; (Mitscherlich).

3. Hydrates of Thorium.—One of the known hydrates,
dried at 100°, corresponds to H,ThO,; another, obtained bya
different method, and dried at the same temperature, cor-
responds to H,T’h,O, (Clere).

4, Manganic Hydrates.—Several hydroxides, formed under |
various conditions, correspond to the peroxide MnO,, and are |
as follows:—H,MnO;, H,Mn,0,, H,Mn;0,, H,Mn,0,, and |
H,Mn,0;. All these dehydrate, when heated, forming MnQ,.

of the Metallic Oxrdes. 99

5. Ferric Hydrates.—The following hydrates of ferric
oxide, Fe,Q3, are known:—

Normal hydrate, H,(Fe,)O, . . Limonite, Fe,0;.3H,0.
5 H, (Fe,) O; ss Fe,O,; . 2H,O.
ie ss H,(Fe.)O, ., . Gothite, Fe,O;. HQ.

a H,(Fe.),0, . . Limonite, 2Fe,03.3H,0.
” ” H,(Fe,) 207 ae Turgite, 2F'e,0; ° HG.
” ” Fyo(Fez)30u. “ 3F'e,0; . 5H,0.

On heating, they are all dehydrated to Ie,Q3.

The brown ochreous precipitate obtained on adding an
alkali to ferric chloride is undoubtedly Fe,(OH)., but when
dried in a vacuum it has the composition (Fe,),H,Q,, or
2Fe,0;.3H,O. Heated to ebullition, it dehydrates and
becomes Fe,H,O, or Fe,03;,H,0. By prolonged ebullition
it can be totally dehydrated.

6. Aluminium Hydrates.—Aluminium solutions, on addition
of ammonia in excess, give the normal hydrate Al,(OH), in

the form of a white gelatinous precipitate. On prolonged
ebullition this is partially dehydrated, like the corresponding
ferric hydrate, and becomes insoluble in acids and in alkalies,
like calcined alumina (Péan de St. Gilles). This new hy-
droxide corresponds to the formula (HO),Al,.O.Al,(OH)s,
which is analogous to pyrophosphoric acid. Besides the
normal hydrate Al,(OH),, hydrargillite, there is also the
mineral diaspore, Al,O,(OH), or Al,O3;.H,O, to which we
shall refer hereafter.

All these oxyhydroxides, both natural and artificial, which
are products of the condensation of the normal hydrates, are
completely dehydrated under the action of heat, at a tempe-
rature above that at which they are formed.

Since the same causes produce the same effects, it is natural
to conclude that the last phase of the dehydration is accom-
_ panied, as in those which precede it, by an analogous molecular
_ condensation. The elimination of water having determined
: the accumulation of the radicals in the products to which
| they give rise, it would be unreasonable and illogical to
| suppose that an entirely different phenomenon occurred at.
the moment of elimination of the last molecule of water,
| resulting in the dislocation of these residues of condensation
previously soldered together by the intermediatory oxygen.
Dehydration and condensation are intimately connected.
| We must conclude, therefore, that at the moment the last
, Molecule of water is given off, not only does the previous
| condensation of the oxyhydroxide remain, but that, con-
/ formably to the general rule, it becomes still more accentuated.
| H

100 Prof. Louis Henry on the Polymerization .
Thus, when crystalline plumbous hydrate, 3PbO . HO, loses —

its last molecule of water under the action of heat, it is not
PbO which is formed, but at least (8PbO).. When H,S8igQ,7
becomes the anhydride, we do not get SiOQ,, but at least
(SigO;,)2. For a similar reason metaphosphoric acid cannot
be HPOs, but is at least (HPOs)o, since it is formed by de-
hydration of H,P,O;. ,

The structure of these oxyhydroxides leads to the con-
clusion that they are, in effect, fragments of the normal
hydroxides, R*(OH),, linked together by atoms of oxygen,

thus :—

H,PO, . . PO(OH); | H,PbO, . . Pb(OH),
H,P,0,. .. PO(OH), | H,Pb,0,. . Pb(OH)
| |
O O-
ieee
PO(OH), Pb
Pb(OH)

The fact of the polymerization of an oxide at the moment
of dehydration of the corresponding hydroxide, is also con-
firmed experimentally in certain cases, thus :—The methylic
and ethylic arsenites (CH;);AsO3 and (C,H;);AsQ3 are decom-
posed by water, and give, not arsenious acid, H;AsQ3, but the
solid anhydride, which is, even in the state of vapour, a
polymer represented by As,Og.

The same thing also holds good with methylene oxide. Its
chloracetate, CH,C1. OAc (Henry, Bull. Acad. Belg. xxxv.
p- 717), its diacetate, CH,(OAc),, and the compounds ob-
tained by Friedel (Compt. Rend. lxxxv. p. 2471) are also |
decomposed by water, yielding, not the hydroxide CH,(OH)., |
but the solid polymerized oxide (CH, : O),. |

The metallic oxides can be obtained from their different
salts by various reactions. ‘These general methods furnisha |
number of facts which tend to prove the polymeric nature of |
these compounds. We will examine successively in this con-_ |
nection the various classes of salts, such as carbonates, sul- ||
phates, Xe.

Carbonates.—As a general rule most carbonates are de- |
composed by heat, leaving an oxide. For this reason these |
salts are perfectly assimilable to the hydroxides, and we may |
therefore confirm by their help the results previously arrived |
at. The chief difference is the degree of heat required. i

It is important, from our point of view, to show that de- |

of the Metallic Oxides. 101

carbonation, like dehydration, does not occur all at once, but
is continuous and progressive according to the temperature.
The course of this decomposition is clearly shown in parti-
cular cases, notably in the decomposition of certain insoluble

carbonates in the presence of water, and at a slightly elevated

temperature, and often at the moment of precipitation. The
decomposition results in the formation of basic carbonates,
anhydrides or hydrates, oxy- or hydroxy-carbonates, of a com-
position more or less complex, and in the smallest molecule
of which are accumulated several molecules of metallic oxide or
hydroxide. A large number of metals, e.g. Mg, Zn, Pb, Cu,
Bi, Hg, Al, Cr, Fe, Co, Ni, &., furnish remarkable examples

of this general fact. These insoluble basic carbonates are
usually obtained by precipitation with soluble carbonates,

either from hot or cold metallic solutions. The composition

of these precipitates for the same metal, frequently varies

between more or less wide limits, depending on the special
circumstances under which the precipitation has taken place.
Tt is evident that these various compounds cannot all be con-
sidered as well defined. There are some however, the che-
mical individuality of which has been well established, and a
number of these, either natural or artificial, are capable of
assuming the crystalline state. The following are examples :—

a Mg . . . 8CO,.4Mg0+4H,0.
4CO,.5Mg0+5H,0. Magnesia alba.

6. Zn. Composition very varied.

ZnCOQ3; . Zn(OH), = rae OD
2ZnCO;. Zn(OH),.
2ZnCO;.3Zn(OH),. Crystalline.
4ZnCO;.7Zn(OH),+ H;0 (the preceding dried at 100°).
dZnCO, S oZ4n(OH), -+ iaeg,
ZnCO;.2Zn(OH),+H,O (natural).
ZnOO;.3Zn0 + 2H,0.
ZnCO, ° 7ZnQ = 2H,0.

e. Pb. The following, according to H. Rose, represent the
composition of the hydrocarbonates of lead formed under
definite conditions :—

by mixing cold and concen-

trated solutions of sodium

6PbCO; . Pb(OH)s+ H:0 carbonate and lead nitrate in
equivalent quantities.

SPbCO;. Pb(OH),. . . ditto, but dilute and cold.

3PbCO;. Pb(OH),. . . ditto, but dilute and boiling.

2PbCO;.Pb(OH),.. . . ditto, but with excess of alka-

line carbonate.

102 Prof. Louis Henry on the Polymerization

d. Cu. Two basic hydrocarbonates of copper are natural
compounds, viz. malachite and azurite, but both may be
produced artificially.

CuCO;. CuO. 2H,0.
CuCO;.CuO.H,O (Malachite).
2CuCO;.Cu0O.H,O (Azurite).

e. Co and Ni.
by precipitating a cobalt salt
2CoCO; . 2Co(OH),+5H,0 by alkali carbonate in the
cold.
the preceding precipitate
2CoCO;.38Co(OH), + HO } after washing at a boiling
heat.
‘ the same boiled with carbo-
CoC0; . 300(OH), + H,0 { ite of oa
by precipitating nickel sul-
2NiCO,; .3Ni(OH), + 2H,0 phate with sodium carbo-

nate in the cold.

As a general rule these precipitates become more and more
basic as the temperature of precipitation is higher, and, like
the corresponding hydrates, they become decarbonated and
dehydrated under a more or less intense heat. The cause
which determines the molecular complexity of these com-
pounds, the accumulation of the metallic radicals in the same
product, leads as a final result to a true oxide. The produc-
tion of a simple monometallic oxide RO, by the action of
heat, would consequently compel us to assume, that at a given
moment an effect was produced quite contrary to all that had
preceded it.

Nitrates, Sulphates, and Chlorides.—The following phe-
nomena are quite analogous to that which we have just
examined. We have seen that certain carbonates, formed
in the presence of water, are partially decomposed by the
latter, with the formation of a hydroxide, and the final
production of a basic hydrocarbonate, the radical CO,
being replaced by HO. The nitrates, sulphates, and chlo-
rides of certain metals, notably those whose oxides are
feeble bases, undergo a similar change in the presence of
water. A salt originally neutral in composition is first con-
verted into a basic salt, and finally, under the action of heat,
in the presence of a large quantity of water (either alone or
slightly alkaline), into a true oxide. Thus :—_

Mercuric Nitrate Hg(NQO3)., under the action of cold water,
gives a basic trimercuric nitrate Hg(NO;), . 2HgO, which

of the Metallic Oxides. 103

is a yellow salt, known as turbith nitreux. By prolonged
treatment with water this salt becomes more and more basic.
Kane has described a salt having the composition Hg(NQ3)..
5HgO. Finally it gives mercuric oxide. Nitrate of bismuth

behaves ina similar manner. A large quantity of water, hot

or cold, converts mercuric sulphate, HgSO,, into trimercuric
sulphate, HeSO,.2HgO, which, in the presence of an alka-
line solution, finally becomes mercuric oxide. The solution
of oxide of antimony in fuming sulphuric acid yields, accord-
ing to Peligot, the crystalline salt Sb,0;.2803. The latter
in the presence of water gives the basic salt H,SO, . 2S8b,0s,
which completely decomposes in the presence of a large
quantity of water into oxide of antimony. Antimony trichlo-
ride, SbCl;, gives, when treated with water, white precipitates
of oxychlorides, which are finally converted into the oxide.
One of these compounds is crystalline, and has the formula
5S8b,0,.2SbCl,. It is well known with what facility oxide
and chloride of mercury combine to form oxychlorides
(HgO),.(HgCl,),. These can be obtained in different ways,
and are of very varied composition. Several of them are
crystalline, one of the latter being 6HgO.HgCl,. They are
all converted into mercuric oxide in alkaline solution.

It is only natural to suppose that, when these various basic
compounds are converted, either by the action of water alone,
or in the presence of an alkali, into the oxide, with the total
elimination of the acid radical and its replacement by oxy-
gen, the bonds, which hold together the several radicals or
atoms of the metal in one and the same molecule, remain
throughout. The action of the water having originally de-
termined them, it would be illogical to suppose that at a
particular moment they cease to act.

Acid Anhydrides.—The same reasoning may be applied to
anhydro-compounds produced by the action of strong acids
on their salts, and especially on their acid anhydrides, as in
the case of the chromates.

Chromic anhydride, CrO;, forms three series of salts :—

Monochromates . K,CrO, or K,O.CrQs.
Dichromates . . K,Cr.Q, or K,Q.2CrQ,.
Trichromates . . K,CrsQ,9 or K,0.3CrQ3.

Under the action of an acid, such as sulphuric, the monv-

_ chromate is converted successively into the di- and tri-chro-

mates, and finally into the anhydride. The latter ought,

therefore, to be represented by some multiple of CrOs.
Finally, we find a direct proof of this polymerization in the

decomposition of the arsenites by acids. Among these salts

104 Prof, Louis Henry on the Polymerization

there are some which contain several molecules of As,O3 in
their molecule, such as :—

K,O . (As,Os)o . H,O.
K,O ° (As,O3)o.
2K.0 . (As,0,), . 3H,0.

All these salts, under the action of acids, give arsenious
anhydride, which, even in the state of vapour, is represented
by (As,O3)2. Here we have a proof of the accumulation of
radicals, which exists in the primitive salt, remaining in the
expelled oxide.

Comparative Density of Oxides.

A comparison of the specific gravities of the metallic oxides
with those of other compounds, containing the same radical,
furnishes a new argument for the polymerization of the
oxides.

The specific gravity of a solid body doubtless depends in

a great measure on the state of aggregation, and is also ulti-

mately connected with the composition of the body and the

size of the molecular weight. This connection becomes.

evident in the case of isomorphous solid bodies, which differ
by elements of which the atomic weights are very far apart
as to their numeric value.

The density, then, is to a certain degree a function of the
molecular weight. If the oxides corresponding to the normal
hydrates, less water, be represented by

R*(OH),=R*°O2—5H,0,
‘ 2

it follows that, having a smaller molecular weight than the

hydroxides, their density in the solid state should also be less
than that of the latter. This is in fact the relation which is
observed between the glycols and their anhydrides, which
latter are the oxides of the radicals C,,H,,, thus :—

Molecular

weight, Density.

O,H,(OH),...... 62 | 1-196 at 0°
fe. ne ee 45- i ”
on(Ow 76 ~=«| 1-051
Oh atgge) oligenge x.
CH, (OH), ..| 104 | 0-987 °
Ces Rein 87 | 0824 ”

A difference of an exactly inverse order exists between the

of the Metallic Oxides. 105

metallic hydrates and their anhydrides, the latter having
generally a greater density. The density gradually and suc-
cessively increases as the hydrate loses its water, until it
attains its maximum in the anhydrides or oxides themselves.

For example :—

Molecular

weight, Density.
{ Bo(OH),.....23.- Brucite «1.06.2... 58 2°34
ees: ...<....:]' Periclase 23.02.42: 40 3°65
{ _ 0 ee Crystallized...... 99 2°677
Zn { Crystallized...... 81 6°00 about
Se ai Amorphous...... 81 5°60
Mn,0,.H,O ...| Acerdese ......... 176 4°325
UE RS See Brawnite i. eo 158 4°750
2 (0): ee Crystallized...... 2 dt GD, 1-48
ao LY... iced Snes cw: 70 1°83
p28 (O,.3H.O ..|. Limonite ......... 374 3°6 to 4:0
jHe.0,. 4,0... « Gothite.....620.1. 178 4:0 to 4:4
Fe.0 { Oligiste «3.0541. 160 5:24 to 5:28
| rene Calemed es. 2i05.: 160 5°04 to 5:17
eee 2oHO ..... Hydrargillite:..|. 157 2°35
me sO. «,..|.. Diaspore ¢......2 121 845

Lael Corundum ...... 108 4-00

It should be noticed that a marked difference in density
is observed in the case of substances which are known to be
polymeric, thus :—

Mole-

es Density.
weight

Acetic aldehyde .........02...- 44 0°801 at 0°.
0,H,0...... | Bthylone oxide AER Ronee fee 44 UFS826 «,
(C,H,O),...| Diethylene oxide ........... 88 E028 vias
err 0) aiah i Parsldeh yd: .Jsc. ice -nabivn ee 132 0998 sab 152:
ve ae Isopropylacetylene............ 68 0-652 at 11°.
Gi blig oii. ' Werdbetthene. 7 Hii. .2... és.00% 136 08767 at 0°.
oe eee SNE 56 Cee ae ee 22.2) LODE) ab 9?:
2S ESSE iol ae ee re 70 0°663 at O°.
ae Te os a res 140 UTE 43
EE FAs) ET EAMIV IGN ore esses ee eee 210 03139 ,,

Relations similar to those which are observed between the
hydroxides and the oxides also exist between the latter and
the chlorides. The densities of the oxides of the radicals
C,H,, are notably less than those of the corresponding
chlorides, and this difference is in some degree similar to that
between the molecular weights of these compounds, thus :—

106 —~Prof. Louis Henry on the Polymerization

sete Density.
weight

5 Aaaic aldchgde:., eee 44 | 0-801 at 0°

C,H,0...... | Bthylene Oxide +... 2400S i» 0898 ss,
O.H.Cl ; Ethylene chloride ............ 99 1256 at 12°.
«ie alee ts | Ethylidene ehloride 72s" mo 1189 at 4°:3.
Propionic aldehyde ......... 58 0:8074 at 21°.

oe 2 6 ae Propylene: oxide. is cess cnenee= 3 fy ONBDS> a6 aa ‘

. PCELONO, A252 zc daaaeeeberaeeee: ke 0-312 5,
+ Aldehydie chloride............ 113 1-143 at 10°.
0.1.0! Propylene chloride ......... . 1-1584 at 0°.
3762 °°"! ) Acetonic chloride ............ ks 1-117 ae, Oe
\ Normal propylene chloride | __,, 1-201 at 15°.
( Valeric aldehyde ............ 86 0-822 at O°.
C.H,,0 Amylono oxide: «)i..:.c.itee. et ie eres: " a
; a :

Various Ketones..........0.06 86 0°8099 ,,
0-813 at 20°.
Aldehydic chloride ......... 141 1:05 at 24°.
CH,,Cl,..:..4 Amylene chloride .....22.:+: : 1:058 at 9°.
Acetonic chloride ............ . 106. ai. See

Now the difference between the density of the chlorides and
oxides of the metals in general is exactly the inverse of the
above, for the density of the oxides is notably greater than
that of the apparently analogous chlorides, thus :—

uae Density.
weight
8 8, Srey 111 2°16 to 2205.
{ oS een ne 56 3°08 to 3°20.
{ a Vs IO gS 136 - 2753 (Bodeker).
i (a 81 5°7334 (Karsten).
{ 6 RR ee 183 3°625 (Bodeker).
ON i iain coeur 128 6°9502 (Karsten).
ROOD i eek 278 5802 (Karsten).
PO dothsssties 223 9:2 to 9:5,
PbCl,+PbO | Mendipite......... 491 7:00 to 7:10 (Berzelius),
PbCl1,+2PbO| Matlockite......... 724 7:21 (Rammelsberg).
a hg @) Rae 471 eae bras. 7:14 (Boullay).
‘95 (Karsten).
ee iis 416 { pe ba tHrerecath,
{ BE reeva sue 271 5'4 (Karsten).
21 See (Crystallized) ...| 216 11-29 (Roger & Dumas),
{ AF tee (Buged ): is i055 aiaus 198 3'677 (Karsten).
5k ¢ eee 143 5749 to 6:098 .
{460° i ieAS abe PMEOO oo gs tiecuys 287 Ros (Schroder).
{ 7-143 (Herapath).
er... 233/41 Bo558 (Karten)
jt 6) ee es 129'75| 2°52 (Schiff).

NIG shigepenuess 74:75; 6398 (Bergmann).

of the Metallic Oxides. 107

A difference of the same kind exists between the oxides
and the corresponding carbonates, thus :—

eee ein) POFIEISSS, (rina dens 3°65

MgCO, .| Giobertite............ 2°99 to 3°15
(Piao Were Crystallized ......... 6-00
ZnCO,...| Smithsonite ......... 4:3. to 4°45
RRsat EN > satwodivn 3°08 to 3:2
OaCOe :|' Calespart’ (25.5) .:5-.-. 2-7

Arravowite ~.2....:.- 29

err ca Sie Miiusths 4°61
SrCO, ...| Strontianite ......... 3°6

beta Cia ES ae ei We 9°2 to 95
BECO, - 2.) CeEusitescs .vonysearscs 6°57

{ 3) 6 Fees reg O ler aae 5°4
BaCO, ...| Witherite ............ 4:28

The great difference observed between the density of certain
oxides and that of their basic anhydrous salts is equally
worthy of note, thus :—

Molecular weight. Density.
= 2) Sere 216 10°69

A study of the oxides, from both a physical and chemical
point of view, leads, therefore, to the general conclusion, that
jor the most part the oxides cannot be compared with the
chlorides ; that the true oxides which are really comparable with
the chlorides are unknown; and that we possess only polymers

thereof (RO;)n.

Value of the Coefficient of Polymerization.

What, now, is the true value of the coefficient n of polyme- »
rization? What is the real molecular formula of these
polymeric oxides? These questions are doubtless of great
interest, but it should be stated at once that it is absolutely
impossible to give an exact answer. I do not know of any
fact which would allow us to assign an absolute value to the
coefficient of polymerization, for we are in exactly the same
position in respect to the metallic oxides as we are in the case
of the solid polymeric ones of organic chemistry, e. g.
methylene oxide, metaldehyde, &c. So far as facts will per-
mit of a conclusion, we may affirm that in most cases this
number is very high, although different for different oxides. The
molecular weight of most oxides should therefore be consider-
able. This conclusion results from a consideration of certain
facts previously referred to, and to which it will be necessary
to return for a moment.

108 Prof. Louis Henry on the Polymerization

Oxygen is an element endowed in reality, in spite of appear- 1

ances to the contrary, with great power of volatilization. It |}
forms oxides which are gaseous, or at least very volatile, with |}

elements which are relatively but little volatile, as sulphur ;
or which are even entirely fixed, as carbon and osmium. To |
these normal oxides, which are regular at least in their physical
properties, there correspond chlorides which are notably less
volatile, as in the case of those of carbon :—

CO, gaseous, b.-p.—78..
CCl, liquid, b.-p.+ 76.

That being so, it may be affirmed that many of the regular |
metallic oxides which correspond to the chlorides should be
gaseous at the ordinary temperature, or at least very volatile.
This should be especially true of SiO,, Al,Q3, &.; for SiCl, is
a liquid boiling at 58°, and Al,Cl, a solid volatile below 100°.
In reality, however, these oxides, which I regard as polymers,
(SiO,),, and (Al,O3),, are fixed solids. Organic chemistry
furnishes examples of the relationships which exist in poly-
mers between the volatility and molecular weight, thus :—

Molec. wt. B.-p. Difference.
Kithylene oxide ©,H,O . 44 liquid 1375 885
Wthylene dioxide (C,H,O), . 88 liquid 102 }
Aldehyde O;H,0 —. 44° "liguid “28 \ 103°
Paraldehyde (C,H,0)3; . 1382 liquid 124

We thus see what is the difference in volatility between
bodies of which the molecular weight in the state of gas is
simple, double, or treble.

Now between a body which ought to be gaseous or very
volatile, such as SiQ,, and a fixed body, such as silica really
is, there is a difference which cannot be compared to that
which separates (as regards volatility) ethylene oxide and
aldehyde from their polymers. If silica is, as I believe,a
polymer (SiOQ,),, then the value of n should be considerable,
and the molecular weight therefore very high. We come to
the same conclusion if we consider the examples which are
furnished by the relative densities of the chlorides and oxides of
the radicals C,,H,, and those of their polymers. As already
remarked above, the density of the chlorides is higher than that
of the corresponding oxides, whereas the densities of the poly-
mers of the latter rise, approaching those of the chlorides, and
even going beyond. The densities of the metallic oxides are
often considerably greater than those of the chlorides. If we

of the Metallic Oxides. ; 109

bear in mind the small difference which exists between the den-

sity of the oxides C, H,,0 and that of their polymers (C,,H»2,0),
or (C,,H;,0)3, we are led to conclude that the coefficient n of
polymerization of the metallic oxides must be very high, in
order to raise the density of an oxide to nearly double that of
the corresponding chloride, as is frequently the case.

Doubtless the molecular condensation is different with

different metallic oxides. A priori we should expect it to be
so. This is proved by the fact that the difference between the
fusibility, volatility, density, &c. of the oxides and the corre-
sponding chlorides is very different in the several cases.

This condensation appears to attain its maximum in certain
fixed and very infusible oxides, such as silica and alumina
&ec., the corresponding chlorides being eminently volatile.
This enormous condensation of their molecules may possibly
be the cause of the greater resistance which they offer to the
action of chemical agents, more especially to simple chemical

agents such as H, C, Cl, 8, &e.

Origin and Mechanism of the Polymerization of the Oxides.

What is the origin, or, rather, what is the mechanism, of this
molecular condensation which is so general in the oxides, and
yet so absent in the chlorides? The diverse chemical nature
of the two elements accounts for the different molecular
behaviour of their compounds.

Chlorine is essentially a monovalent element, whilst oxygen
is divalent. The result of this quality in oxygen is the pos-
sibility of its uniting by its two affinities with two polyvalent
radicals. | |

From the atomic and molecular point of view, polymeri-
zation consists essentially in a simple change in which each
atom of oxygen becomes united to two different atoms of the
radical belonging to different molecules, and maintains them
permanently together as a whole. The conversion of acetic
aldehyde into its triple polymer, paraldehyde, is an interesting
example of this atomic migration :—

CH. eo eee . CH;

CH,;.CH:O becomes ee oa

Similar facts occur, and in a similar manner, in the case of
the metallic oxides, thus :—

110 ‘Prof. Louis Henry on the Polymerization

H,Pb",0, Hg"(NO;)0, -Hg($0,)0,
i. A iy
Pb Hg O
, .* Of
oe a Hg
mY
ane \ rs
O O Hg
rd ha ~*~
Pb Hg O
‘OH ‘o He
He \
S
NO;

The corresponding polymeric oxides (PbO), and (HgQ)»
being probably represented by analogous formule :— |

(PbO), (HgO)n
PU Pew st b. Pb Hg He... he ee
Pe are x Sars ae We Se eee 4 ay Bs
O Qos Cit 4 O
O O O O
nih teh ces LUT eae Yoke Sa \ 7m
Pb Pb... Pb Pb He Hg... . He Be

There is thus formed a kind of closed chain of polyvalent
radicals alternating with n atoms of oxygen. It is thus seen
that the number of atoms of the radical R accumulated in
such a complex molecule may be considerable, and is theo-
retically illimitable. Similar groupings cannot occur in the
case of the chlorides, on account of the nature of chlorine,
which is a monovalent element, and consequently incapable
of uniting directly with two other distinct atoms. All mole-
cular condensation thus disappears, when in a polymeric
oxide the oxygen is replaced by chlorine, as in the case of
(As.O3)., which, under the action of chlorine, is converted
into AsCl;. Also, as in the case of the polymers (C,H»,0),,
more especially those of aldehydes, which are converted by
PCI, into the chlorides C,H,,,Cl,, identical with those fur-
nished by the oxides (C,H»,0) themselves.

Heat of Formation of the Oxides.

The combination of a body with itself, or the act of poly-
merization, is accompanied by the development of heat, as

— + ”
a
—
ei.
,2

of the Metallic Oxides. 111

with every combination of heterogeneous bodies ; and this dis-
engagement of heat may sometimes be considerable, as in the
case of the internal combustion (spontaneous in the com-
plete sense of the word), which certain metallic oxides

undergo on heating (owydes cuits of Chevreul).

The quantity of heat disengaged on the formation of an
oxide RO, in the direct way has a double origin, viz. the heat
of combination of the elements and the heat of polymeriza-
tion ; so that the direct formation of a chloride is not at all
comparable with that of an oxide. A comparison, therefore,
of the phenomena of combination, from the physical point of
view, shows that the heat evolved is an absolute measure of
the difference of the affinities between the elements of the
compound. That being so, we should conclude that the dif-
ference between the affinity of oxygen and chlorine for the
metals is in reality greater than the thermal phenomena, as
determined directly, would lead us to assume.

Action of Heat on the Oxides—their Classification.

The oxides may be divided into various groups, according
to the changes which they undergo on heating.

A. Normal or Regular Oxides, which correspond molecu-
larly to the chlorides. These oxides are few in number, e. g.
SO,, CO., NO, N,O, CO, &e.

B. Polymerized Oxides, the molecules of which are made
up of z molecules of the normal oxide united together. This
group includes most oxides, and may be divided into several
classes according to the action of heat thereon ; viz. :—

(1) Volatile oxides, which are totally depolymerized on heat-
ing, and in the state of vapour are converted into the normal
oxide, The number of these oxides is limited, and they are of
two kinds :— |

(a) Oxides which are completely depolymerized on volatiliza-
tion, yielding a vapour of normal density, e. g. SOs,
OsO,, and probably RuQO, ; also various organic oxides,
such as methylene oxide (CH,0O),, metaldehyde
(CH;.CHO),, and aldehydes and acetones in general.
On condensation the original polymeric oxides are
re-formed.

(b) Oxides which are depolymerized very imperfectly at the
moment of volatilization, and only completed progres-
sively as the temperature rises. The vapour-density of
these bodies, therefore, gradually diminishes with in-
crease in temperature up to a certain point, beyond
which it becomes constant, and corresponds to that of
the normal oxide, e. g. N,O,, fatty acids, and paral-

_S== Se! = . Se -—— — —— — —
© a a ee
a
t

.

———_——=s+
eta

“= |

112 Prof. Louis Henry on the Polymerization

jected; e. g. (As,0O3)n, the gaseous molecule of which is
(As,O3 )o*. : ;
Lactide, or the dioxide of the radical C;H,iv (orCH;. C=)

is perfectly analogous to arsenious oxide, its molecule in the
state of vapour corresponding to (C3;H,O,)o. Hthylene dioxide
(C,H,0), probably belongs to this class, as in the state of
vapour its molecule = (C,H,O)p. |
(3) Oxides which are not capable of depolymerization. Most
inorganic oxides belong to this group ; also the oxides of
certain organic radicals.
They are of two kinds :— 7 |
(a) Oxides which are absolutely indepolymerizable and fixed,
as in the case of many metallic oxides which are not
decomposed by heat; e.g. Si0., BeO3, SnOs, ZnO,
MgO, Fe,Os, Cr,Os, Al,Os, PbO, CuO, Cu,0, Ke.
(b) Oaides which are indepolymerizable, because on heating
they are decomposed eather into their elements or into new
compounds, as in the case of several metallic oxides ;
e.g. HgO, Ag,O, Hg,O, CrOs, MnO,, &e., and also of
certain organic oxides, as glycollide (C,H,O;),, which
carbonizes on heating. | :
The above classification is conveniently represented in the
following table (p. 113). |
It follows from all these facts and considerations that there
are large and important lacunee in the series of oxygenated
compounds of Mineral Chemistry. Shall we ever be able to
obtain the whole series of normal oxides, of which we have at
present so few examples? Iknow not. But to produce them
we should doubtless have to work under conditions totally
different from those which are usually employed. Our actual
methods of preparation are precisely those under which these

unknown bodies polymerize, and thus escape us. If it be—

rash to hope to obtain these bodies in the near future, it would
be still more so to say that their discovery was impossible.
This word is profoundly repugnant to our tendencies and scien-
tific habits. When we measure the progress of science since the
commencement of the century in a domain so vast as Che-
mistry, there are many reasons for not being discouraged.

* Arsenious oxide volatilizes, as is well known, at about 200° without
melting. This temperature is not the boiling-point propel so called,

but is really the temperature of depolymerization of the molecule (As,O,)n
into (As,O,)>.

113

—

Normal oxides, not polymerized ...ssse...ss00e sroreentereeessvstsresssseerevsssesesssvese OO, CO, SO,, NO, NO

( At the temperature of ee SO;, Os0,, (CH,0),, &e.

tion or of volatilization ...

Totally... Imperfectly depo] merized on
Gee ia meets
ebullition, but depolyme-
rized on superheating the (2s; paraldehyde de.
VE pOtire sks tena th chap
: 7
bs

Volatile and
[ depolymerizable ...

Polymerized j Ineoii plete lv.i.cb Garam ies Gedel ative, As.03; C3H,0.; (C.H,O)p.

oxides ... |

Undecomposable and fixed .......0..0.... Si0,, SnO,, Al,O;, Cu,O, &e.

of the Metallic Oxides.

[i 2 MON te . . dope mee eee

Indepolymerizable J ( Totally, into their ele- } Ag,0; HgO; Hg,0, &c.
, Sia tae eh

| Partially, into ao Bal, CrO;, Mn0,, &e.

iS ORION sas piven ies

Phil, Mag. 8. 5. Vol. 20. No. 123. August 1885,

414. ~——~ Prof. Louis Henry on the Polymerization

APPENDIX.

Dehydration and Condensation of the Hydroxides
by eat.

It has been already stated that the hydroxides, under the
action of a gradually increasing temperature, undergo a pro-
gressive dehydration, whilst at the same time their molecules
become more and more condensed by the accumulation of the
residual oxyhydroxides of the radical of the primitive compound.
What is the cause of this general fact? This phenomenon is
analogous, both in appearance and result, to that of direct
etherification, 2. e. to the action of an acid on an alcohol. In
fact, etherification is an example of the production of an
anhydrous oxide from its hydroxide. The most simple case
is that in which we can follow step by step the work of
dehydration, as with ordinary lactic acid.

Lactic acid, C3;H,O3, is the dihydroxide of the radical
lactyl, C;H,O, which we may represent by La. On heating,
lactic acid gives successively two anhydrides, the so-called
lactic anhydride,

2C,;H,O,— H,O=HO s La ° O ° La + OH,

and lactide,

2C;H,O3,— 2 BO La . O ° La ° O.
Pact |

The dehydration thus takes place in two steps, and requires
at least two molecules of acid—a fact which is easily under- |
stood from the usual formula of lactic acid,

CH;. CH(OH). COOH.
The two hydroxyl groups are not equivalent ; the one is acid
and the other alcoholic ; so that the action of heat on lactic
acid is really the action of an acid on an alcohol, which
mutually and successively undergo etherification, giving the

so-called lactic anhydride, which is at the same time an acid,
an alcohol, and an ether,

CH,.CH(OH).CO.0.CH.COOH,
OH,

and which furnishes lactide,

CH;
|
CH, OB .CO.0 GH oooh
Zina Someone OOP et ee Me:
which is only but doubly an ether. If the radical lactyl were

known only en bloc, the reason of this difference in the nature
of these hydroxyls would totally escape us. 2

|

|

of the Metallic Oxides. 115

The molecule of lactide, which in the state of vapour is
(C3H,O,),, ought in its natural state to be still more complex.
The boiling-point of this substance is really the temperature
at which the grouping (C3;H,QO,)o, becomes (C3;H,O,),. This
is the logical conclusion from the molecular constitution of.
carboxylated acids. The molecule of acetic acid must be at
least 7

The formula CH;. COOH, commonly attributed to this acid,
only represents the product of the simplification of its mole-
cule under the action of a high temperature. That being so,
the molecule of lactic acid should be at least

|
CH,

In the formation of dilactic acid and of lactide, two of such
molecules act on one another, and consequently the molecule
of lactide ought to be at least [(C3H,Oz)¢ |o,2. e. the double of its
molecule in the state of vapour. However that may be, the
hydroxides of the polyvalent radicals of mineral chemistry are
assimilable to the above organic hydroxides; and we may

_ therefore reasonably conclude that the hydroxyls which they

contain are of different value. This question may be answered

_ affirmatively and with certainty in some cases, notably that of
_ phosphoric acid. Thomsen, Berthelot, and Longuinine have

shown what is the real nature of this compound.
Orthophosphorie acid, PO(OH)s, is a tribasic acid, but of

quite a different kind from citric acid, in so far as in the latter

the three hydroxyls are perfectly equivalent, as shown by the

_ identity of their heats of neutralization. Citric acid, as shown

by its formula, is equivalent to three molecules of carbonic
acid rolled into one.

CH,—COOH
HO.C.COOH
|
CH,—COOH

12

116 On the Polymerization of the Metallic Oxides.

Phosphoric acid, on the contrary, is equivalent to three
different acids, the heats of neutralization of the three basicities
of the acid by the same metal being unequal.

H,P0O, 4- NaHO = 14°68 cals.
+2NaHO=26°33
+3NaHO=33°59

The substitution of hydrogen by sodium thus causes suc-
cessively the evolution of 14°68, 11°65, and 7:26 cals. As
regards two of its hydroxyls, orthophosphoric acid is strongly
acid, though unequally, the one having an acidity similar to
that of hydrochloric or nitric acid, and the other to carbonic
acid ; whilst the third hydroxy] is more like that of an alcohol
or phenol.

The action of heat on phosphoric acid is really the action of
heat on an acid-alcohol, and ought therefore to exhibit phe-
nomena similar to those observed in the case of lactic acid.
These phenomena ought even to be more complicated, because
for one molecule of PO(OH)s; acting as an alcohol there is
another molecule acting as an acid, and can function by one

9)
9

or other of its hydroxyls, which are not only acid, but un- |

equally acid.
The formation of pyrophosphoric acid seems to be perfectly
analogous to that of phosphoglyceric acid, thus :—

(HO),PO.0.PO(OH), (HQO),..C3;H;.0.PO(OH),.

Two isomers, however, appear possible. Besides the three
hydroxyls of phosphoric acid having a different chemical
value, it follows that the products of the phosphoric etherifi-
cation, from the partial dehydration of H;PO,, ought to be
very numerous and complicated. Our present knowledge of
the action of heat on this acid is therefore very superficial.

Can we refer the condensation of the metallic polyhydroxyls
under the action of heat to the same cause—1. e. to a difference
in the nature and value of the several hydroxyls? It seems
natural and logical to do so, but it is difficult to find sufficient
authority.

Phosphoryl being a multiple radical, formed of hetero-
geneous atoms, one can rigorously account for the fact that
the three hydroxyls, which are grouped around it, are un-
equally placed as regards these elements, and can therefore
have different values. But to conclude that the different
hydroxyls attached to the same polyvalent atom, such as Si,
Sn, &., have not the same value, it would be necessary to —
assume that the polyvalent atoms are not identical at all |

On the Thermoelectric Position of Carbon. 117

points, or at least that the various units of chemical action are
not equivalent, and are therefore of unequal energy. Here
we introduce one of the higher problems of chemical meta-~
physics. I abstain, however, from dealing with it at present.

If the condensation which certain hydroxides undergo on
heating is explicable in certain cases, there are others much
more numerous, which still remain without plausible expla-
nation, and notably is this the case with the hydroxides of the
elements.

On the Oxide of Methylene, CH, : O.

There is no compound more interesting than this, from the
point of view of the polymerization of the metallic oxides. There
is a perfect parallel between the reactions by which it may be
obtained from its ethers and those by which certain oxides
are prepared from their salts. Just as the arsenious ethers
under the action of water give arsenious anhydride, so do the
methylenic ethers, when treated with water, give methylene
oxide, i. e. its solid polymer.

These reactions may also be compared with those which
give certain metallic oxides, as Sb,03, Bi,O3, HgO, &c., by
the decomposition of their salts with an excess of water, or by
weak alkaline leys. Methylene (CH,) acts in all respects
like a metal, and its oxide is strictly comparable to a metallic
oxide, with this difference, that when heated it is completely
depolymerized on volatilization, whereas certain oxides, as
(As,O3),, are only imperfectly depolymerized, whilst others,

~ such as HgO, Bi,Os, &c., are fixed.

In a special paper I intend to return to the oxides and >
chlorides of the dicarbon radicals C.", C,H,'*, C,H,#.

XIV. On the Thermoelectric Position of Carbon,
By Joun BucHANAN™.

[Plate VIII. ]

aN the very interesting paper read by Dr. J. A. Fleming
Z before this Society on March 14, 1885, whilst discussing
the question of the life of incandescent lamps, the author re-
marks that quite a large proportion of the carbon filaments
are found to break at the negative end, that is, near the point
where the current passes from carbon to platinum. It seemed
to me not improbable that the negative end of the carbon
might be subjected to a heating action due to the operation

* Communicated by the Physical Society : read June 27, 1885,

118 ‘3 Mr. J. Buchanan on the

of the ‘ Peltier effect.’’ The local excess of temperature thus —
produced might possibly account for the observed result. In
any case the determination of the position of the carbon line
on the thermoelectric diagram was of interest. I am not
aware that this has been done before.

_ Result—For the specimen of carbon I have tried, the
Thomson effect is of the same sign as in copper, and about
twice the value for that metal given by Prof. Tait, as quoted

in Hverett’s “ Units and Physical Constants,” §186. The
thermoelectric power of carbon is given by my experiments,

¢,= —390—1°87¢.C.G.8, units, |. an

“ots Cmer=

where ¢ denotes, as usual, the mean temperature of the junc-
tions. The neutral point of carbon with lead is therefore
about — 209° C. Mast.
Thermo-couple of Platinum-Carbon.—As I have mentioned,
the series of experiments described below were undertaken
with the object of finding the magnitude of the Peltier effect —
in incandescent lamps. I therefore examined specially the
behaviour of a thermoelectric couple of platinum-carbon.
The platinum was a piece of ordinary commercial wire pro- —
cured from Messrs. Johnson and Matthey. The carbon was ©
in the form of rods 12 millim. diam., such as are made for use
in arc lamps. The direction of the current in such a couple ~
is from platinum to carbon across the hot joint. Thethermo- —
electric power at the mean temperature ¢ is

Cnr_e= +56643'94t C.G.8. units. . . (2)

Due, therefore, to the “ Peltier effect,’ heat will be gene-
rated at the negative end of the carbon filament of an incan- —
descent lamp when the current is passing. We can now —
calculate what proportion the heat thus generated locally —
bears to the whole quantity of heat generated in the filament. |

If T denote the neutral point of a thermo-couple, then the _
Peltier effect for the absolute temperature @ at one joint is |
given by j

m=(K,—K,)(T—0)6;,. 2 we

where K, and Ky, are the numbers that express for each |

material the ratio that the Thomson effect bears to the abso- |

lute temperature. (See Tait’s ‘ Heat,’ § 415.) |
Hence for a platinum-carbon couple,

7394 (14446)0,. . -. -. oS }

As a specific example, take the case of a 100-volt lamp, |

Thermoelectric Position of Carbon. 119

with the filament at a temperature of 2400° C. We get the
ratio
Heat due to Peltier effect 3°94 x 2673(144+4 2400)

eee heat’ lament "100x108"
; 2-7 x 107
” ay aed 2
= 71002:

9? :
The local heating at the negative end of the filament is

therefore only about one quarter per cent. of the whole heat.
This is too small to have any appreciable influence on the
life of the lamp. It does not appear therefore that the Peltier
effect accounts for the giving way of the carbon-filament at
one end rather than at the other. In the course of these
experiments one phenomenon was forced on my attention,
the details of which are given below. It seems that con-
tinuous heating alters the thermoelectric properties of carbon.
This alteration must be the result of molecular change in the
material. That this has any bearing on the destruction of
the filament that ensues, by keeping it for a long time ata
high temperature, my experiments are not extended enough
to decide.
Expervmental Details.

The E.M.F. at various temperatures was determined by the
null method known as Poggendorft’s.

The primary circuit contained a constant cell A, a gradu-
ated wire BC, and a box of resistance-coils R.

On the graduated wire were placed sliding pieces e and /,
which served as the terminals of the secondary circuit. In
the secondary circuit were a galvanometer G, and the thermo
couple T whose E.M.F’. was to be measured, and a key K.

Fige 1.

A

The arrangement of the apparatus used in the platinum-
carbon experiments has now to be described.

120 Mr. J. Buchanan on the

Three rods of carbon of 1°2 centim. diameter, such as are _

manufactured for use in arc-lamps, were taken. One of these —
was thinned down into a conical form at the end destined to

go into the hot bath. This end, and that of another rod

intended to go into the cold bath, were electroplated with a
thin shell of copper. To this copper shell were soldered the
platinum wires, whose behaviour with the carbon I wished
to examine. These two rods were then placed vertically and
clamped in position with coppered ends dipping into their
respective baths. The upper ends of these carbon rods were
cut off square and the third rod laid horizontally, so as to
form with the vertical rods the third side of a vertical rect-
angle. |

io ensure good contact one end of this third rod was bound
by a number of turns of clean iron wire to the upper end of
the rod whose lower end dipped into the cold bath. The
other end of this third rod rested on the top of the rod the
lower end of which dipped into the hot bath. The surfaces
of contact were here scraped flat and kept pressed together
by a weight of 3 or 4 pounds.

The platinum wires that formed the other element of the
thermo-couple were of ordinary commercial material, as was
mentioned above.

The galvanometer used in the secondary circuit is a very
delicate mirror instrument of about } ohm resistance.

The cell used in the primary circuit as a standard was a
Daniell. The graduated wire (BC in fig. 1) was of German
silver.

The baths, both hot and cold, were oil-baths.

The ends of the platinum wires of the thermo-couple that
were exterior to the baths were joined by binding-screws to
copper wires. These joints were insulated from one another
by several layers of calico, and then tightly wrapped up to-
gether. ‘Thus in the secondary circuit the only exposed con-
tacts of dissimilar conductors were those of the copper wires
joined up with the brass binding-screws of the key, of the
galvanometer, and of the sliding piece on the graduated wire.
During the experiment, care was taken to screen these exposed
contacts as much as possible from external radiation. More-
over, all the apparatus was set up the night before, so that on
the days of the experiments no metal joints whatever in the
secondary circuit were touched by hand.

Some preliminary experiments I had made showed the —

necessity for adopting the arrangements just described. The
results of these preliminary trials exhibited very curious
irregularities, not altogether got rid of even by those latest

Thermoelectric Position of Carbon. 121

arrangements. By using the precautions detailed above, I
ensured, however, that the place of any disturbances that
might come into play should be located with certainty in the
thermo-joints under examination.

The proper amount of resistance to be inserted in the
primary circuit having been determined beforehand, the
primary circuit was closed. The current was allowed to
flow during the whole time the experiment lasted. The
temperature of the hot bath was raised very slowly up to the
highest point reached; then as slowly lowered again. Frre-
quent readings of the thermometers in the hot and cold baths
were taken. Simultaneously were noted the positions of the
movable sliders on the graduated wire, for which no deflec-
tion was shown by the galvanometer on depressing the key in

the secondary circuit.

When a temperature reading was intended to be taken,
the temperature was kept as steady as possible for} to4a
minute, accompanied with constant stirring of the hot bath.

_ In this way a series of corresponding readings of temperature

|
: :

and of distances on the graduated wire were obtained. The
results were plotted out, taking centimetres of the graduated
wire as ordinates, and temperatures of the hot bath as ab-
scisse. A regular curve was then drawn amongst the points
thus obtained, so as to give as nearly as could be judged the
mean of the observations. The curves A, B,C (Plate VIII.)
were obtained in this way. The scales for these curves are
not quite the same. Details are given below. The curves A
and B were obtained on successive days.

‘The want of regularity in the individual readings, after all
the precautions taken, I considered could only be due to the
properties of the carbon. Partly to satisfy myself that no
essential precaution had been overlooked, however, and partly
to test an idea I formed as to the cause of the irregularities,
the apparatus was set up again about a month later. Curve
C gives the results of the observations. The heating of the
hot joint gave the curve to the left hand; the subsequent
cooling gave the right-hand curve. These curves have not a
point of intersection at their upper ends shown in the diagram
(Plate VIII.), because the temperature was intentionally raised
to a point higher than that for which a reading on the graduated
wire could be obtained, before cooling down was begun.

Now, as 43 hours were given to the observations, the heat-
ing and cooling took place very slowly indeed. And as a
stirrer was kept almost continuously in operation in the hot
bath, the lagging of the thermometer behind the thermo-
joint must have been very small. I believe that the dif-

122 Mr. J. Buchanan on the

ference between the curves obtained by heating and by ~
cooling is partly due to change taking place in the properties —
of the carbon as the temperature changes. Bearing on this
there is an observation [made. Having calculated the neutral
point of a carbon-iron couple as about +208° C., a glyce-
rine bath was used to verify this result. Having raised the
temperature to about 290° C. and cooled again to 220° C.
several times, I found the H.M.I’. of the carbon-iron couple
rose gradually, until at the time of stopping the experiment
the maximum E.M.F. had increased by more than 50 per
cent. . As is well known, a thermo-couple reaches a maximum
H}.M.F. when the temperature of one joint is that of the

“neutral point.” It was this maximum that was found to

increase with the time during which heat was applied.

The conclusion is, that the thermoelectric properties of
carbon alter with temperature. Principally for this reason I
did not deem it of utility to begin a new series of observa-
tions, and by “loading” the hot joint with metal try to
decrease the irregularities in the readings like those in curves
A and B. Moreover, the results calculated from these curves
agree as well as could be expected. Here are the details:—

Curve A.
Assuming the equation to the curve as

“ta
y=K(a—a)(1-"3*),
I find that ~ =T~=_147° C., K=0-00470

satisfy best the observations as represented by the curve A.
x=temperature of hot bath,
= mS cold bath,
T=neutral point.

The agreement of calculated and observed values of y is

exhibited by Table I.; these values, it will be recollected,
represent centimetres of the bridge-wire. To calculate the
B.M.F. of a carbon-platinum thermo-couple as given by
curve A, we have the following data :—
Resistance of primary circuit. . . . . =803°6 ohms.
Observed E.M.F. of Daniell used. . . . =1°08 volt.
Resistance of 1 centim. of the graduated wire=:00246 ohm.

Denoting the H.M.F. required by E,, we get

B, = 00470 x 1:03 x 108 Tt rf —a)(147 475),

Ba =3'92(. v—ay( 147 Nis +) CG. uaaie

Thermoelectric Position of Carbon. 123

Hence the ote ing power of the couple is
= +577+3°92 xt C.G.S. units,

¢ denoting Np mean temperature of the joints. The direction
of the current would be from Pt to C across the hot joint.

Curve B.
Taking an equation to the curve of the same form as that
just used, T= —140° C. and K=:00406 best satisty obser-

vation.

TABLE I.
| y-
x a. ooo
Calculated. Sis Babee |
|
| 30 9°75 159 15°8
40 9 82 24-4 24-9 |
50 9°85 33°4 33°4
| 60 - 42°9 43°1
70 4 52°9 53-0
80 9°9 63:2 63°4
90 » 74:2 [4:1
100 a 85°5 851
110 Es | 97°4 | 97:7
TABLE II.
| y.
4 & a.
Calculated. Observed.
(@
30 9°6 13°5 | yl
40 9-65 Bes. | aS
50 9°70 27°8 27°8
60 9°75 30°F art
70 a 43°8 44-0
80 vu: 52°6 52-7
90 9-80 61:9 61°9
100 3 Th 7 71:4
110 9°85 81°8 82°4
120 5 92°6 91°6

|

|

Table II. exhibits the agreement of the calculated and
observed values of y when these constants are inserted in the
equation to the curve. ‘To calculate H.M.F. (Kp), the data
are the same as were used in the calculation of Hy, except
that the resistance in the primary circuit was 259°6 ohms.

-, Hy=00406 x 1:08 x 10° x SP (e —a)(1404°4 -)

aU+a

=3-96(2—a)(1404°3 ) C.G.S. units.

124 Mr. J. Buchanan on the

Hence for the thermoelectric power (e,) at the mean tem- |
perature ¢, we get :

ép= +555+3'96¢ C.G.S. units.
The mean of e, and e, being denoted by e,:_,, we get
Cnt—o= +566 + 3'94¢ C.G.S. units,

the result quoted in equation (2) above.

Thermoelectric Power of Carbon.

In order to find the position of carbon on the thermoelectric
diagram, I determined the line on this diagram of the platinum
employed i in the above experiments.

A strip of commercial lead, about 80 centims. long, had the
platinum wires soldered to the ends. The general arrange-
ment of the apparatus and the precautions used were similar
to what have already been described above. Curve D shows
the results of the observations ; the ordinates and abscissze
being the same as in the curves ‘A, B, C. It will be noticed
that the observations lie well on the mean curve, thus showing
that the arrangements were not defective in any essential
point.

Assuming y etn (r- a+ +)

as the equation to the curve, as above, I find that T= —85° O.
and K =-004815 give results agreeing best with observation.
(See Table III.)

TABLE ITI.
y.
Me eR TT
Calculated. Observed.
“a Maa ) meg

30 13°71 8°67 9:05

50 45 20:7 20°6

70 ie 34°6 34°77

90 7 50° 50°38

110 i 68°3 68:2

130 Ms 88:1 88:0
The data for calculation are :—
Resistance of primary circuit . . =1001 ohms.

Mean observed E.M.F. of the Grove rial
ied oe =1°75 volt.
employed .

Thermoelectric Position of Carbon. 125

Hence the E.M.F. of the lead-platinum couple and the
thermoelectric power of platinum are, respectively,

B=-004815 x 1-75 x 10°79" (w—a) (85+ =");

E= 2-07(«—a)(85 + “ C.G.8 units,
.@pt= +176+42:07¢ C.G.S. units.

The current would be from Pt to Pb across the hot junction.

Using this value for e,, we find for the thermoelectric
power of carbon

e,= —390—1°87¢ C.G.S. units.

This is the result quoted in equation (1) above.
In the table of thermoelectric powers given by Everett
(‘ Units and Physical Constants,’ § 186) there are given

for Zn, e, =—234—2°40t,

» AZ, Cag —214—1°508.

The value for e, shows that the place of carbon in the table
is between zinc and silver. In the thermoelectric diagram it
is easy to see that for temperatures below 50° C., the line
for carbon lies below that for cadmium, taking the value
for ¢.g from Everett’s table (loc. cit.).

Concluding Remarks.

The disintegration preceding complete destruction of the
carbon filament in an incandescent lamp, described by Mr. W.
H. Preece, F.R.S., in a paper read before the Royal Society
on March 26, 1885, is doubtless the result of molecular
changes in the structure of the filament produced by the
continued maintaining of the filament at an extremely high
temperature. In my experiments, keeping carbon at a
moderately high temperature altered the molecular condition
of this material, the alteration being manifested as a change
in the thermoelectric power of the carbon.

_ Physical Laboratory, University College,

London, June 1885.

ro 126

XV. On the Use of Daniell’s Cell as a Standard of Electros 4 \
mote Force. By J. A. Furmine, M.A., D.Sc. (Lond.), —
Fellow of St. John’s College, Cambridge, Professor of Elec-
trical Technology in University College, London*.

[Plate V. ]

f Raee extensive use now being made of strong currents of
electricity for purposes of illumination necessitates the

|
|
|
! employment of instruments for measuring current-strength
; and electromotive force of high values. The employment of
electricity for telegraphic purposes, and especially in sub-—
i marine telegraphy, compelled attention to a common accuracy
in measurement of resistance and capacity; but for these
purposes the measurements of electromotive force and currents
il are not of such importance as they are in the recent develop-
ments. This requirement has drawn the attention of electro-
) mechanicians to the subject, and led to the introduction of
| numerous forms of electromotive force and current-indicators,
| variously called potential galvanometers, voltmeters, pres-
: sure-gauges, ampere- or am-meters.
: Considering first. the voltmeters, they may be classified,
| according to Sir W. Thomson’s nomenclature, into two classes
f —idiostatic and heterostatic. In the second class the magnetic
.. force at a point in the neighbourhood of a conductor con-
| veying a current is compared with another magnetic field
| either of the earth or of a permanent magnet, and the differ-
| ence of potential at the ends of the conductor deduced there-
: from. In the first class no magnetic field is made use of other
than that created by the current itself.
| In the case of heterostatic instruments the readings require
} | to be controlled by a special determination of the value of the
auxiliary field; and in the case ofan ordinary tangent- or sine-
galvanometer, the determination of the earth’s horizontal
magnetic force isa task requiring great care and considerable
precautions. In the idiostatic instruments, in many cases,
springs are used to obtain a force with which to compare the
magnetic force in the neighbourhood of a conductor. The
| elasticity of these springs is not necessarily permanent, and is
liable to change by temperature as well as time. Accordingly
in these instruments also, the constant must from time to
time be determined. Pending the appearance of the new
| instruments promised by Sir W. Thomson, in which no
| springs nor permanent magnets are used, it is necessary to |
fall back upon galvanic cells to establish a standard of elec- |
|
|

* Communicated by the Physical Society: read June 27, 1885.

Daniell’s Cell as a Standard of Electromotive Force. 127

tromotive force, and the present paper is a brief collection
of facts concerning the use of Daniell’s cell for this purpose,
and its application in the measurement of currents of
various strengths. In all the experiments and methods cited
below the cell has been used only to create a difference of
potential, and not allowed to send any but a very small
current, the means employed for comparisons being a slight
modification of Clark’s potentiometer, as shown in fig. 1
(Plate V.). Ona stout mahogany board are fastened down
_two boxwood scales, each 5 feet jong, divided into inches and
tenths; these scales are fixed about + inch apart and parallel,
so as to form a groove between them. On these scales are
stretched two fine uniform German-silver wires A B, A’ B’,
about ‘013 inch in diameter, and having a resistance of about
1 ohm to the foot-run: one end of both wires is soldered to
a. thick copper junction-piece J, and the other ends respec-
tively to copper pieces connected to terminals M, M’. The
wire thus forms one length of about 10 feet stretched over
two scales. This forms the potentiometer wire; its length
is divided by the scale into 1200 parts, and each tenth of
an inch can be divided by the eye into 10, making a pos-
sible division of 12,000. To the terminals M, M’ are con-
nected five or six large-gravity Daniell cells, and the poles of
this battery are short-circuited by an Hdison 16-candle lamp
of 240 ohms cold resistance. This enables the battery to keep
a more constant difference of potential between the terminals
M, MW’.

A very constant E.M.F. can be obtained by using small
accumulator-cells in series: the poles being joined by a
carbon-filament lamp, and the leads to the terminals M, M’
taken off from the opposite sides of the lamp.

It is easy to calculate what number of cells are required to
“maintain a given difference of potential, say 2 volts, at the
extremities of the potentiometer-wire. For let n be the
number of cells, e the EH.M.F. of each, and ,¢ the internal re-
sistance, and let R be the resistance of the lamp and poten-
tiometer-wire combined, and v the required difference of
potential desired at the terminals M, M’; then

a neR
nr +R’
which determines n. If » is not known, it can be determined

by a second experiment, in which v is observed in the case of
a given nuinber of cells.

The current along this German-silver wire makes a fall in
potential at about the rate of 1 volt in 5 feet. On one side of

i
| 128 ~=Dr. J. A. Fleming on the Use of Daniell’s Cell

LV the board is fastened a broad copper strap bb, having six ter-
| minals fastened upon it. Between the last of these terminals
| and the end of the scale-wire A’ is inserted a reflecting-galva-
} nometer of 5000 ohms resistance, and an additional resistance
| of 50,000 ohms R. |

‘ To the other terminals s, s on this copper strip are con-
~ nected one pole (like to the pole of the battery B connected
to M’) of each of the cells c, c’ to be compared; and the
| other poles of these cells are connected with sliders 8, 8’
travelling over the wires. These sliders are blocks of wood
sliding in the groove between the scales, and overhanging
the wires. Onthem are German-silver spring-strips as shown
in the figure, and which, when pressed down, make contact
with the wire. The strips are backed with leather to avoid the
production of thermoelectric currents. By using two or more
of these sliders, it is easily seen that several cells may be
balanced at once on the potentiometer; and in particular two
cells c and c’ can be compared in respect of electromotive
force very accurately, even although the E.M.F’. of the main
battery is not quite constant. The introduction of the resist-
-ance R prevents any but the very smallest currents passing
in the cells when the place of balance is being found on the
potentiometer. The German-silver strips g, g on the sliders
make contact only when pressed down; so that in the normal
condition the cells c,c/ are insulated, With the galvanometer
in a sensitive condition it is very easy to read a difference of
Todo Of a volt on the wire, and zp45 can be read with great
accuracy.

Klectromotive forces are read off directly as lengths, since
the H.M.F. of the cells c, c' is directly as the distance of the
constant points of their respective sliders from the end A! of
the wire. Great care has to be taken in the first instance to
stretch the wire uniformly, and to calibrate it if it presents
any want of uniformity of resistance. Provided with this
convenient means of directly comparing electromotive forces,
attention was next directed to a modification of Daniell’s cell
suitable for the experiments required. In Dr, Alder Wright’s
extensive researches on the subject™, the form of cell employed
was that sometimes called Raoult’s form, in which two small
beakers contain the two solutions, and are connected by an
inverted Y-tube, the ends of which are tied over with thin

* “On the Determination of Chemical Affinity in Terms of Electro-
motive Force.—Part V. On the Relationships between the Electro-
motive Force of a Daniell Cell and the Gueasioal Affinities involved in its
Action.” Proceedings of the Physical Society of London, Vol. V. Part I.
p. 44 et seg. { Phil. Mag. vol. xiii, p. 265, ]

as a Standard of Hlectromotive Force. 12¢.

bladder. This form of cell is undoubtedly convenient for
some purposes, but not for others, as when the cell has to be
heated or cooled for comparison at different temperatures.
Moreover, the siphon becomes filled in a very short time with
a mixture of solutions by diffusion; and if the levels of the
liquids are not exactly equal, a siphoning action is added to
that of diffusion in carrying liquid over from one side to the
other. After some numerous trials of all existing forms which
have been proposed for Daniell’s cell, the following was de-
vised, which, though more bulky than others, has yet given
great satisfaction, and has the great recommendation of always
standing ready for use.

A large U-tube, about ? inch diameter and 8 inches long in
the limb, has four side tubes (see Pl. V. fig. 2). The two top
ones, A and B, lead to two reservoirs Z and C, and the bottom
ones Cand D are drainage-tubes. These side tubes are closed
by glass taps. The whole is mounted on a vertical board, with a
pair of test-tubes between the limbs. Suppose, now, a Daniell’s
cell is to be formed with solutions of zinc sulphate and copper
sulphate, and that the zinc-sulphate solution is the denser.
The left-hand reservoir 8 Z is filled with the zinc solution,
and the right-hand reservoir SC with the copper solution.
The electrodes are zinc and copper rods, Zn and Cu, passed
through vulcanized-rubber corks, P and Q, fitting air-tight
into the ends of the U-tube.

The operation of filling isas follows:—Open the tap A and
fill the whole U-tube with the denser zinc-sulphate solution ;
then insert the zinc rod and fit it tightly by the rubber cork P.
Now, on opening the tap C the level of the liquid will begin to
fall in the right-hand limb but be retained in the closed one.
As the level commences to sink in the right-hand limb, by
opening the tap B copper-sulphate solution can be allowed
to flow in gently to replace it; and this operation can be so
conducted that the level of demarcation of the two liquids
remains quite sharp, and gradually sinks to the level of the
tap C. When this is the case, all taps are closed and the
copper rod inserted in the right-hand limb.

Now it is impossible to stop diffusion from gradually
mixing the liquids at the surface of contact ; but whenever
the surface of contact ceases to be sharply defined, the mixed
liquid at the level of the tap C can be drawn off, and fresh
solutions supplied from the reservoirs above.

In this way it is possible to maintain the solution pure and
unmixed round the two electrodes with very little trouble; and
the electrodes, when not in use, can be kept in the idle cells or
test-tubes L and M, each in its own solution. In making

Phil. Mag. 8. 5. Vol. 20. No. 123. August 1885. KK

130) )=—-Dr. J. A. Fleming on the Use of Daniell’s Cell

experiments concerning temperature, the whole U-tube can |
be immersed in a vessel of water or ice up to nearly the top —
of the reservoirs, and the temperature in the solutions taken
by means of a thermometer passing through the rubber-cork. _
Kach of the electrodes can be removed for examination or — .
change without in the least disturbing the surface of contact
of the solutions. If experiments are being made in which
the sulphate-of-copper solution is the denser, the position of
the solutions is interchanged. The bottle R serves to collect —
the waste solutions.

The electrodes are made of rods of the purest zinc and
copper, about 4 inches long and } inch diameter. The zine ~
found most suitable is made from zine twice distilled and cast _
into rods; the copper prepared by electro-depositing on a
very fine copper wire, until a cylinder of the required thick-
ness is obtained. . i

With these appliances to hand, all the facts recorded by
previous experimenters have been carefully repeated and con-
firmed, and the influence of each variable upon the electro-
motive force examined. The results are collected as follows.
For the sake of brevity, a Daniell’s cell in which zine, sulphate
of zine, sulphate of copper, and copper are the elements will
be called a normal Daniell; and the statements here below
refer to the variations in the difference of potential between
the poles of the normal Daniell, when sending either no
current or only an infinitesimal one, caused by variations in
the physical state of the four elements. In each set of
experiments the greatest care was taken to keep all the
elements constant, except the one which was purposely being
varied in order to detect the influence of it on the whole
electromotive force.

1. Lhe Lifect of Variation in the Copper Surface and

Condition.

In these experiments the copper- and zinc-sulphate solu-
tions were sometimes of the same specific gravity, sometimes
different ; and in some cases the zinc electrode was a rod of
the purest cast zinc amalgamated with pure mercury, some-
times a rod of commercial zine wire cleaned but not amalga-
mated. Some scores of observations were made with identical
zinc plate and solutions, but various kinds of copper plates*;

* All the following values are given in real volts, taking Lord Rayleigh’s
value for the mean B.A. unit of resistance as equal to ‘9867 earth-quadrant
per second. i

The determinations have been made against a Clark cell, kindly com- /|
pared by Lord Rayleigh with his No, 1 cell, whose value is given in his |

_n 4
dl a
iy
Me
ire

as a Standard of Electromotive Force. 131

and all observations have confirmed the conclusions of Dr.

Alder Wright that the most uniform results are obtained from
copper plates freshly electrotyped over with a new clean
pinkish surface of virgin copper, when the plate is taken
straight out of the electrotyping bath and put into the cell .
without any delay or touching except a slight rinse with
distilled water. The result of comparing a freshly electrotyped
rod with a rod scraped and glass-papered, was to show that
the scraped copper rod gave a higher electromotive force by
about 4 parts in 629, or 6 parts in 1000. When both rods
were freshly electrotyped in the same bath, and inserted in

_ the same cell, there was absolutely no difference between
them. Cells were compared with solutions of equal density ©

1-2 at 15° C. and amalgamated pure cast zinc, but different
copper rods, as follows :—
Klectrotvped copper rod . . 1105 volt.
Scraped copperrod .. . Ill

7

The exact amount of difference varied slightly in other
experiments ; but with equally clean surfaces, the electrotyped
fresh surface has a lower electromotive force than the scraped.
The effect of amalgamating the copper rod is, like electro-
typing, to lower its value below that of the clean scraped, but
it is not so uniform. ‘The following are values obtained from

equal copper rods in the same cell, one having an electro-

plated surface and the other being slightly amalgamated with
pure mercury :—

Freshly electro-deposited Freshly amalgamated

copper rod. copper rod.
1°108 1°104
1°106 1°104
1°105 1-104
1°106 1-105

The mean of the values of the electro-deposited copper
exceeds by about 1 in 1000 the mean of the amalgamated
copper; but if the amalgamation is done thickly it causes

\ paper in the Phil. Trans. part ii. (1884) as 1-435 true volt at 15°C. My
| standard, called hereafter F,,, is greater than this by 5 parts in 4940, or

is 1°436 volt at 15° C.; and experiments have shown that its variation-
coefficient is nearly that found by Lord Rayleigh, viz. ‘082 per cent. per
degree diminution of .M.F.

Accordingly, the value of this standard Clark cell I*,, has been taken

\ as equal to

1:436(1—:00082 #°) volt at ¢° C.,

| and all values given in the text are in terms of this.

{2

132. Dr. J. A. Fleming on the Use of Daniell’s Cell —

more serious deviations, and though more convenient is not
quite so reliable as electroplating the rod with fresh copper,
but is preferable to employing simply a scraped or glass-
papered surface of rolled copper. A large series of determi-
nations were next made on the effect of oxidation of the copper
surface. If a copper rod, newly electroplated, is left in the
copper-sulphate solution, it gradually gets oxidized and
covered with brown patches of oxide, and if amalgamated,
the mercury sinks in and the surface gets brown and patchy.
When this is the case the electromotive force rises, and by
uncertain amounts, depending on the degree of oxidation. It
is very remarkable how small a trace of brownish oxidation on
the surface raises the electromotive force several parts in
1000, amounting in some cases to nearly 1 per cent. .The
following is one set out of many experiments on this point:—

Two solutions of zinc and copper sulphate were prepared of
the pure crystals. The specific gravity of the zine sulphate
was 1:4, and that of the copper sulphate was 1:1. A rod of
pure unamalgamated zine and one of electrotyped copper was
also taken. Keeping all other conditions constant, the rod of
copper was first cleanly and carefully electrotyped over with
a new fresh surface of copper, and exhibited no trace of oxi-
dation; and the H.M.F. of this cell was then taken against a
carefully tested standard Clark cell of known value; and the
rod was then suffered to oxidize on the surface by successive
exposure to the air, the E.M.F. of the cell being taken at
each stage.

E.M.F. of cell.
Copper, perfectly pure, unoxidized. . 1:072 volt.
is slightly oxidized, brown spots. 1:076 _,,
‘i more oxidized . . . oo ee
» covered with dark-brown oxide
film;., 0. wt. o) 5
if cleaned, replated with fresh )
pinkish electro-surface . . 1:072

9)

Many other experiments of this sort showed the same thing.
A Raoult cell was formed by taking two beakers, one con-
taining zinc sulphate of specific gravity 1:2, and the other
copper sulphate of the same specific gravity, and using amal-
gamated cast zinc and electro-deposited copper rods. The
beakers were connected, when required, by a piece of clean
cotton-wick dipped in water, and connecting both solutions by
dipping into each. The copper pole was first freshly electro- _ |
plated, and the following values obtained for the H.M.F. of —|
the cell as it gradually oxidized :-— |

as a Standard of Electromotive Force. 133

E.M.F. of cell.
Freshly electro-deposited surface of copper 1°105 volt.

Slightly oxidized, faint brown tinge. . . 1:106_,,
Mere oxidized . +. *. Oe CEES,
Guill more oxidized © 2 2 2... 1109 3
Re-electroplated surface... .. . 1:05 ,,

Hence these and many similar experiments all teach that
oxidization, even the slightest trace, of the copper raises the |
j.M.F.; and that, in order to get the real value proper to the
combination, the copper must be electrotyped over with a thin
pure film of copper, and exhibit no trace of brown spots of
oxide, and be used at once.

2. The Effect of Variation in the Zine Surface and Condition.

Numerous experiments have been made to investigate
whether there is any certain difference between zinc amal-
gamated or unamalgamated or cast or rolled. There is very

i hittle, if any, certain difference between perfectly pure cast

zinc unamalgamated and the same amalgamated with pure
mercury. There are greater differences if the zinc is not
pure, and variations are introduced if impure mercury is used
for amalgamation, all of them uncertain in amount.

The effect of oxidation of the zinc is to lower the electro-
motive force. If the bright surface of the zinc becomes
tarnished, it always shows a slightly lower E.M.F. The
smallest deposit of copper upon the zinc, due to diffusion of
the copper salt into the zinc, is indicated by a marked depres-
sion, amounting to 2 or 3 per cent. On the whole, the only
consistent values are obtained from chemically pure zine with
a bright fresh untarnished surface, whether amalgamated or
not ; and the best results are given with pure zinc amalgamated

~ with pure mercury. In Dr. Alder Wright’s memoir the

above conclusions are enforced by tabulated results of most

extensive experiments,

On the question of amalgamation of the zinc we may quote
a note on the subject by M. G. Lippmann in the Journal de
Physique. There is an opinion expressed by some authors
that amalgamated zinc has a higher E.M.F. than unamal-
gamated,evenif pure. Ifa plate of each of the two substances

be immersed in sulphate-of-zine solution, a couple is formed in

which amalgamated zinc forms the negative pole, inasmuch as

| it is the more readily acted upon by oxygen. Such, at least,

is the result obtained if ordinary commercial zine is used fee

? ordinary sulphate-of-zine solution. In a recent work, M. W.

Robb shows that if care be taken to use electro- deposited Zinc,

1384 =Dr. J. A. Fleming on the Use of Daniell’s Cell

and also zinc-sulphate solutions deprived of free acid, the ele- — |

ment zine, zinc sulphate, and amalgamated zinc will not manifest

any appreciable E.M.F. The superiority of amalgamated zinc
over ordinary zinc is easily explained ; but when pure zinc is
used, deprived of any oxide on the surface by slight rinsing in
dilute sulphuric acid before placing it in the sulphate of zine,
there is little or no certain difference, to the extent of 1 part
in 1000 in the case of the normal Daniell, between the pure
zinc amalgamated or unamalgamated.

3. The Influence of Density of the Solutions on the
Llectromotive Force.

The chief reason for the differences in the values assigned _

by various observers to the H.M.F. of the normal Daniell is
due to the great influence that the specific gravity of the two
exciting solutions exerts on the resultant E.M.F. of the com-
bination; and the general law of the effect is this—If a
Daniell cell be formed of amalgamated pure zinc, freshly-
electrotyped copper, and solutions of pure zinc sulphate and
copper sulphate of equal specific gravity, then, taking this
cell as a standard of reference, increasing the density of the
zinc sulphate lowers the electromotive force, and increasing
the density of the copper sulphate raises the electromotive
force, within, at any rate, the limits of density from 1°01 as an
inferior, up to saturation as a superior limit. Zine sulphate
saturated at 15° C. has a density of a little above 1:4, and
copper sulphate similarly saturated 1°2 ; and Dr. Alder Wright
shows by experiments that the increment and decrement of
H.M.F. for identical increment or decrement of density of
both solutions are so nearly equal that for equally dense
solutions, within limits, the H.M.F’. is independent of the
absolute density of both. This fact has been confirmed by
many other observers. Amongst other careful experiments
may be noticed those of Carhart *, whose experiments were
specially directed towards ascertaining the influence which
variation in the density of the zinc sulphate exerts on the
resultant H.M.F.

In using a Daniell cell as a laboratory standard it is neces-
sary, therefore, to prepare and standardize the solutions of
zine and copper sulphates with the same care as if they were
to be used for volumetric analysis. It is a good plan to pre-
pare two large stock bottles of solutions, and having carefully
determined the density to adjust them to known specific

* A paper read before the American Association for the Advancement

of Science, Sept. 5, 1884. Seo ‘Telegraphic Journal and Electrical |
Review,’ vol. xv. p. 250. | |

as a Standard of Electromotive Force. 135

gravity by the aid of the specific-gravity bottle or hydrometer.
For general use two standard solutions of each salt are speci-
ally useful. First, a solution of sulphate of copper very nearly
saturated at 15° C. and having a specific gravity of 1°200, and
a solution of sulphate of zinc of equal density. Secondly, a
‘solution of sulphate of copper of density 1:100 at 15°, and a
solution of sulphate of zinc of 1:4. These last values are
chosen because they were employed by Sir W. Thomson in
his standard gravity-cell ; and they can be used in either his
form of gravity-cell, or in the U-tube form above, or in Raoult’s
form of separate vessels and siphon *.

A very large number of comparisons have been made of the
H.M.F. of cells set up with these solutions and the E.M.F.
of Clark’s cells, whose value has been compared directly with
cells standardized in absolute measure.

If a Daniell cell is carefully made up, either in the Raoult
form or U-tube form, of solutions of pure zinc and copper
sulphate not sensibly interdiffused at the level of contact, and
with pure amalgamated cast zinc and freshly electrotyped
copper plate, which is evenly plated with a new uniform pink-
ish deposit of electro-copper free from all brown spots of oxide,
the E.M.F. of this cell, taken at once, is very close to 1:102
true volt, and the ratio of the H.M.F. of this cell to a cor-
rected Clark cell at 15° C. is very nearly 768 to unity f;
the Clark cell being taken as 1°435 at 15°C.

If, instead of taking the equidense solutions, we take zinc
sulphate of specific gravity 1:400 and copper sulphate of 1:100
at 15° C., and the same plates, the H.M.F. of the cell lies close
to 1:072 volt, and the ratio of the H.M.F. to that of the cor-
rected Clark is :747.

If, however, instead of taking the electromotive force at
once after the freshly electroplated copper pole is introduced
into the cell, the cell is allowed to stand an hour or so, both
the above values will be increased by about *003 volt, pro-

* If 28:25 parts by weight of pure crystallized sulphate of copper
(CuSO,, 50H.) are dissolved in 71°75 parts by weight of distilled water,
the resulting solution will have very nearly a specific gravity of 1:200 at
18° C.; and if 16°5 parts of the crystals are dissolved in 83°5 parts of
water, the solution will have a density of 1:100 at 15° C.

For the zinc-sulphate solutions take 55'5 parts by weight of crystallized
zinc sulphate (ZnSO,, 70H.) and dissolve in 44°5 parts of distilled water.
The resulting solution will have at 20°°5 C. a specific gravity of 1-400.

If 32 parts by weight of the crystals are dissolved in 68 of water at the
same temperature, the solution will have a density of 1200, These are
useful densities.

+ Lord Rayleigh assigns a value very close to ‘770 for this ratio, and |
Dr. Alder Wright ‘765. The figure in the text is derived from about 50
experiments of my own. |

136 ~=Dr. J. A. Fleming on the Use of Daniell’s Cell

vided always that no interdiffusion of solutions has taken
place, and that the zinc retains a perfectly bright untarnished
surface.

4. Influence of Temperature on the Electromotive Force of the
Normal Daniell Cell.

The researches of experimenters who have studied the Clark
cell have established the fact that its E.M.F. diminishes with
rise of temperature. The exact value of the coefficient of
variation seems to depend on the mode in which the cell is
made up ; and very careful examination has been made of the
Daniell cell as above described, to see if its real H.M.F. is
affected by temperature. Some writers have asserted that it
is ; but it is obvious that the variation of internal resistance
must be eliminated entirely by the use of a null method of
observation of the H.M.F. The condenser method is not a
sufficiently accurate one to apply, and the potentiometer
method is the most trustworthy.

Two U-tube cells were prepared with 14 sp. gr. zine sul-
phate and 1:1 sp. gr. copper sulphate ; and these cells were
respectively immersed, the one in a water-bath at 20° C. and
the other in melting ice at 0°. A pair of zine and copper
poles was likewise prepared ; and when the solutions in the
U-tubes had acquired the temperature of the baths, the H.M.F.
of each cell was taken with the pair of zinc and copper plates,
first in one and then in the other, with the following results :—

Plates in the warm cell at 20° C. Plates in the cold cell at 0° C.
E.M.F. of cell. E.M.F. of cell.
1081 1:082
1-079 1°082
1:079 1:082
1:078 1:083

The experiments showed a very small but decided fall of
H.M.F. as the cell is warmed, and at the rate of about 3 parts
in 1000 for 20° C. This is only about 1, of the variation of
a Clark cell for the same range ; and, practically, we may say
that the H.M.F’. of a Daniell cell is independent of tempera-
ture for such range as occurs in the natural temperature of
the air in our climate *. This quality of the normal Daniell

* The same result has been obtained by Mr. Preece in his experiments
on the effects of temperature on the electromotive force and resistance of
batteries, see ‘ Electrician,’ March 3, 1883, vol. x. page 367. He givesa
table showing the E.M.F. at various temperatures, obtained by the con-
denser method, and shows that there is no perceptible change in E.M.F.
of a normal Daniell between 0° and 17°, and a slight fall subsequently of
about 9 parts in 1000 between 17° and 28°, but a rise, however, after
reaching 63°,

as a Standard of Electromotive force. 137

is very important, and goes very far towards helping it to
sustain its position as a standard cell against its rival the
Clark cell.

The various coefficients of temperature-change which have
been found for Clark cells indicate that it is unsafe to rely _
upon any cell taken at random, whose history is unknown,
having a coefficient of variation accurately the same as that
of the cells whose coefficients have been actually determined.

It would be a great advantage if instrument-makers would —
construct these cells in the form suggested by Lord Rayleigh
—enclosed in a test-tube and having gutta-percha-covered
leading-in wires passing through a rubber stopper. The co-
efficients of variation would then be easily obtained for any
cell, just as those of standard resistance-coils are ascertained
and marked on each cell with the range of temperature over
which it is applicable.

As a check on the foregoing experiments on the Daniell, the
coefficient of variation of the standard Clark cell used was

obtained. Two cells were taken, exactly equal at 18° C., and

one of them was immersed for 48 hours in melting ice. The
difference on the potentiometer-wire was then obtained between
this cold cell and its fellow kept in water at 18°C. The
potentiometer-readings were as follows :—

il. lil.

1.
B, cell at 0°C. . . 961 971 974-4
A, cell at 18°C. . . 947 957 960
14 14 14:4

From the experiment 1 the reading for the warm cell

would have been 947+44=9494 at 15° C. And hence the

coefficient of variation is }* in 9494, or ‘0819 per cent. in

the neighbourhood of 15° C. Lord Rayleigh gives :082
as the value for his cells. It is obvious that Clark’s cells
cannot be used for standardizing galvanometers without a
careful determination of their coefficients of variation.

The normal Daniell cell has a certain advantage in that when
null methods for determining electromotive force are employed,
its value is independent of temperature throughout a consi-
derable range.

By the employment of the above two instruments— a poten-
tiometer as described and a standard normal Daniell carefully
constructed with all the precautions named—it is very easy to
make very accurate measurements of strengths of currents of
large magnitude. Ifa resistance is formed of such character

| that the current to be measured can be passed through it

138 =Dr.J. A. Fleming on the Use of Daniell’s Cell

without much affecting its temperature, and the potential

measured at the extremities of this resistance by means of a

comparison on the potentiometer between it and that of the
Daniell, this gives at once the current.

In order to construct a resistance whose value at the in-
stant of passing the current can be accurately measured, the
following device is adopted :—On a board are mounted a
series of copper blocks a, a, b, b, like the connection-pieces on a
resistance-box (see Pl. V. fig.3). | These can be connected by
plugs p, p. At opposite sides and ends are main terminals as
M’; and the blocks a, b are connected by brass or German-
silver wires wound in aspiral; similar coils of equal resistance
connect each pair of opposite blocks. The number of coils
must be an odd number. |

If the plugs 1,11, 3,9, 5, 7 are put in, it is obvious all the
coils R,, R, . . . R, are a series, and their resistance can be
measured: call it r. If, then, the other pegs are inserted, the
coils are now in multiple are or parallel, and if there are n

e e ° r e e
coils the resistance is at By this arrangement the resistance

in parallel can be inferred from the resistance in series ; and
although each coil has a very small resistance, yet by using a
large number, not only is a very easily measurable resistance
obtained when the coils are in series, but a resistance of large
current-carrying capacity when they are put in parallel.

The mode of using it is as follows :—The pegs are all put
in, and the current to be measured is sent through the coils in

parallel. By means of the potentiometer and standard-cell,

the difference of potential is found between the terminals M
and M’ when the current is passing. The alternate pegs are
then removed, the current stopped, and the resistance taken
of the coils in series, and the time noted which has elapsed
between the measurement of potential and of the resistance.
A few observations are taken of the resistance at intervals of
time, by which to construct a curve of cooling; and by projecting
back the curve, it is easy to ascertain very nearly the resistance
at the time when the potential was measured. If small currents
only are employed, it is not necessary to change the resistance;
the current can be sent through it in series, and its resistance
also measured in series. ‘The writer has had constructed a
large resistance of this kind at the Victoria Electric Lighting
Station, for the measurement of dynamo currents up to 500

amperes. In this case the resistance consists of 36 wires of |

brass, No. 14 B.W.G., arranged in parallel, and of equal

resistance. One of the wires can be disconnected and arranged all

instantly on a bridge to measure its resistance ; and in this

as a Standard of Electromotive Force. 139

way, after measuring the potential at the ends of the 36 wires
in parallel, with the current passing through them, one wire is
disconnected, and its resistance taken immediately; from which
observations it is possible to approximate very accurately to
the resistance of the whole wires in parallel. In order to
afford data for constructing such resistances, experiments were
made of passing various currents through coiled spirals of
naked wires of different sizes and materials.

A large number of wires were prepared, of copper, brass,
iron, German-silver, each 25 feet long, and of six sizes re-
spectively, Nos. 10, 12, 14, 16, 18, 20, B.W.G., the diameters
being given below. These wires were coiled into spirals
round wooden rods about one inch diameter, and the turns of
the wire well separated, so that each coil or spiral was about
18 inches long. Measured steady currents were sent through
these for some hours, and so adjusted that after the tempera-
ture had become steady the wires were all at a temperature
just about bearable by the hand, that is near 60° C. The

| currents respectively carried were as follows :—

Negi) 187) 44>): 3614 18. Ih 29
: B.W.G.
Size of wire. 4 | .134 inch | +109 | 083 | -065 | -049 | -035
{| diam.

Currents carried in Amperes.

German Silver ...| 18°75 todo 7,825 6 419° 1.3

BS sees as eves 30 19-79! 15 O Fas). Tae (2h
eee —='18 fP25 | 100.) S25 1) 525 | S75
Ss eee 49°5 38 26°25 |20°25 | 15 9

These currents passed through the above-described naked
spirals bring the respective wires to about a temperature of
60° C., when equilibrium is established; and for the purposes
of measuring currents not more than one third of the above
currents should be used with wires of the size appended.
Thus for 800 amperes, about 50 No. 10 B.W.G. wires will
carry it without much sensible elevation of temperature;
and by arranging 50 wires so that their resistance can be
quickly measured in series, a resistance can be made suitable
for measuring the potential at the ends of a known resistance.

By the use of this method it is easy to standardize any cur~-
rent or potential-galvanometer at any part of the scale, and
obtain the absolute or corrected value of the deflections ; and

———S

=. SSS SS SS —e eee ee
/

140 = Daniell’s Cell as a Standard of Electromotive Force.

these methods have been employed with success for some time

in standardizing the working galvanometers used in the lamp—

factory of the Hdison and Swan United Blectric Light
Company.

In conclusion may be given other results, collected from
various authors, respecting the electromotive force of the
Daniell-cell.

Wiedemann (Galvanismus, vol. i. cap. iv., Bestimmung der
Electromotorischen Kraft, p. 341) gives a description of
Raoult’s cell (see Raoult, Ann. de Chim. et Phys. [4] t. il.
p. 845, 1864), and states that Raoult finds that copper-foil has
higher electromotive force than electro-deposited copper by
about 51,; and attributes it to the oxides of copper contained
in it. He also confirms the invariability of H.M.F. with tem-
perature at least between 10° and 50°; and states that with
pure amalgamated zinc and electro-deposited copper, solution
of saturated sulphate of copper and solution of sulphate of
zinc, containing | part of crystals to 1 part of water=1°35 sp.
gr. at 20°, the electromotive force is 1:124 electromagnetic
units. This, corrected by the Cambridge value of the B.A.
unit, is 1°109 volt.

Sir W. Thomson (p. 245 of ‘Papers on Electricity and
Magnetism’) finds the electrostatic measurements of a
Daniell’s cell to be ‘00374. The nature of the solutions and
electrodes are not given. Taking the value of v as 3 x 10”, we
have the H.M.F. as 1:12 B.A. volt, which, reduced to true
volts, gives 1°105 volt.

Lord Rayleigh gives the ratio of a Daniell set up with
amalgamated pure zinc, electro-deposited copper, and solu-
tions of sulphate of zinc and copper, each of sp. gr. 1:1.
Five observations of the ratio of this cell to that of a known
Latimer-Clark cell at 16° C., taken as unity, gave :7702,
‘7710, °7705, °7698, °7702, mean *7703 ; and since the Clark
is 1485 volt, this gives the Daniell as 1:105 volt.

Latimer Clark (Journ. Soc. Tel. Eng., January 1873) gives
it as 1:11, which, reduced to true volts, is 1:095; and Dr. Alder
Wright (Proc. Phys. Soc. London, vol. vi. p. 292) gives the
value of a normal Daniell set up with solutions of the same
molecular strength, preferably of strength m M SO, 100 OHg,
where m is near 2—that is with copper suiphate nearly satu-
rated, and zine sulphate of equivalent molecular strength, and
pure amalgamated zinc plates and electro-deposited copper—
as 1:114 B.A. unit, or 1:099 true volt.

None of these are very far from the value assigned to the
standard cell described above, viz. 1'102 volt, with equidense
solutions and metals as described.

Bees

XVI. On Molecular Shadows in Incandescence Lamps. By
J. A. Fuemine, W.A., D.Sc. (Lond.), Fellow of St. John’s

College, Cambridge, Professor of Electrical Technology in
University College, London*.

ae presenting a short note on the above subject in 1883

many further opportunities have occurred for observing
the conditions under which molecular shadows are formed in
incandescence carbon-filament lamps, and of correcting one or

LINE OF

| | --"WwO DEPOSIT
COPPER \W | |

oéPosiT \\\\I\

4
<ZEN
ZZ
SI:

—=

two statements then made. It has been observed in an immense
number of cases, that not only do incandescence lamps become
coated on the interior of the glass with a deposit of carbon,
but that the envelope may have deposited upon it a metallic
film, derived from the leading-in wires to which the carbon
filament is clamped. In the Edison lamp the platinum wire
only just passes through the glass, and is connected with
copper wires broadened out into a clamp; the filament is
gripped in these clamps and then electro-plated over with
copper to effect a good junction. In the Swan lamp the
platinum wires are joined directly on to the carbon. It occa-
sionally happens that there is an unusual amount of resis-
tance at the clamps, or that by excessive electromotive force
more current is forced through the lamp and more heat gene-

* Communicated by the Physical Society: read June 27, 18865.

'

142 Dr. J. A. Fleming on Molecular Shadows

rated everywhere. In this case the tendency will always be —
to increase and go on increasing the resistance, and, therefore, _
the temperature at the point of highest resistance. Suppose
this occurs on the clamp or on the leading-in wire, then ex- _
perience shows that the metal is volatilized and deposited asa —
film on the glass. This metallic deposit is not uniform ; it is —
thickest nearest the base of the lamp and gradually thins away —
up to the crown, and ata certain height is thin enough to ~
transmit light. It is not very uncommon to obtain Edison ~
lamps with this copper deposit. ‘The colour of transparent —
copper is a fine sage green inclining to blue; compared with ~
the colour of gold leaf seen by transmitted light, it very —
closely resembles it. On one occasion the writer obtaineda —
lamp with a silvery metallic film deposited over it on the —
inside of the glass. From the outside it had a mirror-like —
lustre; on breaking the lamp this film was seen to be brownish
and not brushed off by the finger, but it could be removed by
scratching. It was not removed by holding in the oxidizing
flame of a Bunsen burner: therefore it was not carbon. It
was not removed by nitric acid; but on boiling a fragment of
the glass, covered with this metallic deposit, in nitro-hydro-
chloric acid the film disappeared. It was therefore probably
platinum. The film was transparent, permitting objects to be
seen through it, and transmitted brownish light.

Now, under certain circumstances, a line of no deposit is
formed on the surface in the plane of the filament, which is,
as it were, the shadow of one side of the loop This indicates
that the process of molecular scattering, which is going on at
some spot on one or other clamp, is not a mere evaporation or
volatilization of the metal, but a projection of molecules in
straight lines in every direction. The trajectory of the mole-
cules will be interfered with in some directions by the carbon
filament; and hence result lines and places of no deposit which
are molecular shadows of the loop. On every other part of
the glass the molecules will inpinge and adhere, forming a
metallic coating. From the facts that the free paths of the
molecules differ in length, and that the clamp is much nearer
to the neck of the lamp than to the crown, it follows that a
much larger proportion of the scattered molecules strike the
glass near the neck, and the thickness of the deposit is there-
fore a measure of the proportion of molecules which have a
free path, equal to the distance of that part of the envelope
from the scattering point. Curiously enough the line of no
deposit, or shadow of the loop, is not always seen in copper-
deposited lamps. This may be because the scattering is
going on from both clamps, and therefore the shadow on one

in Incandescence Lamps. 143

side is covered up by the shower from the clamp on that side.
It has been noticed in one or two cases that small tufts of
carbon are seen on one clamp, and that when a well-defined
shadow exists on that side, this seems to indicate that the
shower-of copper molecules has been partly stopped by the
opposite clamp, which has therefore acted like a target and
become encrusted with a proportion of the molecules shot
at it.

With respect, next, to carbon deposits. Every one knows

_ the appearance of a lamp after it has been burning for some

time or overburnt: it is clouded with smoke-like deposit. In
nearly all cases of copper deposit the molecular shadow exists,
but it is not nearly so often seen in the case of carbon deposits.
After many observations it was found that the molecular
shadow of the filament, or line of no deposit, could be formed
. by suddenly raising the filament to a very high temperature,
| as for instance by placing a 50-volt lamp for an instant on a
|| 100-volt circuit; whereas when the deposit of carbon takes
| place slowly, and as it were in the natural way, the lamp
| exhibits only a general smokiness but no line. Again, it has
been found that when a carbon filament is cut sharply through
at one point, caused by excessive temperature at one spot,
there is very frequently a sharply marked line of shadow

| of the loop on the side of the envelope farthest removed from

| the fracture. These facts seem to indicate that in normal use,
| when the lanip is not being pressed beyond the electromotive
| force at which it was intended to be used, there is a general
| evaporation of carbon going on from all parts of the loop, and
| these molecules, being projected with no abnormal velocity,
| probably collide with molecules of residual air a large number
_ of times before they reach the walls of the envelope, and thus
| get their trajectories very much changed in direction. In
) this case the result would be to cause an irregular deposit
| of molecules of carbon on the glass envelope. But if
we suppose a sudden or very excessive temperature to be
given to part or the whole of the filament, this may cause a
very violent projection of molecules of carbon from the fila-
ment; and these would pass outwards in straight lines, and a
far larger proportion would reach the envelope in the direction
_ In which they were first projected. This would then cause a
| deposit on all parts except those shielded by the loop; and in
| the case when such violent projection went on from all parts

_ of the loop, as when a lamp is overheated, it is easily seen that
'| parts of the envelope not lying exactly in the plane of the
} filament would receive twice as much deposit as those exactly
in the plane. In most cases of carbon deposits the lamps

| 144 Dr. J. D. Dana on the Origin o "a 4
| ‘ g e .
| 3 y
| which give the best shadows are those made with single loop,

but it has also been observed on Swan lamps with double —
: twist. By making lamps with clamps of various metals, it
4 might be possible to obtain metallic films of various kinds. | :
Interesting magneto-optic phenomena might perhaps present ~
i themselves in the case of transparent iron films, if they could —
\ be obtained. e

if !

; XVII. Origin of Coral Reefs and Islands.
By James D. Dana, LL.D.*

[Plate I. |

y ie Presidential Address of Dr. Archibald Geikie, Di-

rector-General of the Geological Survey of Great
/ Britain, before the Royal Physical Society of Edinburgh in
\ 1883 t, reviews the subject of the origin of coral-reefs and
i islands. In the course of the discussion, the author sustains —
: and enforces the objections which have been presented by —
others, and concludes that “ the existence of such reefs is no |
i more necessarily dependent on subsidence than on elevation.” —
‘ The existing state of doubt on the question has led the writer
‘ to reconsider the earlier and later facts, and in the following
| pages he gives his results t. That both sides may be fairly
before the reader, the views of Darwin and the evidences in
; favour of his theory are first considered, and afterwards the
arguments that have been urged against it. Part of the
objections are based on misunderstandings of the facts, and —
. hence a general presentation of the subject has been thought
’ necessary.

Part I.— The Darwinian Theory and its Evidences.

1. According to the Darwinian theory, islands with fringing ~
reefs have been often changed through a slow subsidence of ~
the region into islands with barrier reefs; and, as the last
i summit of the sinking land disappeared, the latter have be-
; come atolls, that is, barrier reefs enclosing simply a piece of
i the ocean (or a lagoon). Darwin added to this conclusion, a
second, in view of the wide distribution of atolls and their

* From an advance proof from Silliman’s ‘ American Journal,’ com-
municated by the Author, to whom we are also indebted for copies of the
plate and clichés of the woodcuts.

+ Proceedings Edin. Roy. Phil. Soc. viii. p. 1 (1883).

{t The writer’s account of his original observations is contained in his
Wilkes Expedition, Geological Report, 1849 (756 pp. 4to), pp. 29-154;
and, less completely, along with a review of facts from other regions, in ~
his ‘ Corals and Coral Islands,’ 898 pp. 8vo (1872, 1875).

Coral Reefs and Islands. 145

relations to other islands: that the subsidence indicated in-
volved the whole central Pacific, besides other large areas.
_ He also expressed the opinion that a Pacific continent had
_ disappeared through the subsidence. The proofs of the first
and the second conclusions are partly different and should
not be confounded. The third is no necessary part of the
general theory, was not adopted in my Report, and need not
be further considered. ,

2. Darwin did not hold that atolls were necessarily evi-
_ dence of a subsidence now in progress, but allowed that in
some regions they may have reached a state of rest, and may
perhaps have undergone an elevation since the cessation of
the subsidence; and also that subsidence and elevation may
have alternated.

3. Darwin found what he believed to be almost certain proof
_ of subsidence in the features of the large barrier-islands and
_ atolls. He perceived in the rocky islets that dot the great
interior lagoon-like waters of the Gambier group, Hogoleu,
' and other similar barrier-islands of the Pacific, and the
| general resemblance of such islands to atolls, strong evidence,
“leaving scarcely any doubt on the mind,” that the islets
| were the emerged points of sunken lands; that such barrier-
‘islands were no less lagoon islands than Keeling atoll (the
| atol] which he investigated); and, if evidence of subsidence;
the atoll was proof of further subsidence, that is, one that had
, continued to the disappearance of the sinking peaks.
| The evidence which had satisfied him was satisfactory to
me when I first learned of his views in Australia (in 1839),
after a cruise amongst the Paumotu atolls and the Tahitian
and Samoan reef-regions; and more decidedly so later, when
I had been among the Friendly, Feejee, and other Pacific
(islands.
| That the argument may be appreciated I here intro-
duce a map of the eastern half of the Feejee Archipelago *.
| Several of the great barrier-reefs in this map, 10 to 20 miles
in length, have but one or two peaks of the sunken land re-
_/maining; Nanuku has but one little point, near its south-
\eastern angle, a mile of peak within a barrier island 200
“square miles in area; Bacon’s Isles are the last two little
peaks of a still larger lagoon; and besides these and other
jexamples of disappearing lands, a dozen of the easternmost
»\islands are actual atolls—the last peak gone.
| 4. To this, the chief of Darwin’s arguments in his own view,

ae

_
en

»| * Reduced from the general map of the archipelago in the Atlas ot
/ the Wilkes Expedition.

| Phil. Mag. 8. 5. Vol. 20. No. 123. August 1885. L
F| |

{

146

" . Korotuna

rite),

Mas
Nuku Levu%<ne.>

Yalangalala

‘Velarara)

Naitamba &
if 3

WN
‘a

® ¢
Kanathia | i?
sp Re

Mango dv)
te

»
aie

@

Chichia sy)
Noten
SAN

Dr. J. D. Dana on the Origin of

¥ n
ue ‘

Pe

mee

- ree

2
> debe vind ated

(Gordon R.
fi "9 Freeman's R
f Hawkins R.
~. abeld’s Isles
oy
a Aa
Bacon’s Isles 9

N
‘ Cac henyatead

Aiva Yee
eu Oneata

©
“a

_
YW

oy

o. Keane sie
a Ss

Fulanga 4.”
aa)

~ ue’

EASTERN PART OF THE FEEJEE ARCHIPELAGO,

Coral Reefs and Islands. 147

another of iike importance is added in my Report—the ewvist-
ence of deep fiord-like indentations in the rocky coasts of islands,
both of those inside of barriers and those not bordered by reefs.

When making the ascent of Mount Aorai, one of the two
high summits of Tahiti (September 1839), where high narrow .
ridges, almost like a knife-edge along their tops, alterna-
ting with gorge-like valleys 1000 to 3000 feet deep, radiate
from the central peaks but die out in a broad even plain at
the shores, | was made to appreciate the consequence to such —
an eroded land of a partial submergence. At any level above |
500 feet, its erosion-made valleys would produce deep bays,
and above 1000 feet fiord-like bays, with the ridges spreading
out in the water like spider’s legs. Observing on the maps
of the Marquesas Islands precisely this condition, it was a
natural inference that the lands had undergone great subsi-
dence, and perhaps were still subsiding.

With this criterion of subsidence in view, the evidence
_ from the Gambier and Hogoleu Islands is doubled in force ;
) and that for the sinking of Raiatea of the Tahiti group, re-

_ presented in fig. 3 of the plate of maps in Darwin’s ‘ Coral-
Reefs,’ is as strong from each of the two enclosed islands as
it is from the great breadth of the reef-grounds; and the same
is true for Bolabola, another Tahitian island on the plate.

So it is also in the Feejees. The demonstration as to sub-
sidence in the large barrier-island called the Exploring Isles,
for example, is made complete by the form of the ridge of
land along one side of the great barrier, and the positions of
adjoining islets.

5. The general parallelism between the trends of coral-
islands and the courses of the groups of which they are part,
and the courses also of the groups of high islands not far
distant, were regarded by Darwin as confirming his view
_ that the coral-islands were once high islands with bordering

|. reefs.

6. Darwin uses also the argument that the large coral-
_ islands have the diversity of form found in the barrier-reefs
_of high islands; and also that they often have such groupings
as would come from the sinking of a large island of ridges
and peaks with encircling reefs. He describes the Maldives
_ as one example of the latter, and the two loops of Menchikoff
island in the Caroline Archipelago as another.

7. The depth of lagoons, and of the channels inside of large
| barrier-reefs, afforded him further evidence of subsidence, it

'| being in many cases two to three times greater than the

limiting depth (120 feet) of living reef-making corals.
L2

148 Dr. J. D. Dana on the Origin of -

8. The great depth of the ocean in near vicinity to the —

atolls is another source of evidence added.

9. He urged also, in supporting his views, the non-existence —
in the ocean now, and the extreme improbability of existence —

at any time, of submarine volcanoes or chains of mountains

having their numerous summits within a hundred feet of the —

surface. |

10. Darwin speaks of smallness of size in coral-islands as
a result of continued subsidence. In my Report I base an
argument for subsidence on smallness in the proportion of dry
land, and on smallness of size, when there is gradation to-
ward either condition, and the seas beyond are free of islands.
The facts bear on the general conclusion with regard toa
Central-Pacific area of subsidence as well as on the funda-
mental point in the theory.

If an atoll-reef is not undergoing subsidence, the coral and
shell material produced that is not lost by currents serves,
(1) to widen the reef; (2) to steepen, as a consequence of
the widening, the upper part of the submarine slopes; (3) to
accumulate, on the reef, material for beaches and dry land;
and (4) to fill the lagoon.

But if, while subsidence is in progress, the contributions
from corals and shells exceed not greatly or feebly the loss
by subsidence and current-waste, the atoll-reef, unable to
supply sufficient débris to raise the reef above tide-level by

making beaches and dry-land accumulations, would—(1) re-

main mostlv a bare tide-washed reef; (2) lose in diameter or
size, because the débris that is not used to keep the reef at
tide-level is carried over the narrow reef to the lagoon by the
waves whose throw on all sides is shoreward; (8) lose in
irregularity of outline, and thus approximate toward an an-
nular form; (4) lose the channels through the reef into the

lagoon by the growth of corals and by consolidating débris; |
and (5) become at last a small bank of reef-rock with a half —

obliterated ]agoon-basin.
The Pacific contains reefs of the three kinds :—(1) atolls
with much of the reef under trees and shrubbery; (2) others,

of large and small size, with the reefs mostly or wholly tide- |

washed; (3) others only two or three square miles in area,

without lagoons. Further, the kinds are generally grouped |
separately and gradationally. (1) The islands of the Pau- ©
motu and Gilbert Archipelagos have usually half or more of |
the reef dry and green; (2) the northern Carolines and the |

northern Marshall Islands, and the eastern Feejees, although
in part of large size, are mostly bare reefs; while (3) the
islands of the Phoenix Group, of the Equatorial Pacific east

Coral Reefs and Islands. 149

of the line of 180°, are, with one exception (Canton or Mary),
not over four miles long. The three more southern of the
Phoenix Islands (see Map, Plate I.), Gardner’s, Hull’s, and
Sydney, between 4° 25’ S. and 4° 35’ §., are two to four
miles long, and have lagoons; five, including Phcenix, Birnie’s,
and Kean’s, between 3° 10’ 8. and 3° 80’ 8., and Howland and
Baker’s, north of the equator, are a mile and a half and less
in length, and have depressions at centre, but no lagoons.
The depressions contain guano, and one of them, Kean’s, has
much gypsum mixed with the guano *; Kean’s and Phoenix
have a foot or two of water at high tide, the tide rising 6 feet.
Another of the number, Enderbury’s, is three miles long and
has a half-dried lagoon, which is very shallow and has no
growing corals t. To the north of these islands for fifteen
degrees of latitude the sea is an open one; and in the next ten
degrees, to the line of the Hawaian Chain, the only islets not
marked doubtful are “ Coral-Reef, A wash” and Johnston
Island.

A similar gradation in size takes place in the Hllice,
Ratack, and many other groups of the ocean. Smallness of
size, and dried lagoon-basins, with occasionally a deposit of
gypsum from evaporated sea-water, are just the result that
should have come, by the Darwinian theory, from subsidence;
and gradation in size from gradation in the amount of subsi-
| dence. ‘The positions of the Union, Gilbert, Ratack, and
_ Ralick groups with reference to the Phoenix group are shown
| on the Map, Plate I. All of the islands on the map are coral-
islands, and nearly all atolls; and the part of the encircling
reef marked by fine dots is under water at high tide.

Adopting this view of the origin of these smallest of coral-
made islands, I readily accepted Darwin’s second conclusion as
to a great central oceanic area of subsidence. The further
inference, also, was deduced, for reasons stated in my Report,
, that the greatest amount of subsidence took place along
| a belt stretching south-eastward from the southern half of

* J. D. Hague, Amer. Journ. Sci. (IIL.) xxxiv. p. 242. Mr. Hague, in
his valuable paper on the Guano Islands of the Central Pacific, mentions
the existence of a bed of xypsum two feet thick under the guano of Jarvis
Island, another small equatorial island, eleven degrees east of the Phoenix

group.
hy Baker's Island has a height of 22 feet, according to Mr. Hague,
showing, he says, some evidence of elevation; and Enderbury’s I
found to be 18 feet in height, from which I inferred someelevation. But
Howland’s, Birnie’s, McKean’s, Phoenix, Gardner’s, Hull’s, and Sydney
are not higher than ordinary atolls would be in a sea of 6-foot tide.

The facts with regard to the “ Reef’? on the map, in long. 175° W.
and lat. 2° 40'S., I have been unable to learn.

150 Dr. J. D. Dana on the Origin of

Japan and passing south of the Marquesas group toward —
Easter Island, and a line was drawn on a map among its ~
illustrations representing the course of “ the axial line of —

greatest depression ”’ *.

These deductions have been apparently sustained by the
soundings of the ‘ Tuscarora’ and ‘ Challenger’ in 1874,

1875, and 1876. The soundings of the ‘ Tuscarora ’ through —

the Phoenix group in 1875, on its route from the Sandwich
Islands to the Feejees (under the command of J. N. Miller,
U. 8. N., by the order of the U. 8. Hydrographic Bureau),
are shown on the map of the central Pacific herewith pub-
lished (Plate I.).

The soundings about these islands prove (1) that the
islands are situated within the deep 3000-—4000-fathom area
of the ocean; and appear to indicate also (2) that along lines
transverse to the trend of the islands (or to the direction of
trend in other groups to the west), mean submarine slopes
of 1:1°5 to 1:7 exist; while in the direction of the trend,

the slopes are much less. The slope of 1:1°5, or that of the —
angle 33° 41’, is nearly the maximum slope of the sides of —
Cotopaxi, Mt. Shasta, and several other volcanic summits of _

Western America.
The facts are these:—

Halfway between Sydney and Birnie’s Islands, 60 English ©
miles apart, a depth of 3000 fathoms (18,000 feet) was —
found. Off Hnderbury’s Island (40 iniles north-east of —
Birnie’s), (1) a depth of 2835 fathoms was obtained 20 miles to —
the south-west; (2) of 880 fathoms 2} miles to the south-west; —

(3) of 1991 fathoms 3 miles to the north-east; and (4) of

2370 fathoms, 11 miles to the north-east. The mean slopes —

to the south-west, calculated from the soundings 1 and 2, are

1:6 and 1:3; and to the north-east, from 3 and 4,1:1°5 and |

1:4; 14 miles south-east of Hull’s Island, at right angles

to the above direction, a depth of 935 fathoms was found, |

which gives the slope 1:13.

Further evidence as to the submarine slopes about equa-
torial coral-reef islands is afforded by soundings, made under _
the direction of the British Admiralty, near the very small |
Swain’s Island, at the south end of the Union group (see |
map); and others, by the ‘Tuscarora’ under Commander |
Miller, in 1876, near the Danger Islands, about five degrees |
east of Swain’s. Off Swain’s Island, two soundings, one |

* Report, pp. 99 and 452, and map between pages 8 and 9

This line 1s reproduced on a chart of the World, in my ‘ Manual of © |

Geology,’ where it is lettered A’ A’,

Coral Reefs and Islands. 151

south of it and the other east (the two directions at right
angles to one another and the latter not diverging far from
the trend of the other islands of the Union group), give the
slopes 1:7 and 1:18. Off Danger Island, as Commander
Miller’s Report states *, the depth of 660 fathoms was obtained
half a mile (nautical) off the south-west corner of the reef
near south-east island, and 985 fathoms one mile east of the
reef—corresponding to slopes 1:1 and 1:0°75. 1:1 is a
steeper slope than occurs even in small dry-made cinder-
cones ; and 1: 0°75 (53° 8’) is steeper still.

The above facts are sufficient to authorize the drawing of
the bathymetric lines for 1000, 2000, and 38000 fathoms
quite closely about the islands of the Phoenix group, and
to give the areas a northwest-southeast elongation, corre-
sponding with that of the neighbouring Pacific islands to the

west, as on the accompanying map, Plate Lt _

Tt follows from the above-mentioned facts that the deep-
water areas adjoining the Phcenix group, named provision-
|, ally by Petermann { the “ Hilgard depths” and the “ Miller
| depths,” are parts of one large area 1200 miles broad. The

' ™* T am indebted for the soundings about Danger Islands to Commander
J. R. Bartlett, Superintendent of the U.S. Hydrographical Bureau.

_ + The line on the map for 1000 fathoms is a simple dotted line ; that
for 2000 fathoms, - - - -; for 3000 fathoms, ~ — — -.

t Geogr. Mittheil. 1877, page 125 and plate 7. The deep areas along
the lines of soundings were named by Petermann on his very valuable
bathymetric map of the Pacific simply to facilitate reference.

The bathymetric lines about the islands on the accompanying Map
(Plate I.) have an unreasonable degree of regularity. But with no facts
to indicate the actual irregularities, none could be reasonably introduced.
The trends given them are the same as on Petermann’s map. The actual
steepness of slope is probably not exaggerated for either of the islands.
If similar slopes exist about the smaller islands in other parts of the
ocean, the final bathymetric map of the Pacific will have a very different
aspect from that presented by the maps hitherto published, and the
Central Pacific a much greater mean depth. About Wakes Island, a
small atoll in latitude 19° 11’, standing alone in the ocean six degrees
north of the Ralick Chain, the width of the area enclosed by the 2000-
fathom line, as drawn on Petermann’s bathymetric map, is nearly 100
nautical miles, while, in view of facts at the Phoenix group, the actual
width is probably not over 10 or 15 miles.

With but four lines of soundings for the part of the great Pacific
Ocean, within 35 degrees of the equator, the author of a bathymetric map
has to rely chiefly on his judgment or conjecture for the larger part of
the surface. There are many great problems in physical, geological, and
biological science that would be elucidated by the facts which a thorough
bathymetric survey of the ocean would afford; and the work is large
and important enough to call for aid from each of the great nations of the
world. Thus far, for the Pacific Ocean, the United States is first in the

amount of work done.

152 Dr. J. D. Dana on the Origin of

greatest depth obtained in the part of the area south-west of —
the group (400 miles broad) is 3305 fathoms, and in the part —
north-east 3448 fathoms. +
Again, the soundings of the ‘Tuscarora’ of 1875 here cited, _
taken in connection with those of the same vessel in 1874,
under Commander George F’. Belknap, along a line from the —
Sandwich Islands westward to Japan (mostly between the
parallels of 20° and 25°), suggest the further conclusion, that ;
the deep-sea area of the central equatorial Pacific, in which
the Phcenix Islands stand, extends north-westward toward :

Japan, and that it was crossed by the ‘ Tuscarora’ between
171° E. and 150° E., where were found depths from 3009
to 3273 fathoms (with some alternations of smaller depths
that isolated areas may account for). It is also probable
that the soundings of the ‘Challenger’ east of Japan be-
tween 153° HE. and 143° H., and just north-west of those E
of the ‘ Tuscarora,’ were within the same deep-sea area. If
this be so, a long deep-water area or trough extends from —
Japan south-eastward through the Central Pacific, conform-
ing well to the suggestion of the Darwinian theory; and ~
corresponding in direction approximately to the “axial line
of greatest depression” referred to above—the line AA on
the accompanying map.

As regards the rest of the Central Pacific between the
above defined 3000-4000-fathom belt and the Hawaian chain,
the depths sounded by the ‘Tuscarora’ are, with few exceptions,
over 2400 fathoms; two thirds of them are over 2750 fathoms;
and a fifth (out of the fifty-five in this area) over 3000 *.

* The same two lines of soundings by the ‘Tuscarora’ suggest the
existence of a second long deep-sea belt or trough in the Central Pacific
just south of the Hawaian chain, This supposed trough was crossed by
this vessel in 1875 between the parallels of 13° N. and 18° N. (“ Belknap
depths”), and in 1874 between the meridians of 1723° W. and 1773° W.
(the “ Ammen depths ”’); the greatest depth found on the former line is
3125 fathoms, and on the latter 3106 fathoms. Should the existence of
these two troughs be sustained, the region between them would be a
Central- Pacific plateau ; having in it, along the 1875 line of soundings,
depths of 2972 to 1825 fathoms, and along that of 1874, depths of 2836 to
1108 fathoms ; the shallower portion is near the middle of each line of
soundings, has a great descent (6000 to 9000 feet) on either side, sug-
gesting the idea of a central ridge. Over this plateau-area there are,
south of the Hawaian chain, two or three small coral-islands; and
further eastward, the Palmyra, Kingman, Washington, Fanning, and
Christmas reefs and islands, which, although coral structures, make the
idea of a central ridge in this part for 600 miles almost a manifest fact.
Further east, the Marquesas Islands are in the same range. The deep belt
lying on the south side of the plateau diminishes in depth to the east-
ward, the ‘Challenger’ soundings from the Sandwich Islands to Tahiti
finding no depth in the course of this belt greater than 2750 fathoms; but

a rt a

2 Ree ee eee ET. ers Sie

SS asst et tn aS
nrenrie

en a a

- ae a oe

Coral Reefs and Islands. 153

11. Since a fringing reef is, by the theory of Darwin, the
first stage in the origin of an atoll, it was naturally regarded
by him as, in general, evidence of little or no subsidence, and
even, at times, of elevation. But since (1) bold shores are
an occasion for narrow reefs and for their absence, (2) sub-
marine volcanic disturbances and eruptions about volcanic
lands would destroy living reefs or retard their progress
where begun, (3) islands of active volcanoes have small or
no reefs, and (4) abrupt subsidences of only 120 feet would
put reef-corals below a surviving depth and so lead to the
beginning of a new reef, I was led to regard the evidence
from a fringing reef for no, or little, change of level as of
very doubtful value. But the doubts, while making such
evidence generally useless, do not affect the value of the pre-
ceding arguments for subsidence. Darwin used the evidence
from fringing reefs only to mark off the limits of the area of
Central-Pacific subsidence to which his coral-island theory
had led him; and the same limits essentially are reached not-
withstanding the doubt. Instead of concluding that the
region along these limits was one of recent elevation or at
least of no subsidence, I was led to speak of it as one either
of no subsidence or of less subsidence than over the central
area referred to. The difference between us is small.

12. The true value of fringing-reefs as evidence in the
question of change of level should be appreciated in this dis-
cussion, as is apparent from the objections to Darwin’s theory
which have been urged; and I mention a few facts from the
Pacific islands in its elucidation.

On Darwin’s map, the Marquesas group is left uncoloured,
which means, doubtful as to subsidence or not. The Tahitian
group (Society Islands) is coloured blue; that is, it is included
within the area of coral-reef subsidence. The Navigator or
Samoan Islands are coloured red, or placed in the area of
elevation ; the Feejees blue ; the Sandwich Islands red. The
facts are these. |

The Marquesas Islands are an example of absence of reefs
to a large extent, with only small reefs where any. But the
shores are mostly too bold for reefs; and hence their small-
ness bears no testimony as to elevation. Along the bold

the belt on its north side may continue eastward of the ‘Challenger’
route. Many more lines of soundings are needed to substitute sure con-
clusions for the above suggestions.

The existence in the ocean of parallel belts of deeper and shallower
waters, such as are here inferred to exist, and such as are actually in-
dicated by the parallel lines of high islands and atolls, is in aecordance
with the facts over the continents.

ie

thy

154 Dr. J. D. Dana on the Origin of

shores there are deep indentations and fiord-like bays, separated
in some cases by narrow ridges, sometimes in spider-leg —
style; so that the proof of subsidence is positive, as explained
in § 4. y

Tahiti presents none of the Marquesan evidence of subsi-
dence. Its erosion-made valleys, as already explained, die
out at the broad shore-plain, and the island is comparatively
even in outline. I found over it, like Darwin before me, no
evidence of elevation beyond one or two feet at the most. It
has broad reefs; and the channel inside the barriers between
Papieti and Toanoa (2 miles) has a depth of 3 to 25 fathoms.
From the width of the reef, and the slope (6 to 8 degrees) of
the land, and of the lava-streams outcropping in the sides of
the valleys, supposing this slope to be continued under water,
I made the probable subsidence 250 or 300 feet. A slope of
6 degrees, and a width of reef of half a mile, gives 240 feet
for the depth of the reef at the outer edge.

The Samoan (or Navigator group) includes (beginning at
the east) Rose Island (an atoll), Manua, Tutuila, Upolu, and
Savail.

Manua has bold shores, a height of 2500 feet, and a narrow
reef where any*. Tutuila is of the Marquesan type in its bold
indented sides, and this suggests a probable subsidence. Pan-
gopango Bay, in which we anchored in 174 feet of water, has
a length of three miles. The coral-reefs are of the Sringing
kind where any occur. Upolu, a few miles west, has bold
shores and small or no reefs for fifteen miles of its north
shore, east of its middle ; but elsewhere there are broad reefs
(mostly 5000 to 8000 feet in width) and a very gentle
slope (three to six degrees) in the land above, which is about
the slope of its underlying lava-streams. The great width of
the reef seemed to be evidence of subsidence. But the ab-
sence on the north side of the island of a channel in the reef
deep enough for any craft larger than canoes made it essenti-
ally a great fringing reef. A calculation from the width and
land-slope gave about 260 feet for subsidence; but I add (on
page 332 of my Report) my doubt as to any subsidence. The
facts known are against any elevation.

* With regard to Manua, Mr. J. P. Couthouy, of the Wilkes Exploring
Expedition for two thirds of its cruise, in his paper on “Coral Forma-
tions” (p. 50 of Proc. Bost. Soc, Nat. Hist., Jan. 1842) reported the
occurrence of fragments of corals at a height of 80 feet “on a steep hill-
side rising half a mile inland from a low sandy plain.’’ I was not on
Manua. I found on Upolu fragments of coral-limestoue and shells in the
tufa of a tufa-cone at a height of 200 feet, which had evidently been
carried up by the erupting action of a slightly submerged vent. (‘ Report,’
p. 328.) The facts on Manua need further study,

Coral Reefs and Islands. 155

Savaii, the largest island of the group, is a gently sloping

volcanic mountain, much like its namesake, Hawaii, in its
features, with lavas looking as if not long out of the fire. It
has a broad reef for only 6 or 7 miles of its east shore; else-
where, on the east and north-east sides, the reefs are fringing
or wanting; and on the southern and western sides mostly
absent. No evidence of elevation, and nothing certain as to
subsidence, has been reported from the island.
- The large Feejee group bears abundant evidence of subsi-
dence in its very broad reef-grounds, barrier islands, and atolls.
Fringing-reefs, or barriers with very narrow channels, occur
about some of the islands of the group; but in view of the
facts that have been stated, these are useless as evidence
either way.

Thus the conclusions as to the changes of level about these
large Pacific groups south of the equator are not far from
Darwin’s, although fringing-reefs and the volcanic character
of the island are thrown out of consideration, and other con-
ditions exist of varied interpretation.

But cases of actual elevation occur in the Central Pacific
about several smaller islands, as proved by elevated coral-reefs.
These occur in the Austral and Hervey Islands south and
south-east of Tahiti, and in the Tonga or Friendly Islands.
In none of these, however, thus far reported is the elevation
over 300 feet: and the amount varies greatly in adjoining
islands of the same group, some affording proof of no elevation.
Hence only local changes of level, not a general elevation, can
be inferred.

To the north of the equator, at the Sandwich Islands some
elevated reefs occur; but the amount of elevation is small, and
is not general in the group. Moreover, the reefs are small,
where any occur; and the largest island of the chain, volcanic
Hawaii, the easternmost, is mostly without reefs, as well as
the larger of the westernmost islands, Kauai, which has partly
bold and indented shores.

13. To give a more adequate view of the changes of level,
or the evidences bearing on the subject, along the “ limits ”’ of
the central area of subsidence, I mention some additional facts.

Tahiti is the large eastern island of the Tahitian group.
To the westward, the islands (1) decrease in size; (2) increase
greatly in relative breadth of reef-grounds; (3) become deeply
indented in shores, as explained; and (4) include an atoll,
Tubuai, as one of the last two of the chain. While the reef
of Tahiti proves comparatively little subsidence at that end of
the group, and its reefs and channels are extensive enough to
make the proof good, the other islands indicate, on the Dar-

156 Dr. J. D. Dana on the Origin of

winian theory, that the subsidence increases much in amount —
westward. ‘The western end of the chain is about a degree ~
nearer the equator than the eastern. |
In the Samoan Islands, the largest island, Savaii, is the
westernmost; and from there the islands decrease in size east-
ward, and end in an atoll, Rose Island. The group is like Tahiti
in gradation as to increase of subsidence, but the direction is
the reverse; and this fact points apparently to a much deeper
area between them *. Moreover, although such broad barrier-
reefs as those of Raiatea and Bolabola do not occur in the
Samoan group, bold shores do in Tutuila and Manua, and
indicate the participation of these islands in the subsidence,
notwithstanding their contracted reefs. Further, the reef of
Upolu is broad enough to be proof of little change in the
region of that island; and there was little, probably, at Savaii,
the larger island west of it. The evidence of increased sub-
sidence to the eastward is strong, and narrowness of reef is no
objection to it.
At the Sandwich Islands the case is similar and yet dif-
ferent; similar in the fact that the largest island of the chain,
Hawaii, makes one of its extremities, the eastern, and a series
of coral-islands the other—the whole length being 2000 miles;
but different in that no great reef exists about the shores of
either of the eastern islands to prove that the subsidence there
was small or none. The elevated reefs are only a local phe-
nomenon, and do not prove the absence of subsidence during
the era preceding the elevation. :
But we have other evidence of importance, derived from
soundings about the group by the ‘ Challenger’ in 1875 and the
‘Tuscarora’ in 1874, 1875. These soundings show that the
deep-sea area of 8000 to 4000 fathoms comes up quite closely
to the eastern end of the chain. It was found within 300
miles of north-eastern Hawaii and 250 of south-western, and
within 80 miles of north-eastern Oahu; and a sounding but
125 fathoms less than 3000 was obtained by the ‘ Challenger ’
within 40 miles of eastern Hawaii (or half its diameter). To
the westward, along the north side of the chain, the deep-sea
area appears to be two or three times more distant, according
to the ‘ Challenger’ results; the condition on the south side is
uncertain. It would seem from the great depth near Hawaii,
that the region of this great island, although itis now actively
volcanic and has little growing coral about it, had undergone
more subsidence than the coral-reef end of the chain, and

* The distance between the remote extremities of these two groups is
nearly 2000 miles, and the interval between the nearer over 800 miles,

Coral Reefs and Islands. 157

that its height and steepness of submarine slopes are due to
the fact that its outflows of lava have kept ahead of the sub-
sidence, and also built up nearly 14,000 feet above the sea. |

This height is large, but the mean pitch of the sides of the
volcanic mountains of the island is between 5° and 7° 45’ ; and
hence it is only the height which successive outflows should
have produced over a vent at the sea-level; and it may be
that the accumulation above tide-level has been made since
the supposed subsidence ceased. The depth of 2875 fathoms
found by the ‘Challenger’ 40 miles east of Hawaii shows a
mean submarine slope to that point of 4° 30’, as if here also
was a slope made by flowing lava. But more soundings are
needed to prove that the slope is a gradual one.

14. The facts reviewed show the uncertainty of evidence as
to little or no subsidence, or as to recent elevation, from (1)
narrow reefs, or from (2) the volcanic character of islands, and
leave untouched the evidence of actual subsidence from the
features of barrier- and atoll-reefs and from deeply indented
coasts.

15. After the above considerations, it is clear that the
theory of subsidence meets well the facts as to the varying
extent of reef among reef-bordered high islands. According
to it, (1) steepness of submarine slope may characterize the
side of a barrier-reef (as well as of an atoll) fronting east or
west, north or south, as is true of high islands; but it is least
likely to occur in the direction of the trend of the island or
group, or that of current drift. (2) Fringing-reefs, or no
reefs, may characterize one side, that of bold bluffs, and wide
barriers the opposite. (3) The barrier-reef may be made on
the submarine slopes of the land, or on a broad plateau or
lowland area between ranges of elevations, one or both of
which have disappeared in the subsidence. (4) By continued
subsidence, the side having a fringing-reef or no reef, may,
later in the subsidence, be that of a very broad barrier-reef,
because of the form of the surface of the subsiding land ; and
vice versa. :

The third of these propositions is well illustrated by the
facts from the Maldives, as reported by Darwin. On account
of its importance I add an illustration from the Feejees.

The great reef-grounds along the north-west sides of the
two large Feejee islands, Vanua Lebu and Viti Lebu, do not
indicate a subsidence proportional to their width.

Kach of these islands is over ninety miles long, and to-
gether the trend is north-eastward*. The north-western

* A map of the Feejee group is contained in my ‘Corals and Coral
Islands,’ and of larger size in Wilkes’s ‘Narrative of the Expedition.’

158 Dr. J. D. Dana on the Origin of

reef-crounds are 10 to 25 miles wide; while on much of —
the south-east side of each island there is (according to the
Wilkes chart) only a fringing-reef. The true explanation of
the great width is found, not in the amount of the subsidence
alone, but largely in the existence there of a broad area of
submerged land at relatively small depths. This inference is
sustained by the fact that the outer barrier-reef, after being
simply a barrier-reef for 125 miles with but two rocky islets
in its course, becomes in the same line westward for seventy
miles, a range of high narrow reef-bordered islands (called the
Asaua Range), and then bends around southward through
other rocky islands to meet the west end of Viti Lebu. The
reef-grounds have thus a chain of islands as their boundary
for a length of 100 miles, and simply a barrier-reef with two
rocky islets for the rest of the line (125 miles).

The following figure illustrates in a general way the above
condition. It is a section across the reef-grounds, i k, and

:
;
£ 4
4

SD ID ER eT | ES at IO Mi iE

z
ae

the outer barrier-reef 7, with a fringing reef at /; and sup-
posing it to have a rocky island at 2, it represents a section
(further to the south-west) across the reef-grounds and the
outer range of islands. The reason for the existence of only
fringing-reefs for much of the south-east side has not been
particularly investigated.

16. Local elevations within the sinking area are not evi-
dence against the Darwinian theory of subsidence. Local
disturbances and faults, as both theory and the rocks of the
continents show, are almost necessary concomitants of a
slowly progressing change of level. Besides this, igneous
conditions beneath a region are a common source of local —
displacements. Such displacements are therefore to be looked
for in the tropical oceans, since the various high islands are
volcanic, and the coral-islands probably have a volcanic base-
ment ; and, moreover, the islands are not unfrequently shaken
by earthquakes. The causes of local displacement by either
method would not necessarily interfere with any secular
change of level in progress.

17. The shore-platform of an atoll, or the “ flat,” as called
by Darwin, situated just above low-tide level, consists usually

SLATE IPO ber te Bt Se

nem ies Di ase:

- ~ _— assays

_ apy tre at ee

—

Coral Reefs and Islands. 159

of the true reef-rock, or the rock made by under-water con-
solidation ; and its height is determined chiefly by the height
of wave-action, its general surface being produced by the
chiseling-effect of the in-flowing waters. When found above
its normal level, it is probable evidence of an elevation ; and
on this kind of evidence the conclusion rests in several of the
cases of supposed elevation which | mention in my Report.
The width of the platform and its evenness of surface vary
with the height of the tides. When the tides are 5 to 6 feet, _
the platform is narrow, more cut up by channels and less
even in surface. After an elevation, if but a foot or two in
amount, the surface of the platform hecomes restored finally
for a large part of its surface to its normal level, and gentle
slopes may connect the newer and older portions. But if
the rise of an atoll is 10 feet, great degradation takes place
along the lifted edge of the reef, which may end in reducing
the elevated coral-barrier to a wall with numerous channels
and broad spaces opening through to the lagoon, as observed.
by the writer (from ship-board) on the south side of Dean’s
Island*.

18. The differences in the kinds of coral-rocks should be
understood (as the recent discussions of Darwin’s theory
have shown) in order to appreciate the structural facts that
bear on changes of level. The beach-made rock is of above-
water consolidation (through calcareous deposition about the
grains as evaporation takes place), and is porous, often oolitic;
and if a conglomerate, it consists mostly of worn masses.
The rock made of drifted sands is similar. But the true

* Our cruise took us from the Paumotu atolls to Australia, and there,
the sandstone blufis making the capes of Port Jackson gave me my first
understanding of the atoll’s “shore-platform.” This bluff has its “‘ shore-
, 00 to 150 yards wide, bare at low tide; it was the lower

ayer of the sandstone, a regularly jointed rock, lying like a loosely laid

ee It seemed strange that it was able to keep its place in the
ace of the breakers. But the first waters of the in-coming tide swelled
quietly over it, and served to shield it from the plunging waters of the
latter part of the flow; the waves, therefore, found nothing to batter
against short of the base of the bluff.

A view of Deans Island from the south is given in Wilkes’s Narrative,
i, p. 842; it fails only in not giving a nearly even top line to the columns.
The view on p. 334 looks as if representing another example of similar
erosion, But, as the text implies, the group of masses of coral-rock was
made by the artist by bringing into a single view the blocks that had
been observed in an isolated way over the platforms of atolls. The size
and shapes of the blocks are exaggerated. But, although isolated, such
blocks are often so united to the coral-platform that they appear to be a
constituent part of it (my ‘Report,’ p. 61) and suggest the question
aa they may not be remnants of an overlying layer elsewhere
removed. .

160 Dr. J. D. Dana on the Origin of

coral-reef rock is of under-water consolidation, and is usually —
very compact, like an ordinary limestone ; and if a conglo- |
merate, it is commonly a breccia, and sometimes a very coarse
breccia. Some masses of it lying on the shore-platform of —
Paumatu atolls (thrown up by storm or earthquake-waves),
100 to 2000 cubic feet in contents, consisted of single pieces
of massive corals—Astrwas, Porites, &c.; and others were —
an agglomeration of fragments of corals. The fine-grained
or impalpable kind made from coral-mud may have few or
no fossils, and be a magnesian limestone. mi

Another variety of the coral-reef rock, made in lagoons and —
sheltered channels, has the corals in the position of growth ;
and when formed of branching corals, the spaces among the
branches are often but partly filled. It is a weak rock ; and
the islets thus made in lagoons and inner channels are some-
times overturned by the heaviest of waves ; and rising banks
(as the experience of the Wilkes Expedition proves) may be
crushed beneath the keel of a passing vessel. .

Owing to the different modes of origin of the beach-made
rock and the true coral-reef rock, the occurrence of the former
underneath the latter would be evidence of subsidence.

Deep borings in atolls with circular drills that would give
a 6-inch core would supply evidence as to the existence or
not of beach-made coral-rocks at levels below the surface.
They would also determine the depth to which true modern
coral-reef rock extends and the nature of the underlying beds,
whether calcareous, volcanic, or of any other kind. This is
hence a sure method for obtaining a final decision of the
coral-island question, and should be tried *.

* The Wilkes Expedition carried out apparatusfor boring. It was put
into inexperienced hands, as Commodore Wilkes states in his ‘ Narrative’
(iv. pp. 267, 268), and at a trial with it on Aratica (Carlshoff Island) in
the Paumotus, it became broken and useless at a depth of 21 feet.
Moreover, the granulated material brought up afforded no satisfactory
evidence as to the kind of coral-rock encountered. The statement in the
‘Narrative’ that “the low coral-islands, as far as they have been inves-
tigated, both by boring and sounding, have shown a foundation of sand,
or what becomes so on being broken up,’ has been quoted and made
more of than the facts warrant. The “soundings” reached only the
sands of the sea-bottom ; and the “ boring,” if it found sand at bottom,
proved only that the beach-made rock may exist at the 21-foot level, in
which case a small subsidence would be indicated.

Commodore Wilkes says on p. 269 of the same volume: —“The elevated
coral-islands which we have examined exhibit a formation of conglome-
rate composed of compact coral and dead shells, interspersed with various
kinds of corals, which have evidently been deposited after life has become
extinct. A particular instance of this was seen at the island of Metia,
and the same formation was also observed at Oahu.” As the corals of a

Coral Reefs and Islands. 161

19. Elevated coral-reefs afford an opportunity to search for
layers of beach-made rock underlying true reef-rock ; and
also, if over 120 feet in height, to ascertain directly the cha-
racter of the rocks below this level.

The elevated atoll, Metia, 75 miles north-east of Tahiti,
whose maximum height (according to the measurement
of officers of the Wilkes Expedition) is 250 feet, I have
described as consisting of the true coral-reef rock. My ex-
aminations were made on the west side, where it presents a
vertical front to the water. The white compact limestone
was, in some parts, almost destitute of fossils, or had only an
occasional mould of a shell or fragment of coral*; and in
others it was a fine or coarse coral-breccia. My notes written
out at the island include the statement that “ large masses of
corals make some lower layers.’’ This observation, though
not as complete as I now see that it should have been, favours
_ the conclusion that the thickness of the reef: rock is at least
twice as great as the depth to which reef-corals grow, in

| which case the elevated reef is proof of a subsidence of 120 feet

| or more.

The island is so near the route to Tahiti that the doubt
which remains could be readily removed.

20. The subsidence indicated, according to the Darwinian
theory, by atolls and barrier-reefs was actual, not apparent
subsidence attributable to change of water-level. The diffe-
|| rence in its amount between the Central-Pacific area of sub-
| sidence and its limits (§§ 10, 11, above), the gradation or
|| variation in amount of subsidence along chains of islands
_ (§§ 10, 12, 13), and the local character of elevations, like those

of Metia, Mangaia, and many others, are proofs on this point.

The preceding explanations have prepared the way for the
consideration of the arguments urged against the Darwinian
|| theory, to which I now pass.
tage [To be continued. |

conglomerate, whether consisting of rounded masses or angular, are
“deposited after life has become extinct,” no inference as to the parti-
cular kind of coral-rock intended can be drawn from the remark. From
|| my knowledge of the island I presume he meant the ordinary breccia
_ conglomerate of the reef-rock, which is one of the kinds of coral-rock of
the elevated island. Commodore Wilkes himself made no examination
of the rock or special study of coral-islands, as might be inferred from
his theoretical views on p. 270 of volume iv. His ‘ Narrative’ was toa
considerable extent made up from the journals of his various officers.

* It was this compact rock, white, flint-like in fracture, clinking under
1, the hammer, that was found on analysis by B. Silliman to contain 38-07
¥ per cent. of magnesium carbonate.

Piul. Mag. 8. 5. Vol. 20. No. 123. August 1885. M

serve three purposes. They might test the truth of certain

Bie aa,
XVIII. On the Specific Refraction and Dispersion of Light by
the Alums. By J. H. Guavstone, PA.D., £RS.* : .

N the Comptes Rendus of November 17, 1884, there appears —

a paper by M. Charles Soret, in which he gives the
refractive indices of a large number of crystallized alums. —
They were determined by the method of total reflection, —
and are measured for the lines of the solar spectrum a, B, —
C, D, E, b, F, and G. As at the same time M. Soret gives ©
the specific gravity of the specimens examined, it is easy to _
calculate out the specific refraction and dispersion of these —
crystals, and it occurred to me that the observations might .

H

physical laws; they might arbitrate between myself and
Kanonnikoff as to the refraction-equivalent of certain metals ;
and they would give values for indium and gallium, which
have not hitherto been obtained. Through the kindness of ©
M. Soret I have received his full papert, with some further |
information, and observations on three additional alums not —
yet published, viz. the Rubidium Gallium, Rubidium Indium, |
and Cesium Chromium alums. |

In order to bring M. Soret’s figures into comparison with —
my own, which are always based on the line A of the spec- |
trum, it was necessary to reduce his figures accordingly. This ©

|

was easily done, as the line A falls short of a by almost exactly

the same amount as that one falls short of B. The following ©
table gives the specific gravity of the selected crystals ; the ©
refractive indices (w) for A calculated to the fourth place of |
decimals, and also those for G ; the specific refraction for A, ©

1. e. the index —1 divided by the density, (a); the spe- :

cific dispersion A to G, 7. e. the difference between the specific —
refraction of these two lines, (Hom) ; and the refraction-

equivalent for A, 2. e. the specific refraction multiplied by the

atomic weight, (P 2 a

is R,, Al,, 480,, 24H,O; the aluminium being replaceable

by indium, gallium, chromium, and iron.

— —

al
i‘
i

) The general formula of the alums

«

* Communicated by the Physical Society: read June 27, 1885.

+ ‘Recherches sur la Refraction et la Dispersion des Aluns crystallisés,”
Archives des Sciences Physiques et Naturelles, vol. xii. p. 553, and vol. xiii.
p- 5, =|

| Specijic Refraction and Dispersion of Light by the Alums, 163

Specific PA—1 |pwG—pA lp pal
Alum. gravity. aN UG wana nets P or

|

Ammonium Aluminium..| 1°631 1:-4542 | 1:4692 27784. 70093 | 252-293
Sodium i. ...| 1667 | 1°43842 | 1:4480 °2604 0083 | 2388°52
Methylamine _,, | 1568 | 1.4496 | 1:-4636 | ‘2867 ‘0090 | 267-68
Potassium af i ale Wa 3 i 14516 | 1:4661 2603 0083 | 246°81
Rubidium af ae G52 <1 4513 |) 41-4662 2437 ‘0080 | 253°69
Cxsium , ape kc S6E 1-4536 | 1:-4682 | -2313 0074 | 262°31
Thallium aa 2°257 1-4914 | 1:5108 | ‘2177 (0086 | 278°34
|| Ammonium Chromium. 1°719 14781 | 1°4959 ‘2781 (0104 | 265°94
| Potassium wg .--| L817 | 1°4754 | 1-4931 2616 (0098 | 261-22
i] Rubidium .... 1946 | 14756 | 14932 | :2444 ‘0090 | 266:69
Cesium s, ...| 2043 | 1:4753 | 1:4928 | +2326 | -0086 | 275-50
Thallium 4 ...| 2386 | 1°5158 | 1:5381 ‘2161 ‘0092 | 286-90
| Ammonium Iron ......... 1-713 | 1:-4783 | 1:4998 | -2792 0126. | 269-15
EEE gv cin neaneee 1806 | 1:4757 | 1:4960 | :2634 ‘0112 | 265-03
) | CO 1916 | 1:4763 | 1:4970 | -2486 (0108 | 273°21
| Cesium gh ges eR 2-061 1-4772 | 1:4984 9315 70103 | 275-95
| Thallium a eee 2-58); | 1°5'Dd | 15411 ‘2161 70108 | 288-85
Rubidium Indium......... 2065 | 1:°4586 | 1°4740 | -2221 ‘0074 | 269-63
)| Rubidium Gallium ...... 1962 | 14606 | 1:4758 "2348 0077 | 264-43

eee ee

Messrs. Topsoe and Christiansen* have aiso published the
refractive indices of three alums for the lines C, D, and F
and as they have given the specific gravity, their observations
are available for the same purpose. ‘The index for A has been
calculated from these data.

Alum. Pa: Fc: ome d
Potassium Iron Sulphate ...... 14751 | 1:4783 | ‘2597 | 261:37
Ammonium Iron Sulphate ...... 14789. | 14821 | -2786 | 268°56

Potassium Aluminium Seleniate| 1°4748 | 1:4773 | :2409 272°98

How far do these figures support the former conclusion,
that a salt has the same specific refraction whether in a solid
form or in solution ?

_ For this purpose determinations were made of the refraction
for A in aqueous solutions of the first two alums in M. Soret’s
list. Deducting the amount due to the solvent, the following
values were arrived at :—

Dissolved. Crystallized.

Ammonium Aluminium alum . . ‘2780 ‘2478

Sodium Aluminiumalum . . . ‘2613 2604

* “Kyystallografisk-optisk Undersoegelser,” Det K. Danske Viden-
skabernes Selskabs Skrifter, 1873, p. 622.
M2

164 = Dr. J. H. Gladstone on the Specific Refraction

These numbers are as close as two different specimens of the |
same salt, even if they were in the same condition, are likely
to be. It did not seem worth while to obtain additional proof _
of a law which has already been established by many cases ;
not only by my previous observations, but also- by those of —
Topsoe, Bedson, and Kanonnikoff. |

Do these figures confirm the law that the refraction-
equivalent of a compound body is the sum of the refraction-
equivalents of its components ?

The alums may be regarded as a compound of the sulphates —
of two metals of different kinds with 24 molecules of water.
Now water in the uncombined state has the refraction-equi- —

valent of 5°926 ; and in regard to the sulphates we have the —
following data :— :

Pract sulphate, (NH,),80,, 39°3. The mean of five
determinations made by myself gives 39°35*: Kanonnikoff
gives 39°27.

Sodium sulphate, Na,SO,, 26°6. Kanonnikoff gives 26°02.

Methylamine sulphate, (NCH ,).SO,, 54°55 awe. ammonium
sulphate with the addition of 2C Hg, or 15:2.

Potassium sulphate, K,S8O,, 32:4. The mean of five deter-
minations gave me 32°6: Kanonnikoff gives 32°22. It may
be estimated at 32°3 from Topsoe and Christiansen’s determi-

nations of the crystallized salt.

Rubidium sulphate, Rb,SO,, 41°3. This is Kanonnikoff’s
determination, which I prefer to my own made in 1869.

Cesium sulphate, Os,SO,, 55°3. Also Kanonnikoff’s.

Being not content with my old observations on the sul-
phates of the trivalent metals, I have recently determined
them afresh, with the following results :—

Aluminium sulphate, Al,(SO,)3, 70:5.

Chromium sulphate, Cro(SO,)3, 82°5 ; the result of old and
new observations.

Ferric sulphate, Fe,(SO,)3, 89°1 ; the mean of old and new
observations.

It is evident that on adding together

Ammonium sulphate . . 393
Aluminium sulphate . . 70°5
24. VV OLOL. <> os. enh te

We obtain th wg Ray

In this way the second column of the following table has
been calculated :-—

* Four of these are given in my paper in the Phil. Trans. 1869, p. 9.
+ Journal of the Russian Physico-Chemical Society, 1884, p. 119,

and Dispersion of Light by the Alums. 165
|

Refraction-equivalent,

d
Substance. Observe hie reduced

Calculated.
Topsoe and
Soret Christiansen.
Ammonium Aluminium alum...... 252-0 252-2
Sodium $3 Mi Hs 239°3 238°5
Methylamine __,, ie peered 267°2 267°7
Potassium 9 an a 245°1 246°8
| Rubidium Fc) Mee Ls 254-0 253°7
| | Cesium A TERI AW ete: 268:0 262°3
Ammonium Chromium alum...... 264°0 265°9
Potassium & Ne SY 257°1 261°2
Rubidium cd Spl Repl 266°0 266°7
Czesium é ee 4140 280°0 275°5
Ammonium Iron alum............... 2706 269°1 268°6
Potassium _s,, I i eka 263°7 265°0 261°4
ye cece eens se. 272°6 273°2
Ceresium 4] Rs Nak? bbs Saad 286°6 276°0

The agreement between the results calculated and found is
as near as might be expected, except in the case of cesium.
It confirms the general law, as the variations only in one
instance amount to as much as 1 per cent., and are some-
times plus and sometimes minus. There is little doubt that
the czesium, in one set of observations or the other, was im-
pure : an old determination of mine, from the chloride, would
give figures lower than Soret’s.

Do these data afford us the means of determining the
refraction-equivalents of the elements with more exactness
than heretofore? It is evident that in the series of aluminium
alums the metal having the smallest refraction-equivalent is
sodium, and that the rest follow in the order—potassium,
ammonium, rubidium, methylamine, cesium, and thallium ;
the same order is preserved in the chromium and iron alums
as far as they extend. This agrees with the order previously
determined both by myself and Kanonnikoff ; but when we
look more closely into the matter it is evident that the figures
are not very exact. Thus, in the aluminium series the differ-
ence between 2NH, and 2K is 5°42; but in the chromium
series it is only 4°72, and in the iron series 4°12, or 7°19 ac-
cording to Topsoe. ‘The values of potassium deduced from
this would vary considerably : assuming the value of NH, to
be 11:1, it might be either 3°4, 8°7, 9°0, or 7:5; which are
wider differences than between my old estimation, 8°1, and
Kanonnikoff’s recent independent determination, 7°75. This
will not be wondered at when it is remembered that all expe-

166 Dr. J. H. Gladstone on the Specific Refraction

_ rimental errors are accumulated on these residual numbers. _
The optical determinations of Soret are so uniform and exact —
as to inspire the greatest confidence ; but he is not satisfied —

;

it
PS ee

|
with the specific gravities. Errors may also arise from im- ~
purity of the salt, or want of homogeneity in the crystal. An —
aqueous solution is probably amore uniform substance thana __
| hydrated crystal, and better fitted for the purpose of deter- —
q mining optical equivalents. |
On this account any determination of the refraction of

indium and gallium made from these alums must be open to
considerable question: the rubidium compounds of these
metals, however, are believed by Soret to be fairly pure salts,
and we can compare them with three other alums of rubidium.
It would appear from the first table that indium has the —
value of aluminium + 8-0, or of chromium +1°5, or of iron
—1°8; and, similarly, that gallium bas the value of aluminium
+ 5:4, or of chromium —1'1, or of iron —4°4. Now alumi-
nium sulphate has already been estimated at 70°5, chromium
sulphate at 82°5, and ferric sulphate at 89-1, while the value
| of SO, is held both by Kanonnikoff and myself to be 17:0. We
may therefore deduce the following refraction-equivalents :-—
Atmimuims 0 Ue ae
Chromium! "2°37." alee
bes aan Prony eet ee eee

From Aluminium | From Chromium
Salt. Salt. From Iron Salt.

fudiwm +) iA a 17-2 17:2

Gallium...... 15°1 14°6 146

From the mean of the above the following constants may be
derived :—

Atomic Specific Refraction-
weight. refraction. | equivalent.

ae os a hese a ie ———_ _

rr

|
|
Rataban 0A. ek 1136 1532 17°4
ART és etd; cee, 69°8 2120 148

These numbers can only be looked upon as approximate.

a ee ae

eee ee

Turning to the matter of dispersion—It has already been —
shown that the refraction-equivalent of an alum is the sum of _|
the refraction-equivalents of its constituents for the line A, —
If this law holds good equally for the more-refrangible part

ial ak

Se

and Dispersion of Light by the Alums. 167

of the spectrum, it follows that the dispersion-equivalent
(P el Bay ee _ , or, which is the same thing, pee)
of an alum is the sum of the dispersion-equivalents of its
constituents. The data by which this can be tested are not
so numerous or so trustworthy as in the former case, but the

following may be accepted.
The dispersion-equivalent for water . . . . 0212

Ne a aluminium sulphate 2°40

os pe ammonium sulphate 1°33

3 a f sodium sulphate . 0°83
from which may be deduced :—

_ Dispersion-equivalent of Calculated. Observed.
Ammonium alum............ 8°82 8-40
- TE i 8°32 765

Though these figures are tolerably accordant, it will be seen
that those in the first column are decidedly higher than those
deduced from Soret’s measurements. The differences are about
5 and 8 per cent. respectively ; but there are known sources
of error in experiment which may affect the first place of
decimals.

The dispersion-equivalents of the different alums may be
thus tabulated :—

Dispersion-equivalents of the Alums.

ba ik: Chromium.| Iron. | Indium. | Gallium.

Ammonium salt ...| 840 9-91 12°11

Sodium _sC_—_, 7:65

Methylamine ,, 8:47

Potassium ,, 7:92 9°73 11°31

Rubidium __,, 8:36 9-87 11:86 | 898 | 873
Cesium a 8:44 10°15 12-28

Thallium se» 10°98 12°13 14°45

It is evident at once :—

First. That the differences due to the replacement of one
metal by another are very considerable ; more considerable in
proportion than the differences in the case of the refraction-
equivalents.

Secondly. That the different compounds of the alkalis pre-

168 Mr. C. Tomlinson on the Bleaching of

serve the same order, and nearly the same proportion, in the —
aluminium, chromium, and iron series: the order is thallium —
far the highest, methylamine, czesium, ammonium, rubidium, ~
potassium, and sodium lowest. This is the same order which —
may be deduced from old observations on sulphates, nitrates, —
chlorides, and acetates. ,

Thirdly. That the order of the other metals is iron far the
highest, chromium, indium, gallium, and aluminium lowest.
This is also in accordance with observations on the simple
sulphates of those previously examined.

Our knowledge on this part of the subject is not yet suffi-
ciently advanced to determine the dispersion-equivalents of the
separate elements.

XIX. On the Bleaching of Iodide of Starch by means of Heat.
By Cuarwes Tomuinson, /.£.8.*

oe bleaching of iodide of starch by means of heat forms
a pretty experiment. An aqueous solution of iodine and
one of starch may be mingled together in a test-tube, when
the well-known densely blue colour is produced. If the tube
be held over the flame of a spirit-lamp, the blue gradually
becomes paler and paler, and disappears long before the liquid
has reached the boiling-point. Ifthe hot tube be now plunged
into cold water, the blue colour immediately reappears,
starting up from the bottom of the tube, where the reduced
temperature is first felt, and quickly spreading through the
liquid up to the surface. The colour returns, but of course
more slowly, if the tube be left to cool in the air. |

Thénard, in noticing this experiment upwards of half a
century ago, remarked that it is, “‘sans contredit, la plus
remarquable de toutes les propriétés de liodure ”’ f.

It is surprising what a large number of papers have been
written on this apparently simple experiment. Gmelin (Hand-
book, Cavendish Society’s Translation, xv. p. 99) has collected
most of them down to 1862 ; and they exhibit various contra-
dictions. For example, some writers maintain that the iodine
and the starch form a definite chemical compound, others |
describe it asa mechanical mixture ; some maintain that iodic |
and hydriodic acids are generated on heating the compound,
others deny that this is the case ; some say that the blue liquid
may be heated many times in a sealed tube and yet recover —
its colour on cooling; this also is denied. Such statements |
as these led Professor Miller, in the ‘Text-book on Inorganic ||
Chemistry’ that he wrote shortly before his death, to remark. _ |

* Communicated by the Author. 4
_ + My copy of Thénard’s Traité de Chinue is a Brussels reprint, dated ‘|
1836 (vol. i. p. 157).

Iodide of Starch by means of Feat. 169

(1871, p. 136) that “ the cause of this change of colour is not
known.”

Some writers, as quoted by Gmelin, maintain that if the
iodine which evaporates in boiling be expelled by blowing air
into the vessel, the iodide of starch remains colourless after
cooling, but if it can reabsorb the iodine-vapours on cooling
the colour is restored. A sufficient answer to this is that the
colour begins to reappear at the bottom of the tube long before
the vapours at the upper part can possibly be reabsorbed. The
vapours may even be blown away and yet the colour reappear.

The usual mode of accounting for the discoloration is that

~ given in Watts’s ‘ Dictionary of Chemistry,’ under Srarcu,

v. 1868, p. 410:— The liquid may be decolorized by ebulli-
tion, whereby the iodine is volatilized ; if, however, the boiling
be not continued for a sufficient time to volatilize the whole of
the iodine, the blue colour reappears as the liquid cools.”

This statement is but partially true; for the colour disappears
long before the boiling-point is reached, and it does not account
for the fact that when the liquid is apparently permanently
bleached and allowed to get cold, the addition of a few drops
of solution of chlorine will restore the blue colour.

In examining this subject experimentally, the difficulty
seemed to lie in the multitude of explanations rather than in
any inberent difficulty belonging to it. My first care was to
determine the temperature at which the bleaching takes place.
For this purpose, fifteen grains of iodine were treated with
half-a-pint of distilled water, and five grains of each of four
varieties of starch were rubbed up and then boiled with about
an ounce of distilled water.

1. One ounce of solution of starch crisp from maize was
mixed with one ounce of the iodine solution ina small globular
flask and heated over the flame of a spirit-lamp. The colour
entirely disappeared at from 150° to 160° Fahr. (65° to 71° C.),
Left to cool in the air of the room the colour began to return
at 120° F. (49° C.), and the full colour was restored at 70° F.
(21°C.). The flask was heated a second and a third time, with
nearly similar results.

2. A solution with rice-starch became paler at 105°, much
paler at 115°, pale at 120°, and the colour entirely disappeared
at from 135° to 140°. The flask was plunged into cold water,
when the solution became reddish at 135°, purple at 100°,
violet at 80°, and blue-violet at 70°.

3. With sago-starch the colour became lighter at 110°, pale
blue at 120°, pale at 140°, and was bleached at 150°. In cold
water the colour returned at 130°, at 85° it was full blue, and
at 80° quite opaque, like ink.

4. The mixture with potato-starch was of a blue-black

170 Mr. C. Tomlinson on the Bleaching of

colour. At 105° it became pale, at 125° of a light red, and at _

140° entirely bleached. In cold water it became reddish a —
little below 140°; at 120° a purplish tinge came over the red;
at 90° the colour was purplish, at 80° violet, at 78° reddish
violet by transmitted and blue by reflected light.

A fifth variety of starch, namely that used for domestic
purposes in my house, was taken ; but this was soon recognized
as potato-starch by the method of Gobley (Journ. de Pharm.
for April 1844), in which various specimens of starch, in
watch-glasses, are arranged ona plate around a central watch-
glass containing iodine, the whole being covered with a bell-
glass. I found that some of the specimens, namely such as
were slightly moist, became coloured by the iodine vapour in
the course of a few minutes; others in an hour or so, according
as they absorbed moisture, there being no action with dry sam-
ples. Ifknown samples are first acted on, the colours assumed
by them may serve to determine other unknown specimens.

The four varieties of starch above referred to were made up
into thin and thick solutions, the thin containing five and the
thick ten grains of starch in about half-an-ounce of water.
Equal quantities of the starch solutions were severally mixed
with equal quantities of the iodine solution, thereby producing
very dark blue or blue-black compounds. Hach variety was
heated in a test-tube over a spirit-lamp flame and boiled during
two minutes, and then cooled by plunging the tube into cold
water. The colour was not in any case reproduced during the
cooling ; but on the addition of a few drops of an aqueous
solution of chlorine the colour was restored in each case, but
with very different degrees of intensity ; for while in the case
of potato- and rice-starch the blue was almost as intense as
before the boiling, it was faint in the case of maize or absent
in that of sago, or exhibited but a mere trace in several trials.
A specimen of arrowroot also behaved like the sago. Hence
it seems that different kinds of starch act on iodine with
different degrees of intensity, in character not like a chemical
compound so much as, according to Liebig and others, a pre-
cipitation of the finely divided iodine on the surface of the
granules of the starch.

M. Personne (Comptes Rendus for 1872, p. 617, in answer
to M. Duchaux, p. 533 of the same volume) refers to a state-
ment of his in the Comptes Rendus for 1861, in which he
regards iodide of starch, not as a chemical compound, but pro-
duced by the fixation of iodine on starch in the same manner
as a colouring-matter is fixed ona tissue. The blue compound,
he says, must be regarded as a dye, not a true lake™*.

* Puchot (Comptes Rendus, 1883, p. 225) has noticed that albumen

poured on iodide of starch suspended in water causes the colour to
disappear. Whey (petzt-dait) has a similar effect. Here, again, I found

Lodide of Starch by means of Heat. 171

It is evident from the preceding details that, during the
heating of the tube containing the so-called iodide of starch,
the solvent power of the liquid for iodine increases with the
temperature, and if the tube be watched at the moment when
the last trace of colour disappears, the tint of iodine in solution
becomes apparent ; but this aspect soon disappears, not only
on account of the small quantity of iodine present and the
rapidly increasing solvent power of the liquid as the tem-
perature rises, but also from the loss of iodine in the steam
which escapes from the mouth of the tube. ‘his view of the
case is supported by the fact that if boiling solutions of starch
and iodine be mingled no colour is produced ; for at this high
temperature the starch and the iodine seem to be dissociated
and incapable of mutual action. This appears to be what is
meant by Briicke (as stated in the Chemical Society’s Ab-
stracts for 1884, p. 576), who makes the bleaching to depend
on the affinity of warm water for iodine being greater than
that of cold water. |

Thus far the reasoning seems to be justified by the facts ;
but the restoration of the colour on the addition of chlorine,
apparently after all colour had been destroyed, remains to be
accounted for. In such case the small quantity of iodine that
remains after the boiling seems to be disguised under the form
of hydriodic acid, formed partly, as Thénard suggested, at the
expense of the starch ; and the action of the chlorine is to set
free a minute quantity of iodine, sufficient, however, for the
starch to restore the blue colour. The fact just given, that
boiling solutions of the starch and iodine may be mingled
without the production of colour further supports the above
view ; nor does the colour appear when the tube has become
cold; but the addition of a few drops of chlorine solution
immediately starts the colour, or if to any of the tubes con-
taining the cold bleached solution a few drops of nitrate-of-
silver solution be added, a faint indication of the presence of
hydriodic acid is obtained.

The conclusions arrived at seem to be:—

1. That the blue colour disappears in consequence of the
dissociation of the iodine and the starch under the action of
a high temperature.

2. That when the liquid is apparently permanently bleached,
and yet the colour is restored by the action of chlorine, a
minute quantity of iodine is present under the form of
hydriodic acid.

the effects to vary with different kinds of starch, the blue colour dis-
appearing more quickly in the case of rice- than with potato-starch.
Casein (cheese, for example) causes a precipitation of the blue compound
upon itself, leaving the liquid clear above. Milk also bleaches, apparently
from the presence of casein.

Gees

|
|
|

XX. The Production of Monochromatic Light, or a Miature —
| of Colours, on the Screen. By Capt. W. DE W. ABNEY, —
| RE, P.RS.* i
. BZ‘OR some time past I have been making experiments on —
: the illuminating power of various sources of light, and —
" it became necessary in some investigations to produce fair-
sized patches of different monochromatic, and combinations of —
monochromatic lights, upon a screen, all other colours being
absent. To obtain this result I had recourse to a modification
of Clerk-Maxwell’s arrangement, as used for colour-mixing.
By his plan slits were inserted at different parts of the
spectrum as formed by an ordinary spectroscope. These slits
were then illuminated by light reflected from a white screen,
and the prism was viewed through the slit of what is ordi-
narily the collimator. The prism was then seen to be coloured

Sqn
pe ee ee Sea ae ee

with light from those rays which would Vv

have fallen on those slits had the spec- ‘ee

troscope been used in the ordinary man-
ner. In this arrangement only one
observer could be utilized at one time,
and my object was to allow several ob-
servers simultaneously to view the colour.
In order to effect this the following appa-
ratus was employed:—A collimator C, ,,
the aperture of the lens Ly, of which
was 13 inches, and the focal length
about 12 inches, was employed. On fae
the slit 8, was cast an image of the LoS
source of light A by means of the con- * Re
densing lens L,. This was of such an
aperture as entirely to fill the colli-
mating lens L, with the rays entering
through the slit.

Two prisms, P, and P,, of 24 inches by
1? inch side, and of angles of 62°, gave
the necessary dispersion to the parallel
beam, from the collimator. The rays

were brought to a focus by means of the 7
: LL
lens L;, of about 14 inches focal length, F

* Communicated by the Physical Society : read June 27, 1885.

<< ee
we at ae we
eS —

a ge
a

we

Le

(

On the Production of Monochromatic Light. 173

on to the screen belonging to a camera, B. ‘This screen was
placed at an angle with the axis of the lens L; as shown, so that
a fair focus of every visible ray was obtained uponit. (It may
be worth mentioning that a hair placed across the slit or a
little particle of dust is a good means of obtaining a focus
when Fraunhofer or bright lines are not observable. The
black streak produced by it should be sharp along the whole
of the spectrum.) A card, D D, with a slit 8, or slits cut in
it, replaced the ordinary dark slide, and, by moving it along
the spectrum, any colour or colours can be allowed to pass.
Before using the apparatus the whole of the spectrum was caused

L- to fall on a convex lens, Ly, of about 24 inches focal length and

about 5 inches diameter. This collected the dispersed beam of
light, giving an enlarged image, I’, of one surface of one of
the prisms on a screen, EH. By placing this lens at an angle
with the axis of the lens Lz, the blue and red fringes can be
made to disappear almost entirely, and a practically white
patch of light is seen on the screen EH.

I may say that the lenses used are white flint of medium
density and almost colourless, even in great thickness.

When the adjustments are complete, as the slit is moved
along the spectrum every patch of colour or colours will suc-
cessively occupy the same position on the screen and have
the same area very nearly. We thus can have patches of
monochromatic light of any colour or combinations of any
colours, all other colour being absent,
| I have also obtained the same results by substituting mir-
|| rors for all the lenses and a reflecting-grating for the prisms;
| but I do not see any particular advantage in this plan, as the
white light is more tinged (with the colour of the metal)
than when prisms and lenses of white glass are employed.

When the source of light is the arc light, if an image of the
crater of the positive pole be thrown on the slit 8, of the
collimator, the intensity of the light is such (when the slit
is fairly open) that the patch of nearly pure light may be
made 1 foot square, and yet be sufficiently brilliant to be seen
by a fairly large audience, and for an ordinary lecture-room
it is very effective. The mixture of colours to imitate any
colour in the spectrum may be shown by placing a narrow
_ slit in a small card in the colour required to be imitated,
and fixing in front of it and in contact with it a portion
of a cylindrical lens, the axis of the cylinder being of course

| parallel to the slit. This throws the image of the particular

colour to be observed at one side of the image that would be
obtained were the cylindrical lens absent, and any amount of
deviation can be given the patch by using the different parts

: 174 On the Production of Monochromatic Light.

of thelens. Thus the greater the deviation required the nearer |
to the margin of the lens the part of it employed should be. |
In other words, the cylindrical lens acts as a series of prisms |
of varying angles. To obtain mixtures of colours to corre- |
spond to the deviated patch, movable slits, and capable of being |
narrowed or widened, are placed on each side of the fixed slit. |
By this plan two patches of light of equal size and equal inten- |
sity can be readily produced. When measurements are to be |
} obtained, scales are attached to the various slits, by which any |
| part of the spectrum can be identified ; and the widths of the |
| slits are measured by a gauge.
| Since this apparatus was described I have referred to a |
paper by Helmholtz, which appeared in Poggendorff’s An- |
nalen in 1855, in which one of the methods he used for the |
combination of coloured light to produce white light is de- |
scribed. The general principle he adopted is the sameas that |
| described above ; but in several important details the latter |
| | differs considerably from Helmholtz’s apparatus. For instance,
the apparatus now described is suitable for the comparison of
colour-mixtures with monochromatic light of any colour and
for their exhibition on the screen for lecture-purposes, and an |
illumination is secured which is very largely in excess of that —
usually obtained. Ina paper read before the Physical Society
(Phil. Mag. June 1885) Lord Rayleigh shows how a mono- |
chromatic image of an external object may be seen by placing

| a concave lens immediately behind the slit of the spectroscope
Pa of such a power as to throw an image of that object on the
prism. I have found that by altering the distances apart of
the collimating lenses and viewing lens, a monochromatic
image of the sun may be thrown on the screen. If such an
image be coloured with the light of the blue or violet hydrogen
lines, it should be possible to photograph the solar prominences

| en bloc. I may mention that the apparatus as described was

| employed in two Cantor Lectures at the Society of Arts in the

beginning of April, but without describing it in detail. :

tale { am at present engaged in using this apparatus for inves-
tigating some phenomena existing in colour-blindness, and

| obtaining curves of illumination of different lights, and also
ae in some photographic researches. These results are not yet
= ripe for publication ; but I have thought it might be well to

. | publish the method employed, as it is one of great conve-

7. nience, and very easily carried out by any one who has a

| spectroscope and a photographic camera.

r 175

XXI. Mechanical Integration of the Product of two Functions.
By WILLIAM SUTHERLAND, J.A., B.Sc.*

Li a communication to the Royal Society (February 3,

1876), “ On an Instrument for Calculating \ $(2)v(2)da,
the integral of the product of two Functions,” Prof. Sir
William Thomson has shown how, by the use of Prof. James
Thomson’s disk, globe, and cylinder integrating machine, the
integral of the product of two functions can be found. The
operations involved consist, first, in plotting the curvey=~7(z),
then in making the instrument, by a rather difficult attach-

ment, yield a trace of the curve y= "ab («e) dar, then in plot-

0
ting the curve y=¢(2): these two latter curves have to be
wrapped round one or two cylinders. When the cylinders are
caused to revolve, a pointer capable of moving parallel to the
axis of y has to be kept on each curve, the motions of the two
pointers being communicated to the disk and globe respectively

of the integrator: the amount of the angular movement of

the cylinder gives the integral. :

In view of the very important part that the analysis of an
arbitrary function into its harmonic constituents is destined
to play in extracting law out of the immense mass of physical
and chemical measurements that are being accumulated, it
seemed to be worth while to look for a simpler method of
obtaining b(2)p(2)da than the above. The following may
be found to be such, and seems likely to be capable of more
immediate application; for the disk, globe, and cylinder

integrator, despite its kinematical elegance, has not yet come

into general use.

The following method is merely a mechanical realization of
the operations which Fourier so carefully describes in his
‘Analytical Theory of Heat,’ art. 220, chap. iii. sect. vi., in
order to give as concrete an idea as possible of the meaning of
the coefficients in his expansion. ‘‘ We see by this that the
coefficients a, b, c,d, e, f which enter into the equation

37$(v)=asin«+bsin2z+csin3e+dsin4ze+ Ke.,
and which we found formerly by successive eliminations are
the values of the definite integrals expressed by the general
term {sin ich(a)dz, t being the number of the term whose
coefficient is required. This remark is important, because it
shows how even entirely arbitrary functions may be developed
in series of sines of multiple arcs. In fact, if the function
(xz) be represented by the variable ordinate of any curve
* Communicated by the Author.

Describe a circle of radius R

| geen

-_

i176 = Mr. W. Sutherland on the Mechanical Integration

whatever, whose abscissa extends from «=0 to w=7, and if
in the same part of the axis the known trigonometrical curve
whose ordinate is y= sin x be constructed, it is easy to repre-
sent the value of any coefficient. We must suppose that for
each abscissa 2 to which corresponds one value of ¢(#) and
one value of sinw, we multiply the latter value by the first,
and at the same point of the axis raise an ordinate equal to the
product ¢(z)sinw. By this continuous operation a third
curve is formed, whose ordinates are those of the trigono-
metrical curve reduced in proportion to the ordinates of the
arbitrary curve which represents ¢(#). This done, the area
of the reduced curve taken from e=0 to z=7 gives the exact
value of the coefficient of sin 2.”’ |

To take the more general case, let it be required: to find

{4 O)(0)d0. Plot the two curves whose polar equations

are r=$(0), r=wW(@), using the same pole and the same initia |
line for angular measurement in both cases. Let 8; and 8, bi |
the points on the two curves corresponding to any value of 6 |.
(see figure) ; then

r(8)6(0) = O8; . OS.

passing through §, and &, ;
through O draw the tangent OT
to this circle, then

OT =05, « O8e.

Now if the arm O§,S, is turned
into any other position, cutting
the curves in two fresh points 8’;
and 8’y, and if the circle of con-
stant radius R is again brought
into position so as to pass through
these points, the new length of
the tangent OT’ is always such
that OT!?= OS, «.O8/,: therefore
generally OT’=W(0)¢(@).

In O8,8, mark off O8S,=OT ;
then if, as OS,S, is revolved
from the initial to the final position, or from that given by
6=a, to that by 0=a,, the motion of T is by any mechanical
device communicated to 83 unaltered, the locus of 8, is a curve

such that always OSs=4(6)W(0) ;
3 ( $(O)(0)d0 = (“os: do

a)

of the Product of two Functions. 177

= twice the area. of the figure enclosed by the third curve
traced out by 83, and by the lines 0=a,, O=ay.

The arrangements for describing the third curve mechani-
cally are obvious. An arm, O8,8,, capable of turning freely
in the plane of the paper round an axis through QO, bears two
collars, S; and 8,, free to move along O8,8,; and to these,
two equal arms, 8,0, 8,C, are freely jointed, being at the same
time each free to turn’in the plane of the paper round an axis

through C; the two collars bear pointers with which to follow
the two curves. A second long arm, OT, can turn round the
axis through O; at T is a collar free to move along OT, and -
to it is rigidly attached an arm, CT, equal to CS, and C&,,
and always at right angles to OT: this arm bears the axis
through C. ‘Then as 8; and 8, are guided along the corre-
sponding curves, the two arms OT and CT move in such a
_ manner that T is always the point in which a circle of radius
| C§,, CS,, or CT is touched by OT; thus always OT?=O8,. O8,.
As regards the transmission of the motion of T to a point
- $3 on the axis OSS, different methods are possible, perhaps
the simplest being that represented by the dotted lines in the
figure. ATBS; is a rhombus with hinges at the four angular
_ points; the hinges at T and 8; are attached to the collars
| there, while the hinges at A and B bear two collars free to
' move along an arm OBA, which in its turn is free to move
| round the axis through O, obviously always OS;=OT. A
pencil attached to the collar 8; traces the required curve.
All that is necessary, then, for finding the integral

{ “*$(8)(0)d0

| is to join the two ends of this curve by straight lines to O and
take the area of the resulting figure by means of any planimeter.

But a very simple attachment to the above mechanism
_makes a planimeter of it. Suppose the area of any figure to
\ be required. Fix the collar 8, at a certain suitable distance b
from QO, so that when the arm O§8,8, moves, 8; must describe
a circle of known radius b. (Guide 8; along the outline of the
| figure, then S, moves always in such a way that

( { OS340= f OS, . O8,d0 = f 08,00 ;

but this last integral is the length of the path travelled by §,.
Thus all that is necessary is to attach at S, a wheel which will
\record the length of the path traversed by 8).

The whole operation, then, of finding { b(8) (8) de
Phil. Mag. 8. 5. Vol. 20. No. 123. August 1885. ie
)

V3
|
L
,

178 Mr. S. Bidwell on the Sienilttaieatone

resolves itself into the plotting of the two curves r=¢(@),
y=r(0), between the limits @=a,, @=a,, the fixing of the |

instrument so that its axis passes through O, the pole chosen
in plotting the curves, the driving of the two pointers 8; and
S, along the two curves (which is easily done by one person)
causing 8; to trace a third curve between the same limits ;
then the clamping down of 8,, the forcing of 8; to follow
back its former track, and the reading of the revolutions of
the wheel attached at 8, during this last part of the perform-
ance. This reading gives the required value of the integral

- when the seales on which the two curves were drawn are taken

into account. These scales must be chosen so that the maxi-

mum difference between OS, and OS, shall at least not be |

greater than the diameter of the circle whose radius is CS).
In the above description, the word collar is used to mean any
form of connection that allows with as little friction as possible
one degree of relative freedom. ie

Melbourne, June 1, 1885,

FORD BIDWELL, J./4., OL.B.*

HE remarkable property apparently possessed by crystal-

line selenium of having its electrical resistance varied

by the action of light, a property which was first announced |
by Mr. Willoughby Smith in 1873, has been the subject of |
many investigations}. Of these the best known, and by far |
the most exhaustive, are the researches of Prof. W. G. Adams |
and Mr. R. H. Day, an account of which is published in the |
Phil. Trans. of 1877. As the result of numerous experiments, |
these gentlemen were led to form the opinion, that “the |
electrical conductivity of selenium is electrolytic” {. The |
principal reasons given for this conclusion are:—(1) that the |
resistance of the selenium-bars used appeared to depend upon >

* Communicated by the Physical Society ; having been read at the

Meetings on May 23 and June 13. 7

+ Willoughby Smith, Journ. Soc. Tel. Eng. ii. p. 31; Earl of Rosse,
Phil. Mag. March 1874, p. 161; Sale, Proc. Roy. Soc. 1873, p. 283; Phil.
Mag. March 1874; Werner Siemens, Phil. Mag. November 137 5, p. 416;
Draper and Moss, ‘Chemical News, xxxili. p. 1; Adams and Day, Proc.
Roy. Soc. 1876, p. 118; Phil. Trans. 1877, p. 318; C. W. Siemens, Proce.
Roy. Inst. 1876, p. 68; Sabine, Phil. Mag. June 1878, p. 401; Graham
Bell, ‘ Nature,’ xxii. p. 500; Shelford Bidwell, Phil. Mag. April 1881, and
January 1883; Fritts, ‘ Electrical Review,’ March 7, 1885, p. 208.

{ Phil. Trana, vol. 167, p. 328; Proc. Roy. Soc, 1876, p. 115.

XXII. On the Sensitiveness of Selenium to Light, and. the
Development of a similar Property in Sulphur. By SHEL- |

i
if

f
Ie

of Selenium and Sulphur to Light. 179

the electromotive force of the battery employed, being gene-
rally diminished as the battery-power was increased ; (2) that
the resistance of a bar AB was generally not the same for

current in the direction AB as for a current in the direction
BA; (8) that the passage of a battery-current was always
followed, when the battery had been disconnected, by a
secondar y or polarization-current in the opposite direction, it
being clearly proved that this secondary current was not due

_ to any thermoelectric action, either in the selenium itself or

|
|
:

a_i

in any other part of the circuit.

The authors do not, however, appear to have considered
that the observed behaviour of selenium was to be explained
by actual electrolysis, but rather that the molecular structure

or crystalline condition of the substance was altered or modi-
fied by the action of a current of electricity in such a manner

as to produce effects analogous to those which would have
occurred if the selenium were an electrolyte and actually de-
composed by the current. As to the possible influence of
light, the following are their words *:—“ Light, as we know,
in the case of some bodies, tends to promote crystallization,
and when it falls on the surface of such a stick of selenium,

tends to promote crystallization in the exterior layers, and

therefore to produce a flow of energy from within outwards,

which, under certain circumstances, appears in the case of

selenium to produce an electric current. The crystallization

produced in selenium by light may also account for the dimi-

nution in the resistance of the selenium when a current from
a battery is passing through it, for, in changing to the
crystalline State, selenium becomes a better conductor of
electricity.”

Attention has lately been again directed to the subject of
selenium, and its behaviour under the influence of light, by
the publication, by Mr. C. E. Fritts of New York, of a new
and extremely ingenious method of constructing selenium
cellsf. He melts a thin film of selenium upon ‘a plate of

-metal with which it will form a sort of chemical combina-

mom 2... During the process of melting and crystallizing,
the selenium is compressed between the metal piate upon

- which it is melted and another plate of steel or other substance

with which it will not combine..... The non-adherent plate

being removed after the cell has become cool, | he] then covers

that surface with a transparent conductor of electr icity, which

may be a thin film of gold-leaf..... The whole surface of
* Proc. Roy. Soc. 1876, p. 117.

+ Proc. American Assoc. 1884.. Reproduced in the ‘Electrical eh ie
March 7, 1885, p. 208. )
N 2

180 Mr. S: Bidwell on ‘the Sensitiveness

the selenium is therefore covered with a good electrical con-
ductor, yet is practically bare to the light, which passes through
the conductor to the selenium underneath.” The sensitive-
ness to light of cells constructed in this manner seems to be
far in excess of anything that has been previously obtained ;
and the “ photoelectric’ currents which (like the selentum
bars of Messrs. Adams and Day) they are capable of origina-
ting, are said to be strong enough to be actually useful im

practi eal work.

It is impossible to read Mr. Fritts’s paper without being |
impressed by the resemblance of some of the phenomena —
which he describes to those of electrolysis. The mere arrange-
ment of the apparatus—two metallic plates with a third sub-
stance between them—is in itself strongly suggestive ; while
the unequal resistance offered by the two surfaces, and the
generation of an independent electromotive force, in conjunc-
tion with the polarization-effects above referred to*, make it

hard to believe that the conduction of selenium (in the form

used in experiments) is not truly and literally electrolytic.
The only considerable difficulty in the way of this hypothesis
arises from the fact that selenium is not an electrolyte. Hyver
since its discovery in 1817, selenium has been regarded as an
element, and very strong evidence indeed would be necessary |
to deprive it of its elementary character ; this is perhaps the |
reason why the electrolytic theory has not previously been
proposed. But there is a possible way out of the difficulty,
which was suggested to me by the first words in the above
quotation from Mr. Iritts’s paper. He spreads the selenium
upon a plate of metal with which i will form a chemical
combination. Now selenium will, I believe, combine more
or less easily with all metals, forming selenides; and in ex-
periments upon the conductivity of selenium, it has been
usual to submit the substance to prolonged heating in contact |
with metallic electrodes. This prolonged heating (generally |
followed by slow cooling) has hitherto been called “annealing;” |
and the undoubted fact that it diminishes the specific resist- —
ance of the selenium and increases its sensitiveness to light, |
has been explained by supposing that the process is favourable |
to perfect crystallization. |
I venture to suggest, as the true explanation of the effect, |
that heating is favourable to a chemical combination between |
the selenium and the metal forming the electrodes, that a |
selenide is thus formed which completely surrounds the elec- |

“ “The existence of polarization,” says Clerk Maxwell, “may be |
regarded as conclusive evidence of electrolysis,” ‘ Electricity,’ vol. 1. |
363, |

p:

of Selenium and Sulphur to Light. 181

trodes, and is perhaps diffused to some extent throughout the
mass of the selenium*; and that the apparently improved
conductivity of the selenium, together with the electrolytic
phenomena which it exhibits, are to be accounted for by the
existence of this selenide.

I have sometimes been tempted to think it possible that the
apparent conductivity of selenium may in fact be entirely due
to the impurities which it contains, and that perfectly pure

selenium would be as good an insulator when in the crystal- ©

line form as it is in the vitreous condition. Vitreous selenium |
might contain a large percentage of conducting particles
without sensible increase of its conductivity, but that this
would not be the case with crystalline selenium, is rendered

more than probable by the results of some experiments which
I have described in a former communication. If a conduct-
ing powder, such as graphite, is mixed with melted sulphur,
even in small proportions, the mixture when cold is found to
conduct electricity; while if a very large proportion of the
same powder is incorporated with melted shellac, the shellac
when cold remains sensibly as perfect a nonconductor as if
it were pure. The explanation which I have given of these
facts, and in support of which a number of experiments are
quoted, is as follows:—The first mixture does not consist of a
uniform structureless mass of sulphur, having particles of
carbon imbedded in and completely surrounded by it: it is in
fact an aggregation of little crystals of sulphur with carbon
packed between them like mortar between bricks. The con-
duction thus takes place entirely through the carbon particles,
which may be considered as extending in a series of chains
from end to end of the mass. In the case of the shellac
mixture, though the proportion of carbon may be larger than

in the sulphar experiments, the resistance is still sensibly

infinite, because the structureless shellac penetrates between
and completely surrounds the carbon particles.. Just in the
same manner, selenium, when in the vitreous condition, would

_ completely surround any particles of conducting selenides

>
ee ST

which it might contain; while, when the selenium was
erystallized, the conducting particles would arrange them-
selves in the form of a network, capable of conveying a current
of electricity.

Selenium which is free from impurities appears not to be
an article of commerce. An analysis of samples collected by
Professor Graham Bell from different parts of the world

* The selenium is pecpornily for some time in a liquid state,
+ Phil. Mag. May 1882, p. 347.

182 Mr. 8. Bidwell on the Sensitiveness

disclosed the presence of the metals iron, lead, and arsenic’,
all of which would form conducting selenides. Nevertheless
I thought it would be worth while to ascertain roughly the
specific resistance of a piece of selenium which, since it has
come into my possession, has never been in contact with metal.
The selenium (which was supplied by Messrs. Hopkin and
Williams) was melted in a mould built up of slips of glass,
erystallized and ‘‘ annealed ”’ in the usual way ; but, contrary
to the general practice, it was not fitted with metallic electrodes
before annealing. A plate of crystalline selenium was thus
formed, having a thickness of about 2 millim. and a superficial
area of 1 square centim. The two opposite surfaces were
rendered smooth and clean by rubbing them upon a flat board
covered with fine glass-paper, and the plate was placed between
two layers of thick tinfoil which were pressed into good con-
tact with it by a weight of 500 grammes. When this arrange-
ment was connected in circuit with 6 Leclanche cells and a
reflecting galvanometer, a deflection was produced indicating
a current of about ;, micro-ampere. Assuming the electro-
motive force of the battery to have been 10 volts, the-resistance
of the plate would be 500 megohms; and therefore the resistance
of a cubic centimetre of the selenium between opposite faces
(i. e. its specific resistance) would be 2500 megohms. From
the dimensions and resistance of a good selenium cell with
copper electrodes, which I have in my possession, I calculated
that the specific resistance of the selenium contained in it was
about ‘9 megohm. Thus, so far as the result of a single rough
experiment can be trusted, it appears that the conductivity of
selenium which has been annealed in contact with copper is
nearly 3000 times greater than that of selenium which has
undergone similar treatment without the presence of a metal.
Whether selenium, when perfectly pure, is altogether a non-
conductor, would be an interesting question for an expert
chemist to determine J. It is sufficient for the theory which
I am at present advocating that its specific resistance should
be very high. |
By assuming the admixture with the selenium of metallic
selenides, an explanation is attorded of the following facts:— |
(1) The diminished resistance produced by annealing. |
(2) The fact, first pointed out by Graham Bell, that the |
resistance of selenium appears to depend oreatly upon the |
nature of the metals of which the electrodes are formed. For |
obtaining low resistance he recommends the use of brass in |

* Paper read before the National Academy of Sciences, April 21, 1881.
+ On more mature consideration I am inclined to think that it is not,

_ p. 590.

of Selenium and Sulphur to Light. 183

preference to platinum, and expresses his belief that the che-
mical action between the brass and selenium contributes to
the low resistance of his cells, “ by forming an intimate bond
of union between the selenium and brass.”’ *

_ (8) The fact observed by Adams and Day that there is
generally a “‘ diminution of resistance in the selenium as the
battery-power is increased.’’ ‘he same phenomenon occurs:
in the mixtures of sulphur and carbon before referred to. . It

points to the existence of imperfect contact between conducting

particles, the conduction partaking of the nature of disruptive
discharge, and is consistent with the supposition that particles
of conducting selenide are imbedded in the selenium f.

(4) The apparent production by a current through a piece
of selenium of a “ set of the molecules which facilitates the
subsequent passage of a current in the opposite, but obstructs
one in the same direction ’’{. This would be accounted for
by the electrolytic deposition of selenium (from the selenide)
upon the anode.

(5) The polarization-effects, which would also proceed from
electrolysis.

(6) ‘‘ A slight increase of temperature of a piece of annealed
selenium is accompanied by a large increase of electrical re-
sistance ’’§. This also occurs in the mixture of sulphur and
carbon, and is explained by supposing that the heat-expansion
of the medium draws apart the conducting particles contained
in it, causing them to have fewer points of contact with each
other, and thus increasing the resistance of the whole||. A
more considerable rise of temperature so greatly diminishes the
specific resistance of the selenide (and perhaps of the selenium)
as to more than counterbalance this effect; and thus it happens
(as I have shown in a former communication 4) that selenium.
cells have a ‘‘ temperature of maximum resistance,’ which is
generally a few degrees above the average temperature of the

| al F;

(7) The resistance of prepared selenium is generally greatly
diminished by the action of time. Prof. Adams found that
the average resistance of a number of pieces of selenium was

* Lecture to American Assoc, 1880. Reprinted in ‘ Nature,’ vol. xxii.

+ See “On the Electrical Resistance of Carbon-contacts,” Proc. R. S.
Feb. 1, 1883; and “On Microphonic Contacts,” Journ. Soc. Tel. Eng.
April 12, 1883.

t Adams and Day, Proc. R. 8. 1876, p. 114.

_. § Adams and Day, Phil. Trans. 1877, p. 342. See also Phil. Mag. Jan.

1885, p. 31.
|| Phil. Mag. May 1882, p. 351.

(| Phil. Mag. April 1881.

5S NTs ees |

184. Mr. S:; Bidwell on the Sensitiveness :

reduced to less than one fortieth in the course of a year *.
During this period the selenium had been in contact with the
metallic electrodes; and it seems possible that a larger quan-
tity of selenide than was produced in the first instance by the -
process of annealing was slowly formed. This would espe-
cially occur at the “marked end,” or anode, where there
would naturally be a quantity of free selenium. |
In the above argument it has been assumed that selenium
will combine directly with any metal with which it is brought
into contact, the combination being facilitated by the appli-
cation of heat. In the case of such metals as copper, brass,
and silver this is undoubtedly the fact. Indeed, an attempt
to make a selenium cell with silver wires was attended with
failure in consequence of the complete destruction of the |
metal after contact with the melted selenium for only two or
three minutes. Itis, however, questionable whether platinum
(which was the metal used by Adams and Day) is, in any
sensible degree, attacked by selenium either at the ordinary
temperature or at that reached in the process of annealimg.
With sufficient heat the two substances will undoubtedly
unite; and I have found that the surface of platinum-foil upon
which melted selenium has been kept for an hour or two ata_ |
temperature probably of about 250° C. acquires a bluish-grey_ |
colour which may be due to selenide. But whether any ap-
preciable quantity of selenide is formed im the ordinary pre-
paration of crystalline selenium is a question only to be settled
by the aid of refined chemical operations which I am incom-
petent to undertake, and in the meantime the suggested theory
is left without direct confirmation. |
But certain indirect evidence in support of my views has
been forthcoming. Selenium is an element which, in its
properties, closely resembles sulphur, and attempts have from
time to time been made, hitherte without success, to develope
in sulphur that peculiar sensitiveness to light which is sucha |
remarkable characteristic of selenium. It oecurred to me |
that if this property of seleniam were really due to the acci-
dental existence of metallic selenides, then the admixture |
with sulphur of metallic sulphides might be expected to lead |
to similar effects. Itis not possible to ‘‘anneal”’ a stick of |
sulphur or a sulphur “ cell” previously furnished with metallic |
electrodes, because sulphur does not, like selenium, solidify |
and crystallize at a higher temperature than that of its first |
melting-point, But if it is true that the virtue of annealing |
really lies in the fact that a chemical union of the two elements |
is promoted by the action of heat, it is clearly immaterial |

* Phil. Trans. loc. cit. p. 348.

of Selenium and Sulphur to Light. 185

whether the substances are heated together before or after the
formation of the cell. Sulphur containing sufficient metallic
sulphide to render it a conductor of electricity might be used
in the construction of a cell which might be expected to be
sensitive to light without any preliminary annealing. This
turned out to be actually the case.

Silver was the metal chosen for the experiments on account

of the facility with which it combines with sulphur.

Cell No. 1.—Five parts of sublimed sulphur and one part
of precipitated silver were heated together in a porcelain cru-
cible for about two hours. The mixture was from time to
time stirred with a glass rod and was finally allowed to settle,
so that the bulk of the sulphide and any free silver which might
remain fell to the bottom of the crucible. When the tem-
perature was slightly above the melting-point the liquid
sulphur, which was perfectly mobile, though black with
minute suspended particles of sulphide, was poured off for
use. ‘Two wires of fine silver* were then coiled side by side
around a strip of mica 50 millim. long and 27 millim. wide ; the
wires were about 1 millim. apart, and care was taken that they
did not touch each other at any point. Some of the melted
sulphur was spread evenly over one surface of the mica, the
two wires being thus connected with each other through half
their entire length by a thin layer of the prepared sulphur.
When cold, this cell was connected in circuit with a battery
and a galvanometer. It was found to conduct electricity,
but its resistance was very high, being probably between
2() and 80 megohms. With the object of partially bridging
over the intervals between the wires, the sulphur was melted
by laying the cell upon a hot plate, and a piece of very thin
silver-foil, measuring 25 millim. by 10 millim., was laid upon
its surface: this was probably entirely converted into sulphide
before the cell was again cold. The cell was now found by
a bridge-measurement to have a resistance of 900,000 ohms f.
Once more it was connected with a Leclanché cell and a
suitably shunted galvanometer; the deflection was noted,
and a piece of magnesium wire was burnt at a short distance
from the sulphur. The deflection was immediately more than
doubled; and when the magnesium was extinguished, the spot
of light at once returned to very nearly its original position.
The effect was almost as great when a glass trough containing
a saturated solution of alum was interposed between the
sulphur and the burning magnesium f.

* Supplied by Messrs, Johnson and Matthey.

1 This resistance was afterwards found to be very variable, and it was
never the same with a direct and a reverse current.

{ This cell was exhibited in action at the Meeting of the Physical
Society on May 23rd, and at the Soirée of the Royal Society on June 10.

/

186 Mr. 8. Bidwell on the Seriwiblobiees

Now it is well known that the resistance of sulphide of

silver is greatly diminished by heat *, and it was therefore
important to ascertain whether the effect just described was
due to light or to heat. To speak more accurately—Is it an

effect of radiation or of temperature? Exposure to radiation,

whether visible or invisible, is of course always accompanied
by a certain rise of temperature, and confusion has sometimes
arisen, especially in discussing the properties of selenium, from
failure to distinguish between the direct effects of radiation,
and the indirect effects which are primarily due merely to a
rise of temperaturet. In the photographic processes it is
radiation per se that produces the observed results: in the best
known processes, the effective rays happen to be those which
correspond to the most-refrangible part of the visible spec-
trum together with the invisible rays beyond it. But by
more recently discovered methods the ‘‘ obscure heat-rays,”
as they are sometimes called, have been made available for
photographic purposes {; and these do not act by virtue of
any rise of temperature which they may cause, but exert
direct chemical action upon the sensitized plate. Again, if a
thermo-pile is exposed to radiation, an electromotive force is
generated. Here, however, the effect of radiation is indirect ;
it acts only through the medium of the heat which it pro-
duces’; and if an equal and similarly distributed amount of
heat were communicated to the thermo-pile by any other
means (as by conduction), exactly the same effect would
follow. In an ordinary selenium cell radiation acts both
directly and indirectly, tending to produce opposite effects.
The direct. effect of the radiation, whether it be visible or
infra-red or ultra-violet, is a diminution of the resistance of
the cell; at the same time the radiation slightly raises the
temperature of the cell, and so indirectly tends to increase its
resistance. If a selenium cell in circuit with a battery and
a galvanometer is suddenly exposed, by withdrawing a screen,
to the radiation of a black-hot poker, a momentary swing of
the galvanometer-magnet will at first indicate a fall in the
resistance; but this will be almost immediately followed by
a rise which will increase up to a certain limit as the tempe-
rature of the cell becomes higher. The same kind of thing
occurs when the cell is exposed to the infra-red or red por-
tions of the spectrum; but in the latter case the temperature-
effect merely diminishes, instead of overpowering, that directly
due to radiation. Ifthe bridge method is used for measuring
* Faraday, Exp. Res. §§ 452-439.
+ See Moser, Proc. Phys. Soc. 1881, p. 348.

{ Captain Abney is said to have obtained a photograph of a kettle of
boiling water by means of the invisible radiations which it emitted.

of Selenum and Sulphur to Light. 187

the resistance, it may easily happen that the effect of gradu-
ally rising temperature escapes notice, a balance not being
obtained until the temperature has become constant ; and thus,
probably, is to be explained the fact that different observers
have attributed the most powerful action upon selenium to
different parts of the spectrum, ranging from infra-red to
greenish yellow.

The resistance of the sulphur cell which has been de-
-seribed, unlike that of most selenium cells, was diminished
by a rise of temperature*. When in circuit with a Le- ©
elanche cell and a galvanometer, the effect of holding a nearly
red-hot brass rod at a distance of 3 centim. from its surface, |
was a gradual fall of resistance, which in 15 seconds was
indicated by 23 scale-divisions. When the rod was removed,
the spot of light slowly returned to its original position,
occupying several seconds in doing so. It is certain that the
temperature of the sulphur must in this experiment have
been much higher than when it- was exposed to burning
- magnesium, with a solution of alum interposed, yet the effect
was very much smaller; moreover, it was gradual instead of
instantaneous.

Another experiment seems to prove conclusively that the
resistance of the cell is diminished by the direct action of
radiation, quite apart from any effect which may be produced
by an incidental rise of temperature. On a cloudy day the
cell, with the alum trough before it, was placed at a distance of
16 feet from a small window, all the other windows in the
room being darkened. With the same battery and galva-
nometer as before, it was found that closing the window-shutter
caused an instantaneous swing of the spot of light through 90
scale-divisions in the direction indicating increased resistance;
and when the shutter was again opened, there was immediately
an equal swing in the opposite direction. A delicate thermo-
pile of 54 pairs, connected with an astatic reflecting-galva-
nometer of low resistance, was then put in the place of the
sulphur cell, and the alum trough placed before the open end
of the conical reflector attached to it. On opening the window-
shutter, a deflection occurred indicating a current which was
found by trial to be equal to that produced by the radiation
of the human body at a distance of 10 ft. 6 in. Itis needless
to say that such a minute change of temperature as this
implies was without sensible effect upon the resistance of the
sulphur cell. There can then be no doubt whatever that the
whole of the observed effect of the light upon the sulphur

_ * This was not so with all the cells subsequently made. See descrip-
_ tion of cell no. 3 below, ,

SS

188 Mr. S. Bidwell on the Sensitiveness

was due to the action of radiation as such, any change of —

resistance resulting from the incidental rise of temperature
being quite inappreciable. |
~ Cell No. 2.—This was constructed in a somewhat tlifferent |
manner. A piece of silver-foil was laid upon the surface of ©
the mica before the two wires were wound round it, and
instead of having prepared sulphur spread upon one face, the
whole was immersed in pure melted sulphur for a few minutes,
and then carefully drained. Before this treatment the silver
wires were of course short-circuited by the foil, but the liquid

‘sulphur penetrated between them, forming a film of sulphide;
-and when cold, the resistance of the cell was about 100,000

ohms, Though this cell turned out to be somewhat less sen-
sitive than the other, it seemed likely that, on account of its
comparatively low resistance, it might be successfully used
for a photophonic experiment. It was therefore connected in
circuit with a battery of ten Leclanché cells and a telephone,
and exposed to a rapidly interrupted beam of light. The
telephone at once gave out a musical note, which was nearly
us loud as that produced by a good selenium cell under
similar circumstances.

The behaviour of this cell sadtiae changes of temperature
was the same as that of the other.

Cell No. 3.—A mixture, consisting of equal parts of sub-
limed sulphur and precipitated sulphide of silver, was melted
and spread on one surface of a slip of mica, around which two
silver wires had been wound as before. No foil was used in
this case. The resistance of this cell was diminished by radia-
tion, but increased in a very marked manner by rise of tempe-
rature. <A paraffin lamp, at a distance of 18 inches, produced
a steady diminution of the resistance. Wher the lamp was
placed at a distance of 10 inches, the galvanometer-needle first
moved in a direction indicating a further fall of resistance ;
but after a few seconds, when the temperature began to rise,
it turned in the opposite direction. On moving the lamp
6 inches nearer, there was at once a large deflection in the
direction of increased resistance, the temperature-effect com-
pletely predominating over that of radiation.

Cell No. 4.—A strip of silver-leaf was attached to a glass
plate by means of gold size, and the middle part of it was
exposed to the vapour of boiling sulphur until both surfaces
were completely blackened. The resistance of this cell was
high, but for a few days it was extraordinarily sensitive, a
reflaoted beam of sunlight instantly effecting a diminution of
80 per cent. in its resistance. About a fortnight after it was
made, its sensitiveness had greatly fallen off.

- of Selenium and Sulphur to Light. 189 _

All these cells resemble selenium in giving. polarization-
currents after being detached from the battery.

Supposing it to be true, then, that it is not in the selenium
or sulphur itself, but in certain metallic selenides or sulphides,
that the sensitiveness to light is resident, does it become easier
to explain the phenomenon, or, rather, to deprive it of the
unique position which it has hitherto appeared to hold, and
assign to it a place among a class of analogous effects?

I believe that, at all events in the case of the sulphur-silver —
cell, it is principally at the surface of the electrodes that the
effecis of radiation are to be looked for.

If a current of electricity is passed through a mass of
sulphide of silver having silver electrodes, silver will be
deposited upon the cathode and sulphur upon the anode. The
accumulation of silver upon the cathode will clearly produce
no appreciable effect upon the conductivity of the arrange-
ment, and need not be considered. But sulphur has an enor-

_ mously high resistance, and the deposition of a mere film of

free sulphur upon the anode would be sufficient to stop the
current altogether. The current is not in fact stopped, because
the deposited sulphur at once combines with the silver of the
anode, merely adding a new layer to the electrolyte. Thus
the metal of the anode gradually combines with the sulphur
of the electrolyte ; and the conductivity of the arrangement
will depend to a great extent upon the facility with which this
combination is effected, the quantity of electricity which can
pass in a given time being limited by the quantity of sulphur
which is capable of uniting with the electrode in the same
time.

Sulphur combines with silver far more readily than with
iron. If, therefore, my views are correct, we should expect
a cell with an iron anode to offer a much greater resistance
than one which had an anode of silver, the material of the
cathode, so long as it was a good elementary conductor of
electricity, being of comparatively little importance. To test
this idea, a cell was made consisting of electrodes of iron and
silver imbedded ina mixture of sulphide of silver and sulphur.
The cell being connected with a battery and a galvanometer,
the deflection was 115 divisions when the current passed from
the silver to the iron through the electrolyte, and only 4 divi-
- sions when the direction of the current was from iron to silver*.
_ The resistance was therefore nearly 30 times as great with an

* The resistance of the galvanometer was 3483 ohms, and it was shunted
with a coil of 20 ohms. The resistance of the Leclanché cell was about
5 ohms. |

190 Sensitiveness of Selenium and Sulphur to Light.

iron anode as with a silver anode. It is clear that this was
not the result of bad contact between the iron and the elec-
trolyte (such as was supposed by Graham Bell in the analogous
case of selenium to account for the high resistance of a cell
with platinum electrodes as compared with one in which the
electrodes were made of brass), because such an effect would
be independent of the direction of the current. Rather it
seems that the resistances of the two anodes afford data for
measuring the relative facilities with which sulphur combines
with silver and with iron.

Assuming it to be thus experimentally proved that the re-
sistance of a sulphur-silver cell depends largely upon. the
readiness with which sulphur unites with the anode, it follows
that any cause which would assist this union would at the same
time diminish the resistance. Now it is well known that certain
chemical combinations are accelerated by the action of radia-
tion—the explosive union of chlorine with hydrogen under the
influence of sunlight being a familiar example. The question
then suggests itself, Does sulphur combine with silver more
readily when exposed to radiation than it otherwise would ?
There is, I believe, direct evidence that it does.

A glass plate, covered with silver leaf, was placed, with the
silvered side downwards, over a crucible of boiling sulphur.
One half of the plate was covered with a piece of black cloth,
and the arrangement was exposed to bright sunshine. In a
short time the visible portion of the silver was darkened,
owing to its partial conversion into sulphide ; the cloth was
then removed, and the silver beneath it was found to be
scarcely discoloured. There was a distinct line of demarcation
between the two halves. The experiment was repeated with
the same result. |

Since this effect might possibly have been due to other
causes than the action of light (such as the unequal conden-
sation of sulphur vapour upon the covered and uncovered
portions of the plate), the experiment was made in another
form. A piece of silver leaf attached to glass was brushed
over with a solution of sulphur in bisulphide of carbon ; and
in order to keep the temperature low and uniform, the silvered
glass plate was immersed in a basin containing cold water,
which was placed in the sunshine. A board was laid across the
top of the basin so as to shade one half of the plate, the other
half being exposed to the direct rays of the sun. In a quarter of
an hour the exposed portion of the silver had acquired a dark
brown colour, while that which had been protected was of a
pale yellow tint, the outline of the shadow of the board being
sharply defined. I think we have here the strongest evidence

On the Law of Density-Numbers. . £91

that the combination of sulphur with silver is assisted by
radiation.

But it is not perhaps necessary to assume that the effective
action of light is confined entirely-to the surface of one of the
electrodes. If, as is commonly believed, electrolytic conduc-
tion involves a series of decompositions and recompositions
throughout the electrolyte, any cause which assists either the
separation or recomposition (or both) of the components of
the electrolyte might be expected to increase its conduc-
tivity ; and it seems reasonable to suppose that the. same
influence which would assist the union of two substances when
they have a tendency to unite would also be favourable to
their separation when they have a tendency to separate. It
is not impossible, therefore, that radiation, acting upon the

. surface of a thin layer of sulphide of silver through which an

electric current is passing, might, by facilitating the mole-
cular rearrangement of the atoms of sulphur and silver, exert
a material influence upon the conductivity of the sulphide*.

So far as regards the explanation of the effect of light upon
the resistance of selenium, I am aware that this paper contains
little more than speculative suggestions, which are at present
almost entirely unsupported by experimental evidenceft. It
is, however, noteworthy that these speculations led to the con-
struction of a cell which, without containing a particle of
selenium, behaved almost exactly as if it were composed of
that substance. How far this may be considered to prove
anything with regard to selenium | do not know ; but in any
case the discovery of another substance possessing the same
remarkable property seems in itself to be a matter of some
interest.

XXIII. On a New Law, analogous to those known under the

names Law of Avogadro and Law of Dulong and Petit. By
J. A. GROSHANST.
[ Concluded from p. 30.]

? T T may be said that, as yet, the study of the causes mame

influence the value of z has to be commenced, and that
all that pertains to this constant is still uncertain. Still I

* There are some experimental reasons, into which I am not at present
prepared to enter, for believing that the admixture with the sulphide of a
certain amount of free sulphur is necessary for the development of sensi-
tiveness to radiation.

+ It is especially desirable to ascertain experimentally whether the com-
bination of selenium with the metals used as electrodes in selenium cells
is assisted by light.

{ Communicated by the Author. Translated by W. W. J. Nicol, M.A.,
DSe. | |

192 M. J. A. Groshans on the.

may be permitted to embody in the following pages the prin-
cipal observations that I have been able to make on this
subject ; and as in doing so I shall have frequently to refer
to glycerine, it is perhaps as well to make a few remarks on
its boiling-point.

There is but little agreement to be found among the figures
given by various experimenters who have attempted to deter-
mine the boiling-point of this compound, C; Hz O03. The older
experiments point to 9 as the value for 2, the more recent
to 9°.

In this connection I may add that Wurtz’s butylic glycol,

C, Hy) Oo, boils at 183°-184°, that of Kekulé at 201°-204° :

sV8-5=182°9, sV¥9=196%2. :
In the following table I have adopted 9 as the value of a for
glycerine.

Glycerine exhibits a peculiar property with regard to e—
it is this :—In glycerine, C;H;(OH)3 one, two, or three of
the hydrogens in hydroxyl groups may be replaced by ethyl,
_C,H,, to form ethylines. The original glycerine and its three

derivatives form a species of homologous series the members

of which have the common difference C,H,, and yet the
increase of molecular weight is attended by a lowering of the
boiling-point, exactly the opposite of what is observed in the

case of other homologous series: this anomalous behaviour 1.

attribute to the “ constancy of x.” In fact glycerine and its
derivatives have all the same constant Tn/a, and consequently
the same value of z—i. e. 9. The theoretical value of Tn/a,
when z=9, is 83°41, as will be seen from the table. Other
derivatives similar to the ethylic ones also possess the same
constants.

TABLE XV.

Name of subst F oe

ame of substance. ormula. a. 1. ohana Tn/a.
CREVCCTIND 55 <capnnescdse C, H, O, 92 14 277 83°69
Monoethyline ......... O;H,.0, 120 20 227 83°33 |
RTMBTONWAIND ois rsiovasni CG, H;,.0, 148 26 191 81°51
Triethyline ............ O, £10, 176 32 185 83°28
fi Eee eee Orit, | ele 35 232 83°37

* Mendelejetf gives 290° at 760 millim.—W. W. J. N.

s/9=275'1 and sV9:5=290'1*. ;

former :—

Law of Density-Numbers. 193

Methyl glycerine, CH(OH),, is not known, nor is methy|
siirite (methylphycite), C(OH),. But there are two com-
pounds w hich have the composition of the ethylines of these
two unknown bodies, and a third, which is the allyline of the

(1) C, Loa e ae CH(OC,H;)s3,
(2) OF Hy)0,=C(OC.Hs).,
(3) C6 oe CH(OOC3H;)3.

And these have all the same value of Tn/a, z=7. This pro-
perty (the constancy of x) is to be met with in very many —
other cases. I ought, however, to observe that in reality this
constancy of x is but a special case. When substitution-
products are formed each has a particular value of a, which
may be the same as that of the body from which it was formed,
or may be greater or less than that by one or more units,
according to circumstances.

In instances where the value of & for the original substance
cannot be determined, it is frequently found that the products
of substitution, or some of them, have the same x. This is
the case with urea, CO(NH,).. According to Table IL, the
value of Bis 12. Table XVI. contains four substituted ureas;
x=16 in three cases, but in the fourth = 11; or a difference
of 5—a number which we shall find later on is a very frequent
difference in the value of « for analogous bodies. In the
following table I have, instead of the constant TB/a, inserted
the value of 2, =1(27: 8)? x (TB/a)’, which is the same thing,
but in this instance more convenient. The boiling-points are
the mean of those given in Fehling’s dictionary.

FABLE XVI.

N | Formula a. |B.| nia (Se
ame, | pace ; | ’ lobserved. found.

es Se ee SS sae OSES

ten... 5... oe oe COOGNEL), o4 60. 12

x;
calculated.

NHOH,

Dimethyl urea ...| CO(NHCH,), | 18] 269 | 15:90} 16
. | .
Methylethy! urea. CONF, i, que 21) 267 | 15 29 16 |

Diethyl urea ...... | CO(NHC,H,), 116| 24} 263

equate Lae, me (CH;).) . med 24 176 11

In the last column of. the table is given ne theoretical value
of x, which is _ simply the whole number nearest to the found

Phil. Mag. 8. 5. Vol. 20. No. 123. August 1885. O

194 M. J. A. Groshans on the

value. In Table XVII. the constancy of w is given in + the

ease of the hydrazines, bodies resembling in formule urea 5
and the hypothetical first member, NH, differs from that in
the ureas, sas io by the group CO.

TaBLe XVII.
é 5. at, 2,
Scone Forms. a B. observed.| found. | theory.

‘Hypothetical NH } 2

- first member NH, = 7”
= talib he | 108 20 083 | “SG | oie
I bs a wien NH.

Diethyl hy- NHO,H; | | £
adage ie NHOH | 88 22 965| 1110 | 1
Phenylmethyl {| NHC,H; | | 4
hydrazine... { NHCH, f 128 a0 223 | Irs 1

Here, again, is the value of #=11, as was the case in one
of the bodies in the last table*.
Table XVIII. contains ethylene diamine, C,H,N.Hy, and

its substitution-products, along with an analogous compound,

TABLE XVIII.

Ethylene diamine and its substitution-products.

52; ry z,

Formula. te B. observed.| found. | calculated.
N,H,(C,H,) ...... 60 16 117° | -75-98 14
MOH ACH,),....5- 86 20 170 13-73 14
N,(C,H,),(C,Hs),| 142 32 185 | 13-72 14.
ee Sea 112 24 210 | 13-86 14
OK: 103 26 208 | 1907 | 19

* We find in the aromatic series bodies metameric with those in
Table XVII. ; thus :—
Metamers of C,H,N,.
Three phenylene diamines, with the constants z respectively =12, 13, 14,
Metamers of C,H,,N,.
Three cresylene diamines, with z respectively 14, 18, and 14.

; |

Law of Density-Numbers. 195

In ammonia, NH;, z=9; and thus we find that in the
three series of bodies of which the three following are types
there is a steady increase of x apparently without limits:-—

Ammonia, NH,, e e 2 e ° ob = a

Hthylene diamine, N,0,H,, . #=14,
5
Triethylene triamine, N3;C,Hy3, 2#=19;

the difference NC,H; therefore corresponds to z=5, or in the
atomic weight =43, in density numbers = 10, numbers com-
parable with those observed in the case of the elements (see
above).

It is also possible to use the property of the constancy of «
in substitution-products to determine the value of some
density-numbers, especially that of chlorine. It is found
that, in general, 2 does not change when 1, 2, 3, or more
atoms of chlorine are substituted for 1, 2, 3, or more atoms

of hydrogen in a body of the formula C,H,O,. Table XIX.

TABLE XIX,

Toluol and its Chlorine substitution-products.

| |
; eerie oe

| Formula. | a. B. - observed. T. B/a
a 92 th ode 62:1
i 126°5 8 | 1641 62°18
_ 2 ae | 161 pot 206 W 62:48
Ot ke 19595 geet Say il * Saas
A 3 via ae 230 27 | 255 L | 61-98

N=Noad; L=Limpricht ; W=Wicke.

shows this clearly in the case of the four chlorine deri-
vatives of toluol, C;H,; for when the density-number of
chlorine is taken as 4, the value of the constant Tn/a is all
through 62:2, and «2 is equal to 5. In the case of benzol,
C,H,, <=4, and so also for C,H;Cl, C,H,Cl., and C,H;Cl,;
while e=4°5 for C,H,Cl, and C,HCl;. But =5 for C,Cl,.

I have published, elsewhere*, extended tables showing
clearly this constancy of # in chlorine substitution-products.

* Annalen, Wiedemann, 1879, p. 184; J. A. Groshans, Hin neues
Gesetz, Leipzig, Barth, 1882, p. 43; J. A. Groshans, De la nature dea

_ Eléments, Haarlem, héritiers Loosjes, 1875, p. 47.

O 2

196 M. J. A. Groshans on the

It follows from this that numerous analogous compounds
have the same value for z. In the following examples I have,
for the sake of conciseness, omitted the numerical data, giving
only the value of w deduced from the observed boiling-
points :—

Examples of the Constancy of x.
(a) Methylic and Ethylic compounds, with Oxygen, Sulphur,

Selenium, and ‘Tellurium.

(CH. )) O; 2: (C,H,), O, | #3

(Gls )6 153 sata: (CoH;)98,. ae

es: Sp) w= a. (C,H; )3 85) 29

(CHs). Se, a=8. (C,H). Se, «=o
Vo x

)
(CH; Te, v= ae (CH). Seo
i in every case =—™.

(8) Analogous compounds of Carbon and Silicon.

CCl, oe
SiCh, pa (CH), C. 28
CHCl,, ise
MCs owe 2 (CH,), 81, . 22

Here again «=m.

(y) Analogous compounds of the members of the triad, Phos-
phorus, Arsenic, and Antimony.

(CH;); P, C=5, PO (CLH,0), ae
(CH.), As, . #=0. AsO (C,H;0),, @=30
POCO). sg). as i ® &
AsO (CH,0), a2=5. SbI;, “set.

2=7, also for HgCl,, HgBr,, Hgl,, and Hgoly.

On the Temperatures T (273°+8°) considered as proportional
to the Molecular Volumes of Substances in the Gaseous State
at the Boiling-point S°.

We are accustomed to regard the absolute boiling-points
as quantities expressing degrees of heat, and this no doubt
they are; but they are at the same time proportional to the
molecular volumes of the vapours, analogous to those of the
liquids at the boiling-point (vs=a/d,). "When the vapour-
density D, is expressed in terms of the volume of a gram of
water at 4° C., the absolute boiling-point T (affected by a

lias me 5 da,
ES ah Sn :

Law of Density-Numbers. 197

constant coefficient) expresses the volume in cubic centimetres
of the molecular weight in grams.

In explaining this I may be permitted to restate a few
well-known facts.

Two grams of hydrogen at 760 millim., and at 0° C
occupy 22327 cubic centim.

a grams of any substance in the state of gas under the same
conditions of temperature and pressure also occupy 22327
cubic centim., both by convention and in accordance with
Avogadro’s law.

This constant volume 22327 cubic centim. is multiplied at

_ the boiling-point S° by the fraction T/273; and consequently
the volume (molecular) of a substance in vapour at its boiling-
point is

3

v,== 22327 T/273=81'8 T cubic centim.,
and the vapour-density is
D,=a/v,=a/81'8 T.

_ As the coefficient 81:8 is the same for all bodies, it may be
'___ omitted if it be thought fit to do so.

Relation between the Gaseous Molecular volume V, and the
Liquid Molecular volume vs.

When we take equal volumes (e.g. cubic centims.) of dif-
ferent liquids at 8° and convert them into vapour by heating,
we obtain a certain number of cubic centim. of vapour at
760: millim. and at 8°; this number is for any substance what-
eyer.

81:3 T/v, (cubic centim.).

In this connection I wish to draw attention to a special
case which appears of particular interest : it is as follows :—

We find that in certain groups of analogous substances

81-8 T/v,=constant.

This phenomenon is independent of the new law, and may be
expressed as follows :— :

~_

273
‘ -
(The constant 81°8 is omitted here.)
I shall call the constant v, the “reduced volume’”’ of the
substances in question.
- When a cubic centimetre at s° of the liquids (which have
the same v,) is volatilized, one obtains a definite volume of
vapour which is the same for all these substances, and is

= eonsiani=t.

~ vol. const. = 22327/v, cubic centim.

198

M. J. A. Groshans on the
Table XX. gives the value of v, for some groups of these

substances, and also the volumes (constant) of vapour.

TABLE XX.
Name Formula. v,. Ss’, Autho- | y,, | ¢.¢. of
° _|observed.) rity. vapour
oo ae CH, 11957! 108 | No. | 857| 260
Methyl benzoate ...| C,H,O, 151-46 | 199 Pe. 87°'7| 255
Methy] salicylate .... C,H, O, | 157°23 223 Kp. 86°5| 258
ea C,H,,0 | 10597} 35 93:9| 288
Ethyl carbonate C.H,,0, | 140:56 | 126 Kp. | 96:2] 232
Ethyl propionate ...| C;H,,0, | 12808 100 93°7| 238
Ethyl valerate ...... O71 ,0, |. 1763 143 Kp. 1158} 193
Ethyl succinate... C,H,,0, | 20971 | 217 Kp. | 1168} 191

No=Noad; Pe=Peligot; Kp=Kopp.

It is clear that the phenomenon of the constancy of vr
points to equality at the same time of the constants Tn/a and
v/a; whence it follows that

Bie ee
Sy = 7"
Here is another consequence of the above :—
Since it has been shown that the law of density numbers is

applicable to the volumes of vapour T and T’ of compounds,
it must also be applicable to the liquid volumes vs and v’,.

Application of the New Law to Volumes of Liquids at their
Boiling-points.

The properties of the liquid volumes are different from those
of the gaseous volumes. ‘These last, when of isomeric bodies,
can differ greatly; the former, on the other hand, are equal,
or almost exactly so. But it seems to me a little premature
to make any extended remarks on the subject of liquid
volumes in general.

If d, be the density of any liquid whatever at the boiling-

point S°, the volume v, = ~; and in applying the law to

ds”
these volumes we obtain the constants v,n/a analogous to
Tn/a (this constant vsn/a or ysB/a is the same as n/ds or
B/d,). Instead of writing v,n/a, I shall write k;, which is
more convenient. ‘The following Table contains three groups,
each of three members from Tables XLX. and XX. I have not
been able to form larger groups, comprising 8-10 compounds,
for the data relating to the volumes of liquids are too scanty.

‘Law of Density-Numbers. 199
TABLE XXI.
| Volume| Ob-
Names Formula q ”- | observed.|. servers. he,

_ 2 eee C,H,,0 74 £5 106°1 Kp. 21-51
Ethyl propionate .... C;H,,O, | 102 17 128°1 We. 21°34
Ethyl carbonate .../ C,H,,O, | 118 | 18 | 1388 | Kp. | 21-17
Walael -2....5...2.2.- C,H, 92 15 119°5 Lo. 19°49
Methyl benzoate .... C,H,O, 136 18 136°2 Kp. 19-89
Methyl salicylate ...| C,H,O, | 152 | 19 | 1522 | Kp. | 1963
Propylic oxide ...... C,H,,0 102 | 21 1516 | Za. 31°21
-Ethy! valerate ...... C,H,,O, | 180 23 1735 Kp 30°70
O,H,,O, | 174 26 209°0 Kp. 31:23

Ethyl succinate......

Kp.=Kopp; We.=Weger ; Lo.=Longuinine ; Za.= Zander.

Table XXII. contains several substitution-products re-
sulting from the action of chlorine, bromine, and iodine on
the alcohols and fatty acids.

Name.

Methyl alcohol ......
Methyl bromide ...
Methyl iodide

Ethyl alcohol
Ethy] chloride
Ethyl bromide
Ethyl iodide

Normal

a
Propyl bromide

Hermal.: 1.2/0.0.
| Propyl iodide
Normal

eeereecteseee

Amyl alcohol.........
Amyl chloride ......
Amy! bromide ......
Amyl iodide

TABLE XXII.
Ob-

Formula. | 4. | observed.| servers. | hs. |
m=2. | bas |
CHO ) 32 | 6 | 422| Kp. | vox |
CH.Be | 65 | 18 | 589.| - Pi. | 7-96 |
CH,I 142 | 18 Go) Fi | Sed
rT oad | |
C.H,O | 46 9 620 | Kp. | 1212 |
C,H,Cl | 645] 11 71-2 | Pi. 1214 |
C,H;Br | 109 | 16 (HA Ba 11-51 |
C,H,I | 156 | 21 86-1 | Pi. 11:59
m4. |
C,H,O 60 | 12 |
RES aR Ree 3 81:3 | Za. 16:28
sere ee ce Ea 82°7 Za. 16°59 |
C,H,Cl | 785] 14 |
gee a NaS Sane 916 | Za. 16-40 |
es ete Shy 93:8 | Za. 16-78
C,H,Br | 123 | 19
AC we cen es 974 | Za. | - 15-03 |
CRE RE Fee 99-4 | Za. 15°36 |
GH | 170. }, 24 |
Reni hh 8 eed oleae 1072 | Za. 15-13 |
bee Bd) .| 108-7 | Za. 15°34
m=6, ,
O20 }. 88 (| 18); 4944.1. Ke ) > 2a
C,H,,Cl | 106-3] 20 | 137-0 | Kp. 25°73
OH, Br | 151 | 25 | 1492 | Pi. 24-70 |
C;H,I | 198 | 30 | 1588 | Kp. 24:06

Kp.= Kopp; Pi.=Pierre; Za.=Zander.

200 M. J. A. Groshans on the

On the Variability of Physical Properties.

I have given the above name to the following phenomena:—

One frequently finds differences (according to Kopp reaching
sometimes as much as 2 per cent.) in the density of the same
substance at the same temperature. Similar differences are
to be observed in the fusing- and boiling-points. Such dif-
ferences present themselves when these physical properties
have been determined by different persons or by the same
observer at different times. ,

Space does not permit me to cite more than a very few
instances. ,

I shall begin with crotonic acid, C,H,O.; a=86, n=12.
According to the law of density-numbers, and in consonance
with analogy (which I need not further refer to), this acid
should have the value of Tn/a corresponding to x=5, that is,
the same constant as that of water.

There are three known isomers of this substance, which
differ in chemical properties, with which I need not trouble
my readers. One of these isomers, 8, or isocrotonic or
quartenylic acid, has the following physical properties ; it is
a liquid, not solidifying even at —15° C., and boiling at 171°
173°; from S°W5 is calculated 172°°5. There is thus complete
concordance between the calculated and observed boiling-

oint. |

With regard to the other isomers, their physical properties,
as given in Fehling, are as follows :— |

(1) a. acid—solid crotonic acid, tetracrylic acid : melting-
point 71°—72°, boiling-point 187° (corr.).

(2) Metacrylic acid: melting-point +16°, boiling-point
160°°5,

We have seen above that the 8 or iso-acid has a=5. For
its two isomers we find respectively e=5°330 and a=4°734;
the Beit value is x=5'038, which is essentially the same as
pH:

The particular fact that suggested to me the idea of the
mean value for # in various isomers was as follows :—

“The liquid acid, which does not solidify even at —15°, is.
changed, by heating to 180° in a sealed tube, into the solid
modification melting at 71°5 and boiling at 187°” *.

We may infer from this that the physical properties depend,
among other things, on the treatment to which the compound
has been subjected.

* Wurtz, Diet, Suppl. p. 561,

Law of Density-Numbers. 201

In this connection I shall take some other case from the
paper by Thorpe on the molecular volumes of liquids, Chem.
Soe. Journ. 1880.

Properties of the Compounds Br, and ICl.

‘The molecular weights of the above substances are respec-
tively 160, 162°5, being very nearly equal, and both have
B=18. We would expect to find their densities (at the boil-
ing-points) equal, but they differ by U1, or about 3°5 per
cent. With regard to their densities at 020, 5), these are
practically equal, notwithstanding the fact that their boiling-
points differ by 42°. The following are the data relating to
the above :—

—_ B. s°. a. d,. ds.
[Se ee 160 18 59:3 181 3°1883 2°9822
a Ee eee 162°5 18 101°5 2-22, 31822 2°8812

The values of x differ from the whole number 2: the one is
smaller, the other greater, the mean being z=2°01. Where
x==2 for each body the respective boiling-points would be
76°°5 and 81°-9, which lie close together ; and the two densi-
ties d, and d’, would have been equal. I have calculated from
the dilatations observed by Thorpe, what the densities would
be at the calculated boiling-points: they are respectively
279244 and 2°9404.

It appears, then, that there exist in the case of these sub-
stances certain unknown causes which affect the boiling-
points, raising the one 19°-4 and lowering the other 17° 5.

The existence of such an influence in very many cases will
be strange to no one, for it has been suspected by many. It
is the new Law, however, that points most clearly to it.

This also explains why the values of z, calculated from
observed boiling-points, do not always present themselves as
whole numbers, but are often affected by fractions, for
Instance :—

I have for valid reasons assumed that e=3 in the case of
the four following bodies (see above) :—

Peed p= S98. OH cate.
C,HSe. . 2=3:05. O,HiTe . . 2=3-27.

_ The experiments of Thorpe (/oc. ett.) furnish a another
example. i

202 —M. J. A. Groshans on the

Properties of the Substances 8,Cl, and SO.Cly.

These two bodies have the same molecular weight, the
density-number is the same,=12. The data are as follows:—

a. B. S° obs. x obs diy. ds.
Sob) eee 135 12 138°1 1°73 1°7094 1-4920
ere 185 le 69:9 1:20 1-7081 1:5602

Mean: .-, @=3l 46.

_ The boiling-points differ by 68°2, but the almost perfect
equality of the two densities dy) points to the equality of 2;
and we may admit that the true mean value is z=1°5, which
gives 110°-1 as the common boiling-point.

In the two sets of cases that we have examined above, it
appears that the disturbing cause does not operate till above
O°C., the values of d, are the same in each pair. But this is
not always the case: for instance, in two isomers examined by

Thorpe (loc. cit.),

a. B. S° obs. x obs. Aye ds.
DE 9) eee 99 14 83°5 3°28 1:2808 1:1563
Oris *. Sosa 99 14 59°9 2°86 1:2309 1:1092

Mean . . w=3'07. 1\/3=67°5.

Such a disturbing influence on the boiling-points and
densities of liquids is of a nature to hinder the progress of
the study of the physical properties of substances,

Compounds Liquid at the Ordinary Temperature.

The best method of applying the law of density-numbers
to liquids is perhaps that based, as above, on the comparison
of the volumes at the boiling-points ; but the data necessary
for this comparison are but seldom available. When this is the
case, we are still able in very many cases to verify the density-
numbers simply by comparing together the values of d, or d:
(t being the ordinary temperature 15°—20°C.) of analogous
bodies, when the values of /, will exhibit an agreement more
or less complete. 7

In Table XXIII. an example is given of the above. The
substances are arranged in the order of their boiling-points.
By thus employing corresponding temperatures (slightly
raised), it is possible to obtain satisfactory concordance in the
values of k. When t=0 the constants k; arrange them-

Law of Density-Numbers. 203

TABLE XXIII.

Formula. S°obs. B. dt. £2, hee:

hie ed 140 | 29 | 1-151 167 (1) | 25-20
6.11 Bbc. 1585 | 34 | 13244 | 15 (1) | 25:67
C,H,,0,As...| 167 | 32 | 1-224 0 (2) | 2614
O,H,,0,P ...| 191 |} 1-075 0 (3) | 2512
C,H,,0,P ...| 215 | 28 | 1-086 0 (4) | 2578
O,H,,0,4s.... 236 | 33 $264 | O (2) | 2488

(1) Landolt ; (2) Crafts ; (3) Williamson; (4) de Clermont.

selves in the contrary order to that of the points of ebul-

lition S°.

I have thought it interesting to compare with the above six
compounds another, the formula of which resembles those of
the bodies in the table: this is ((C,H;)3Sn).. Its properties

are as follows :—

S°obs. iB. dy. hy.

CypHaSng . . 237° (Ca.) 70 1°4115 (La.) 49°59.

Ca=Cahours; La=Ladenberg, who has found S=267°.
We have thus for

Casitiagans ° ° baa x 24°79,
C,H,,;0,As : > hgze 24°88,

A most complete agreement.

The value of k; for liquids appears to depend on the value
of p and g in the general formula C,H,O,, for, as seen
above, the atoms of O, 8, and of metals, appear to exercise no
influence. With regard to Cl, Br, I, they play the part of so
many atoms of hydrogen: for instance, ‘the three following
metallic ethers should be considered to contain 16 atoms of
hydrogen and not 15, for the constants /; are a little higher
than those of the compounds of As, P, &., given above.

Formula. S°obs. dt. ae B. Kt.
GT, CL Sp. 66. «<3 2 209 1:428 8 39 27°32
Cele Orbe 0.5.5 223 1:630 ? 44 26°98
3k AS 235 1°833 22 49 26°74

These are the experiments of Cahours.

I regret that space does not allow of my treating of cor-
responding temperatures and of the relative expansion of
liquids, or of the application of this new Law to solutions.

[ 204 ]

XXIV. Notices respecting New Books.

Diary of a Magnetic Survey of a Portion of the Dominion of Canada,
chiefly in the North-western Territories, executed wm the years
1842-44. By Lieutenant Lerroy, R.A., now General Sir J. H.
Leroy, C.B., K.C.M.G., FRS., &c.

HE late Sir J. Herschel, in an article on “ Terrestrial Magne-
tism,” which appeared in the ‘ Quarterly Review, June 1840,
strongly advocated the carrying out of Magnetic Surveys of the
Colonies. The late Gen. Sir E. Sabine, then Major R.E., in
November 1840 addressed a letter to Sir J. Herschel, as Chairman
of the Committee of Physics of the Royal Society, referring to the
‘Quarterly Review’ article, and containing proposals for a Mag-
netic Survey of the British Possessions in North America. The
Committee laid this, and other letters on the subject, before the
Council, with a strong recommendation that Government should be
urged to have the proposed Survey executed. The Lords of the
Treasury acceded to the proposal, and on July 20, 1842, Lieutenant,
now General, Sir J. H. Lefroy, left England ‘to take charge of
the Observatory at Toronto, tats the duty of carrying out the
Magnetic Survey annexed.

This duty was completed by the close of 1844, and the general
results only having been embodied in Sabine’s memoir “ Contri-
butions to Terrestrial Magnetism ” (Nos. 7 and 11, Phil. Trans. 1846,
1872), Sir J. H. Lefroy has been led by the renewed attention
directed to the Distribution and Periodical Changes of the Earth’s
Magnetism, to present the observations of his Survey with fuller
explanation and in a form more convenient for reference.

A Preface of pp. xxii gives the correspondence, out of which the
Survey originated, as well as other particulars ; concluding with some
remarks on the subject of local anomalies in Magnetic Elements,
reference being made, among other publications, to the recent
Report of the U.S. Coast Survey for 1881.

Part J., pp. 58, is devoted to an account of the Instruments
employed, and a general description of the observations made with
them, followed by a discussion of the probable errors involved and ~
other connected matters. To this succeeds the Diary of “a
journey which was specially laid out for scientific work and almost
excluded sport or excitement of any kind,” containing the records
of observations made from day to day between May 1843 and
November 1844. Starting from Montreal the observers’ route was
up the Ottawa and down the French River to Lake Huron, along
the N. shore of Lake Superior to Fort William, by way of Lake of
the Woods, Lake Winnipeg, and Hill River to York Factory on
Hudson’s Bay, thence by Lake Athabasca to Fort Chipewyan
(September 23), where the author “was warmly received by the late
Mr. Colin Campbell, and resided with him until March 4, 1844.
The arrival created a sensation in the small community, for it was
twelve years since any traveller had arrived by canoe from the

Geological Society. 205

south; and the self-importance of his voyageurs, who had thus
happily completed a journey of 3600 miles, was conspicuous in
their gala attire and gay chansons as the landing was approached.”
On the approach of spring the journey was resumed by way of
Great Slave Lake to Fort Simpson on the Mackenzie River. At
daybreak on March 28 the temperature sank as low as —41° F.,
and from the next day to May 25 Lieut. Lefroy remained at the
Fort, awaiting the break-up of the ice, when the boats started for
Fort Good Hope, lat. 66° 16’ N., long. 128°31' W. The return to
Lake Athabasca (30th June) was by the same route, but from
thence a détour was made up the Peace River to Fort Dunvegan, and
by the Lesser Slave Lake, to Edmonton, from which point the River
Saskatchewan was descended to Cumberland Harbour, reached on
August 29. Useful Tables, Charts, and an Index complete the
volume.

These records of field-work patiently and perseveringly carried
out (a kind of work comparable in the labour involved with that of
the Geodesic or Geological Surveyor in a new country) are not, of
course, light reading, but are sufficiently interspersed with entries
of difficulties encountered and overcome, risks run, physical and
other features of the route observed, to impart much general
interest to the narrative. The name of the author will now be
inseparably associated with the few who have devoted themselves
to build up the existing store of knowledge on the subject which
has been secured for Science; while “ Mount Lefroy,” round the
N.W. slopes of which the Canadian Pacific Railway will wind,
when completed, will worthily perpetuate that name on the scene
of those labours: ‘“ternumque locus . . . nomen habebit.”

XXV. Proceedings of Learned Societies.
GEOLOGICAL SOCIETY.
{ Continued from p. 74. ]
May 27, 1885.—Prof. T. G. Bonney, D.Sc., LL.D., F.R.S.,
President, in the Chair.

et following communications were read :—-

1. **On the so-called Diorite of Little Knott (Cumberland), with
further Remarks on the Occurrence of Picrites in Wales.” By Prof.
T. G. Bonney, D.Sc., LL.D., F.R.S., Pres. G.S.

The Little Knott rock and its microscopic structure were briefly
described by the late Mr. Clifton Ward, who named it a diorite, but
called attention to its abnormal character. The author gave some
additional particulars, and showed that although the rock varies in
different parts of the same outcrop, and is not one of the most
typical representatives of the picrite group, its relations on the
whole are with this rather than with the true diorites. He also
called attention to the extraordinary number of boulders which have
been furnished by this comparatively small outcrop, and discussed
the relation of their distribution to the former extension and effects

206. Geological Society :—

of ice in the Lake District. He briefly noticed the occurrence of
additional boulders of picrite in Anglesey, and described specimens
from two localities (Caemawr and Pengorphwysfa) where a similar
rock has been discovered in situ by Professor Hughes. Hence it is
probable that the Anglesey boulders are derived from localities in
that island, and not from Cumberland. From a re-examination of
specimens collected by the late Professor Sedgwick and Mr. Tawney,
preserved in the Woodwardian Museum at Cambridge, the author
showed that the rock must occur tn situ in two localities in the Lleyn |
peninsula, in the neighbourhood of Clynnog and of Aberdaron.
Lastly, he described a very remarkable picrite boulder, discovered
by Dr. Hicks, which rests on “‘ Dimetian” rock at Porthlisky near
St. Davids.

2. “Sketches of South-African Geology.—No. 2. A Sketch of the
Gold-Fields of the Transvaal, South Africa.” By W. H. Penning,
Ksq., F.G.S. |

The Gold-Fields of the Transvaal have been defined as covering
nearly all the eastern and northern districts of the State, though but.
a small portion of the area is productive. In this paper the author
described only the Lydenburg and De Kaap gold-fields, leaving those.
of Pretoria and Marabastadt for a future communication. The
auriferous region is known to extend 350 miles to the northward
beyond the Limpopo river, so that the gold-bearing rocks are found
throughout at least 73 degrees of latitude and 3 of longitude.

The area of the two gold-fields mentioned, comprising together
about 3000 square miles, was defined; and the Author, after noticing
some old gold-workings, proceeded to give an account of the physical
features of the country. He especially called attention to the cir-
cumstance that most of the rivers rise to the west of the highest
range, and flow eastward through it.

The oldest gold-bearing rocks consist of unfossiliferous schists,
shales, cherts, and quartzites, classed by the author as Siltrian.
Amongst these a great mass of ccarse granitic rock is intruded, con-
sisting of quartz and felspar, with but little, if any, mica. This
granite, in the De Kaap valley, forms an ellipse 17 miles long by
10 broad, with a narrow northerly prolongation. Both the granite
and the stratified rock are traversed by intrusive dykes, chiefly of
diorite.

These beds have been much disturbed, and then cut down, pro-
bably by marine denudation, to a level plain 1700 or 1800 feet above
the sea. Upon them rest unconformably a great sequence of con-
glomerates, sandstones, and shales, the ‘‘ Megaliesberg beds” of a
former paper, but now provisionally classed as Devonian. These
rocks also are traversed by dykes of diorite and other kinds of trap.
The ** High Veldt beds” overlie the ‘** Devonian” with some uncon-
formity.

Several sections and observations illustrative of these facts were
described, and details were given of the different gold-mines in each
of the great systems noticed, and also in alluvial deposits. It was’

*
3
.
{
3
y.
£

ae

Im ates q “

y*

EE

—

ae

oad

————_——_— .
%
*

ie Nin igang eRe Ne

™ shige AG cuit

=" Bute Gib: , m

”
MO ll si i __a =
i

pees _ a

On some Erratics in the Boulder-clay of Cheshire. 207

shown that much gold was derived originally from veins in the older
or Silurian rocks, and that some of that met with in the newer
system occurred in conglomerates or other detrital beds. But there
are also gold-bearing quartz-veins intersecting the latter.

3. “On some Erratics in the Boulder-clay of Cheshire &c., and
the.Conditions of Climate they denote.” By Charles Ricketts, M.D.,
F.G.S.

The Author stated that the glacial phenomena of the valley of
the Mersey indicate that the country has been entirely covered with
ice and snow, resulting solely from the snowfall on its water-slopes
and those of the tributary valleys. The glacial striz coincide in
direction with that of the respective valleys, or are in direct con-
nection with the contour of the ground. The bottoms of the valleys
are usually filled to some extent with irregularly stratified sands
and gravels, containing erratic pebbles from which all strie have
been removed, probably by currents of water holding sand in sus-
pension. Above these there is a boulder-clay containing a larger
proportion of sand and gravel than the boulder-clay proper. The
flanks of the valleys are covered with unstratified sand or fragments
of sandstone derived from the Trias, probably left by glaciers as
submarine moraines. The whole is overlain by the true boulder-
clay, an unstratified reddish-brown clay containing erratics derived
from different and distant localities. This clay originated in the

grinding-action of the glaciers upon the neighbouring rocks, and was

carried out in the form of mud by subglacial streams of water. The
contained pebbles, many of which are smoothed, flattened, scratched,
and striated, were carried by and dropped from icebergs and floating
ice ; they are so abundant as to indicate that the bay of Liverpool
was densely packed with ice.

The Author noticed the occurrence in these beds of masses of
contemporaneous sands, gravels, &c. caused by changes in the exten-
sion of the glaciers, and described a large series of erratics derived
from granitic, voleanic, Silurian, Carboniferous, and other rocks
covered with striz and other glacial markings, and also affording
evidence of subsequent exposure to weathering before they were
floated away and dropped into the clay. In connection with this
weathering of the boulders, the Author remarked that in the case of
the granitic and volcanic rocks the process differed greatly in degree,
extending in some granites to the separation of each individual
grain.throughout the whole mass, and he called attention to the
occurrence in Ireland of fragments of disintegrated granite and trap
imbedded in moraines, eskers, and Boulder-clay, and to that of
Wastdale-Crag granite similarly decomposed in the moraine in the
neighbourhood of Shap, where also rocks of volcanic origin have
become weathered in the same way as some in the Boulder-clay of
Cheshire. Fragments of limestone also show traces of erosion,while
others have been split into two or more pieces since their glaciation,
phenomena also observed in moraine-accumulations in limestone-
districts, Similar phenomena occur in the case of slaty and other

208. Geological Society :—.

stratified rocks. Some limestone pebbles have been puter by
Mollusca and other marine animals.

The inference drawn by the Author from the facts recorded in his
paper is that these weathered boulders once formed portions of
moraines on land from which, for a time, the glaciers had receded,
and that, after a succession of seasons sufficient to disintegrate these
blocks more or less, an increased snowfall caused such an extension
of the glaciers that the blocks were carried down to the sea and
conveyed away in icebergs and by floating ice to the spots where
they are now imbedded. As they occur at different horizons there
must have been a repetition of the advance and retreat of the glaciers
such as now occurs in Greenland.

June 10.—Prof. T.G. Bonney, D.Sc., LL.D., F.R.S., President,
in the Chair.

The following communications were read :—
1. “ Note on the Sternal Apparatus in Jguanodon.” By J. W.
Hulke, Esq., F.R.S., V.P.G.S.

2. “The Lower Palzozoic Rocks of the Neighbourhood of Haver-
fordwest.” By J. E. Marr, Esq., M.A., F.GS., and T. Roberts,
Esq., B.A., F.G.S.

The authors in this communication described the sequence of the
Lower Palozoic rocks lying to the north of Haverfordwest and
Narberth. Their work is founded on that published by the Geolo-
gical Surveyors in their maps, sections, and memoirs.

To the north of the ridge of rock running eastward from Rock
Castle, claimed as Shes by Dr. Hieua; they have discovered
Lingula Flags, with Olenus spinulosus, Wahl, and Agnostus spisi-
formis, Linn. These beds are seen underlain by conglomerate,
resting upon older rocks, near Trefgarn Bridge.

South of Dr. Hicks’s Archean ridge, a great fault brings beds of
Bala age in juxtaposition with the rocks of the ridge; hence, in the
tract described, no rocks of Tremadoc ard true Arenig age have been
met with. In the area south of the ridge the rocks are thrown
into a complex synclinal, with a complex anticlinal to the south-east,
near the town of Narberth.

The succession which the authors attempted to establish in this
area is as follows (in ascending sequence ) :—

1, Didymograptus shales, with Murchisoni-form Graptolites.

ii. Llandeilo limestone, with Asaphus tyrannus, Murch. &e.

il. Dicranograptus shales, with Dicranograptus, Climacograptus
bicornis, Diplograptus foliaceus, &c., having a zone at the summit
marked by the abundance of Orthis argentea, His.

iv. Robeston Wathen Limestone, with many corals and Brachio-
pods, and few Trilobites.

v. Trinucleus-seticornis beds, characterized by abundance of
Trinucleus setwornis, His., and its variety 7’. Bucklandt. These are
subdivided into three stages, viz. :—(a) Sholeshook limestone, with
Cystideans and an abundant Trilobite fauna, including Agnostus

Lower Paleozoic Rocks of Haverfordwest. 209°

trinodus, Salt, Trinucleus seticornis, His., Styyina latifrons, Portl.,
Phillipsia parabola, Barr., Cheirurus parvus, Salt., Encrinurus sex-
costatus, Salt., Phacops Brongmarti, Portl., &c.

(6) Redhill beds, blue-grey shales, generally poor in fossils, but
containing here and there a fair abundance, especially of Phacops
Brongniarti, Portl., and Trinucleus Bucklandi, Barr., and many
Lamellibranchs and Gasteropods.

(c) Slade beds, consisting of gritty green shales with calcareous
bands crowded with fossils. Glauconome disticha, Phyllopora Hisin-
gert, McCoy, Phacops Brongniarti, Portl., Trinucleus seticornis, His.,
Calymene trinucleina, Linn., Orthis testudinaria, Dalm., are abundant.
Climacograptus, sp., also occurs.

vi. Conglomerate, containing many quartz pebbles, succeeds the
beds of the Slade stage in many localities, and does not seem to
mark a great discordance, as the authors have nowhere found it
resting on lower beds.

vil. Lower Llandovery beds. Green gritty shales, with grit and
very fossiliferous calcareous bands, characterized especially by Nidu-
lites favus, Petraia subduplicata, var. crenulata, Stricklandinia
lirata, &c., and containing Phacops elegans, Boeck and Sars, Phacops
mucronatus, Ang., and Deiphon Forbesi, Barr.

The authors attempted a correlation of the Haverfordwest rocks
with those of other areas :—

Conglomerate and grit of Trefgarn =Harlech?

Lingula Flags of Trefgarn &c. = Dolgelly beds.
Didymograptus shales =Llanvirn.

Llandeilo Limestone = Lower Bala.
Dicranograptus shales = Lower and Middle Bala.
Robeston Wathen Limestone = Bala Limestone.
Trinucleus-seticornis beds = Upper Bala.
Conglomerate a. }
Fossiliferous Lower Llandovery beds } ie CES

In conclusion the authors notice two points which require further
elucidation. The first is the separation of the Robeston Wathen
and Sholeshook limestone. This is made on paleontological and
lithological grounds, as the whole thickness of the two limestones
had nowhere been met with in the same section. The second is the
relationship of the conglomerate to the fossiliferous Lower Llandovery
beds. In the only place where the two appear in juxtaposition the
conglomerate series appears to underlie the fossiliferous Llandovery
beds ; this the authors explained by a faulted overfold. These dit-
ficulties do not, however, prevent the authors from hoping that the
sequence they have established will be of use in assisting to deter-
mine the character of the very remarkable folds by which the district
has been affected.

3. “On certain Fossiliferous Nodules and Fragments of Hematite
(sometimes Magnetite) from the (so-called) Permian Breccias of Lei-
cestershire and South Derbyshire.” By W.S. Gresley, Esq., F.G.S.

In this paper the author described certain pebbles of hematite

Phil. Mag. 8. 5. Vol. 20. No. 123, August 1885. l*s

~ he 4 f
:

210 Geological Society :—

and magnetite which occur in the so-called Permian breccias on the
western margin of the Ashby-de-la-Zouch Coalfield. These pebbles,
which are largely collected for sale and used as ‘“‘ burnishers,” vary
in size from a diameter of ;4, inch to the size of a man’s fist. They
present many varieties of form, have sometimes an agate-like struc-
ture, and occasionally exhibit well-marked magnetic polarity.
Sometimes they show grooving and striation resembling those pro-
duced by ice-action, while at other times they seem to have been
crushed and recemented. Many of these pebbles contain fossils of
various kinds, chiefly plant- and insect-remains, but with a few of
Annelids, Mollusca and Fish. All the fossils are of Carboniferous
age.

From the consideration of all the facts detailed in the paper, the
author concluded that these nodules were originally composed of
clay ironstone, and that they were derived from Carboniferous strata.
He considered that the pseudomorphie action by which they have
acquired their present composition must have taken place in situ
since their inclusion in the breccia.

June 24.—Prof. T. G. Bonney, D.Sc., F.R.S., President,
in the Chair.

The following communications were read :—

1. “Supplementary Notes on the Deep Boring at Richmond,
Surrey.” By Prof. John W. Judd, F.R.S., Sec.G.S., and Collett
Homersham, Esq., F.G.S.

Since the author’s former communication to the Society on the
subject, this boring, in spite of the strenuous efforts made by the
Richmond Vestry and the contractors, Messrs. Docwra and Co., has
had to be abandoned, after reaching a total depth of 1447 feet from
the surface. This depth is 145 feet greater than that of any other
well in the London Basin, and, reckoning from Ordnance Datum,
reaches a lower level by 312 feet than any other well in the district.

Before the termination of the work temperature-observations
were obtained, which, generally, confirm those previously arrived at.

The strata in which the boring terminated consisted of the red
and variegated sandstones and marls previously described, which
were proved to the depth of 208 feet. Although it was demon-
strated that these beds have a dip of about 30°, complicated in
places by much false-bedding, no conclusive evidence could be
obtained concerning their geological age. They may be referred
either to some part of the Poikilitic series, or to the Carboniferous
(for similar strata have been found intercalated in the Carboniferous
series at Gayton, near Northampton), or they may be regarded as
of Old Red Sandstone age. !

Some interesting additional observations have been made since
the reading of the former paper, on the Cretaceous rocks passed
through in this well. Mr. W. Hill, F.G.S., of Hitchin, has found
the exact analogue of the curious conglomerated chalk met with at
a depth of 704 feet at Richmond. His observations entirely confirm

Igneous and Associated Rocks of the Breidden Hills. 211

: the conclusion that we have at this depth the ‘‘ Melbourne Rock”’
. with the zone of Belemnites plenus in a remanié condition at its
\ base. Some new facts concerning the state of preservation of the
fossils in the Chalk Marl are also recorded.

With respect to the conclusions arrived at by the author concern-
ing the distribution of the Jurassic rocks on the south side of the
London Basin, an important piece of confirmatory evidence has been
supplied by a deep boring made at the Dockyard-Extension Works
at Chatham. ‘This section, for the details of which the authors are
indebted to the officers of the Geological Survey, shows that under
the Chalk and Gault, with normal characters and thickness, there
lie 41 feet of sandy strata of Neocomian age, and that these are
directly underlain by blue clays of Middle Oxfordian age, as is
proved by the numerous fossils which they have yielded. We have
now, therefore, direct evidence of the existence and position of strata
of Lower, Middle, and Upper Oolite age, respectively, beneath the
Cretaceous rocks of the south-east of England.

2. * On the Igneous and Associated Rocks of the Breidden Hills
; in East Montgomeryshire and West Shropshire.” By W. W. Watts,
| Ksq., F.G.S.
He The author, in this paper, described the succession of rocks in the
} small tract near the Breidden Hills situated between Welshpool and
Shrewsbury. The Cambrian rocks are :—

(1) Criggion Shales, dark and barren, much penetrated by in-
| ee trusive diabases and about 2700 feet thick.
} (2) Andesitic lavas and ashes, followed by conglomerates of the
r same materials.
fe (3) Ashy grits and shales containing Climacograpsus antiquus ?,
C. bicornis ?, C. Scharenbergi, Cryptograpsus tricornis, Diplograpsus
foliaceus, Lepto grapsus flaccidus?, Beyrichia com plicata, Trinucleus
concentricus, Orthis testudinaria, Bellerophon bilobatus. The rocks
are thus of Bala age, the fossils indicating that the ashy grits and
shales are on the horizon of the top of the Glenkiln or bottom of
the Hartfell series.

These are followed by Silurian strata.

REE NE ae “=
ieee > =| 1 - 4
soit the ict Kress LN = NE athty
ATA ; :

(5) Lower Ludlow Shale. M. colonus, M. Nilssoni, M. Salweyt,
M. lintwardenensis.
The paper concluded with microscopical descriptions of the igneous

rocks, of which there are two sets :—
(1) An older set interbedded with the Cambrian and consisting of -
andesites bearing a large percentage of a mineral allied to enstatite,

(1) Pentamerus-beds. Soft sandstones and mudstones yielding

Pentamerus globosus ?, P. oblongus, P. undatus, Leptena transver-
| salis, Strophomena rhomboidalis, Petraia subduplicata.
ie (2) Purple shales, unfossiliferous.
He (3) Lower Wenlock Shale, with Monograptus vomerinus?, Crypto-
¢ grapsus, Sp. M. priodon, var. Flemingt. These graduate into: a
3 (4) Upper Wenlock Shale, with M. priodon, M. vomerinus?,
i: M. basilicus, M. Nilsson, M. Haare

212 3 Geological Society.

together with augite and a small quantity of hornblende and mica.
These are chiefly lavas, but some few are perhaps intrusive rocks
and dykes.

(2) Intrusive rocks of a diabase type, generally, however, contain-
ing a variety of enstatite identical with that in the andesites. These
are intrusive in the Cambrian rocks, and from their relations appear
to be most probably of Post-Silurian age.

3. “Note on the Zoological Position of the Genus Mierochoerus,.
Wood, and its apparent Identity with Hyopsodus, Leidy.” By R.
Lydekker, Esq., B.A., F.G.S.

4. **Observations on some imperfectly known Madreporaria from
the Cretaceous formation of England.” By R. F. Tomes, Esq., F.G.S.

5. “Correlations of the Curiosity-Shop Beds, Canterbury, New
Zealand.” By Capt. F. W. Hutton, F.G.S.

The ‘Curiosity Shop” is a locality on the River Rakaia in the
Canterbury Plains, and has been thus named on account of the
numerous fossils found in some calcareous sandstones cut through
by the river. The section exposed consists of

. River-gravels.

. Loose grey quartz sands.

. Soft calcareous sandstone with glauconite, passing downwards
into tufaceous clay.

. Calcareous sandstone without glauconite.

. Loose grey or yellowish brown sands.

By Mr. McKay, of the Geological Survey, No. 2 had been referred
to the Pareora series (Miocene ?), No. 3 to the Upper Eocene series,
and Nos. 4 and 5 to the Cretaceo-Tertiary series. The author, who
was inclined to class all these beds in a single series, pointed out
that the only difference between the fossils found in Nos. 3 and 4,
the most important fossiliferous beds, consisted in the presence of a
greater number of forms in No. 3, all found in No. 4 being identical
with those in the overlying bed. He then gave a complete list of
the species of Vertebrata, Mollusca, Brachiopoda, Echinodermata,
Bryozoa, and Coelenterata, from the locality, 48 in all, and compared
them with those from the Weka-pass stone, 26 in number, and the
Ototara fossils from Oamaru, to show that a large proportion were
identical. He gave reasons for not agreeing with the views of Dr.
Hestor and Mr. McKay, who held that unconformity exists between
the beds referred by them at the Curiosity Shop, in the Weka-pass
district, and north of Otago, to the Upper Eocene and Cretaceo-
Tertiary series respectively, and showed both from paleontological
and stratigraphical data that all these rocks must be included in
one system, the Oamaru system of Dr. von Haast and himself.

Onore

Or

6. “On the Fossil Flora of Sagor in Carniola.” By Constantin,
Baron yon Ettingshausen, F.C.G.S.

ae ee

XXVI. Intelligence and Miscellaneous Articles.

ON HYGROMETRY. BY M. JAMIN.

N MASCART publishes every vear, in the Annales du Bureau

* metéorologique de France, observations made in nearly one
hundred stations all over France, and the same is the case in all
European States. I have taken from them the pressure of the air
at the base and on the summit of the Puy de Dome, and I now
desire to make a general remark on hygrometry.

On looking at the tables of relatwe moisture, it is surprising to
see them so uniform. The means are almost the same in each
moment and at all stations. Thus we find at Clermont Ferrand,
at midday in the year 1880,

February. June. August. September. October.
593 599 570 569 622

It is, however, clear that there are great differences in the hy-
grometric conditions of the months of February and August, and
that if these are not brought out in the tables it is, probably, that
the system adopted in reducing the observations is erroneous.

This system consists in expressing the ratio i of the elastic

FE ?
force observed f to the maximum force F, which air would have
at the same temperature if the air were saturated ; this is what is
called the relative moisture. But for any given air of constant

composition the quotient F varies : Ist, with the proportion of

vapour ; 2nd, with the height and the barometric pressure, since f is
proportional to this pressure ; 3rd, and more particularly, with the
temperature which changes the value of F; it is then a function
of three independent variables; and we may hope that it brings
out the variations in the quantity of vapour. We must for this
eliminate the disturbing influences of pressure, of height, and of
temperature, but that can easily be done.

When chemists analyze air, they determine the quantities of
oxygen, nitrogen, and carbonic acid; to complete the analysis, it
would be logical to add the proportion of aqueous vapour. As
this vapour is a gas subject to the same laws of compression and
expansion as other gases, there is no reason to measure it other-
wise.

Let f be the tension of vapour, H the total pressure of the
atmosphere, H—/ that of the dry air, we have

. v(1°293)(0°622)f
Weight of vapour P = Taomio0.
. . a, v(1°:298)(H—
Weight of dry air P= TT rag

ee ie

214 Intelligence and Miscellaneous Articles.

= is accordingly the ratio of the weight of vapour to that of dry
air; it is independent of pressure and temperature, for f and H—f
follow the same laws; it expresses the hygrometric richness in

weight, and nee measures its volume.

It may be remarked that the observations do not directly give
the ratio ordinarily used, 4 In fact, the condensing hygro-
meter measures f, which is a function of the pressure and there-

fore of the height; the psychrometer determines ei ; In order

to pass from the measurement made to the function 4 a calcula-

tion is made which serves no further purpose. By introducing
the denominator EF’, we just introduce into the tables disturbances
arising from the temperature which mask the influence of the
vapour and complicate the result. It is no more difficult to

calculate and keep the result, than to calculate and retain

ited
, this would be replacing a complicated function, from which

nothing can be deduced by an exact gravimetrical or volumetrical
measurement of the hygrometric composition of the air.

I propose therefore to suppress the relative moisture . in me-

teorological tables, and to replace it by the hygrometric richness

a In order to justify this substitution, 1 may show the

effect by an example. I take the determinations made b
M. Alluard at Clermont Ferrand in 1880, and published in
the Annales Météorologiques. No very appreciable differences are
seen in the different months ; the numbers decrease from the morn-
ing until 3 p.m., which shows the influence of temperature as
by M. Angot. Nothing in it indicates change of moisture.

Clermont.— Relative Moisture.

Intelligence and Miscellaneous Articles. 215
The following Table is the above modified, in which we have

3

replaced 7 by as In order to avoid decimals, the results have
been multiplied by 100,000.

October ... Base aes

Base ......
November es ath

- F
6 ) 12. 3 6 9. ba 7

Baa Ce 410 | 444| 526! 560] 459] 459} 380

ery -. Aneatan | 473 | 395| 489] 458| 487 | 446] 514
Base 606 | 660 | 7271 715 | 727 | 696 | 646
Earanry: - t deceit ..| 609 | 607] 671] 657| 626] 626] 617
het os ee 635 | 770| 8621 7491 7951 8141] 666
Summit .... 663 | 676| 7151} 733 712 688 179

ae Be Bernat 842 | 93821 946] 945] 9671 8981 876
Pe es Summit 740 | 956| 7911 807 | 775 | 53851 687
Ma cr 981 | 1019 | 1001 | 1012 | 979} 969] 917
ee Summit = 831 | 868] 877] 883] 868] 830] 823
iio ize 1248 | 1268 | 1291 | 1290 | 1278 | 1497 | 1106
eae Summit 973 | 1036 | 1099 | 1130 | 1094 | 1023; 944

Jul a Barr: 1534 | 1607 } 1566 | 1567 | 1668 | 1599 | 1377
that Summit 1240 | 1304 | 1885 | 1427 | 1402 | 1380 | 1854
ow i a 1482 | 1588 | 1581 | 1611 | 1808 | 1663 | 1318
) Summit 1267 | 1348 | 1883 | 1428 | 1426 | 1452 | 1250
Base <2): 1269 | 1472 | 1460 | 1485 | 1499 | 1418 | 1137
September } gammit .... 1129 | 1173 | 1128 | 1264 | 1275 | 1263 | 1125

It will be at once seen from the last Table—

(1) That the hygrometric richness increases from the morning
towards midday, or 3 o'clock, and then diminishes as the sun sinks
and during the night ; this arises from two causes, evaporation
during the day and the expansion of the air.

(2) That the richness increases from January to July and
August, and then decreases; that it varies from 0:005 to 0°018,
that is to say, it is three or four times as much in summer as in
winter. We find in like manner that it increases in hot countries,
that it is greater even at Laghouat than in Marseilles; and we

216 Intelligence and Miscellaneous Articles.

must conclude therefore, with Dove, that there is less dry air in
summer than in winter in the northern hemisphere.

(3) The hygrometric capacity of the air, that is the maximum |

vapour which it can contain, 1s expressed by a=3% But accord-

ing as we ascend, the temperature and F diminish, while on the
other hand H becomes less also. There are two inverse causes
of variation; experiment proves that the capacity decreases but
very slowly. In like manner the richness is always a little less on
the summit than at the base of the Puy de Dome.

(4) The last column of the table measures the total capacity
during the night when the temperature is lowest. It is in general
greater than the richness at 6 a.m.; but it is sometimes smaller,
from which it may be inferred that the air is saturated at any
height when its temperature is a minimum. This explains all the
condensations of steam which generally take place in the night.

In fine, the values of i only tell us the relative degree of dry-

ness or moisture; they do not measure the quantity of vapour.

On the other hand, the ratio aS gives the hygrometrie compo-
sition of the air; it shows the changes which take place at day
and night, in summer and winter, at various altitudes and in
different latitudes.—Journal de Physique, vol. iii. p. 469.

ON THE DEFORMATION OF THE LUMINOUS WAVE-SURFACE IN
THE MAGNETIC FIELD. BY E. VON FLEISCHL.

The results of this investigation are summed up by the author as
follows :—

1. That the wave-surface in the magnetic field has undoubtediy
in general the form described by me of two surfaces of rotation
which intersect one another.

-2.. That the form of these two surfaces differs to an extremely
small extent from the spherical form; but that notwithstanding

3. The spherical form of both parts of the double surface which
results from Verdet’s law of the cosine, and which with this law is
implicitly assumed in all previous inquiries, is only a faint approxi-
mation to the trueform. And

4. That the form of the wave-surface in the magnetic field is a
double surface made up of two ellipsoids of rotation, which are so
displaced in the direction of their greatest axes that they nearly
coincide with each other.— Wiedemann’s Annalen, No. 6, 1885.

Intelligence and Miscellaneous Articles, ye

ON A NEW METHOD FOR DETERMINING THE MECHANICAL EQUI-
VALENT OF HEAT. BY A. G. WEBSTER.

In 1867 Joule published the results of his experiments for deter-
mining the mechanical equivalent of heat, by means of observations
on the thermal effect of an electric current. In his experiments a
calorimeter was used holding over a gallon of water, the tempera-
ture of which was taken by a thermometcr. The method about to
be described differs from Joule’s in that the temperature is measured
by the change of resistance of a wire, which is heated by a current,
and no water is employed. The idea of the method was suggested
by Professor John Trowbridge. Accuracy is not claimed for the
results which follow, as the experiments were undertaken only with
the view of ascertaining the practicability of the method.

The method of conducting the experiments was as follows :—A
thin ribbon of steel, about 45 cm. in length and 1 mm. in breadth,
and weighing *23 gr., was included in one side of a Wheatstone’s
bridge, by which its resistance was measured. It was then thrown

into another circuit, and a transient current from twelve large

Bunsen cells was passed through it. The quantity of electricity
transmitted was measured by a ballistic galvanometer, and the
difference of potential of the ends of the steel strip was compared
with the electromotive force of a Daniell’s cell by means of a
quadrant-electrometer. The rise in temperature of the steel was
found by immediately measuring its resistance again. It had been
previously found, by a series of experiments made between the
temperatures of 90° and 10° C., that the resistance of the steel
used was represented by the equation

R=a (1+ 00503 6),

6 being the temperature.
If then R, be the initial resistance of the strip, and R, the resist-
ance after the passage of the current,
R,=a (1 +B 4,),
R,=a (1+ 4,),
and the rise in temperature is
; R, 7 R,
ap
If w be the weight of the strip, and s its specific heat, the quantity
of heat imparted to it by the current is
ws (Rk, —B,)
a 2 |
But if Q is the quantity of electricity transmitted, and E the
Phil. Mag. 8. 5. Vol. 20. No. 123. August 1885. Q

0,—0,=

h=ws (0,—9,) tt es a (1)

218 Intelligence and Miscellaneous Articles.

difference of potential between the ends of the strip,
Jh=Q E,

where J is the mechanical equivalent of heat. We have

Eee sin = cs baled S cee (2)

where a is the first swing of the needle of the ballistic galvanometer,
G the galvanometer constant, T' the period of a single vibration of
the needle, and H the horizontal component of the earth’s magnetic
force. G was determined by comparison of the deflections on the
scale of the ballistic galvanometer with the readings of a tangent-
galvanometer whose constant was calculated, included in the same

. « « Pre ob . .
circuit. In the experiments, 2 sin 5 Was considered as proportional

to 6, the deflection on the scale, and the value of G for 6=1 em.
was found to be 769°4. A shunt was used with the galvanometer,

so that the value of Q above given is to be multiplied by =

r being the resistance of the galvanometer, and 8 that of the gee
The arrangement of the apparatus was as follows :—

R. The steel strip enclosed in a glass tube to protect it from
draughts of air.

W. Wheatstone’s bridge.

G,. Thompson astatic galvanometer.

S,. Shunt for the same.

G,. Ballistic galvanometer.

S,. Shunt for the same.

E. Quadrant-electrometer.

B,. Battery of twelve Bunsen cells.

B,. Battery of two Leclanché cells.

B;.- Do. do. do.

K,. Key for battery B, and galvanometer G.,.

K,. Key for passing current from B, through strip.
K,. Key in auxiliary circuit with commutator C, and a second

coil of galvanometer G,, for bringing the needle quickly to
rest without heating strip h.

The two galvanometers were arranged to throw their spots of
light on the same scale. The key K, was first depressed, R being
then in the bridge circuit, and the spot of G, was brought to zero
by adjusting the resistance c. a was always 1000 ohms, and 6
one ohm. On K, being raised, a suflicient extra resistance was
inserted in c, so that when K, was momentarily depressed, and K,
was immediately afterwards “again depressed, the spot G, did not
move. ‘The hand soon became accustomed to pressing ‘i just
long enough to accomplish this result. In the experiments, ¢ had
to be increased from 1167 ohms by the amount of 50 ohms, and the
temperature of the strip accordingly rose about ten degrees. As

ORR Nt BONY RA

Intelligence and Miscellaneous Articles. 219

_ the resistance was measured almost simultaneously with the passage

of the current, the rise in temperature could be very exactly
known, and the effect of radiation could be very easily determined.

sys

Combining equations (1) and (2), we have
ya Bh 5 (r+S)a Bp
Ga Sws(R,—R,)

HE, as measured by the electrometer, was about one volt, =10°
C.G.8S. units. H was ‘171; 7, the resistance of the ballistic
galyanometer, was 3296 ohms; S8,, the shunt, was 1025 ohms; T,
the time of a single vibration of the needle, was 12°6 sec.; a, the
resistance of the strip at 0°, was 1:072 ohm; £ was ‘00503, the
weight of the strip was ‘230 gr.; its specific heat, ‘114; the gain
in resistance of the strip was ‘05 ohm, and 6, from 20 experiments,
was 26°8 cm.; G was 769°4 for é=1. |

i 10° x 171 x 12°6 x 26°38 x 4°321 x 1:072 x 00503
769°4 7°05 x 1:025 x 230 x°114
= 4:14 x 10’ ergs per gram-degree. = rad
In.Joule’s experiments, the process of heating was continued for
nearly an hour, whereas here it lasts for less than a second. In

the former method, it was necessary that the current should remain

sensibly constant throughout the experiment, and the calorimeter
was radiating heat throughout that time. In the short time

required by the latter method, the radiation must be. very small,

and the error from the inconstancy of the curent is avoided. -I

220 Intelligence and Miscellaneous Articles.

intend to undertake a further course of experiments in order to
obtain an accurate determination, the purpose of the present paper
being merely to show the method.—Proceedings of the American
Academy of Arts and Sciences, May 26, 1885.

NEW FORM OF HYGROMETER. BY M. BOURBOUZE.

When vapour begins to condense on a glass plate interposed
between the observer and the source of light, concentric rings begin
to appear at this point. They are like those seen round the cloud
in foggy weather.

The apparatus we have constructed to produce these rings con-
sists of a small rectangular tube, with small holes on opposite faces
closed by very thin glass plates. A very delicate thermometer is
placed so as to dip in only a small quantity of the liquid. When
a current of air is produced in the liquid either by blowing or by
aspirating, a deposit of dew rapidly forms on the glass. If the
apparatus is placed between the eye and a point of light, concentric
rings are seen about this point, red outside and violet within. The
appearance of the rings as well as the reading of the thermometer
may be made from a distance by a telescope.—Comptes Rendus,
June 22, 1885.

ON THE DISENGAGEMENT OF HEAT IN THE SWELLING AND
SOLUTION OF COLLOIDS. BY EK. WIEDEMANN AND (CH.
LUDEXING.

The conclusions of this research are as follows :—

The solution of the dry colloid is made up of two processes :—
(1) hydration, which is attended with disengagement of heat;
and (2) solution, which is attended with absorption of heat.—
Wiedemann’s Annalen, No. 6, 1885. |

ON A MERCURY-GALVANOMETER. BY G. LIPPMANN.

A mercury-manometer is placed between the poles of a fixed
magnet, which lie right and left of its horizontal arm. The current
traverses the latter in a vertical direction parallel to the axis of the
tube, by which the mercury rises in one limb.

If i is the strength of the current, H the strength of the mag-
netic field, 7 the length of a small rectangular parallelopipedon in
the direction of the current, e its thickness in the direction of the
lines of force, the force which displaces the parallelopipedon is Hh,

Hi Hi
and the pressure p= Toe Boer
€

Hence the sensitiveness increases with the strength of the mag-
netic field and the thinness of the column of mercury. Accordingly
the lower bend of the manometer is replaced by a chamber filled
‘with mercury at right angles to the current, and only 4, millim. in
thickness. —Comptes [endus, 1884, pp. 1256-57; Berblatter der
Physik, No. 6, 1885. We

= “ RG his +t rage PHS : a % ee ' os ii nd

THE
LONDON, EDINBURGH, ayn DUBLIN

PHILOSOPHICAL MAGAZINE

AND

JOURNAL OF SCIENCE.

22 ee eee

[FIFTH SERIES.]

SHPTEMBER 1885.

XXVIII. Problems on the Distribution of Electric Currents in
Networks of Conductors treated by the Method of Mazwell.
By J. A. Furemine, M.A., D.Sc. (Lond.), Fellow of St.
John's College, Cambridge, Professor of Electrical Tech-
nology in University College, London*.

[Plates VI. & VII. |

§1 : ia any number of points in a plane be joined together

by linear conductors such as metallic wires, we
have an arrangement of conductors which is called a N etwork.
If at any point in the network a current of electricity be
allowed to flow in and is drained off at some other point by
conductors, called respectively the anode and kathode con-
ductor, then, after a short period, depending on the self and
mutual induction coefficients of the various conductors, the
total quantity of electricity arriving by the anode will distri-
bute itself throughout the network and setile down into a

_ steady flow. When this is the case there is a certain definite
_ difference of potential between the anode or source-point and
_ the kathode or sink-point, and there is also a certain definite
_ and constant strength of current in the anode conductor and
_ in every mesh or branch of the network. Call « and y the
_ potentials of these source- and sink-points, and # the strength
_ of the current in the anode lead, that is the whole quantity
_ of electricity flowing per second ‘through the network, then

(y—a)/wx measures the resistance of the network. We can

F imagine the network replaced by a single linear conductor or
_ wire of such sort that if the anode and kathode conductors

* Communicated by the Physical Society: read June 27, 1885.
Phil. Mag. 8. 5. Vol. 20. No. 124. Sept. 1885. R

222 Dr, J. A. Fleming on the Distribution of

are applied to its ends, the difference of potentials at the ends
of this simple conductor and the strength of the current flow-
ing through it have the same numerical values y, a, and 2.
The resistance of this single conductor is then the same as
that of the complex network. |

The resistance of the network is obviously some function
of the resistances of the separate conductors or wires which
compose it, and is capable of being calculated from them.
Experimentally, the resistance of a complicated network would
best be determined by the measurement of the current-
strength in the anode lead and the difference of potential
between the source and the sink. Theoretically, it is in-
teresting to examine the law of distribution of currents in a
network, and to reduce to a function of the separate resist-
ances the total resistance of the whole network between any
two points. 4 oe

§ 2. In his larger Treatise on Electricity, Clerk Maxwell
has treated the general case to determine the differences of
potentials and the currents in a linear system of m points con-

nected together in pairs by : n(n—1) linear conductors”, and

has shown how to form the linear equations, the solution of
which gives the condition of the network when given electro-
motive forces acting along some or all of the branches have
established steady currents in them.

The usual method of obtaining a solution for the distribu-
tion of currents is the application of Ohm’s law round the
everal currents of the network, controlled by the condition
of continuity that there is no creation nor destruction of elec-
tricity at the junctions. |

Since the publication of the first edition of his Treatise,
Maxwell reduced these two sets of equations to one set by
the simple device of regarding the real currents in the meshes
of the network as the differences of imaginary currents round
each cycle or mesh of the network, all directed in the same
direction, and thus obtained by the application of Ohm’s law
a single set of linear equations, the solution of which gives
the required currents in each branch. Maxwell’s method is
as follows t:—If we have p points in space and join them
together by lines, the least number of lines which will con-

: Seana on Electricity and Magnetism,’ 2nd edition, Vol. i. § 280
and § 347.

+ This method was first given by Clerk Maxwell in his last course of
University lectures. It is alluded to in the second edition of his larger
Treatise and in the Appendix of his smaller Treatise by their respective
editors, Mr. W. D. Niven and Professor Garnett, to whom it was com-
municated by the present writer.

Electric Currents in Networks of Conductors. 223

nect all the points together is p—1. If we add one line more
we make a closed circuit somewhere in the system ; that is
to say, a portion of space is enclosed and forms a cell cycle
or mesh. very fresh line added then makes a fresh mesh;
and hence if there are / lines altogether joining p points, the

number of cycles or cells will be k=/—(p—1). Now let

such a system of points and lines represent conducting wires
joining fixed points, and forming a conducting network. Let
a symbol be affixed to each point which represents the elec-
trical potential at that point, and also a symbol affixed to
each line representing the electrical resistance of the con-
ductor represented by it. In such a diagram of conductors
the form is a matter of indifference so long as the connections
are not disturbed and lines are not made to cross unless the
conductors they represent are in contact at that point.

Consider a network, Pl. VI. fig. 1, formed by joining nine
points by thirteen conductors. Then there will be 13—(9—1)
=5 cycles or cells. Now let an electromotive force E act in
one branch B, and give rise to a distribution of currents in the
network. Let a, 8, y, 6, &e. represent the potentials at the
points, and A, B, C, D, &c. the electrical resistances of the
conductors joining these points, and imagine that round each
cycle or circuit an imaginary current flows, all such currents
flowing in the same direction.

A circuit is considered to be circumnavigated positively
when you walk or go round it so as to keep the boundary on
your right hand. Hence, going round an area A in the di-
rection of the arrow is positive as regards the inside if you
walk inside the boundary-line, and negative as regards ex-
ternal space B if you walk in the same direction round the
outside. We shall consider a current, then, as positive when
it flows round a cycle in the opposite direction to the hands
of awatch. Returning then to the network, we consider that
round each cycle flows an imaginary current in the positive
direction. The real currents in the conductors are the
differences of these in adjacent cycles or meshes, and the
imaginary currents will necessarily fulfil the condition of con-
tinuity, because any point is merely a place through which
imaginary currents flow, and at which therefore there can be
no accumulation nor disappearance of electricity.

Let x, y, z, &c. denote these imaginary like-directed cur-
rents. Then x#—y denotes the real current in the branch I,
and similarly e—z that in branch H. Then a, y, z, &c. may
be called the cyclic symbols of these areas. The cyclic symbol
of external space is taken as zero; hence the real current in
branch B is simply zw.

Let an electromotor act on the branch B, bringing into

~

224 Dr. J. A. Fleming on the Distribution of

existence an electromotive force in that branch. Let the
internal resistance of the electromotor be included in the
quantity B, representing the resistance of the branch A. Then
apply Ohm’s law to the cycle « formed by the conductors B,
I, H; we have E—Be=y—z«.

x is the actual current in this case flowing in the resistance B,
and the potential at the ends of B is equal to the effective
electromotive force acting in it less the product of the resist-
ance of the conductor multiplied by the current flowing in it.
For the conductor I we have similarly
y—B=(#—y)L.

Hence w—y represents the actual current in I: it is the dif-
ference of the imaginary currents flowing round the w and y
cycles in the positive direction. And for the conductor H

we have also B—a=(«—z)H.
Add together these three equations,
K= v= 2 -|- De,

O=B8—y+(2-y)L,
=a—B+(ae—z)H ;
and we have, as the result of going round the cycle x formed
of conductors B, I, and H,
H=a2(B+I1+H)—yI-—-H. ... . Q)
a, 8, y have disappeared in virtue of these opposite signs.

This equation (1) is called the equation of the w cycle; and
we see that it is formed by writing as coefficient of the cyclic
symbol wz the sum of all the resistances which bound that cycle,
and subtracting the cyclic symbol of each neighbouring cycle
multiplied respectively by the common bounding resistance
as coefficient, and equating this result to the effective electro-
motive force acting in the cycle, written as positive or nega-
tive according as it acts with or against the imaginary current
in the cycle. This is Maxwell’s rule.

Since there are & cycles or meshes we can in this way form
k independent equations, and by the solution of these deter-
mine the & independent variables, x, y, z, &. The value of
the current in any branch is then obtained by simply taking
the difference of these variables belonging to the adjacent
meshes, of which the conductor or branch considered is the
common boundary.

§ 3. Let us now consider the most general case possible, in
which we have a network composed of linear conductors suf-
ficiently far apart to have no sensible mutual induction, and
let there be electromotive forces acting in each branch or

Electric Currents in Networks of Conductors. 225

_ conductor. Let the system be considered to have arrived at

the steady condition. Let w, y, z, &c. be the cyclic symbols

or measure of the imaginary current circulating counter-

clockwise round each mesh. Let A, B, O, &e. (fig. 3) be the

resistances, and ¢, é, €3, &c. the electromotive forces acting

In each branch. These are reckoned positive when they tend

to force a current round the mesh counterclockwise, and

negative when they act in the opposite direction. Then the

equation to the « cycle will be
“w(A+J +L)—yJ +O0z+ Out Ow=e,.

The symbols of all the cycles are written down, putting in

those of z, vw, and w with zero coefficients, as they are not

adjacent cycles to that of w. We shall have five equations
similar to the above for the other cycles, y, z, w, and w.

Now it can very simply be shown from the theory of deter-
minants, that if there are » linear equations of the type

Sepa Oe Ge a es BSP),
| O12, + bois + = ° . ° : bn Ln = P2
j ° ° °
(3 E ;

kya -- Ko®> + : - : ° : - k,, Ly aa ee

the solution for any variable x, is the quotient of the deter-
minants

| P1 As e - = 2 a i
| P2 bs ° ° ° ° b,
| P n hy kin
N=
a ag mG)
by de by |
ky ky Kn |

The only difference between the numerator and denominator
is that the solution for 2 is given by writing as numerator the
determinant of the n equations having the column p,, po... p”
substituted for its nth column, and then writing down as
denominator the determinant of the m equations simply.

226 Dr. J. A. Fleming on the Distribution of

Thus, for example, the solution of the three linear equations
ax +by +cz =d,
autby+ez=dy,
gt + bey + 6.2 = de,

is
die Diatee
dp Oy 4
| ee
L =. es GO oe
| i Ws ee |
| ty: te, |
ee aad ig eae

with similar expressions for y and z, differing only in having
as numerators respectively

a she 16 ub ae
a Oa es and ad, 0; “ae
| ag dy C9 9 bs dy

denominator being the same.

In this case the evaluation of these determinants is easy: a
simple symmetrical process of taking products, according
to the rule,

Vee CMe, 6
de f | =(aei+b/g+cdh)—(ecg+bdi+afh).
Gera

§ 4. The properties of determinants enable us, however,
very easily to evaluate a numerical determinant of any order.
The process consists in the gradual reduction of the determi-
nant in order by such transformations as will render all the
elements of the first row or column zero except the first. The
determinant is then reduced to the product of its leading
elements and the corresponding minor. <A repetition of this
lowers the determinant one degree at each stage ; and finally,
when it is resolved into a numerical two-row determinant, a
simple cross multiplication gives its value.

The process of evaluation of a numerical determinant is
dependent on four principles :—

(1) That the value of a determinant is not altered if rows
are changed into columns.

(2) The interchange of two rows or two columns reverses
the sign of the determinant.

Electrie Currents in Networks of Conductors. 227

(3) Ifevery constituent in any row or column be multiplied
by the same factor, then the determinant is multiplied by that
factor.

(4) A determinant is not altered if we add to each consti-
tuent of any row or column the corresponding constituents of
any other row or column multiplied respectively by an iden-
tical factor, positive or negative.

For example, suppose that the solution of a series of network
equations with numerical coefficients of resistance yield the
determinant

we proceed to operate on this as follows :—Subtract the
second column from the first and write the remainder. Asa
new first column we get

Oe ala 46
—1 8 9 2
Lito 3
Shier anid ar

Subtract the third row from the first and put the remainder
as a new first row. Also add the third row to the second for
a new second row, and we get

1 2-3 38
Ge eer dD
1 neues Seah: Sams
E ieey Meee ane

Again, subtract the first row from the third for a new third,
and subtract three times the first row from the fourth row for
a new fourth row, and we have

1 2-3 3
931-9 phostao
O—1Ie 20
O 1 14-2
which is equivalent to the third order determinant
oh ae 4 1}
—l1 7 O

1 ,14,.=2

228 Dr. J. A. Fleming on the Distribution of

And a similar series of operations reduces this to
76 4)
a —s |
which is equal to
—76x 2—5 x 21= —257.

Accordingly a series of simple subtractions and multiplica-
tions will effect the evaluation of any numerical determinant,
and enable us to solve a series of linear network equations for
the currents in all the branches when the numerical values of
the resistances of the conductors are given. The equations as
written above give as solutions the values of the cyclic sym-
bols or imaginary currents round each mesh. To obtain the
actual current in any branch, we should have to obtain the
values of the cyclic symbols or imaginary currents, for the
adjacent meshes of which the given branch is a common
boundary. Maxwell ingeniously saves labour in this opera-
tion by taking as the symbol for one mesh say x+y, and for
an adjacent mesh y (fig. 4), and then the real current in the
branch AB is ety—y=e. .

And the simple rearrangement and solution of the network
equation gives at once as value for x the current in the resist-
ance AB, which is the common partition of the two meshes.

§ 5. Returning now to the case when there is only one
impressed electromotive force in one branch, we see that in
forming the cycle equations only one will be equated to an
electromotive force, viz. the equation for the mesh containing
the impressed electromotive force in one of its branches. All
the other equations will be equated to zero ; and accordingly
the equation for the current in any conductor will be of the
form ae HAS.

EDI i
where A,, is a determinant of the nth order, and A,_, isa first
minor of this. Referring to fig. 1, we see that, by writing
down the five equations of the cycles x, y, z, u, w, we obtain
equations by which to calculate the currents in any of the
thirteen branches, and the current in branch B will be
HA,-1

Xv ss elas Sige
ev
where A, is the determinant formed of the coefficients of the
five equations, and A,_, is the first minor corresponding to
the coefficient of 2 in the equation of the x-cycle.

We also saw that if y and « are the potentials at the ends

of the branch B, H—Br=y—a.

Electric Currents in Networks of Conductors. 229

Now consider that part of the network which remains if the
conductor B is removed, and let us imagine that a current n
continues to be forced into it at y and drained out at a; the
total resistance of that part of the network, not counting B, is

d Re

but this is equal to oa
Bop.
v

Now since the resistance of B may be anything, let it be zero;
then the total resistance of the network between y and e will be

but

= [BO

where the suffix and bracket denote that after the determi-
nants are formed from the cycle equations, according to
Maxwell’s rule, then in them B is put equal to zero.

If we denote the determinant of all the n-cycle equations
under the condition of B=0 by d,, and by d,_, the first minor
of this or the minor of its leading element corresponding to the
coefficient of « with the resistance of the circuit containing
the effective electromotive force put equal to zero, we have
for the total resistance R of the network between the points at
which the current enters and leaves, the expression

d

iS
| An—1

Since, then, as we have seen, the linear equations for the
cycles can always be solved by evaluating the determinants,
it follows that in all cases, no matter how complicated, the
resistance of any network can be calculated by simple arith-
metic processes from the given resistances of the branches or
conductors which compose it. We have therefore an interest-
_ ing extension of Maxwell’s method of calculating the currents
in a network and the potentials at the junctions to a method
of calculating the combined resistance of a number of con-
ductors forming a network ; which method consists, as seen
above, in forming a certain determinant whose elements
are formed of the separate resistances of the branches, and
dividing this determinant by another of an order next below,
viz. the first minor of its leading elements; and we find
that the resistance between any two points of any network
of conductors, however complicated, is expressible as the
quotient of a certain determinant by another formed from it.

230 Dr. J. A. Fleming on the Distribution of

§ 6. We shall proceed to illustrate this method by a few
examples. |

1. Find the resistance between the points 1 and 3 (fig. 5)
of a network consisting of five conductors, whose resistances
are A, B, C, D, H, joining four points, 1, 2, 3, and 4.

Connect 1 and 3 by an imaginary conductor of zero resist-
ance, and having an electromotive force, e, supposed to act in
it. Let wv, y, z denote the cycles or imaginary like-directed
currents in the three meshes so formed, and write down the
current equations, according to Maxwell, for these three
cycles :—

(A+B)ez —Ay — Bz a
—Ax +(A+E+D)y —Ez =e
—Bu —Hy +(B+C—E)z=0.

Then, by what has been shown above, the resistance R be-
tween the points 1 and 3 of the network is given by the
expression

(A+B), cin —B
Si ee a
Bo =i ahi (B+C+E)

reeet ny ii
em Pom G2 9)

In dealing with numerical cases we need no longer intro-
duce any notice of imaginary electromotive forces, but proceed
according to the following rule.

To determine the resistance of a network of conductors
between any two points on the network. Join these two
points by a line whose resistance is supposed zero, and give
symbols to the meshes of the network so formed; calling
this additional mesh produced by the added zero conductor
the added mesh. Then write down a determinant whose dexter
diagonal has for elements the sum of the resistances which
bound each mesh, beginning with the added mesh; and for
the other elements of each row the resistances which separate
this mesh respectively from adjacent meshes, and having the
minus sign prefixed, zeros being placed for elements corre-
sponding to nonadjacent meshes.

More explicitly, if we denote by 2, y, z, &c. the meshes,
z being the added mesh, and by }R,, 2R,, [R,, &c. the sum
of the resistances which bound each cycle, then these will be
the elements along the dexter diagonal of the determinant.

:
ry

——- ee

aes

grATNea*

TR Rn cero

sahil
a ee = aA “2 ?

Electrie Currents in Networks of Conductors. 931 —

And if # and y are adjacent meshes, and *R represents the
resistance of the common boundary, then —*R will be the
element in the «th row and yth column, and also in the yth
row and zth column ; but if w and z are nonadjacent meshes,
then 0 will be the element in the eth row and zth column, and
also in the zth row and wth column. Having formed this
determinant, which we call the network determinant, we
divide it by the first minor of its leading element; and the
quotient is the resistance of the network between the two
points, joined by the zero-conductor forming the added mesh.
It is seen that, owing to the mode of formation of the network
equations, the network determinant is a symmetrical deter-
minant—that is, one half of the determinant is the reflection,
as it were, of the other half in the diagonal considered as a
mirror.

§ 7. As a means of comparing the results of this method
with other known results, let us take the exceedingly simple
case of three conductors joining two points in what is com-
monly called multiple are.

Let 1, 2, and 3 (fig. 6) be the three conductors joining
two points A and B; let their respective resistances be
71, 79, 73; then join A, B by a dotted line so as to make one
added mesh, and let the resistance of this added circuit be
zero. Then, without writing down the equations to the cycles,
we see that the network determinant is |

it ie i 5
btn ae Ty +s

The elements 7, 7;+72, 72 +73 of the dexter diagonal are the
sums of the resistances which bound each mesh, 2, y, and z,
taking the added mesh 2 first.

The other elements of the first row are the resistances, with
minus sign prefixed, which separate the mesh x from mesh y
and mesh z; or are common to w and y and « and gz, viz. ry
and zero, because w and z are nonadjacent. And, similarly,
if m and n are any two meshes, then the element in the nth
row and mth column is the resistance separating or common
to the two meshes ; and the element in the nth row and mth ©
column is identical with that in the mth row and nth column:
zero being placed as an element if these meshes, m and n,
have no common boundary or circuit.

The above determinant is easily evaluated. By adding the
first row to the second for a new second row, and this new
second row to the third for a new third row, we transform the

232 Dr. J. A. Fleming on the Distribution of

determinant easily into

Las | —— 0
0) Te. ae
) O 73
which is equal to
111 oP3.

The first minor of the leading term of the network deter-

-minant is

T+ eee:
hs san Reet

| ——¥) pees
which is equal to
PyP2 +73 Pst 5
and hence the resistance of the network between A and B is
An __ WTe"3

ny Pyrat Pers trey
which is a known result. In these simple cases the above
general rule is, of course, a less easy method of finding the
combined resistance than the direct application of Kirchhoff’s
corollaries of Ohm’s law ; but whereas the general method is
alike applicable to the most complicated as well as to the most
simple cases, the simple direct method requires twice as many
equations, and does not determine the direction as well as
magnitude of the current in each branch.

§ 8. As asimple numerical example we may take the case
of a crossed square of wires. Let 12 conductors join 9 points
(fig. 7) so as to form a square divided into four squares, or a
four-mesh network of conductors. Let the resistance of each
branch, as ab, be unity. It is required to find the combined
resistance between A and B. Number the meshes 1, 2, 3,
4,5; 1’ being the added mesh formed by joining AB by a
dotted line, making an additional fifth mesh, the resistance of
this additional ideal conductor being zero. Then the network
determinant is

4 —l -—2 —-1 0
ee ee: : ths se ican ie ee
—2 —1 4 —1 0
—1 O —l 4 —]

_ Electric Currents in Networks of Conductors. 233

The dexter diagonal has for each element 4, viz. the sum of
the four resistances, each to unity, which form each mesh or
cell. And all the other figures, say, in the nth row, are the
resistances (with minus sign prefixed) separating the nth
mesh from all other meshes, zero being placed in the column
corresponding to any mesh which has no common conductor
or branch with this nth mesh. The order in which the columns
stand and also the rows correspond to the order in which ak
meshes are numbered in fig. 7.

The numerical value of this determinant is easily found 6
be 288=3 x 96=d,. Now if we take the first minor of its
leading element, we get a determinant formed of the elements
included in the dotted rectangle ; and taking this as a separate
determinant and evaluating it, we have its value

dy = 92 2c 96

hence the resistance of the network between the points A

and B is

§ 9. One more simple numerical case may be taken and
compared with the results of known methods.

Let a hexagon of conductors be taken (fig. 8) having
crossed diagonals all meeting in the centre. Let the resistance
of each side, as ab, be unity, and also let the resistance of each
semidiagonal, as Oa, be unity. - Then required the combined
resistance of this network of 12 conductors between the points
A and B diametrically opposite. Join the points A and B by
a dotted line of zero resistance, making an added mesh 1.
Mark the other meshes 2, 3, 4, 5, 6, 7. Then by forming the
network equations it is easily seen that the network deter-
minant d” is

d—-l1-1-—-1 0 0 0
ile alien ESM re Oye
pee re he SO
are bee OS ii F

Ce Ot ee BD

Get Oo Oe bo |

BS inde oh shy od: |]. dye

The value of this determinant is 256.

234 Dr. J. A. Fleming on the Distribution of

The first minor of the leading element of d, is dae
—_ D1 oe Te 0
Sao. eS trre A, - Ch aR Sort
0—-1 8 0 0-1
= eG) Ye re
OF 0) 0 eo eae ed
0 O-—1 O-1 38

The value of this last is 320.

Hence the resistance of the network between the points A
and B is
at 200 foe

BG ae

We can easily verify this result in the above symmetrical case,
for the hexagonal framework in fig. 8 is traversed symme-
trically by the current flowing through it; and hence no
disturbance of the distribution of currents will take place by
separating it, as in fig. 9. We break the connection between
the semidiagonal conductors a, 6 and the mean diagonal A B,
whilst keeping them in contact with each other, the resistance
of each branch still remaining unity. It is then easily seen
that the hexagon so arranged must offer exactly the same
resistance between the points A and Bas in its original form.

Now the combined resistance of a, b, and 7, each equal to
unity, between the points C, D is 3, and the combined resist-
ance of this with e and g in series is 22; and hence the total
resistance of the whole network between A and B is equal to
that of three conductors in multiple arc whose resistances are
respectively 22, 2, and 22, which is equal to

1 pend
Bair ak Sa
a le Me!

the same result as obtained above.

These numerical examples show conclusively that, in cases
in which the resistance of a network can be obtained by simple
direct methods, the results coincide, as should be the case,
with those obtained by the employment of the general method ;
but at the same time the general method is capable of con-
ducting easily to a solution in the most unsymmetrical cases.
The general rule will, for instance, just as easily give the
determinants when the selected points between which the
resistance is required are not symmetrically placed, but are,

5 has ees hese i
| emEE St Se er eh

—

WG RIOR GES FORE RU IT

Electric Currents in Networks of Conductors. 235

say, adjacent angles of the hexagon, in which case no such
simple direct method as employed above can be used.

§ 10. The following example will give a good illustration
of Maxwell’s method of treating network problems, viz. the
case of Sir W. Thomson’s resistance-balance for small resist-
ances. In this arrangement (fig. 10) 9 conductors join 6
points and form 4 cells. B is the battery-circuit in which
operates an electromotive force HE. Let the four cycle cur-
rents be denoted by #+y, y, z, and w. These are the imagi-
nary like-directed currents round the circuits, and the real
currents in the branches are the differences of these.

The problem is to determine the current in the galvano-
meter branch G, and the relation of the resistances when this
current through G is zero. Let P,Q,8, T, R, 7, D be respec-
tively the resistances of the branches, and G the resistance of
the galvanometer circuit, and B the resistance of the battery
circuit. Then «+y and y being the imaginary like-directed
currents in the two adjacent meshes of which the galvanometer
branch is the common boundary, then «+y—y=~ is the
current through the galvanometer.

Proceeding to write down the cycle equations, according to
Maxwell’s rule, we have

—_—_——

(P+G4+Q+R)e+y—Gy—Q:—Rw=0,
(T+r+8+4+G)y—Gety—Sz—rw=0,
(Q+8+4+ D)z—Sy—Qe+y—Dw=0,
(R+0+7r+B)w—Ret+y—Oz—ry=H.

Rearranging these equations and solving for 2, we have the
ollowing value :—

—Q+58, D, —D
T+S+7, T+r, —r |;
P+Q+R, P+Q, —R

&
Hae
P| &

in which A is the determinant of the four equations in 4, y,
z, and w, and whose specific value does not concern us. ;
This gives the current in the galvanometer-branch ; and if
this is zero, then the determinant in the numerator of the equa-
tion giving « must be zero. Hence, when is zero, we have

(+5, D, O

P-S+7, (ebb, T.)-=0,
P+Q+R, P+h, P

236 Dr. J. A. Fleming on the Distribution of

this determinant being derived from the one in the equation
for « by adding the second and third columns for a new third
column.
This last determinant equation writes out into
Pee dS ES
PR P, Q |
Hence the condition that the current in the galvanometer-
branch shall be zero is that both determinants in this expression
shall be simultaneously zero, or

+ D =<)

(Q+8+D)

Ohi af | Tos a
PR =0, and P, Q | =
that is, ‘ig ha;
a ee ee

Hence this condition expresses the relation which must hold
good between the magnitudes of the resistances T, P, Q, 8, 7,
R, in order that the galvanometer-branch G may be conjugate
to the battery-branch B. |

The above example shows well the symmetry of the method
when dealing with a case of distribution of currents in a net-
work.

§ 11. Asa final illustration, let us consider the case of a
circular wire A P B Q, with a diametral wire P Q across it.

Take any two points A, B, at the extremities of a diameter
not coinciding with P Q, but separated by an angular distance
@ from it, and let us obtain the resistance of the circular wire
so crossed between the points A and B.

Join the points A, B by a dotted line of zero resistance.
Call the three meshes so formed w, y, and z; let r be the
radius of the circle ; and let p be the electrical resistance of

the wire per unit of length. Then the
Resistance of branch PQ =2pr,
AP=pr0,
” ” AQ=pr(7—8@), —
and
Resistance of branch BQ=resistance of AP,
” ” PB= ,, » AQ.

Then the network determinant d,, is
pear, —pr(n—0), —@
—pr(m—0), —pr(w+2), —pr2
—9, —pr2, pr(m +2)

Electric Currents in Networks of Conductors. 237

_ Removing the common factor rp, we have to evaluate

7 —7r—0 —6
—7—-0 7+2 —2
a9 —2 7+2
This is very easily reduced to
8 tar
T Soh Seed See aed
—O0 2m

which is equal to
2a( m+ Or —6"),
and therefore
dyn=1' p2ar(m + Or — 6),
The value of the minor of the leading element of the network
determinant, viz. d,—1, 18
| rp" {(7+2)?—4} =r*p*ar(r +4).
Hence the resistance of the network between A and B, =R, is

\ dn yp. 2a(7+0r—86")
P| ma R=rp peer ae
(ie Sse 2a + 270 — 26"
& q = i 4 ;

We can check this in the extreme cases when 0=O or =.

When @=0, the network-resistance is simply that of three
conductors whose resistances are 2pr, mpr, and mpr joined in
multiple arc, as in Plate VII. fig. 12, because PQ now coin-
2ar
m+ 4
once that the above value for R becomes this when @ is put
equal to zero. Now, when @ =5 , the diameter PQ joins
points at equal potential (fig. 13), and is not traversed by any
current at all; and hence its removal will not affect the
resistance between the points A and B. —

Hence the resistance of the network simply reduces to that
of a circle measured at the ends of a diameter, or to two con-
ductors of resistance wpr joined in multiple arc, and this 1s
T
a
for R above, we get it reduced to rp = ; and accordingly this

Phil. Mag. 8. 5. Vol. 20. No. 124. Sept. 1885. S

It is seen at

cides with AB; and this is simply 7p

| equal to pr By putting 0= ~ in the general solution

238 Dr. J. A. Fleming on the Distribution of

formula agrees, as it should do in these reduced cases, with the —
results of the direct method based on first principles. If a
value of @ be found which will make the expression

Ir + 278 — 20"

equal to 7+4, then for such a position of the diameter AB
relatively to PQ the resistance of the circle and its diagonal
PQ would be exactly equal to the resistance of half the dia-
metral wire or to its radius, assuming both the circle and
diagonal to be made of wire of equal conductivity per unit of
length. To find the value of @ for which this is the case, we
have to solve the quadratic

Qor + 270 — 20? = 7 + 4.
If we put 0= a «°, where w° is the number of degrees equi-
valent to the angle 0, we find, as a solution for this quadratic,
that the positive root is nearly

171°°804.
Now 38 radians, or 3 unit-angles in circular measure, are nearly
171°°887.

Hence, for a position of the diagonal PQ as in fig. 14, when
the arc AP is nearly equal to 7—3, or to the fractional part
of 7, the resistance of the circle and diagonal PQ measured
between the points A, B is very nearly equal to that of half the
diagonal PQ; or, which is the same thing, the resistance of PQ
alone is nearly double the combined resistance of the circle and
diagonal measured between the points A and B at the extremity
of a diameter removed 171°°804 from PQ.

§ 12. A small practical application of this last example may
be made in constructing a variable resistance.

Let PAQB (fig. 15) be a narrow circular canal cut ina slab
of wood or ebonite and filled with mercury. Let PDQ be a
bent copper wire balanced on a pivot CD, and having its ends
P and Q dipping in the trough at opposite extremities of a
diameter of the circular trough PAQB. |

The total resistance between any two points A and B in the
trough, which are also diametrically opposite, can be varied
within limits by changing the position of PQ relatively to AB.

When PQ is turned so that it is at right angles to the dia-
meter AB, it does not affect the total resistance between A
and B, and may be removed. ‘The resistance is then just that
of the circular band of mercury taken at opposite extremities
of its diameter. When PQ is coincident with AB it reduces
the resistance, and in intermediate positions the joint resistance

PCS TTT AME

P.
wy
I ia
*
_,
ee
P
4 ay
a
s
=
¢
me
on.
f
ib

Electric Currents in Networks of Conductors. 239

of trough and diagonal wire is intermediate between the greatest
os least when it is in position removed either 90° or 0° from

Bie.

By using a circular glass canal filled with sulphate-of-zine
solution, and a zinc diagonal electrode and amalgamated-zine
electrodes at A and B, a variable resistance may be constructed
capable of being varied over considerable ranges perfectly
gradually and with no imperfect contacts.

§ 13. Having illustrated, by the foregoing examples, the
methods of calculating both the currents in and resistances of
networks of any complexity, we return for a moment to some
general considerations.

Consider a function formed of the sum of each separate
resistance in a network multiplied by the square of the current
strength flowing through it. This expresses the heat gene-
rated per second in the whole network by that distribution of
current. This is called the Dissipation Function of the net-
work. It represents the rate at which energy is being trans-
formed into heat or rendered unavailable. —

Write down the dissipation function for the network in

fig. 1. Callit H. Then
H=Be2*+ Iz—y" + He—r’ + Cy? + Lz—y' + Az?+Ju—y’
+Kz—w’+(D+E)u?+(F+G)w?+ Mu—wu".
Now the cycle equation for the cycle or mesh y is, by Max-
well’s rule, 3
(C+1+L+4+Jd)y—Iv—Lz—Ju=0,
which is the same as.
Cy—1x—y—Lz—y—Ju—y=0.
And this is at once seen to be identically the same as the first

partial differential of the dissipation function with respect to
the cyclic symbol y, or is the same as

1DH_,
2 OY ’
where 9 represents partial differentiation ; and by writing

down the other cycle equations for each cyclic symbol or

imaginary current, z, y, z, &., we can show that these cur-
rent-equations are respectively

25x ’ 25 ’ 252 ’ dC.,
S 2

240 Dr. J. A. Fleming on the Distribution of

each equated to the effective electromotive force in that cycle
or mesh. :

Let us assume now that 2 is constant, but that y, z, u, w, &.
are independent variables and are arbitrarily changed. This
is equivalent to supposing that a given quantity of electricity
per second is pushed into the network, but that its distribution
is supposed to be varied. We see that the equations which
we write down, according to Maxwell, to determine the real
distribution of currents in the network, according to Ohm’s
law, are the same equations as would be written down to find
the values of y, z, u, w, &e., which make the dissipation func-
tion a minimum under fixed conditions of total current flow-
ing into the network, viz. equating to zero the first partial —
differentials of H with respect to the variables y, z, u, &e.
The same holds good generally, hence we see that this is
another way of arriving at the theorem of which Maxwell has
given a proof on page 375, § 284, vol. i. of his large Treatise,
2nd edition, viz.:—‘ In any system of conductors in which
there are no internal eleciromotive forces the heat generated —
by currents distributed in accordance with Ohm’s law is less
than if the currents had been distributed in any other manner
consistent with the actual conditions of supply and outflow of
the current.”

The exact proof that the partial differentials of the dissipa-
tion function equated to zero gives the condition that the dis-
sipation function shall be a minimum is not complete without
an examination of Lagrange’s conditions. It is obvious that
the second partial differentials of the dissipation function are
quantities which are resistances, viz. the coefficients of the
current symbols in the cycle equations, and that the conditions

2

e ° e a e fo) H .
for a minimum are complied with, since =, Nec. are posi-

tive ; and the discriminant of the quadratic function of the
currents or symmetrical determinants formed of these second
partial differentials is what has been called above the network
determinant. This and all its successive minors are positive
quantities *. |

§ 14. In the foregoing sections the problems have been
treated under the limitations that the various meshes of the
network of conductors have no mutual and no self-induction. :
The introduction of these inductive actions will affect in a
considerable way the treatment of the problem ; and the dis-

* See Williamson’s ‘ Differential Calculus,’ p. 408, “‘ On the Conditions
for a Maximum and Minimum of a Function of any number of Variables,”
§ 163, and Appendix.

Electric Currents in Networks of Conductors. 241

tribution of the currents in, and the resistance of, the network
will be affected by them during the time taken by the cur-
rents to become steady.

In those pages of his Treatise in which Clerk Maxwell
worked out his splendid dynamical theory of electromagnetism,
he starts with the explanation of the methods Lagrange and
Hamilton employed to bring pure dynamics under the power
of analysis, and the results of Lagrange are embodied in the
equation :

er Sar a re

wa ie eee

in which X is the impressed force tending to increase the
variable x, and T denotes the visible energy of the system of
bodies at that instant.

This equation establishes a relation between the kinetic
energy of a material system at any instant, the force im-
pressed upon it in a certain direction, and a quantity called a
variable, which expresses the state or condition of the system
with respect to that direction. Maxwell, by a process of ex-

traordinary ingenuity, extended this reasoning from materio-

motive forces, masses, velocities, and kinetic energies of gross
matter to the electromotive forces, quantities, currents, and
electrokinetic energies of electrical matter, and in so doing
obtained a similar equation of great generality for attacking
electrical problems.

In the electrical problem the variables are the quantities of
electricity 2, y, z, &e. which have from the beginning of the
epoch flowed past any points, and the analogues of the velo-
cities are the fluxes of these, x, y, z, &c., or the currents.

The electrokinetic energy is measured by the quadratic
expression

T=LzitiLowe + ... Myaat, &e.,
where the coefficients L,, L,, My, are functions of the geome-
trical variables, but into which the electrical variables do not
enter.

If now, as before, z,, 2, represent the imaginary like-directed
currents round each mesh of a network, in which currents
are beginning to flow, then

represent the electrokinetic momenta of these circuits. De-
note them by 7, pe, &e., and accordingly

242 Dr. J. A. Fleming on the Distribution of
Pi Lia, + Moto, &e.

If Ei is the impressed electromotive force in the circuit or
mesh arising from some cause, battery, thermopile, dynamo
machine, &c., which would produce a current independently
of magneto-induction, then, if R be the total resistance
round the mesh, and #& the cyclic current, Ra is the electro-
motive force required to overcome the resistance of the circuit,
and E— Rz is the electromotive force available for changing
the electric momentum of the circuit.

Accordingly, by Lagrange’s equation,

<i apoedD
where T is the electrokinetic energy. As T does not contain 2,
that is to say it is a function of currents, not quantities, the
last term disappears, and we have

i Gid®
or 7
ieee °

The electromotive force is therefore expended in two
things: first, overcoming the resistance R; and, secondly,
increasing the electromagnetic momentum p. Now if there
is no electromagnetic momentum, we have seen that the
cyclic equations are of the form

where H is the dissipation function of the system, and EH’ is
the acting electromotive force concerned in overcoming the
resistance of the circuit.
If, then, we substitute for Rez in equation ee we have
x
as the general equation for the electromotive force in any
mesh or cycle 2, 2 -
d d

A ee — Bi:

dt dx “dx
This most important equation is Maxwell’s general equation
for determining the current 2 in any circuit when the dis-
sipation function, and kinetic energy, and impressed electro-

Electric Currents in Networks of Conductors. 243

motive force are known. We shall proceed to apply it to the
solution of some network problems, in which the self and
mutual induction of the branches is taken into account to
determine the distribution of currents and combined resistance
at any instant during the variable state.
$15. Consider, first, the case of a galvanometer with a
coefficient of self-induction L and resistance G, and shunted
by a shunt of resistance 8, but wound so as to have no co-
efficient of self-induction, and let the shunt and galvanometer-
coils be so far removed that there is no coefficient of mutual
induction. ‘This is the ordinary practical case.
Let a battery be joined up and let the battery and connec-
tions have a resistance B and electromotive force H (see fig. 16).
We have then a two-mesh network. Call the current in
the galvanometer- and shunt mesh y and the current in the
shunt and battery mesh «+y. Then the current through
the galvonometer is y, the current through the shunt is z, and
the current through the battery is x+y.
The dissipation function H is

Be+y°+S8a? + Gy’?=H,
which may be written
B+S aty/+G+8 y—Wetyy=H;
and the electromagnetic energy is
tif? =F.
Hence, by the general equation,

age ede.
dt dy 3 dy =H,
we have the two cycle equations for the y and x+y cycles,

Sly + G+ 8y—Sety=0

and
B+8 «+y —Sy=R8,
or
(1 id +G)y—Se=0
dt :
and

By+B+S82=H.-

244 Dr. J. A. Fleming on the Distribution of

The solution of these for x and y is

EK B+S8
Bae
y= = eurrrrent through
pe B+58 [ galvanometer,
d
Lo+ G -—S
and
B i
d
L aah a 0
= SE SS eee
B B+8 [the shunt.
d
Lat G —S
Writing out this differential equation for y’ we have,
LZ -G| _| B B+8| |B —B+8
y abe I= :
0 B+s Aaa Dak
or

B+SL& + (BS+RG+8G)y=ES,
or
dy _BS+BG+S8G. -Hid
dé (B+ 8a? 3 Bae
The solution of this differential equation is
ES ( __BG+G8+B8 |
Y= BGTGSTES\I—@ FH )
This gives the value of the current through the galvanometer

at any time, ¢t, after starting the flow by making the connec-
tion with the battery.

When ¢=0, then y=0, and as ¢ increases y increases, and
H

finally, when t= o, y= BGLGSTBS % 48 it may be
iD

“4 oh
written, y= G79 a Sq
G+S8
This last is the ordinary formula given for the current
through a shunted galvanometer ; but we see that when self-
induction is taken into account, it is not until after an infinite
time that the current rises to this value,

Electric Currents in Networks of Conductors. 245

By the cycle equation, By + B+Sc¢=E : hence

H—By
2 Bust
ire _ BG+GS+BG ,
And if we write N for the factor (1—e (B+8)L i then
3 n_ __EBSN
BG+BS+8G

o—

B+
_E(BG+BS+8G—BSN)_
— (B+8)(BG+BS+8Q) °

which gives the current through the shunt at any instant.

§ 16. Consider now the combined resistance of the galva-
nometer and shunt at any instant.

The self-induction of the galvanometer acts like a spurious
resistance during the commencement of the current and drags
- out or prolongs the rise of current in the galvanometer-coils ;
accordingly, during this period the combined resistance is a
function of the time ¢ from the commencement of the flow.

To calculate the combined resistance of galvanometer and
shunt at any instant, we proceed as in the cases above exem-
plified. Form the cycle equations

(B+S)¢+y—Sy=8,
frie bere \ a‘
—Sety+(G+ S+L7)y=0.

Write down the determinant of these equations with the bat-
tery-circuit resistance put equal to zero, that is put B=O, and
the combined resistance R required is the quotient of this
determinant by its first minor, viz.

S 8
9 16572.
= #8 Tes,
ee
d
G+S8+h7
or
eAitorisG
ung digipey, © d
li S(G+L5,)
Be ma

= —___“..
d S+(G+L—
B+ GH, ( dt

246 Dr. J. A. Fleming on the Distribution of

We have now to see what is the meaning of G+L9 as an
operator in a determinant.

If we consider the formation of a current in a cireuit of —

resistance R and coefficient of self-induction L by an electro-
motive force E, we have the equation for the current 7

Write thus
d !

(L5+R)i=E:
or, by notation of the calculus of operations,

d ts

i=B(LS+ R) .

But now the solution of the above differential equation

under the conditions t=0, 7=0, and t= », i= 5, Is

R?
(ie ')
EM.

Comparing these two expressions for i together, we have

Hence we may substitute in the expression for the combined

resistance of galvanometer and shunt for L c+ G,

and we have as a result,
R SG
SS aa
G+ s(1 —e-l ‘)
GS
We see that when t=0, R=§8, and when t= o, R=qy5°

Hence the result shows that at the first instant of starting a
current through a shunted galvanometer, when the shunt
has no self-induction and the galvanometer a considerable
one, the galvanometer behaves as if it had a high spurious
resistance, which in time dies away, allowing the total current,
after an infinite time, to be divided between the galvanometer

and the shunt in the ratio of to

S
G+S° G+

Electric Currents in Networks of Conductors. 247

§ 17. We may apply the same methods to the examination
of the case when the current sent through the shunted galva-
nometer is not generated by a source of constant electromotive
force, but is a discharge from a condenser.

Let K (fig. 17) be acondenser connected up with a shunted
galvanometer, so that when the key & is pressed a discharge
passes through the galvanometer and shunt. Call the two
cycles wand y. Let G be the galvanometer-resistance and 8
the shunt, and let L, and L, be their respective coefficients of
self-induction ; the coefficient of mutual induction being zero.

Let g be the quantity of electricity in the condenser at any
instant t. Counting the time from the instant of commencing
the discharge, let C be the capacity of the condenser, and let
g, and gz be the quantities of electricity which have, since the
beginning of the epoch, flowed respectively through the gal-
vanometer and the shunt.

If T be the energy function and F the dissipation function,
we have, as above, the fundamental equations

2T= ly’ + L(a—y)’,
2k = Gy’ + S(a—y)? :

and

. 2T= ly’? + Lx’? + Lay’? —2L, ay,
2F =Gy’+ S82? + 87?—282y.
By the fundamental equation
ddT , dF
dida* da
4
Ci

Writing, then, the cycle equations, we have

For e we must write

d

f (Lay + Lay —Lye) + Gy + Sy—S2=0 5

_ from which we deduce easily

dy By
Ly dt -- Gy= @
and

d |
Ly @—y) + Sey) =G

eee TIE: =

248 Dr. J. A. Fleming on the Distribution of
or

d. d
Ly — Ls (#—y) =S(@—y) — Gy.

Now y and w—y represent the strengths of the currents flowing
through the galvanometer and the shunt at any instant.

If we integrate both sides of the equation from 0 to o, we
have

[Ly —L,(#—y)|o = sf 2 —ydt— af ydt.

Now the left-hand side of the equation is zero because quan-
tities of the form of Ly represent the number of lines of force
which are added into the circuit of the galvanometer, and the
discharge may be divided into two parts, during one of which
lines of force are being added to, and in the other of which sub-
tracted from, the circuits of the galvanometer and shunt; and

the sum of these is zero. Again, { (e—y)dt and\ ydt re-
0 , 0

present the whole quantities g, and gq, of electricity which
have flowed respectively through the galvanometer and the
shunt. Hence we arrive at the conclusion that

Sgo2—Gq=0,

or
G_ @.
S 1

that is, the total quantity of the discharge is divided between
the two circuits inversely as their resistances. We see there-
fore that self-induction does not affect the ratio of division of
a discharge in a divided circuit, provided that no external
work, such as the moving of magnets or circuits conveying
currents, absorbs current energy. Hence, if a ballistic gal-
vanometer is shunted and a discharge sent through it, if the
needle has sufficient moment of inertia and the discharge is
sufficiently short, so that the needle has not perceptibly moved
from its position before the discharge is over, then the whole
quantity of electricity is divided between the galvanometer
and the shunt in the inverse ratio of their resistances.

§ 18. To complete the solution we have to calculate the
current flowing through the galvanometer and shunt at any
instant.

Taking the two cycle equations

d j
L,! +Gy=4,, i.

Electric Currents in Networks of Conductors. 249
and
oe *
L, me—yt Se—y=~ Hite see CHE, )
we get
d
k Ly a - — Gy,
and

dea eee
LL, a= (Ly + La) @—L, Sz—y—-L.Gy

=(L, + Ly) 4—L,Se+(lyS—L.@)y. (ii)

Differentiate this last equation with regard to ¢ and eliminate

= by the help of the equation above it, and we arrive at
ax dx _L,S—L.G
: L,L pn LS Gt (§- Gy).
ue gee bee (algae

| Kliminating y between the last and equation (iii.) and re-
ducing, we arrive at

2
Lily 92+ (L,S+ L,G) + GSe=(G+8) f

| but now «= —%. Making this substitution we have
3 2

L,L,0 9440 (L,8+1,G) 92+ 04844 (G+8)¢=0, (iv.)

an interesting equation, the solution of which gives us

the quantity of electricity in the condenser at any instant, ¢,

after starting the discharge. According to the equation above

OE wits i
/ Ly dt + Gy= Cc
This equation gives us a value of y or the current through
the galvanometer at any instant when we know g, or the

quantity left in the condenser at that instant. The above
may be written

d =.
y= (Lajt G) =
and the final equation (iv.) may be written

oe us oo4+ces2 +c(4+8)) 0;
a= L, LC ap t (uS+L, ) apt aaah (G+S8) ;

250 Dr. J. A. Fleming on the Distribution of

and accordingly we have the following equation for the value |
of y at any instant
;

d = | da? dq? d s
y=(Ing+ G) (IuL.0" —+(iS+L@CG+Cass +0(G+ 8))
which may be written

4 : a |
y= (LiLo + LgSC?+ 211 L,0°G) + 2L,COS+ 146°C)

seh See ees
+(GFSCL, + CGS) Z +GOG48) 0.

This linear differential equation is solved when we know
the roots of the auxiliary biquadratic; and according as they
are all real or partly imaginary, so will be the nature of the
solution.

If the roots are all real the solution is a sum of exponentials,
whose total value first increases and then dies away as ¢
increases, indicating that the discharge produces a wave of
electricity through the galvanometer always in one direction;
but if two or all of the roots of the auxiliary biquadratic are
unreal, it indicates asthe form of solution a function of sines
and cosines which will have periodic values, and points to the
fact that the discharge is a series of alternations. The general
case, when both the galvanometer and shunt have coefficients
of self-induction, when treated to determine the conditions for
an oscillating discharge, leads to an expression of considerable
complexity and not much practical use. The reduced case, in
which the galvanometer is wound to have self-induction and
the shunt so as to have no coefficient of self-induction, is,
however, a practical case, and can be treated without much
difficulty.

Taking the differential equation for g, equation (iv.), and
writing in it L,=0, we have

“ig l
CL,S33+0GS9 + (G+ S) ¢=0,

—

or
S dq Chee yr Oe, oe |
G+8 d?*L(G+8) a * CL,” |
The discharge will be oscillatory if the auxiliary quadratic |
SS cate GS 1 |
ore lee CL, |

has unreal or imaginary roots.

Electric Currents in Networks of Conductors. 251

Solving it we have
Ging GG?) +5
eee oe
acon Vraaae PAlae” | Pe mea< (1 WR
or
( ,G)\_, VGC —4G + SL,80
<3 sr.) Seen et jean Bel:
Hence, for the roots to be imaginary,

41, G+8 SC must be greater than G’S?’C’,
4L'_ G.GS

C GES.

If this relation holds good, then the discharge is oscillatory
in the condenser; and accordingly we see that to prevent
electrical oscillation in the galvanometer circuit, the product
of resistance of the galvanometer and combined resistance of
galvanometer and shunt must be equal to or greater than
four times the self-induction of the galvanometer divided by
the capacity of the condenser.

We may write the solution of the quadratic above,

or

G+S8
pre Sue
Wee OTO. 40
=—at V7—1£,

where G+S8
Mey Ge By. S Ge
ere CEC Tae

and accordingly when 8 is real, that is when

G+S8

NZ

Lc * 7407
we have, for solution of equation,
g=Ae—“ cos Bt + Ber sin Bt.
When ¢=0, g=Q=the original charge of the condenser, and
“70 when t=0;
dt :
therefore QA, andi: 3=B ;

and a—Ae oo (cos Bi+ asin At).

252 Dr. J. A. Fleming on the Distribution of

Having now the value of the quantity of electricity left in
the condenser at any instant, we can find easily, from the
cycle equation (i.), the value of the current through the
galvanometer. For

dy
L,7 + Gy= rst
or s G
ops URS SEG Fotan
A Lt ae
d e (Ga
y=e Sr ent Ee Mi

and the constant C! is determined by the condition y=0 when
PU. ‘
Substituting the value of g above, we have

say i
ee) Fh

rad {+5 Gt, ler(B cos Bt +a sin st) }
a | C!+ Q e* sin ae} ;

BCL
but C’=0,
° = —at ay °
i y— ger sin Bt ;
and since
G+S8
a E S (72
Enos, and | oe Oe =| aL?
a RM Ste- aa)
fe JGin. Come SL,C 413 /”
ieee

which gives the value of the instantaneous current in the
galvanometer-circuit at any instant ¢ after starting a discharge —
from a condenser of capacity C and original quantity Q through
a shunted galvanometer, the shunt being wound without self-
induction, and the galvanometer having a coefficient of self-
induction Ly.

§ 19. Two concluding examples of this method of treating

network problems will now be given, which are in Professor
Clerk Maxwell’s own words*.

* In the May term 1879, Professor Clerk Maxwell lectured at Cam-
bridge on Electromagnetism, and in the two last lectures of the Course he

Electric Currents in Networks of Conductors. 253

Theorem.—To compare the induction between one pair of
coils and any other two. |
Let a, 8, y, 5 be four coils of wire. |
It is required to compare the mutual induction of a and vy
with that of 8 and 6.
Join up a and 6 coils in series with a galvanometer, and
join up y and 6 coils in multiple arc with a battery, as shown
in fig. 18.
Place the coils in position.
Let S and R be resistances of the primaries y and 6; let 0
be resistance of the two secondaries and of the galvanometer.
Let Ly, L., N;, No be the coefficients of self-induction of
the coils, and M,, M, the coefficients of mutual induction of
aand y,@and 6. Let I be the coefficient of self-induction
of the galvanometer.
Call x the cycle current of y, y that of 6, and z that of the
circuit formed of a, 8, and the galvanometer.
The kinetic energy T of the system is
2T=a2°N,+a2M,4+27(L,4+ 1,41) +y’No+2y2Mp,
___ and the dissipation function F is
2F=27R+y84+2’Q.
Then, by the formula
a8 dk

da?’

d dl
dtd
aN, +2zM,+cR=E.
yN,+ 2M, +y8=—E.

aM, +yM,—2(L,+ L, +1) +2Q=0.

Now x—y is the current through the battery; hence if we
put z+y for z# in the above, we shall get x as the battery-
current. Hence, making the change, we have

+4

gave this method of obtaining the equation for the currents in a network
of conduction. In the last lecture of all he applied the method to cases in
which self and mutual induction was taken into account, and gave the two
illustrations in § 19. At the conclusion of this lecture he had ended his
professorial duties for the term, and a melancholy interest attaches to the
subject which occupied his mind on the last occasion on which, uncon-
sciously to himself or his pupils, he was to perform them. Those who
enjoyed even for a brief period the privilege of being taught by him, ever
cherish a vivid remembrance of the intellectual treat afforded by Professor
Maxwell’s lecture-teaching, and the profound suggestiveness and interest
of it.

The two examples in § 19 and § 20 are taken from my notes of Prof.
Maxwell’s lectures, with some little alterations to make them clearer.

Phil. Mag. 8. 5. Vol. 20. No. 124. Sept. 1880. fi

Shae. ds

204 Dred A. Fleming on the Distribution of

(444)N, 42M, +2+yR=E, . . . ee
jN, +2M,+y8=-B,. . . . 37
(+ y)My 4 yxM.—2(L + Le + T) + 2Q=0; . Git)

d
add equations (i.) and (ii.) and arrange, putting n for Tp
(Nn+R)x2+(Nin+Non+ R+ 8)yt hh + My)nze=O,
Mnx + (Mn+ Mon)y+4(Iy+L.+T)n+ Q§2=0.
Eliminating y, we have

§n(Nyn-+ R) (My + M,) —Myn(Nin + Nan + B+ B) fa

+( (My + My)’n?— §(Ly +1. +T)n+ Q} {Nin+ Non+ R48} )e=O.
Hence we get |
{n2(M,N, —M,N,)—n(M,S—M,R)} a

JOE LE ae, ot
a denominator which does not concern us

If matters are so arranged that z=0, or the galvanometer
shows no current,

(M,N, —M,N,) = =(M,S—M,R);

hence if there is no “ kick.”’ on the galvanometer on making
the current, then

M, &

§ 20. Theorem.—To determine the capacity of a condenser
by means of a Wheatstone’s bridge (fig. 19).

Let «, B, y, 6 be the four points of a Wheatstone’s bridge;
and let the branch between a and § be interrupted at ab, and |
a Leyden jar or condenser inserted provided with some rapid
commutator, such as a tuning-fork, so that whilst the outside
of the jar is kept permanently attached to 6, the inside is
alternately joined to a and 6. |

If a tuning-fork is used and its prongs have small metal
styles which just come down to the surface of the mercury in
two little cups, when the fork vibrates, as the prongs come
together, the upper point dips in; and as they separate, the
lower one dips in; hence the shank of the fork is alternately
connected with one and the other cup. The interval between
the time of connection being exactly half the time of a com-
plete oscillation of the fork.

Now let the meshes of the network be called 2+2z, 2, and y;

Electric Currents in Networks of Conductors. 2595

then z is the current through the galvanometer, and y is the
current through the battery. When the arrangement is made
as in the diagram, and the fork set vibrating, the vibrating
fork and the condenser act together like a resistance, and let
through so much electricity per second.

Now, as the condenser gets its charge by electricity flowing
into it, it builds up an opposing electromotive force in the z

circuit which at any instant is equal to the value of i) Eta) where

K is the capacity of the jar, the integral being integrated from
the instant when the charging commences up to the instant
considered. Now, if the fork makes x vibrations a second
when the steady state is set up, the current z which flows into
the jar has a mean value z; and therefore a is the opposing
electromotive force in that branch. |
Accordingly, the condenser and associated commutator
behave like a voltameter inserted in the branch @£, or like a
resistance with a counter electromotive force in it. Only such
a combined jar and fork differs from an ordinary metallic
resistance in this, that its apparent resistance is not constant,
but depends on two things, the speed of commutation or
charge and recharge, and the capacity of the condenser; whilst
the counter electromotive force depends on the current z, and,

oi

being represented by ae is dependent not only on n and K,
but also on the values of all the other resistances in the branches,
In the first place, we require an expression for the electro-
motive force charging the condenser. Let the difference of
potential between a and b be callede. Then consider the net-
work formed by the five conductors R, 8, Q, G, and B with
the electromotive force in the branch B; write down the net-
work equations for this z mesh network.

(B+R+S8)y—S(e+z) =H,
—Sy+(Q+8+4+G)(e+z)=0.

H(Q+8+G).
os ile Se ME Me

Hence

and

T 2

256 Dr. J. A. Fleming oz the Distribution of
which is
8(Q+ G)+(R+B)(Q+8+G).

Now the difference of potential e between a and 6 when the
condenser is just beginning to be charged is

G(e+z)+Ry=e;

_ ESG BR(Q+8+G)

ee habs 4 a
_-¢,°
fe GE ae) tt Q+8+6 Sm

S(Q+ G)+(R+B)(Q+8+ G)~ 8(Q+G)+(REB)Q+S
or
=G

Now if the electromotive es e be employed m times in a
second to charge a jar of capacity K, the average current
flowing into the jar is nKe=z.

Now to find z we have to consider the distribution of cur-
rents when the fork or commutator is in operation, and the
_ condenser allowing a flow of electricity to take place through it.

Let P be the resistance which could equivalently replace
the jar and fork—that is, would allow an equal quantity of

electricity to pass per second ; then, since —*_ is the opposing

nk
electromotive foree in this branch, we have the following
equation for the three cycles v7, «+z, and y:—

—Se+(R+8+B)y—(R+8)z=H
—Ge—Ry+(P+R)z= —
(Q+84+G)2—S8y+(Q4+S8)z=0.
Now let A stand for the determinant
—S, R+8+B, —(R+8)
1 |
—G, —R, P+ aie +R],
Q+8+G, —S, Q+58
Then the solution of the above equations for z and « are
E —G, —R
MME: BS ao al
A 3

Electric Currents in Networks of Conductors. 257

and
1]
ny) ~& P+aR +B
i reels aes ue
t= A

z is the average current flowing through the condenser, and «
is the current through the galvanometer. Now let the resist-
ances R, 8S, and Q be so varied that the current through the
galvanometer is zero, then «=0; and therefore

1
“BR, P+ +R |
oe Aces
—8, Q+8
or
ea 1 )
R(Q+8)=8(P + — +R),
or 0
R 1
Sones
Now insert this value for P+ ss in the determinant A

nKk

above and calculate its value, and we arrive at the expression

a (BQ+S)+ Q(R+S)HGR+8)+R(Q+S)}
sf A ;

We have now, by substitution of this value of A in the value
obtained above for z, an expression fer the value of the average
current through the condenser when the bridge is balanced,
and it is
—G —R
KS Pee
~ (B(Q+S8)4+ QR+8)}{GR+8) + RQ+S8)}
Hquating this to the other value for z, namely,

LG pia 9o Al
zenKe=nKB_ | QtS+G,—-S5 | _
S(Q+ G)+(R+ B){Q+8+ G}
we have
aK =. Si8(Q+G)+(R+ B)(Q+8+G)}

{BQ+8)+ QR+S)HER+FS)+RQ+SP
which gives us a value for nK in terms of

a Fe BG

258 Electric Currents in Networks of Conductors.

Now it is interesting to note that we may otherwise write
the above expression for nK,

2s
nK~ 6’
where A is the determinant,
R+B+8, —§, —R+8
—§, Q+8+G, R(2 +1)|
—R, —G, Q+5

and 6 is its first minor,
R+B+8, -—-8
8, Q+S+G

1 J ’ e es e
and ae is of the dimensions of a resistance.
n

The value for nK writes out by a simple transformation into
another form,
G2
s{1-

oe (Q+84+G6)(R+B+8)

+) /

B14 Gee rey {1+ EgsEFe)

which is the form in which it is given by Prof. J. J. Thomson
in his paper, and quoted by Mr. R. T. Glazebrook in his
memoir on a Method of Measuring the Capacity of a
Condenser™.

The above examples are amply sufficient to exemplify this
method of treating problems in networks of conductors, and
show how it enables calculations to be made with considerable -
ease, not only of the distribution of currents and potentials,
but of the resistances between any points on a network, the
branches of which consist either of simple resistances or of
wires having self- and mutual induction with other branches,

or of electromagnets, or condensers associated with appropriate
commutators.

* This method of Maxwell’s, of obtaining the capacity of a condenser has
been practically employed, with most excellent results, by Mr. R. T.
Glazebrook, F.R.S. ; and the full details of the tests to which he subjected
the method are gv en in his paper in the ‘ Proceedings of the Physical

Society,’ vol. vi. part i. p. 204 (June 28, 1884). [Phil. Mag. for August
1884, p. 98. ]

See

XXVIII. The Periodic Law, as Illustrated by certain Physical
Properties of Organic Compounds.—Part I. The Alkyl
Compounds of the Elements. By THomMAs CARNELLEY,
D.Sc., Professor of Chemistry in University College, Dundee™.

(a a previous communication (Phil. Mag. [5] xviii.p.1) I

have shown how the truth of the Periodic Law may be
illustrated by means of the melting- and boiling-points and
heats of formation of the normal halogen compounds of the
elements. I shall now endeavour to show how the same law
is further confirmed in a similar manner by the physical pro-
perties of certain compounds of Organic Chemistry. For this
purpose we shall make use of the normal alkyl compounds
(methides, ethides, propides, &c.) of the elements; and in the
sequel we shall find that exactly the same relations hold good
with these compounds as in the case of the corresponding
chlorides, bromides, and iodides.

As the melting-points and heats of formation of but very
few of these compounds have been determined, we shall neces-
sarily be limited to a consideration of the boiling-points and
specific gravities.

As pointed out by Mendeljeff (Ann. Chem. Pharm. Suppl.
1872, p. 151), elements belonging to even series (except
series 2) do not combine with alcohol radicals to form methides,
ethides, &c., whilst those belonging to odd series generally do
so combine. In what follows, therefore, we shall merely be
able to take into consideration elements belonging to odd
series and to the first even series only.

Table I. contains the experimental data, with the authorities,
on which our conclusions are based. Tor the purpose of
avoiding minus signs, all temperatures are reckoned from the
absolute zero (— 273°). The specific gravities employed were
those corresponding as nearly as possible to 15° C.; but it
was not possible to take them all at a uniform temperature
throughout, as in many instances there is only one determi-
nation on record, and that was made at different temperatures
in the several cases ; whilst scarcely any of them have been
determined at, or equally distant from, the boiling-point. The
following abbreviations are employed :—Me = methyl, CH, ;
Ht = ethyl, C,H; ; Pr = normal propyl, C;H,; Bu = normal
butyl, C,H, ; Ph = phenyl, C,H;.

* Communicated by the Author.

260

Prof. T. Carnelley on the Periodic Law.
TABLE I.
Boiling Authority Specific - Authority
point gravity.
BMe, Wen ak b. 257 Frankland.
CMe, ...... 282 Lwow.
NMe, ...... 281 Vincent.
DE abe nk 249 - pe ee 5
uckton an
AIMe,...... 403 |{ _ dling.
SiMe, ....-. iA Gaara ag
Hofmann and
<< Saeee 314 { ee
Pie. i30553. 314 Regnault. —*845 (21) Regnault.
ClMe ...... 249 5 ath a E ihe
rankian eS rankland an
2nMe, 319 { sep eee \ 1386 (11) { Dope
AsMe, bole Cahours.
SMe, 351 Jackson.
BrMe 21 Perkin. 1-664 (0) Pierre.
SnMe, 351 * Ladenburg. 1314 (0) Ladenburg.
SbMe, 354 Landolt. 1523 (15) Landolt.
- Wohler and
TeMe, 305 { Dean. :
Pie ...55.; 317 Haagen. 2°264 Haagen.
HeMe, 367 Buckton. 3069 Buckton.
PbMe, 433 t Cahours. 2°034 (0) Butlerow.
Bekt, 459 F ’ aT d a
rankland ) : Frankland an
Bute 4... 309 || ana Began een ae oer
Lt ene 362 Brihl. 728 (20) Brihl.
ORES «5%: 308 Kopp. ‘735 (0) Kopp.
a eae 283(?) ae.
uckton an
Alt, ...... 467 cling. ; aitaies
riedel an ; riedel an
ae 405 | ort | 766 (23). 4.) oe
ofmann an ‘ Hofmann and
ae 401 { snot } 819 (15) { peter
ae 364 Pierre. ‘837 (0) Pierre.
“Oh 285 Linnemann. 925 (0) Linnemann.
ZnWty -....- 391 Frankland. 1-182 (18) Frankland,
413-45
AsEt, ... 4 a jaa |} Landolt. 1151 (17) | Landolt.
SeEt, 381 Pieverling.
Brkt ...... 312 Linnemann. 1468 (13 Linnemann.
SnHEt, ...... 454 Frankland. 1187 (14) Frankland.
=) ee 432 Lowig. 1:324 (16) Lowig.
lh ae b. 873 Wohler.
(oo ee 345 Linnemann. 1-944 (15) Linnemann.
Hgkt, 432 Buckton. 2°444 Buckton.
PbEt, 473 Buckton. 1585 -
Bieeet math Wisvases ad Tih ene 1°820 Breed.

* 413 Cahours,

t 373 Butlerow.

Prof. T. Carnelley on the Periodic Law. —-261
Table I. (continued).

Boiling- ; Specific .
joe Authority. gravity. Authority.

Bee..| | Bis eho
lS 429 Zander.
REE bes ot BOT Linnemann.
a re 523 Cahours.
PEs v2. 486 Pape. “762 (15) Pape.
BE... 405 Cahours. 814 (17) Cahours.
eee... 319 Linnemann. "896 (19) Linnemann.
gmPr, ...... 432 Peeear 1-098 (15) { ee
2 344 Rossi. 1:388 (0) Rossi.
Serr,:..... 498 Cahours. 1-179 Cahours.
aa 375 Linnemann. 1-747 (16) Linnemann.
HgPr, ...| 464 oa 2°124 (16) Cahours.

‘ leben an gd | Lieben and
Wes... 436 || Rone } 77s (20) {| epee?
Se 413 7 "768 (20) G
plane <-. 5: 455 Grabowsky. "839 (16) Grabowsky.
fia... 351 Reaennee 887 (20) { ee
fame, <....\) 46) Cahours.

Grbo ...... 373 Linnemann. 1°299 (20) Linnemann.
Misa. \....2- 403 - 1580 (18) se
HgBu, ...| 479 Cahours. 1°835 (15) Cahours.
m.p. 218 Behr.
= { b.p. a. 633 |Friedel and Crafts.
‘m.p. 400 | Merz and Weith.
mph { b.p. a. 573
| OPh, = .. af bs } Hoffmeister.
m.p. b. 253 || Paterno and
FEL ...4|"5 p. 358 Oliveri.

(| mp. 501

SiPh, 1| bp. a. 633 Polis.
2 See b.p. 565 Stenhouse. 1-119 Stenhouse.
De
CIPh ... { = vor | | Sungfleisch. | 1:219(0) — | Jungfleisch.
- . . J

: m.p. 331 ee Coste and
AsPh,...{| SP gaz |} “Michastis
BrPh ... { as ee \ Adrienz. 1502 (12) | Adrienz.

{ |m.p. b. 255 | Schutzenberger. ;
IPh ' be dél | Kekulé. 1640 (15) | Ladenburg.

m.p. 393 Dreher and
daa { b.p. a. 578 Otto.

RELATION 1. (a) Jf the elements be arranged in the order
of their atomic weights, then the boiling-points of their alkyl
compounds rise and fall periodically. (b) Under similar con-
ditions, the specific gravities diminish up to the middle member
and then increase to the last member of each series. (See Table II.)

The exceptions to this rule, in the case of the boiling-points,
occur either at or near the maxima or minima (i. e. at the

262

Prof. T. Carnelley on the Periodic Law.

turning-points) of Meyer’s curve of the elements (Mod. Theor.
der Chem.), in which respect they resemble those correspond-
ing halogen compounds of the elements which are exceptions _
to the same rule.

The above facts are shown in the following Table :—

TaB.e I1.—Illustrating Relations 1 and 2.

METHIDES. ETHIDEs.

GP. | 5p. er.|| B.P. | Sp. gr
Li
io: Ti eee ADD wise
ee nee 569 | °696 |
scent BOeiectess || gs-seth | leak
a ts) aoe 362 | “728
Bdecks 2 or 308 | “735
ee eee Daa ceva.
Na
Meg.
Al Es ae Bf © | ovate
Si BN ites nie: 425 | *766
ae al: Si ee AOL |) GAZ
Be uct. 310 | °845 || 364 | :837
Cl ra! al eee 285 | °925
Cu
Zn ...| 319 | 1°386 |} 391 | 17182
Ga ...|
EKka-Si
BB... S10 | «sees 432 |1:151
|e es se ee 38]
Br 277 | 1-664 || 312 | 1-468
Ag
Cd
in.
Sn ...) 351 | 1314 || 454 | 1187
Sb ...| 354 |1:523 || 432 | 1:324
Te Be sax ak b. 373
Ee O17 | 2°264 || 345 | 1:944
Au |
He 367 | 3:069 || 4382 | 2-444
Tl |
Pb 433 | 2-034 || 473 | 1-535
og Oe ERO eee [ sseees | 1-820 ||

|

| a-PROPIDES.

a-BUTIDES.

PHENIDES”*.

Sp. gr.) B.P. | Sp. gr.J M.P. | B.P. | Sp. gr.

BP.
518
B90) iat
B57 | ect
523
486 | -762
405 | -814
319 | -896
432 | 1-098
344 | 1-388
498 | 1-179

| 875 | 1-747
464 | 2-124

esgits nee A418 Ja. 633
486 | ‘778 | 400 |a. 573
413 | "fos eo 519
ay era ake b. 253 | 358
ee 501 | 633
455 | 1850 7 Nee 565
351 | ‘887 | 233 | 404
461

ute Gir eeiseee ddl | 633
373 | 1-299 fb. 253°| 427

| fC.

1-502

ee

479 393 -|a. 573

1°835

* The phenides of course are not comparable with the methides, ethides, &c.,
though they appear to obey the same rules.

RELATION 2. The boiling-point increases and the specific
gravity diminishes as we pass from the methide to the ethide and
thence to the propide, butide, §c. (See Tables II. and IIT.)

Revation 38. The boiling-points and specific gravities of
the alkyl compounds of any one group increase as the atomic
weight of the positive element increases. (See Table IIL.)

Prof. T. Carnelley on the Periodic Law. 263
TasB_eE [1].—Illustrating Relations 2 and 3.

MeETHIDES. ErHIpEs. | a-PROPIDES. a-BUTIDES. PHENIDES.

—EE 1 EEE Hie eee

EP. ‘Sp. er.|| B.P. | Sp. gr.|| B.P. | Sp. er.|| B.P. | Sp. gr] M.P. | BP. | Sp. er.

oo

. 1 ae o85 | -925 || 319 | -896 || 351 | -887 1 233 | 404 11-129.
Br ...| 277 |1-664 | 812 |1-468 }) 344 [1-388 || 873 | 1-299 p.253 | 427 | 1-502
eee 317 | 2-264 || 345 |1-944 || 375 |1-747 || 403 | 1-580 fb.255 | 461 | 1-640
pes. ee ton | ee 4ee- f ver. | | 501 | 633

Sn ...,| 301 {1314 || 454 |1:187 || 498 |1:179
Pb ...} 433 | 2°034 || 473 | 1°585

rs

Zn ...| 319 | 1386 |} 391 | 1:182 || 432 | 1-098 || 461
Hg ...| 367 |3°069 || 432 | 2-444 || 464 | 2-124 || 479 | 1:835 393 a. 573

a i aie 401 | -812
eee | 432 |1-151

Sb ..| 354 |1-593 |] 432 | 1-394 |
0 ae ae 1-820 | |

os .... 310 | 845 || 364 | -837 |

en est | ...... 381

ee ae [b.373(2) |

RELATION 4. The differences between the boiling-points (and
also between the specific gravities) of the methides and ethides,
of the ethides and propides, of the propides and butides, &c.,
increase algebraically from the beginning of each series up to the
fourth or middle (tetrad) member and then diminish to the
seventh or last member. (See Table IV., which gives the avail-
able data for the Second Series.)

Taste TV.—Diif. of Boiling-points. Illustrating Relation 4.

Ethide | Propide | Butide | Propide| Butide | Butide
minus | minus | minus | minus | minus }| minus
Methide.} Ethide. | Propide.{| Methide.| Ethide. | Methide
Diff. in atomic weight} ~—————--+-—_ —j ——— —
me alkyl radieals...).....:..s.: 14 42
ee
_ Ee ee ee 7
Ee 2h 5 wine's nino an 65 BIG PES we Hee
+L SUS ee 122 GENS eer
= 2 87 Satie Se
1 AOR Bae pie ieee 54 4l 50 95 91 145
> Lee eee 3 34 32 70 66 102

RELATION 9. Lhe differences referred to in (4) increase on
the average in nearly the same proportion as the difference
between the atomic weights of the alkyl radicals (see Table V.) ;
whilst in the case where the difference between the atomic weights
of the alkyl radicals is the same, the difference between the boiling-
points either tends to become equal, or diminishes as the atomie
weights of the alkyl radicals increase (see Table 1V.; also com-
pare Kopp’s Law of the Boiling-points of Homologous Series).

Prof. T. Carnelley on the Periodic Law.

264

L¥T lel= =6XG-EL §8T 681=6XG-16
_ 96h GEL =1XG-Eh _ 16 G16 =1X¢-16
.——————————_ wo. ~~ Y
"ug Ig
G0G POG=E X89
OGT OGL=c X09 ET 9E1=G X89
09 09 =T X09 89 89 =1 X89
SS Se ye
TY “AT
crl IVL=EX LP CFL IVIL=EX LP POL CIT=EX FG
66 v6 =GX LP &6 16 =C6X LP LOT SOL=ZX FE
Li Ly =1XLy CP Ly =1XlP PG FG =I XFG
‘ Ee eh peek Oe et re ee —
"UZ iS ‘O
18 18=€ X 66 96 96=€ XGE GOL COL=EX FE
8g SS=GX 6G 9 TI=CXGE - 89 89 =ZXTE
66 66=1 X66 oe CS=1XGE VE VS =LAPE
‘uvoul : "UvOUl “UvoOTL ;
jeyuountzedxry | POFPINOTPO [equowtiedxq “peye[noyen jequounsedxgy peyeorey

‘sjulod-suljiog Ut soueteyIq

‘sjulod-surpiog ul soueLey ICT
~~ ——— ee ee

= _— \~—

T 1g

—_—

—

‘syulod-Surplog Ul eoue1egIG

—sV-. ——

10

Gh=EXFL a Se ee ee eee oe

SG=SXFL [t+ Uae XO) y—¥(s+usp] StU) y

PIS=LXFL [ite O)y—"(et+ueq ttup)y
=X

or= - opeexer " S(I+UaHy UD )y —8(Ltuszy 8+UD)y

SC=SXPFL | (+My XO) x— "(s+ ue t+ug)y

PISLXPL [oe (tl +ueeup)y —*(st+usqy tt+ug)y
=YX

GP=EXFL ft (+e) x—(2+ ue &+UQ)X

SG=GXFL | (+4 YO) Kx — "(s+ 4a s+%O)y

PI=ITXPL oft “(tee op) x — “(8+ ueq t+up)y
=

LPH EK Pp [serene I+UsHy MQ x —L+ueH &+uy ¥

SG=ZXPFI - or Tee “Oo sg

PSs UTeeHy 49 Xe ep oy

‘sTvotpea [Aye JO

SJYSIOA OTULOFL

‘UL SDUEIOTIC,

=.

San Sec ec 8 Ne es a ee ee i a

"Gg uoTyLOY Jo yaed Suryeaysn]]T— A ATAV,

Profs 2. Carnelley on the Periodic Law. 265

RELATION 6. The differences referred to in (4), for both
boiling-points and specific gravities, diminish algebraically as
the atomic weight of the positive element increases. (See

Table VI.)

: TABLE V1L.—Illustrating Relation 6.

14 | 28 49

2 es Se

a eo ae Ee ON |e Se So
ee ote.) C.H,—OF-./\C,H'—C.H, || C,H,—CH,. || C,,—C.H, || C,H,—CHy

ee. jap. gr. | B.P. Sp. er.|| BP. |Sp.gr.|| B.P; |Sp. gr/|| BP... Sp, gr. || B.P. | Sper
34 |—-029|| 32 |—-009 66 |—-038
32 |—-080|| 29 |—-089 | 61 |—°169

— 320, Bit |= 197 |) 98. \—-167 —-517|| 58 |—-364|| 86 | —-684

RELATION 7. (a) The differences between the boiling-points,
and also between the specific gravities, of the methides, or ethides,
or propides, or butides, Sc. of the elements of the same group
diminish as we pass from the methides to the ethides and thence
to the propides, &e.

— (b) The above differences, as regards the boiling-points, in-
crease as the difference between the atomic weights of the two
positive elements increases. (See Table VII.)

266 Prof. T. Carnelley on the Periodic Law.

Taste VII.—Illustrating Relation 7.

| Diff. of |
atomic MeETHIDEs. ETHIDES. PROPIDES. ButivDEs. PHENIDES.
I weights |
: of ele- re ee eens |
mt | BP. Sp. gr. B.P. | Sp. gr.|| B.P. | Sp. gr.|/| B.P. | Sp. gr.J BP. | Sp. gr.| |
eer) 445 || 28 | ...... 27 | 543 || 25 | -492 || 22 | 412 | 23) aay
Past 2.4 470 40 ‘600 oo ‘476 ob “O09 30 281 34 "138
Pa). 9175 68 pes Os ol 2 feaad Pia AF ie. cee 5? ice 57
Pb-Sn...; 89 82(?)} °720 19 398
Sn-Si ...) 90 ona Eee 29 491 12 “417
Pp si -:.| 179 | 130 uanoss 48 "784
Hg-Zn..| 135 48 | 1:683 4l | 1-262 32 |1:026 18

RELATION 8. (a) The differences between the boiling-points
(and also between the specific gravities) of the methides, ethides,
propides, &¢. of the elements of the seventh or halogen group and
those of the elements of groups il. to vil. respectively merease
algebraically from the methides to the ethides, and thence to the
propides &c. (See Table VIII.)

(b) With members of the second group (Zn, Hg, &e.) these

differences as regards the specific gravities dinunish algebraically
from methides to butides. (See Table VIII.) Whilst as regards
the boiling-points of the second group, the first part of the
rule (a) seems to hold good, except in so far as a tendency to
a reversal of the rule appears as we approach tbe butides and
higher alkyl compounds ; and this reversal becomes the more
pronounced the greater the difference in the atomic weights of
the metal and the halogen. Data are entirely wanting for
the first group.

RELATION 9. The differences referred to in Relation 8, for
both boiling-points and specific gravities, increase algebraically
as the algebraic difference between the atomic weights of the
positive elements increases. (See Table VIII.) 3

Prof. T. Carnelley on the Periodic Law.

7 267
i‘
; TapLeE VIII.—HIllustrating Relations 8 and 9.
Diff. of
atomic || MeETUHIDES. ETHIDEs. PROPIDES. BUuTIDES. PHENIDES.
I weights
ig of ele- ||— |
ag B.P. | Sp. gr. || B.P. | Sp. gr. || B.P. | Sp. gr. | BLP. Sp orf B.P. |Sp. gr.
—95 — 7 |—1419)| +19 |—1:107) +30 | —-933 || +52 |—-741)+104 |—-521
—48 2 te 36 :
—44-5 ye eee aS, 52 t= Gor 61 ee. | 82 |—-4604 138 |—°383
= ch a Se anee 79 }— -088 86 | —"082 | 104 |—-048]7 161 |—-010
— 2 a 2g een 28?
— | SY ih a 69
Bass 82| ...... 96
i ee 61?
89:5|| 106 | ...... 88?
— 96 2: eee +56 |—1:132
= 92 ib. +56) ...... Oh = Ta oe te Se Eee amremek AR
—49 | Se eee 89 |— °656
=7 || 37|—741 | 87 |— 620
eae GP LL... HGHes tS
— 5 OL 1) Res a ee ee | een rere 206
eos i be E24 be 5 e.:.s BB aay ee Te ce Ie Ste 229
40 771-141 120° }—._ *144
84:5|| 105 | ...... 147 |+ -399
== 99 =) 2 eee + 80 |—1:178 PEt = OS both. prea 172
ae 6 | == | Pie fe ee oe ee 206
...| — 9 34 | —-950 109 |— °757 123.) —°568
Jer A? een $40 1 1595) 1671 = 7184 4h Aol Ae el BO
+38 TA | —-350 142 |— -281 154 | —°209
SO 116 | —°230 128 |— “359
82:5 st 169 |+ °242 179 | +:283
127 156 |'-1-370 161 |+ °117
ae Post be on. 188 |+ °640
ey = 30° |) + 417 |=F-916 By 83 |—*802 fa. 112
— 66 | + 41] ...... 50 |— °740 Sob h > ze veres 113 |—°521 fa. 146
=2l-b =e See 77 |— °197 fh A ee ea 135 |—°109 fa. 169 -
—Jil || —68 | ...... ip te tS Pe +10 |—°812 58
Ee We AO lek wanes = FAM Foot! bho b isese 40 |—°531 92
—19°5 Obl x), se + 23 |— °190 Sie eee 62 |—119] 115
Zn-1 == 62 2 | —‘878 46 |— ‘762 57 |— °649 58
Znu-Br er 42 | —°278 79. |— 286 88 |— °290 88 |
Zn-Cl +-99-7 Te es 106 |+ :237 1138 |+ °202 110 :
Heg-I 10 50 I+ °805 87 |+ °500 89 |+ °377 76 |+°255 fa. 112
/Hg-Br 120 90 |+1°405 120 |+ ‘976 120 |+- -736 106 |+°536 fa. 146
164°5)) 118 147 |+1-499 144 |+1:228 128 |+-948 fa. 169

HPO] eee

The above 9 Relations can at present be applied in 942
cases, of which 54 (or about 5°7 per cent.) are exceptions.
These are distributed as follows :—

268 Prof. T. Carnelley on the Periodic Law.

Boiling-points. Specific gravities.
Number of Number of
cases in cases in
which the hae of which the Number of
Pace exceptions, wale: 4a exceptions.
applicable. applicable.
Belstion 13.9... 2:. is 8 30 1
LD ee 492 ) 35 3
: es ae 46 2 at 0
a Ne ee 42 1 16 al,
a Bene 7 51 0 ee ase
$ adn) Tapes 43 0 25 0
ey MON vans 35 0 14 )
5, WOT ssc: 33 3 Hi. J
zs oh eee 104 3 45 S
EF SO 2s cINER | Ratdekes > oars pakke 19 2
ES Sa ae 180 20 %8 7
Matar sc. ....0. 649 37 _ 293 17
=5°7 per cent.| =5°8 per cent.

For the reasons stated in my former paper (ibid. p. 10) this
is avery small proportion of exceptions ; and even of these no
less than 14 are due to the boiling-points of TeHt, and SbEt;
being too low, and 8 due to the specific gravity of PbHt,
being too low and those of NBu; and Sikt, too high. ‘In the
case of the specific gravities, some of the exceptions are
undoubtedly owing to these data not being all strictly com-
parable with one another, since they have neither all been
determined at the same temperature, nor at equal distances
from the boiling-points.

The above facts show, therefore, that the physical properties
of the alkyl compounds of the elements (so far as they have
been investigated) obey exactly the same rules as those of the
corresponding halogen compounds, and would consequently
allow of general conclusions being drawn and practical appli-
cations being made similar to those indicated in my previous
paper (ibid. pp. 11, 14, and 19), in connection with the latter
compounds, and which it will not be necessary to repeat here.

In my next communication I hope to extend these investi-
gations to the halogen compounds of the hydrocarbon radicals,
and subsequently to show how the facts thus obtained may
throw light on the nature and raison d’étre of the Periodic
Law.

rg
ie

[ 269 ]

XXIX. Origin of Coral Reefs and Islands.
By James D. Dana, LL.D.

[Continued from p. 161, ]

Part II.— The Objections considered.

5 haa objections to the Darwinian theory may be considered
in the following order :—

I. Darwin’s insufficient knowledge of the facts bearing on
the subject. |

IT. Subsidence not ordinarily a fact, because methods of
producing barrier reefs and atolls have been brought forward
that do not require its aid.

ITI, The occurrence of cases of elevation in regions of atolls
and barrier-reefs inconsistent with the subsidence-theory.

IV. No ancient coral-reefs in the geological series have the
Sey thickness attributed by the subsidence-theory to modern
reefs.

V. Other methods of explanation and their supporting
evidence. :

The adverse remarks directed against the idea of a sinking
continent in the Pacific as the initial condition in the coral-reef
subsidence are outside of the present discussion for the reason
stated on the first page of this paper. In the following pages
the objections are first explained, under the above-mentioned
heads, and then follow, in paragraphs lettered a, 6, c, &e., the
writer’s discussions of the several points.

I. Darwin’s Insufficient Knowledge of the Facts.

In the Address referred to in the opening page of this
article, Dr. Geikie, speaking of Darwin, observes :—“ lt
should be borne in mind that, compared with more recent
explorers, he did not enjoy large opportunities for investi-
gating coral-reefs.” “He appears to have examined one
atoll, the Keeling Reef, and one barrier-reef, that of Tahiti.”
“By a gradually widening circle of observations a series of
facts has been established which were either not known, or
only partialiy known, to Darwin.’’—The authors appealed to
for the views that are presented as a substitute for Darwin’s
are Prof. Karl Semper, who has examined and described reefs
of the Pelew and Philippine Islands; Dr. J. J. Rein, who
has published on the physical geography of the Bermudas ;
Prof. Alexander Agassiz, who has written on the Florida
reefs and others in that vicinity; and Mr. John Murray, of
the ‘Challenger’ Expedition, whose investigations were made
at Tahiti: all able men in science, whether more learned or

Phil. Mag. 8. 5. Vol. 20. No. 124. Sept, 1889. U

270 | Dr. J. D. Dana on the Origin of

not than Darwin on the special subject under discussion.
The facts from ‘‘a widening circle of observations ”’ referred
to comprise the physical and biological results of deep-sea
exploration. The writer is mentioned as one of the “ com-
petent observers’ who had given “‘ independent testimony ”’
in favour of Darwin’s views after “at least equal opportunities
of studying the subject,” and as he has, in these latter years,
looked into the new facts, he has at least a claim to a hearing.

As to Darwin’s knowledge, it appears to the writer that the
apology offered in the above citations was not needed. In
his detailed investigation of Keeling atoll (a good example
of atolls, and like all the rest in its principal features) and
in his examination of the Tahitian reefs, followed up by a
careful study of other atolls and reefs of the ocean through
the maps and descriptions of former surveying-expeditions,
he had a broad basis for judgment and right conclusions.
When the second edition of his work was published in 1874,
many of the important facts from deep-sea exploration were
already known ; and later he learned of the more recent
results ; and he did not recant. A letter of his, of October
2nd, 1879, published by Mr. Semper, while admitting with
characteristic fairness the interest of the facts collected by the
latter, expresses his continued adherence to the opinion “that
the atolls and barrier reefs in the middle of the Pacific and
Indian oceans indicate subsidence.”

The writer, as his expositor, may be excused for adding
here that his own “independent testimony ’’ was based on
observations among coral reefs and islands in the Pacific
during parts of three years, 1839, 1840, 1841; that, besides
working among the reefs of Tahiti, the Samoan (or Navi-
gator) Islands, and the Feejees (at this last group staying
three months), he was also at the Hawaian Islands; and in
addition, he landed on and gathered facts from fifteen coral-
islands, seven of these in the Paumotu Archipelago, one,
Tongatabu, in the Friendly Group, two, Taputeuea and Apia,
in the Gilbert Group, and five others near the equator east of
the Gilbert Group—Swains, Fakaafo, Oatafu (Duke of York’s),
Hull, and Enderbury Island*. The writer may therefore be
acquitted of presumption if he states his opinion freely on the
various questions that have been brought into the discussion
by other investigators. Sympathizing fully with the senti-
ment expressed in the words, “* The example of Darwin’s own
candour and over-mastering love of truth remains to assure

* These five islands are on the map of the Central Pacific accompany-

ing Part I. of this paper, Hull’s Island is “Sydney” of the writer’s Expe-
dition Report. |

Coral Reefs and Islands. 271

us that no one would have welcomed fresh discoveries more
heartily than he, even should they lead to the setting aside of
his work ;’? and knowing that we are all for the truth and
right theory, he has reason to believe that those who have

_ been led to object to Darwin’s conclusions will be pleased to

have their objections reviewed by one who has a personal know-
ledge of many of the facts.

II. Subsidence not ordinarily a Fact, because Methods of Origin
have been brought forward that do not require its aid.

It is urged that, while subsidence may have happened in
several cases, it is not at all necessary to the making of
barrier-reefs and atolls: that “subsidence has been invoked
because no other solution of the problem seemed admissible ;’’ ©
that the “solution’’ by subsidence “is only an inference
resting on no positive proofs” *.

a. Darwin’s usual methods were not such as these words
imply, and we think that he was true to those methods in his
treatment of coral-island facts. Darwin can hardly be said to
have “‘invoked’’ subsidence. Subsidence forced itself upon
his attention. He saw evidence that it was a fact, and the
theory came ready-made to him. The proof of subsidence from
the relations in form, structure, and history between atolls
and the large barrier-islands, like the Gambier Group, Raiatea,
Bolabola, and Hogoleu, scarcely admitted, he says, of a doubt;
and other facts were all in harmony with it. This, his chief
argument, with the enforcing evidence in my Report (see §$
4 and 10 of Part I. of this paper) is not set aside and not men-
tioned in the Address from which the above sentences are
cited.

b. Darwin observes that ‘from the nature of things it is

scarcely possible to find direct proof of the subsidence,”

recognizing the fact that subsidence, unlike elevation, puts
direct testimony out of sight. But still it has left evidence
which he perceived and thought convincing ; and this stands,
whatever virtue there may be in other explanations.

Moreover, we have now direct testimony for subsidence
from the facts brought forward (for another purpose) by Mr.
Murray, as is set forth further on.

Ill. The Occurrence of cases of Elevation in Regions of Atolls
and Barrier-Reefs.
The fact that elevated reefs and other evidences of eleya-

* Address, page 24.

aie Sa a

— 272 Dr. J. D. Dana on the Origin of —

tion occur at the Pelews, a region of wide barrier-reefs and

atolls, has been presented by Prof. Karl Semper*, after a
study of those islands, as an objection to the theory of subsi-
dence ; for we have thereby (in the words of the Address)

‘a cumbrous and entirely hypothetical series of upward and
downward movements.” Prof. Semper reports the existence

of reefs raised 200 to 250 feet above the sea-level in the
southern third of the larger of the islands, while the other
two thirds exhibit evidence of but little, if any, elevation.

a. Such facts are of the same general character with those
of other elevated reefs and atolls discussed in $§ 12,13, 16 of
Part J. and the same explanation covers them. The Pelew

region is one of comparatively modern volcanic rocks, and this’

4

renders local displacements a probability.
6. The occurrence of great numbers of large and small

masses of coral-rock, in some places crowded together, upon

the western or leeward reef of the several Pelew Islands, and
of none on the eastern reef, is mentioned as evidence against
subsidence and in favour of some elevation; because, Professor
Semper says, the strongest wind-waves on the western side

are too feeble to break off and leave on the reef such large

masses, some of them (as his words imply rather than

distinctly state) ten feet thick.
But the difficulty does not exist in fact ; for earthquakes

may have made the waves. The region just west of the

Pelews is one of the grandest areas of active voleanoes on the
globe. It embraces the Philippine Islands, Krakatoa and.

other volcanic islands of the Sooloo sea, Celebes, &c. The
agents that could do the work were there in force. To the

eastward, in contrast, lie the harmless islands of the Caroline

Archipelago, mostly atolls, serving, perhaps, as a breakwater

to the Pelews.

The small elevation referred to is therefore not proved by
the evidence adduced ; and yet it may be a fact without
affecting the theory of Darwin, as I have fully illustrated fT.

It is important to have in mind that the coral-reef era
probably covered the whole of the Quaternary and perhaps
the Pliocene Tertiary also; and hence the local elevations that

* First in 1868, Zeitschr. Wissensch. Zool. xiii. p. 558; additions in Die

Philippinen und thre Bewohner, Wirzburg, 1869; and still later in his
‘Animal Life,’ published in Appleton’s International Scientific Series in
1881.
+ Mr. Semper’s objection to the theory of subsidence based on the co-
existence of all kinds of reefs in the Pelews—atoll, fringing, and barrier—
with no reefs about one island, and from the relative steepness of the
submarine slopes on the east and west reefs of an island, have been
sufficiently met in Part I.

a a

Coral Reefs and Islands. 273

have taken place in the ocean were not crowded events of a
short period.

- Moreover, these local elevations in coral-seas are spread
over an area of 25,000,000 square miles. As an example of
the long distances, the Paumotu Archipelago, consisting of

_ more than eighty atolls and two barrier-islands, and covering

about 450,000 square miles, contains only three or four
atolls that are over twelve feet high; and of these, Metia is
250 feet in height, Elizabeth 80 feet, Dean’s, probably where
highest, 15 or 20 feet. Metia is one of the westernmost, near
148° 13’ W. and 15° 50’ S.; Dean’s is 60 miles to the north-
north-east of Metia, and Elizabeth is far to the south-east, in
128° W. and 25° 50’ S., or nearly 1450 miles distant from
Metia. Locate these points on a continent, and Pacific
distances and the length of Pacific chains of atolls will be
appreciated.

IV. No Ancient Coral-reefs have the Thickness attributed by the
Subsidence- Theory to Modern Reefs.

An argument against the subsidence-theory is based by
Prof. J. J. Rein* on the alleged fact that the thickness
attributed to modern reefs is far beyond that of any such
reefs in earlier time ; that is, the thickness is unprecedented.
The argument decides nothing. The question is one of
geological fact, not to be settled by a precedent. Whether,
then, there are precedents or not it is not necessary to
consider. |

Besides this, it implies a distinction between coral-made
and shell-made rocks which does not exist. The coral-reef
rock is largely made of shells, and the process of formation
for a limestone of shallow-sea origin is essentially the same
whether shells or corals are predominant or the sole material.
No thick formation of any kind of rock was ever made, or
could be made, by shore or shallow-sea operations without a

slowly continued subsidence or a corresponding change of

water-level.

V. Other Methods of Explanation, and their Supporting

Evidence.

A. Mr. John Murray, one of the able naturalists of the
‘Challenger ’ expedition, reports the following important results

* Dr. Rein’s first memoir on Bermuda appeared in the Senckenberg
Ber. naturforsch. Gesellschaft, 1869-70, p. 857, and the later in the
Verhandlung des I. deutsch. Geographentages, 1881, Berlin, 1882. The
above argument is from the latter paper, and is given here from the
citation by Dr. Geikie, the publication not being accessible to the writer.

274 Dr. J. D. Dana on the Origin of

from soundings off northern Tahiti, made under his super-
vision and that of the surveying officers *. :

Along a line outward from the edge of the barrier-reef there
were found :—(1) for about 250 yards, a shallow region _
covered partly with growing corals, which deepened seaward
to 40 fathoms ; (2) for 100 yards, between the depths of 40
and 100 fathoms, a steeply but irregularly sloping suriace,
which commenced with a precipice of 75° and had a mean
angle exceeding 45° t+; then (3) for 150 yards a sloping bottom
30° in angle ; (4) then a continuation of this sloping surface,
diminishing in a mile to 6°, at which distance out the depth
found was 590 fathoms (3540 feet). Over the area (2), or
the 100 yards between 40 and 100 fathoms, the bottom was
proved to be made of large coral-masses, some of them “ 20
to 30 feet in length,” along with finer débris ; outside of this,
of sand to where the slope was reduced to 6°; and then of
mud, composed “ of volcanic and coral-sand, pteropods, pelagic
and other foraminifers, coccoliths, &c.”’

These observations have great significance. They show (1)
that the feeble currents off this part of Tahiti carry little of
the coral débris in that direction beyond a mile outside of the
growing reef ; (2) that a region of large masses of coral-rock
and finer material occurs at depths between 240 and 600 feet ;
(3) that, a mile out, the bottom has the slope nearly of the
adjoining land, and in this part is covered with the remains of
pelagic life. |

From the second of these facts—the great accumulation of
coral-blocks below a level of 240 feet—Mr. Murray draws the
conclusion that, in the making of fringing, barrier, and atoll
reefs, the widening goes forward (a) by making first upon
the submarine slopes outside of the growing reef a pile of
coral-débris up to the lower limit of living reef-corals ; and
men (b) by building outward upon this accumulation as a

ase.

He also announces, after speaking of other causes influencing ~
the growth of corals, the more general conclusion that “it is
not necessary to call in subsidence to explain any of the
characteristic features of barrier-reefs and atolls;’? and con-
cludes that his views “do away with the great and general
subsidences ” appealed to by Darwin.

a. Ihe widening-process, in the first conclusion, had
previously been a part of the Darwinian theory; for, as

* Proc. Edinburgh Roy. Soc., Session 1879-80, p. 505.
Tt Dr. Geikie gives in his paper a section of the soundings, “ on a true

scale, vertical and horizontal,” and in it the upper steepest part of this 100
yards has a slope of about 75°.

Coral Reefs and Islands. 275

stated in § 10 (Part I.), a fringing reef, where no subsidence
is going on, widens above and steepens its seaward-slope, and
it could do this only by the process described: that is, by
building out upon a base of débris, or, more correctly, upon
true coral-reef rock made by the gradual consolidation of the
débris *.

6. The broader conclusion Mr. Murray does not sustain by
a mention of special facts from the soundings, tending directly
to meet the question of change of level, but by attempting to
show that through the eroding action of currents and other
means (as had been argued by Prof. Semper), in connection
with the process already explained, reefs of all kinds can be
made from submarine banks without aid from subsidence.

In this place I confine myself to the question as to the fact
of subsidence. The only direct argument presented against
subsidence is contained in the statement, that the very broad
shore-plain of Tahiti shows that “ the island has not in recent
times undergone subsidence,” and may indicate a slight
elevation ; and in this he sustains the earlier statement of my
report, which says (p. 293) that the broad shore-plain of
Tahiti probably overlies in some parts the fringing reef ;
and (p. 300) the shore-plain, if built upon reefs, as I was
assured, may afford proof of a rise of one or two feet.” But
this admission, as [ have explained for other cases of local
elevation, is in no way opposed to the theory of subsidence.

ce. The kind of submarine slopes to be looked for off reefs
is illustrated by the soundings, as Dr. Geikie indicates. But
it is interesting to note that the facts, while very important,
sustain instead of correcting those announced by earlier
observers. Beechey and Darwin make the mean slope about
45°, and my report says 40° to 50°. I have assumed for the
slope of the bottom. outside of the reef-limit the same angle as
for the surface-slope of the island just above the water-level :
5° to 8° off Tahiti, of which 5° is accepted as most correct,
and 3° to 5° off Upoluf ; and the assumption as regards Tahiti
is sustained by the ‘ Challenger ’ soundings. My Report states
(from the Expedition surveys) that off Upolu the bottom
“loses more and more in the proportion of coral-sand till we
finally reach a bottom of earth,” and introduces this as an
argument against the indefinite drifting of coral-sands into
the deep ocean { ; and this argument the Tahiti soundings
sustain.

With reference to the occurrence off some shores of precipi-

* My Expedition Geological Report, pp. 181, 182, where figures are

given illustrating the effect of widening.
+ Page 47, { Page 154.

276 — Dr. J. D. Dana on the Origin of

tous submarine slopes, the ‘ Challenger ’ soundings give definite
facts as to one case. They leave undisturbed the previously
reported cases of like steepness at greater depths: for
example, the sounding of Captain Fitzroy at Keeling atoll
(while Darwin was there), 2200 yards from the breakers, in
which no bottom was found at a depth of 1200 fathoms, but
the line was partly cut at a depth between 500 and 600 fathoms ;
the sounding by the Wilkes Expedition off Clermont Tonnere
(Paumotu Archipelago), where the lead, brought up an instant
at 350 fathoms and then dropped off again, descended to 600
fathoms without reaching bottom, and came up bruised, with
small pieces of white and red coral attached ; a sounding by
the same Expedition, a “ cable’s length” from Ahii, in which
the lead struck a ledge of rock at 150 fathoms and brought
up finally at 300 fathoms”*. All the older soundings need to
be repeated ; but there must be enough truth in those quoted
to warrant the remark that the force of Darwin’s argument
for subsidence from the steepness of the submarine slopes
about atolls is not weakened by the ‘ Challenger ’ results.

d. But the chief interest of the ‘ Challenger ’ soundings con-
sists in their affording “ direct’’ proof, “ positive’ proof, of
much subsidence ; a kind of proof that subsidence sinks out of
sight, and which soundings may yet make available in many
similar cases.

That belt of coarse débris—including “masses 20 to 30
feet °’? long—was found over the steeply sloping bottom at
depths between 240 and 600 feet. These depths are far
below the limit of forcible wave-action. They are depths
where the waters, however disturbed above by storms, have
no rending and lifting power, even when the bottom is
gradually shelving ; depths, in this special case, against a
slope which for 100 yards is 75° in its upper part, and in no
part under 45°, the vertical fall being 360 feet in the 100 yards.
Strokes against the reef-rock thus submerged, and under such
conditions, would be extremely feeble. Waves advancing upa
coast, whether storm-driven waves or earthquake waves, do
little rock-rending below the depth to which they can bare
the bottom for a broadside plunge against the obstacle before
them, although the velocity gives them transporting power
to a greater depth. It is the throw of an immense mass of
water against the front, with the velocity increased by the
tidal flow over a shelving bottom,—the rate sometimes
amounting, according to Stevenson, to 36 miles an hour or
52°8 feet a second,—together with the buoyant action of the
water, that produces the great effects.

* Ibid, p. 59.

Coral Reefs and Islands. 277

A vertical surface below the sea-level of 20 feet made bare
for the broadside stroke is probably very rarely exceeded even
in the case of earthquake-waves ; and with storm-waves, or
recorded earthquake-waves, the displacement of the water at
a depth of 240 feet would be at the most only a few inches.
I saw on atoll reefs no upthrown masses of coral-rock over
ten feet in thickness and twenty feet in length or breadth.
It is therefore plainly impossible that such a belt of débris
should have been made at its present level, or even at a depth
of 20 feet ; and hence the débris affords positive proof of a
large subsidence during some part of the reef-making era.

The existence of the belt of débris may be explained as
follows :—-If the reef now at a depth of 240 feet were at the
sea-level as the sea-level reef, and subsidence were not in
progress for a period, the very steep front of the reef now
just below the 240-foot level might have resulted from the
widening that would have gone forward. And under such
conditions, the action of the occasional extraordinary waves
might have torn off masses from the front which would have
tumbled down the steeply sloping surface until the belt of
debris had been formed. Then, with a renewal of the slow
subsidence, the thickening of the reef would have been
resumed and gone on to its final limit, and the rendings of
the great waves found lodgment at higher levels. The masses
now on atoll reefs must be from comparatively recent up-
throws. .

This direct evidence of subsidence from Tahiti renders it
reasonable to make subsidence in atoll-making a general
truth. It is nevertheless desirable that facts of the kind
should be multiplied. The abrupt descent in the submarine
slopes of reefs detected by Fitzroy at a depth below 3000 feet,
and those reported by the Wilkes Expedition at depths of
2100 and 900 feet, seem to indicate a similar rest at the sea-
level, and consequent reef-widening in the course of a pro-
gressing subsidence ; and proof of this may yet be found in
belts of coarse coral-rock débris at the foot of the precipices.
Such a period of rest would lead to the forming of submarine
precipices in different regions contemporaneously at different
depths according to the rate of subsidence of the part of the
subsiding area.

B. From facts observed about the Florida reefs, Lieutenant
KE. B. Hunt, U.S.N., announced, in 1863%*, the conclusion
that these reefs had received their westward elongation

* Silliman’s American Journal [2] xxxv. p. 197.

oS a) a
in w
rere oct
Sy
we »,
«

278 Dr. J. D. Dana on the Origin of

through the westward “sweep’’ of an eddy current to the
Gulf Stream. The subject, nearly twenty years afterwards,
was more thoroughly investigated by Mr, Alexander Agassiz,
and the same conclusion reached *. Mr. Agassiz made also
another important observation—that this current is an abundant
carrier of marine life for the feeding of the coral animals, and so
accelerates the coral-growth and accumulation in its direction.
Combining with these effects others hereafter considered, Mr.
Agassiz expresses, like Mr. Murray and Mr. Semper, the
further conclusion, that all kinds of reefs—atoll, fringing, and
barrier—may be made without aid from subsidence.

a. The facts presented by Lieutenant Hunt, and more
fully by Mr. Agassiz, with regard to the effects of the eddy
current of the Gulf-Stream, show that coral-reefs may be
elongated, and also that inner channels may be made, by the
drifting of coral-sands. But the action with coral-sands is
essentially the same as with other sands ; and illustrations of
this drifting process occur along the whole eastern coast of
North America from Florida to Long Island. We there
learn that drift-made beaches run in long lines between broad
channels or sounds and the ocean ; that they have nearly the
uniform direction of the drift of the waters, with some irregu-
larities introduced by the forms of the coast and the outflow
of the inner waters, which are tidal and fluvial and have much
strength during ebb tide. The easy consolidation of coral-
sands puts in a peculiar feature, but not one that affects the
direction of drift accumulation.

b. The great barrier-reef off eastern Australia, a thousand
miles long, has some correspondence in position to the sand-
reefs off eastern North America. ut it is full of irregulari-
ties of direction and of interruptions, and follows in no part
an even line. In the southern half, it extends out 150 miles
from the coast and includes a large atoll-formed reef; in the
northern half, the barrier, while varying much in course, is
hardly over 30 miles from the land. ‘There is very little in its

* On the Tortugas and Florida Reefs, by A. Agassiz, Trans, Amer.
Acad. xi. 1883.

Professor Louis Agassiz’s account of the Florida reefs was published
in the U.S. Coast Survey Reports of 1851 and 1866, and reproduced in
vol. vii. of the Memoirs of the Museum of Comparative Zoology. It gives
an excellent description of the Florida reefs, and of the action of boring
animals and other injurious agents on corals, and reaches the conclusion
that the reef has been raised to its present level and thickness by wave
and current action, without the aid of elevation or subsidence. The
argument is based on such observations as could be made over the surface
of the reefs and the adjoining sea-bottom, and bears on the question of
the necessity of subsidence, and not on the fact of subsidence.

Coral Reefs and Islands. 279

form to suggest similarity of origin to the drift-made barriers
of sand.

ce. In the Pacific Ocean, the trends of many of the coral-
island groups and of the single islands do not correspond
with the direction of the oceanic currents, or with any eddy
currents except such as are local and are determined by them-
sélves.

Near longitude 180°, as the map of the Central Pacific (see
Part I.) illustrates, the equator is crossed by the long Gilbert
(or Kingsmill) group, at an angle with the meridian of 25°
to 80°, and not in the direction of the Pacific current, which
is approximately equatorial. This obliquely crossing chain of
atolls is continued northward in the Ratack and Ralick Groups
(or the Marshall Islands), making in all a chain over 1200
miles long ; and, adding the concordant Ellice Islands on the
south, and extending the Ratack line to Gaspar Rico, its
northern outlier, the chain is nearly 2000 miles long. Nothing
in the direction of the long range, excepting local shapings
of some of the points about the atolls, can be attributed to the
Pacific currents. Moreover, the diversified forms of the atolls
have no sufficient explanation in the drift process.

d. Further, drifting by currents may make beaches and
inner channels whether subsidence is going on in the region
or not, and are not evidence forj or against] either a move-
ment downward or upward. Sandy Hook, the long sandy
point off the southern cape of New York harbour, has been
undergoing (as the U. 8. Coast Survey has shown) an increase
in length, or rather variations in length, through the drifting
of sands by an outside and an inside current ; and this is no
evidence that Professor G. H. Cook is wrong in his conclusion
that the New Jersey coast is slowly subsiding.

e. But even in this region of Florida we have strong evidence
of a great subsidence during the coral-reef era, and all the
subsidence that the Darwinian theory demands.

In a very valuable paper by Mr. Agassiz, published in 1879
in the Bulletin of the Museum of Comparative Zoology *, the
author points out that the South American continent, in com-
paratively recent geological times, had connection ‘with the
West-India Islands through two lines: (1) one along a belt
from the Mosquito Coast to Jamaica, Porto Rico, and Cuba ;
and (2) the other through Trinidad to Anguilla, of the Wind-
ward Islands. He sustains the conclusion by a review of the
soundings made by the steamer ‘ Blake,’ under the command of

* An abstract of the paper is contained in Silliman’s American Journal
[3] xvili. p. 230 (1880).

280 Dr. J. D. Dana on the Origin of

J. R. Bartlett, U.S.N., and a consideration of the facts con-
nected with the distribution of marine and terrestrial species.
As the soundings show, the former of the two connections
requires for completeness an elevation of the region amounting
to 4060 feet over the part south of Jamaica, 4830 feet between
Jamaica and Hayti, and 5240 feet between Hayti and Cuba.
The other line of connection requires an elevation of 3450
feet. An open channel, as he observes, would thus be left
between Anguilla and the Virgin Islands, where there is now
a depth of 6400 feet. The close relations in the existing
fauna of the Gulf to that of the Pacific waters prove that it
continued to be a salt-water gulf through the era of elevation.

Mr. Agassiz infers that the connection of the West-India
Islands with South America existed before the Quaternary
era. But there are other facts which seem to prove that it
was continued into, or at least was a fact in, the Quaternary.

The opinion as to a connection of the Windward Islands
with South America in the Quaternary was presented by
Prof. E. D. Cope in 1868, and earlier, as he states, by Pomel,
on the ground of the discovery in the caves of Anguilla of a
species of gigantic Rodent related to the Chinchilla, as large
as the Virginia Deer, and nearly equalling the Quaternary
Castoroides of Ohio*. Further, De Castro, as cited by Dr. J.
Leidy in his ‘Mammalian Fauna of Dakota and Nebraska,’
1869, announced, in 1865, a gigantic Sloth of the “ Quater-
nary,” from Cuba, which he referred to the genus Megalonyz,
and Dr. Leidy named Megalocnus rodens, proving a Quater-
nary connection between the continent and Cuba.

The fact of an elevated condition of the region sufficient to
make Cuba and Anguilla part of the continent during the
earlier Quaternary, if not in the Pliocene also, is thus made
quite certain. This is fully recognized by Wallace tf. Such
a condition could hardly have existed without a large elevation
also of Florida, though probably not, as Mr. Agassiz holds, to
the full amount of the depression between it and Cuba (nearly
3000 feet), because Cuba is most closely related in fauna to
South America. The subsidence which brought the region to

* Proc. Philad, Acad. Nat. Sci. 1868, p. 315, and Proce. Philad. Amer.
Phil. Soc. 1869, p. 183; also ‘Smithsonian Contributions to Knowledge,’
30 pp., 4to, with 5 plates, Washington, 1885. The last paper (prepared in
1878) contains descriptions of the following species from the Anguilla
bone-cave :—Amblyrhiza mundata, Cope (the large Rodent announced in
1869), A. guadrans, Cope, A. latidens, Cope, an Artiodactyl apparently of
the Bovide and a little smaller than Ovis aries, With them was obtained
an implement (“a spoon-shaped scraper or chisel”) made of the lip of the
large Strombus gigas.

Geograph, Distrib, of Animals, 11. pp. 60, 78.

Coral Reefs and Islands. 281

the present level was consequently within the coral-reef
period. It is hence hardly to be doubted that the making of
the Florida, Bahama, and other West India coral-reefs was
going on during the progress of a great subsidence. None
of the facts mentioned by observers are opposed to this
view.

It is of interest to note here that on Cuba and Jamaica
there are elevated coral-reefs, the highest on Cuba 1000 feet
above the sea, according to Mr. Agassiz, and probably at one
point 2000, according to Mr. W. O. Crosby’s observations *,
and on Jamaica 2000 feet, according to Mr. Sawkins ;
indicating that there have been upward movements subsequent
to the downward. Mr. Crosby argues that the great thickness
of the now elevated reefs could have been produced only
‘during a progressing subsidence ;” so that ‘“ we have
apparently no recourse but to accept Darwin’s theory.”

C. It has been urged by Mr. Semper, Dr. J. J. Rein, Mr.
A. Agassiz, Mr. Murray, Dr. Geikie, and others, that since
the growing calcareous deposits of the sea-bottom are slowly
rising toward the surface by successive accumulations of the
shells and other débris of marine species, they may have been
built up locally in various regions of the deep seas (as they
actually are now about some islands) until they were near
enough to the surface to become next a plantation of corals ;
and that in this way atolls became common within the area
of the tropical oceans. The method is regarded as setting
aside subsidence.

a. The advocates of this hypothesis have not pointed to
such a mound now approaching the ocean’s surface on the
western border of the Gulf-Stream, where the depth over the
remarkably luxuriant region is least ; and none over any part
of the tropical Pacific. It is suggested that the Chagos Bank
may be one example ; but it isnot knowntobeso. Professor
Semper states that he found evidence of pelagic life, instead
of modern corals, in the lower part of the elevated reefs of the
Pelews. Dr. Geikie cites from letters by Dr. Guppy in
‘Nature’ of Nov. 29, Dec. 6, 1883, and Jan. 12, 1884, the
fact that in elevated reefs on the Salomon Islands, 100 and
1200 feet high, the coral-rock forms a comparatively thin
layer over impure earthy limestone abounding in foraminifers
and other pelagic organisms such as Pteropods. Such obser-

* Proc. Boston Soc. Nat. Hist. xxii. p. 124, (1882), and in abstract in
Silliman’s American Journal, xxvi. p. 148 (1883).

282 Dr. J. D. Dana on the Origin of

vations have great interest, but they only prove that, in coral-
reef seas, corals will grow over any basis of rock that may
offer where the water is right in depth, and do not nullify any
of the evidences of subsidence. This point should be kept
before the mind in all future study of coral-reef regions.
Borings in coral-islands, as recommended on a former page,
are the true means of investigating it.

6. The old hypothesis that atolls may have been built upon
the summits of submerged mountain-peaks, or volcanic cones
at the right distance under water for growing reef-corals, or,
if not at the right level, brought up to it by other organic
depositions, or down to it by abrasion, is urged by Mr.
Murray.

This writer observes that “ the soundings of the ‘ Tuscarora ’
and ‘Challenger’ have made known numerous submarine
elevations ; mountains rising from the general level of the
ocean’s bed at a depth of 2500 or 8000 fathoms, up to within
a few hundred fathoms of the surface.” But “a few hundred
fathoms,” if we make few equal 2, means 1200 feet or more,
which leaves a long interval yet unfilled *.

It is also urged that some of the “ emerged volcanic moun-
tains situated in the ocean-basins’’ may have been wholly
swept away and left with a few fathoms of water above them.
But this is claiming more from the agents of erosion than
they could possibly have accomplished, as the existence of an
atoll in the ocean and the examples on coasts of wave and
tidal action prove.

D. To give completeness to the hypothesis which makes
barrier and atoll islands out of submarine banks (whether
these banks have a basis of volcanic or other rocks, or of
calcareous accumulations), it is necessary to show that the
waters of the waves and currents can make barrier islands and
atolls out of such banks without subsidence ; and explanations
to this effect have been given.

It is urged, in agreement with Darwin, that the outer por-

* The actual depths over the elevations in the ‘Tuscarora’ section between
the Hawaian Islands and Japan, numbering them from east to west, are
as follows :—1, 11,500 feet; 2, 7500 feet; 3, 8400 feet; 4, 12,000 feet ;
5, 9000 feet (this seven miles west of Marcus Island); 6, 9600 feet.
Whether ridges or peaks the facts do not decide; probably the former.
No. 1 has a base of 185 miles, with the mean eastward slope 40 feet per
mile (=1: 182), and the westward 128 feet per mile. No. 2 has a breadth
of 396 miles, with the mean eastern slope mostly 37 feet per mile, but 51
feet toward the top, and the westward, 55 feet per mile (=1: 96). No.3
was the narrowest and steepest, it being about 100 miles broad at base,
and having the mean eastern slope 192 feet per mile and the mean western
200 feet.

Coral Reefs and Islands. 983

tions of reefs increase faster than the inner, owing to the
purer water about them and the more abundant life for food ;
that the inner parts are not only at a disadvantage in these
respects, but suffer also from coral débris thrown over them.
They add to these causes of unequal growth mentioned by
Darwin, the solvent and abrading action of the waters.

It is hence concluded that, under these conditions, the
simple bank of growing corals may have a depression made at
the centre, which, as the process continues, will become a lagoon
basin, and the reef thereby an atoll with a shallow lagoon ;
that the atoll, so begun, may continue to enlarge through the
external widening of the reef and the further action of
current-abrasion and solution within: or, in the case of
fringing reefs, that the change may go on until the reef has
become a barrier-reef, with an inner channel and inner reefs.
It is admitted that subsidence may possibly have helped in
the case of the deepest lagoons.

Dr. Geikie expresses his opinion on the subject thus :—“ As
the atoll increases in size the lagoon becomes proportionally
larger, partly from its waters being less supplied with pelagic
food, and therefore less favourable to the growth of the more
massive kinds of corals, partly from the injurious effects of
calcareous sediment upon coral-growth there, and partly also
from the solvent action of the carbonic acid of the sea-water
upon the dead coral.”

Mr. Semper gives examples of the effects of currents at the
Pelew Islands, stating that, by striking against or flowing by
the living corals, they make the reef grow with steeper sides
and determine its direction, and urging that abrasion and
solution have made, not only the deep lagoon-like channels,
but the deeper channels between the islands. He holds that
in Kriangle, which he describes as a true atoll with no
channel leading into the lagoon from the sea, the lagoon
may have been “the result of the action of currents on the
porous soil during a period of slow upheaval*. He says,
further, that the large channel in the main island of the group
‘forty fathoms deep and many miles wide,” “ finds an easy
explanation on the assumption of an upheaval ;’”’ it became
‘‘ wider in proportion as the enclosed island, consisting of soft
stone [tufa], was gradually eaten away; and during slow
upheaval it would continue to grow deeper in proportion as
the old porous portions of the reef and the rock in which it
was forming were more and more worn down by the combined
action of boring animals and plants, and of the currents

* “Animal Life,’ pp. 269, 270.

284 Dr. J. D. Dana on the Origin of

produced by the tides and by rain.’”’ Mr. Semper refers to
the dead depressed tops of some masses of Porites near tide-
level, as the effects of the deposit of sediment over the top of
the living coral, and of erosion by the waves and exposure to
rains while the sides continued to grow ; and the fact is made
an example on a very small scale of atoll-making. Other
examples of the action of currents, sediments, boring species,
and the solvent action of carbonic acid in the waters are
mentioned by Mr. Agassiz, in his excellent account of the
“ Tortugas and Florida reefs.”

a. The theory, if satisfactory, accounts not only for the
origin of an atoll, but for the origin of atolls of all sizes,
shapes, and conditions, and for great numbers of them in
archipelagos and chains; not only for channels through
fringing reefs, like those that abrasion in other cases makes,
but for all the irregular outlines of barriers, for the great
barriers reaching far away from any land, and for the positions
and indented coasts of the small included lands. Is it a
sufficient explanation of the facts ?

b. The currents that influence the structure of reefs are :
(1) the general movement or drift of the ocean, in some parts
varying with seasonal variations in the winds ; (2) the
currents connected with wave-action and the inflowing tide
over a shelving bottom; (3) the currents during the ebb,
flowing out of channels ; together with (4) counter-currents.
Each region must have its special study in order to mark out
all the local effects that currents occasion. Such effects are
produced whether a secular subsidence is in progress or not,
and hence a particular review of the subject in this place is
unnecessary.

The shaping of the outside of the reef and the determination
of the width and level surface of the shore platform are due
chiefly to the tidal flow and the accompanying action of wind-
waves, as explained in § 17 of Part I.*

* Since the first part of my paper was published I have observed in an
article by Mr. A. R. Hunt, in the Scientific Proceedings of. the Royal
Dublin Society, iv. p. 254, January 1885, the remark, referring to a state-
ment of the above fact in my‘ Manual of Geology,’ that the “ statement,
though strictlyin accordance with Mr. Russel’s theory, has, so far as I can
ascertain, no foundation in fact.” The statement, as I have said (and as
I illustrate in my ‘ Geology ’), was but the statement of a fact observed by
me first in 1839 on the coasts of Australia and New Zealand, without a
thought of any theory; and part of.the explanation is overlooked by Mr.
Hunt. I observed that the first waters of the incoming tide swelled over
the sandstone platform (which was a hundred yards or more wide off the
Port-Jackson Heads), and became thus a protector of the sandstone
platform from breaker strokes; and that the lower part of the sandstone
bluff to a height a little above high tide was hollowed out by the strokes

Coral Keefs and Islands. 285

‘The current that accompanies the ebb is locally the strongest.
Owing to the great width of many barrier-reefs and of the
channels and harbours within them, the tide flows in over
a wide region. At the turn in the tide the waters escape at
first freely over the same wide region; but, with a tide of
but two or three feet, there is but little fall before the reef—
which lies at low-tide level and a little above it—retards it by
friction ; and thus escape by the open entrances is increased
in amount and in rate of flow.- The facts are the same in
atolls where the lagoons have entrances”*.

of the breakers. A similar erosion near high-tide level of the great coral
masses standing on the coral-rock platform of atolls I also observed while
among the Paumotu Islands. Prof. A. I. Verrill informs me that he has
seen examples of the same action on a grand scale about the island of
Anticosti in the Gulf of St. Lawrence. The observations do not appear
to me to be at variance with the principles laid down in Mr. Hunt’s
valuable paper; they require only his recognition of a tidal effect which
he does not fully consider, and which British seas cannot illustrate.

To produce a platform, (1) the rock-material exposed to the flow of
the tide and the breakers must be firm enough to resist wear during the
early part of the flow, and, at the same time, soft enough to allow the
striking breakers to cut into the base of the bluff, or shear off the projec-
ting ledge; and (2) the region must not be one of very high tides or
stormy seas, for, in such regions of forceful waves and tides, the move-
ments are too often of the destructive kind through the whole continuance
of the flow leaving no chance for the protection a platform needs. Loose
sand-depusits are too soft; they are worn off below the sea-level and
changed in surface by storms ; but some firmer kinds may make a low-
tide flat in a bay where the tidesare small. Coral-reef rock, the material
of the atoll platform, has the hardness and solubility in carbonated sea-
water of ordinary limestone. The rock of the Port-Jackson Heads is a
friable sandstone. At the Bay of Islands, New Zealand, the platforms
occur in an argillaceous rock, which becomes soft avd earthy above by

weathering, but is unaltered and firm below because kept wet (doc. cit.

p. 442). At the Paumotus the tides are two to three feet high, and the
platform usually 100 yards or more wide; at the Phoenix Group the tides
are five to six feet high and the platform mo-tly 50 to 70 yards wide ;
at the Port-Jackson Heads, the ordinary tides are six feet high and the
platform 50 to 150 yards wide; at the Bay of Islands (in the sheltered
waters of the bay), the tides are eight feet high and the platform is under
30 yards wide.

* The currents of the tropical Pacific Ocean are of very unequal rate in
its different parts, and very feeble in the Paumotu Archipelago and the
Tahitian and Samoan regions. Capt. Wilkes reports that in the cruise of
the Expedition through the Paumotu Archipelago to Tahiti, a distance
of a thousand miles, during a month from August 13 to September 15,
1839, the drift of the vessels was only 17 miles; and that during fourteen
days in the first half of October, between Tahiti and Upolu of the
Samoan group, nearly 1800 miles, the drift was only 43 miles.

The ‘ Challenger,’ on her route from the Hawaian Islands to Tahiti, found,
between the parallel of 10° S. and Tahiti, “ the general tendency of the
eurrent westerly, but its velocity variable ;” between the parallels of 10° 5,

Phil. Mag. 8. 5. Vol. 20. No. 124. Sept. 18385. x

I ho) Sn rere. aie
Via Te
a 2

286 Dr. J. D. Dana on the Origin of

c. Examples of massive corals having the top flat or de-
pressed and lifeless, while the sides are living, are common in
coral-reef regions, wherever such corals are exposed to the
deposition of sediment, and where they have grown up to the
surface so that the top is bare above low tide. A disk of
Porites, having the top flat and the sides raised (owing to
growth), soas to give it an elevated berder, is figured on plate
lv. of my Report on Zoophytes. Many such were found in
the imptre waters of a shore-reef at the Feejees. At Tonga-
tabu one flat-topped mass of Porites was twenty-five feet in
diameter; and both there and in the Feejees, others of
Astreids and Meandrinas measured twelve to fifteen feet in
diameter. |

Over the dead surfaces, as Mr. Semper observes, the coral
may be eroded by the solvent action of the waters, and espe-
cially where depressions occur to receive any deposits, and
boring animals may riddle the coral with holes or tubes. But
generally the erosion is superficial ; the large masses referred
to showed little of it. Such dead surfaces in corals are
generally protected by a covering of nullipores and other
incrusting forms of life, and the crusts usually spread over
the surfaces pari passu with the dying of the polyps.

d. Hivery stream, says Mr. Semper (when explaining, as
cited on a preceding page, the origin of the deep channel of
the large Pelew island, whose depth is “ 35 to 45 fathoms”),
‘has a natural tendency to deepen its bed.’””? But there is a
limit to this action. The eroding or deepening power of a stream

and 6° N., the direction was westerly with “the average velocity 35
miles per day, the range 17 to 70 miles per day,” the maximum occurring
along the parallel of 2° N. Further west, about the Phoenix group, the
equatorial current as described by Mr. Hague (Joc, cit. p. 237), has “a
general direction of west-south-west and a velocity sometimes exceeding
two miles per hour.” At times it changes suddenly and flows as rapidly
to the eastward. The drifting of the sands about Baker’s Island (in
latitude 0° 13’ N., longitude 176° 22' EK.) has much interest in connection
with this subject of current-action, and the facts are here cited from Mr.
Hague’s paper. The west side of the little island (1x 2m. in area). trends
north-east, and the southern east-by-north, and at the junction a spit of
sand extends out. During the summer the ocean swell, like the wind,
comes from the south-east, and strikes the south side ; and consequently
the beach sands of that side are drifted around the point and heaped up
on the western or leeward side, forming a plateau along the beach two or
three hundred feet wide, and eight or ten feet deep over the shore-plat-
form. With October and November comes the winter swell from the
north-east, which sweeps along the western shore; and in two or three
months the sands of the plateau are all drifted back to the south side, -
which is then the protected side, extending the beach of that side two
or three hundred feet. This lasts until February or March, when the
operation is repeated.

¥

Coral Reefs and Islands. 287

through abrasion and transportation is null or nearly so below
the level of its outlet. A basin or channel 45 fathoms (270
feet) deep, with an outlet of much less depth, could not be
deepened by such means or protect itself from shallowing.
The depth of the outlets is not stated, except that they are
said to be ship-channels. Moreover, with a tufa bottom,
solution could not contribute to the removal, since carbonated
waters, although decomposing the tufa, dissolve very little of
its ingredients. An elevation in progress would result in
making of the channel a closed lake, and finally dry land.

Wor the same reason, the small atoll, Kriangle, having, as
described, a closed lagoon, could have no deepening of the
lagoon from abrasion by tidal currents or wave-action during
the progress of an elevation. And if a lagoon have an out-
let, the rapid current of the ebb would be confined to the
narrow passage-way and a portion of the bottom near it ;
through the larger part of the lagoon, as in any other lake,
the waters would have scarcely perceptible motion, and
therefore slight energy for any kind of work. Hence a
lagoon would lose very little by this means, and shallowing
would go on unless there were great loss through the solvent
action of the waters. An elevation would only hurry the
shallowing and end in emptying the lagoon.

e. Hrosion through solvent action is promoted by the
presence in the waters both of carbonic acid and organic
acids. ‘The material within reach of the tides or waves ex-
posed to this action is dead corals and shells, or their débris,
and bare coral rocks, occurring over—(1) the outer region of

living corals and for a mile or so outside; (2) the shore

platform and the reef, bare at low tide, on which there is
comparatively little living coral; and (3) the lagoon basin.
There is nothing in the material within the lagoon to favour
solution more than in either of the other two regions ; in fact,
the platform and bare reef are most exposed to the action
because of the small amount of living corals over them. The
outside waters take up what they can through the carbonic
acid they contain, and supply thereby the wants of the lime-
secreting polyps, shells, &c., and carry on the process of
solidification in the débris; the same waters move on over
the atoll reef and take up more lime as far as the acid ingre-
dient is present ; and then they pass to the lagoon for work
similar to that outside, with probably a diminished amount of
free carbonic acid, on account of the loss over the reef-ground
previously traversed. ;

The lagoon. basin is not, therefore, the part of the atoll that
loses most by solution, any more than by abrasion and trans-

A 2

ri 4° Bait es TS er: i
u rah a Su

288 Dr. J. D. Dana on the Origin of

portation. The outer reefs suffer the most; and yet, if the
island is not subsiding at too rapid a rate, they keep extending
and encroaching on the ocean, instead of wasting through
the drifting into the ocean at large of calcium carbonate in
grains and solution; and the shore-platform also preserves
its unvaried level notwithstanding the daily sweep of the
tidal floods, and the holes that riddle its outer portions.

The remarks, “It is a common observation in atolls
that the islets on the reefs are situated close to the lagoon
shore ;” and such “ facts point out the removal of matter which
is going on in the lagoons and lagoon channels,’ * I know
nothing to sustain. The width of the shore-platform on the
seaward side is always greater than that on the lagoon side ;
but the outside shore-platform has its width determined by
tidal and wave action, and this action is powerful on the
ocean side, and feeble on the lagoon side ; it produces a high
coarse beach on the outside as the inner limit of the platform,
and a finer, lower, and much more gently sloping beach on
the inside. ‘The amount of erosion is far greater, as it should
be, on the side of the powerful agencies.

j. The loss to the lagoon by abrasion and solution is reduced
to a minimum, in the majority of atolls, by the absence of lagoon
entrances, which leaves them with only concealed leakage
passages ior slow discharge.

Nine tenths of atolls under six miles in length (or in longer
diameter), half of those between six and twenty miles, and
the majority of all atolls in the Pacific Ocean, have no en-
trances to the lagoon a fathom deep ; and the larger part of
those included in each of these groups have no open entrances
at all.

For evidence on this subject, I refer to the Wilkes Expedi-
tion Hydrographic Atlas. This Atlas contains maps of nearly ©
sixty coral-islands from the surveys of its officers, drawn on
a large scale (one or two miles, rarely four, to the inch).

Out of the number, nine, ranging from 14 to 3 English
miles in the longer diameter of the reef, have no lagoon, but
only a small depression in its place ; two of these take in
water at high tide, and the rest are dry.

Of those under six miles in length having lagoons, seventeen
in number, sixteen are represented as having no entrances to
the lagoon at low tide; and the one having an entrance is
5x4 miles in size. The smallest is about a mile in diameter.

Of those that are siz miles or over in length, twenty-nine
in number, seventeen have channels and twelve have none.

* Mr. Murray, loc, cit. p. 515.

Coral Reefs and Islands. 289

Those having channels are mostly over ten miles in length.
A list of them is here given with their sizes, and also the
proportion of the reef around the lagoon which is under water

above third tide, and bare at low tide, a feature of much

interest in this connection.

_ Hiiuic—E Group.—Depeyster’s: 6x6 m.; three fourths of
the encircling reef bare. LEllice’s: 9x5 m.; three fourths
bare.

GILBERT GRroup.—Apia: 17x7m.; half bare. Tarawa:
21x9m.; half bare. Taritari: 18x11 m. two thirds bare
Apamama: 125 m.; half bare. . Taputeuea; west side
mostly submerged.

MarsHAtt Is~tAnps (northern).— Pescadores : 10x 8 m.;
four fifths bare. Korsakoff: 26 m.; four-fifths bare.

Paumotus.—Peacock ; 15x7 m.; nearly all wooded.
Manhii: 13 x5 m.; nearly all wooded. Raraka: 6x9 m.;
three fourths wooded. Vincennes: 18 x9 m. ; mostly wooded.
Aratica: 18x11 m.; three fifths bare. Tiokea: 18x4m.;
two thirds wooded. Kruesenstern’s: 16x10m.; mostly
wooded. Dean’s (or Nairsa): 53x18m.; half er more
bare.

g. The absence of open channels in so large a proportion of
lagoons, and especially in lagoons of the smaller atolls, appears
to be fatal te the abrasion-solution theory. The method of
enlarging atolls through currents and solution can act only
feebly, if at all, where waters have no free outlet ; and this is
eminently so with the smaller atolls which have been assumed
by the theory to be most favourable in purity of water and in
abundant life for progress; if the small cannot grow, the
large lagoons cannot be made from them by the proposed
method.

Reverse the method, letting the large precede the small (as
under the subsidence theory), and then we have a consistent —
order of events. We have large atoll reefs with several
large entrances (like the great barrier-reef about a high island
in this and other respects) gradually contracting, and the
entrances concurrently narrowing through the growing corals
and the consolidating débris, in spite of the efforts of abrasion
and solution to keep them open and make them deeper ; and,
afterwards, the atoll becoming still smaller until the entrances
close up; and, finally, the lagoon-basin is reduced to a dry
depression with nothing of the old sea~-water remaining except,

perhaps, some of its gypsum.

h. Instead of small lagoons having the purest waters, the
reverse is most decidedly and manifestly the fact, and this

290 Dr. J. D. Dana on the Origin of

accords with the reversal in the history just suggested. Since
atolls of middle and larger size commonly have one third to
two thirds of the encireling reef covered with the sea at one-
third tide, making the ocean and lagoon for more than half
the time continuous, the large lagoon in such a case has
as pure water as the ocean, and commonly as good a supply
of food-life, and sometimes as brilliant a display of living
corals. But in the smaller atolls, the area of the lagoon has
little extent compared with the length and area of the encir-
cling reef; coral-sands and other calcareous material con-
sequently have possession of the larger part of the bottom, and
the waters, since they are less pure than those outside, contain
fewer and hardier kinds of corals and less life of other kinds.
They are exposed, also, to wider variations of temperature
than the outer, with injury to many species ; and at lowest
tides may become destructively overheated by the midday
sun, as many a plantation of corals with dead tops for a foot
or more bears evidence. In the smallest atolls, the lagoons
are liable also to alternations of excessive saltness from evapo-
ration and exeessive freshness from rains, and consequently
no corals can grow inside, though still flourishing well in the
shallow sea about the outer reef. The above are the facts,
not the suggestions of theory.

2. We read :—“So great is the destructive and transporting
influence of the sea under the combined or antagonistic work-
ing of tides, currents, and wind-waves that the whole mass of
the reef, as well as the flats and shoals inside, may be said to
be in more or less active movement’’*., ‘This description of
the Tortugas reefs 1s not applicable to the atolls of the Pacific.
Notwithstanding the testimony of Captain Beechey and others
about occasional catastrophes (which are mostly catastrophes
to the islets and banks within the lagoons), I was led to look
upon a coral-island as one of the most stable of structures.
The waves and currents have shaped its reef, shore-platform,
and beaches, fitting it well in all respects for its place by
means of the forces that were to assail it; and an air of placid
repose, as it lies amid the breakers, is its most impressive
feature. Through the wind-made and tidal movements the
loose sands are drifted along the shores and over the reef ;
edges of the reef are broken off in gales or by earthquake
waves ; and occasionally a mushroom islet in the lagoon,
where growing corals are not compacted by wave-action, is
overthrown by the same means ; but beyond this the structure
is singularly defiant of the encroaching waters. Harthquakes
may bring devastation ; and so they may to other lands.

* Address, p. 23,

Coral Reefs and Islands. 291

VI. Conclusion.

With the theory of abrasion and solution incompetent, all
the hypotheses of objectors to Darwin’s theory are alike weak ;
for all have made these processes their chief reliance, whether
appealing to a calcareous, or volcanic, or mountain-peak base-
ment for the structure. The subsidence which the Darwinian
theory requires has not been opposed by the mention of any
fact at variance wth it, nor by setting aside Darwin’s arguments
in its favour; and it has found new support in the facts from
the ‘ Challenger’s’ soundings off Tahiti that had been put in
array against it, and strong corroboration in the facts from
the West Indies.

Darwin’s theory therefore remains as the theory that
accounts for the origin of coral reefs and islands.

VIL. Central-Pacific Subsidence.

Darwin, as has been said, took a step beyond direct obser-
vation in his inference that the subsidence attested to by each
atoll extended over the intermediate seas and characterized a
large central area of the ocean. He may be wrong here (and
the writer with him), while not wrong in his theory. But,
considering the distribution of the Pacific atolls in the ocean,
their relation in this respect to the distribution of other
Pacific lands, and the facts connected with the history of
coral reefs and islands, the generalization appears to be well
sustained. The question is here left without further argu-
ment, to be considered over the best geographical map of the
ocean to be had, and the best bathymetrical map that can be
made, only asking that the doubts which physical theory has
set afloat may not be allowed by the geologist to warp the
judgment or cripple investigation”.

My own agreement with Darwin as to the area of cora!-
reef subsidence was promoted by an early personal study of
the oceanic lands. For more than five years previous to pass-
ing my third decade I was ranging over the oceans, receiving
impressions from a survey of the earth’s features. I was

* One point often encountering an @ priors doubt is the slowness of the
required subsidence. The subsidence over the Appalachian region which
preceded the making of the Appalachian Mountains amounted, according
to well-ascertained facts (as stated by Hall and Leslie), to at least 30,000
feet. The great trough, nearly a thousand miles long, was in progress
through all of Palwozoic time. If the Paleozoic ages covered only
20,000,000 years (a low estimate) the mean annual rate was 0-018 inch,
which is less than half a millimetre per year. Such a fact is no evidence
as to the rate of the atoll-making subsidence; but, whatever the cause to
which the Appalachian subsidence is to be attributed, it 1s suggestive as
to possibilities and probabilities connected with the earth’s movements,

292 Mr. J. Hopkinson on an unnoticed Danger in

made to see a system of arrangement in the Pacific islands,
instead of a “labyrinth ;” to appreciate the vast length of the
island-chains in the great ocean with their many parallelisms,
and the accordant relations subsisting between them and long
lines of atolls. Iwas thence led to observe the corresponding
system in the features of the continental lands, and the more
fully so when afterwards it was proved that Geology was not
in America merely the study of strata and fossils, but of the
successive stages in a growing continent. Thus a conception
of the earth as a unit became early implanted, and the idea
also of its development as a unit under movements as compre-
hensive as the system in its feature-lines. My faith in any
mountain-making theory hitherto proposed is weak. But
that idea of system in structure and progress stands, and,
however much ignored by students of the earth’s stratigraphy,
it must have its explanation in a true theory of the earth’s
dynamics.

XXX. On an unnoticed Danger in certain Apparatus for Dis-
tribution of Electriaty. By J. Hopxinson, £.R.S.*

\ [ ANY plans have been proposed, and several have been
: to a greater or less extent practically used, for combi-
ning the advantage of economy arising from a high potential
in the conductors which convey the electric current from the
place where it is generated with the advantages of a low
potential at the various points where the electricity is used.
A low potential is necessary where the electricity is used ;
partly because the lamps, whether arc or incandescent, each
require a low potential, and partly because a high potential
may easily become dangerous to life. Amongst the plans
which have been tried for locally transforming a supply of
high potential to a lower and safer, the most promising is by
the use of secondary generators or induction-coils. It has
been proved that this method can be used with great economy
of electric power and with convenience ; under proper con-
struction of the induction-coils it may also be perfectly safe.
It is, however, easy and very natural so to construct them
that they shall be good in all other respects but that of safety
to life—that they shall introduce an unexpected risk to those
using the supply.

In a distribution of electricity by secondary generators, an
alternating current is led in succession through the primary
coils of a series of induction-coils, one for each group or -
system of lamps. The lamps connect the two terminals of the

* Communicated by the Author.

certain Apparatus for Distribution of Electricity. 293
secondary coil of the induction-coils. It is easy to so con-
struct the induction-coils that the difference of potential
between the terminals of the secondary coils may be any
suitable number of volts, such as 50 or 100; whilst the
potential of the primary circuit, as measured between the
terminals of the dynamo machine, may be very great, e. g.
2000 or 3000 volts. If the electromagnetic action between
the primary and secondary coils, on which the useful effect of
the arrangement depends, were the only action, the supply
would be perfectly safe to the user so long as apparatus with
which he could not interfere was in proper order. But the
electromagnetic action is not the only one. Theoretically
speaking, every induction-coil is also a condenser, and the
primary coil acts electrostatically as well as electromagneti-

cally upon the secondary coil. This electrostatic action may

easily become dangerous if the secondary generator is so con-
structed that its electrostatic capacity, regarded as a condenser,
is other than a very small quantity.

IT
E

ro

Imagine an alternate-current dynamo machine, A, its ter-
minals, B, C, connected by a continuous conductor, B DQ,
on which may be resistances, self-induction-coils, secondary
generators, or any other appliances ; at any point Is a con-
denser, Hi, one coating of which is connected to the conductor,
or may indeed be part of it, the other is connected to earth
through a resistance, R. Let K be the capacity of the con-
denser, V the potential at time ¢ of the earth-coating of the
condenser, U the potential of the other coating, w the current
in resistance R to the condenser from the earth, being taken
as positive, and the earth-potential as zero. We have

r= 7, K(U'-V)=2;
whence, since
U=Asin 27nt,

where A is a constant depending on the circumstances of the
dynamo circuit as well as the electromotive force of the
machine, and 7 is the reciprocal of the periodic time of the

294 Danger in Apparatus for Distribution of Electricity.
machine, we have
KRe+2=27nKA cos 2rnt,

QamnKA t
c= (KR2QanyP+1 { —IarnKR sin 2 arnt + cos 2ant } ’
moan sauare 01 C= — eee RE oat mean square of A
Lane / (K R270)? cy oe d

Let us now consider the actual values likely to occur in
practice. Let the condenser K bea secondary generator ; let
the resistance R be that of some person touching some part of
the secondary circuit, and also making contact to earth with
some other part of the body ; » may be anything from L100 to 250,
say 150; K will depend on the construction of the secondary
generator—it may be as high as 0°3 microfarad or even more,
but there would be no difficulty even in large instruments in
keeping it down to one hundredth of this or less. The mean
square of A will depend on the circumstances of other parts
of the circuit; it might very easily be as great, or very nearly
as great, as the mean difference of potential between the ter-
minals of the machine if the primary circuit were to earth
at C. Suppose, however, that the circuit BDC is symme-
trical, that E is at one end, and that another person of the
same resistance as the person at E is touching the secondary
circuit of the secondary generator F at the other end of the
circuit. In that case, if 2400 be difference of potential of the
machine, mean square of A will be 1200; in which case we

have, taking Ras 2000 ohms,

2er x 150 x 0:3 x 10-°
V (Qa x 150 x 0°3 x 107 x 2000)? +1
= about 0°3 ampere.

mean square of v=

x 1200

Experiments are still wanting to show what current may be
considered as certain to kill a man, but it is very doubtful
whether any man could stand 0°3 ampere for a sensible length
of time. It is probable that if the two persons both took firm
hold of the secondary conductors of EH and F, both would
be killed. Ifthe person at F be replaced by an accidental
dead earth on the secondary circuit of F, the person at EB
would experience a greater current than 0:3 ampere.

It follows from the preceding consideration that secondary
generators of large electrostatic capacity are essentially dan-
gerous, even though the insulation of the primary circuit and
of the primary coils from the secondary coils is perfect. The
moral is—for the constructor, Take care that the secondary
generators haye not a large electrostatic capacity, say not

a4

2 a 4

On Supersaturation of Salt-Solutions. 295
more than 0:03 microfarad, better less than ;¢,5 microfarad ;

for the inspector, Test the system for safety. The test is very
easy. Place asecondary generator of greatest capacity at one
end of the line and connect its secondary circuit to earth
through any instrument suitable for measuring alternate cur-
rents under one ampere ; put the other end of the primary to
earth; the reading of the current-measuring instrument
should not exceed such a current as it may be demonstrated a
man can endure with safety.

XXXII. On Supersaturation of Salt-Solutions. By W. W. J.
Nicou, M.A., D.Sc., F.R.S., Lecturer on Chemistry, Mason
College, Birmingham*.

; qN my previous paper f on this subject I described experi-

ments which proved that a so-called supersaturated
solution was able to dissolve more of the salt it already con-
tained, provided only that the salt was added in the dehydrated
state, and under conditions which precluded the possibility of
access of crystals of the hydrated salt. I also expressed the
opinion that a supersaturated solution is merely a solution of
the anhydrous salt which may, or may not, be saturated; and
that it differs in no way from an ordinary solution so long
as no disturbing cause operates to bring about the formation
of a hydrate which exists only in the solid state.

Until I made the experiments described in the following
paper, I was of the opinion that although a hydrated salt
existed in solution in what may be termed an anhydrous state,
still, when a dehydrated salt is dissolved in pure water,
solution always took place in two stages—the first consisting
of the hydration of the salt to form the solid hydrate, and that
this was followed by solution accompanied by the decomposi-
tion of the hydrate thus formed ; but the experiments I am
about to describe are, I think, conclusive, and show clearly
that solution of the dehydrated salt in pure water is unattended
by the formation of any hydrate, solid or liquid.

Two similar wide-mouthed bottles, capable of holding some
60 cubic centim. of water, were taken ; in each were placed
25 cubic centim. of the same sample of distilled water, and in
each a glass bulb containing 5 grms. of dehydrated Na,SQ,.
One of the bottles (No. 1) was placed at once in the constant-
temperature bath at 20° C.; the other (No. 2) was placed in
a water-bath which was raised to the boiling-point ; 1t was
then, after cooling, placed alongside the first bottle. After

* Communicated by the Author.
+ Phil. Mag. June 1885.

296 Dr. W. W. J. Nicol on

some time the two bottles were taken out and shaken, so as to
break the bulbs and thus bring the dehydrated Na,SO, in
contact with the water. After a few seconds’ shaking, the
salt was completely dissolved in bottle No. 2, which had been
heated ; but the salt in bottle No. 1 had caked into hard
masses, which dissolved only very slowly. In this case the
solid hydrate was formed ; while in the other, solution had
taken place without hydration.

Another experiment with the same quantity of water, but a
larger amount (10 grms. of Na,SO, in both bottles), was
attended with a like result.

A third experiment, with the same quantities of water and
salt as in the previous case, differed to some extent from it.
In both bottles solution was complete at once; but the con-
tents of the bottle which had not been heated crystallized out
again in a few seconds.

A fourth experiment gave results identical with the second
experiment. In each of these three last experiments the
solution in bottle No. 2 was distinctly supersaturated, crystal-
lizing on removing the stopper.

I have made numerous other experiments similar io the
above, and have never failed in obtaining a supersaturated
solution in this way. I noted that when special care was
taken in cleaning the bottles and fresh distilled water was
employed, it was frequently the case that the solution behaved
as in experiment No. 3 above; but I shall return to this later.

A saturated solution at 20° of the hydrated salt contains,
according to Mulder*, 19°5 parts of the anhydrous salt per
hundred parts of water. Thus, the supersaturated solutions I
prepared as above described contained more than twice as
much Na,SQ, as a saturated solution prepared in the ordinary
way. lt remained to be seen to what extent this supersatu-
ration may be pushed—in other words, to find the amount of

salt in a true saturated solution of Na,SQ,.

A bottle sommennine 25 cubic centim. of water and 15 grms.
of Na,SQ, sealed up im a bulb was treated as above, and rafter
shaking was allowed to remain in the bath at 20° for several
hours. The stopper was then removed and some of the clear
solution was rapidly poured into a previously weighed platinum
dish, and in this way the percentage of sali was obtained.
Two trials gave :—

(1) 35°87 per cent., | corresponding to 55°63 parts
(2) 35°61 per cent., per 100 parts of water.

* Buydragen tot de pe nas van het scheikundig gebonden Water
(Rotterdam, 1864).

Supersaturation of Salt-Solutions. 297

Another experiment at 25° gave
35°63 per cent. =55'35 per 100.
Another pair at 30° C. gave
1) 34°25
‘- 31g ¢ O94 per 100;
while another at 83° C. gave
D109 per 100.

Before proceeding further I wish specially to lay stress on
the approximate nature of the above figures. They are all
necessarily high, for two reasons in particular. One is, that
it is impossible to filter the solution, for the instant it comes
in contact with the air crystallization commences ; thus, as it
is very dense and far from mobile, it is impossible to obtain

the solution perfectly clear. The second cause that raises the

percentage is, that a very considerable evolution of heat ac-
companies the solidification, and thus some of the water is
driven off, even when the dish is covered. In spite of this
the figures show clearly enough a decrease in solubility as the
temperature rises.

As is well known, the solubility-curve of Na,SO, ascends
rapidly to a point about 33° C. (Mulder 82°75, Kopp 32°-93)
and then descends slowly to 100° C., after which it rises
again*. ‘This change of solubility at 33° C. is usually ascribed
to the dissociation of the decahydrate at that temperature ;
but I shall return to this. The most reliable data relating to
the solubility of Na,SO, are those given by Gay-Lussacf and
Muldert. From the results of Gay-Lussac’s experiments
Kopp§ constructed the following interpolation-formule.

Below 30° C. The quantity dissolved in parts per 100 is

(I.) Q=5-02 +0°30594¢—0-00041 2 + 0:00099772.
Above 40°, }
(II.) Q=58°5 —0°27783¢+ 0:00069 ¢ + 0:0000049802 2°.
And the point of intersection of the two curves is given by

— 0-0009927 # + 0:001122—0°58377t453:-48= 0;
whence = 32°93.

~ Now the agreement of the figures found by both Gay-Lussac
and Mulder with the values calculated by Kopp’s formule is
very complete, as is shown by the table, where Mulder’s results
are compared with Kopp’s figures.

* Conf. Mulder, loc. cit. page 120.
+ Ann. Chim. et Phys, [2] x1. p. 812.
* t Loe. cit. page 123.
§ Ann. Chem. und Pharm. 1840, xxxiv. p. 271.

298 Dr. W. W. J. Nicol on

ia Mulder. Kopp.
100 42°5 42-597 Formula II
90 43°71 42-717 -
80 43°7 43°242 iA
70 44-4 44:840 %
60 45:3 45°391 4
50 46:7 46-958 53
40 48°8 48-811 Pe
30 40°9 AL13a Formula I.
Nicol
30 51:09 50:263 Formula II.
30 51°94 50-921 =
25 55°39 52:064 a
20 55°63 53°260 *

The lower half of the table compares the results I obtained
as described above, compared with the figures obtained from
Kopp’s second formula extended below 33° C.; and the con-
cordance here is, I venture to think, thoroughly satisfactory
when the difficulties attending such determinations are taken
into account. My results will also be found to agree well
with those of Loewel*, who determined the solubility at 20°,
and 25°, and 30° of the anhydrous salt by heating an excess
of the decahydrate and water above 33°, and then allowing
the solution thus obtained to cool in contact with the separated
anhydrous salt.

These experiments conclusively prove the truth of my con-
tention, that a so-called supersaturated solution differs in no
way whatever from an ordinary solution. The arguments |
advanced in the previous paper on this subject were incom-
plete in one point. They left room for the objection that a
solution of sodium sulphate might, for all we know, have a
different constitution, after heating and cooling, from that
possessed by a similar solution that has not been heated. To
such a change of constitution has been ascribed the possibility
of the existence of supersaturated solutions; but now that
supersaturated solutions have been prepared from cold water
and cold dehydrated Na,SQ,, and have been found pre-
cisely similar to those prepared with the aid of heat, this
theory falls to the ground, and there is no reason for regard-
ing them as anything but ordinary solutions of the anhydrous
salt.

Why, then, do these solutions crystallize when brought in
contact with the air? or why is it that they cannot be pre-

* Ann. de Chim, et Phys. {3} xlix.

Supersaturation of Salt-Solutions. 299

pared from water and dehydrated salt without heating these
out of contact with the air? and what effect has this heating ?

There is no doubt that one, if not the sole, cause which
brings about crystallization of a supersaturated solution is the
presence of a crystal of the hydrated salt; and it is for this
reason that supersaturated solutions can exist only out of con-
tact with the air of the laboratory. This also is the reason
that the water and anhydrous salt must be heated before they
are brought in contact ; for minute crystals of the solid hy-
drate may lie hid in the scratches in the neck of the bottle or
in the mass of dehydrated salt, and these must be heated above
53° before they are decomposed, and so lose their power of
determining the union of the dehydrated salt and the water
with which it is broughtin contact. The fact that no hydra-
tion of the salt takes place before solution is conclusively
proved, not only by the formation of a supersaturated solution
which cannot exist in the presence of undissolved hydrate,
but also by the fact that the finely powdered dehydrated salt
dissolves at once ; or if excess be present, the portion remain-
ing undissolved retains its powdery form, no caking together
or change in appearance taking place.

Finally, a very considerable rise of temperature attends the
solution of the salt quite equal to—so far as it is possible to
judge without quantitative determination of its amount—if not
actually greater than, that attending the solution of the dehy-

drated salt under ordinary conditions. Now, as shown above,

there is no hydration, and the question remains—To what is
this change of temperature due ? This cannot be fully answered
until our knowledge of the volume and temperature changes

_ attending the solution of the dehydrated salt is more complete

than at present. The probable explanation however is, I ven-
ture to suggest, to be found in the great contraction attending
the act of solution. According to the most reliable determi-
nations, the density of Na,SO, (solid) is 2°6618 at 20° C.
This gives the molecular volume 53°35. But, as shown in my
previous paper, the molecular volume of Na,SQ, (dissolved)
in a solution containing 6°244 molecules of salt (that is, 49°26
per 100) is 32°79. A saturated solution, as shown above,
contains not more than 55:6 per 100, and the molecular volume
in such a solution cannot exceed 33°5. ‘The contraction, then,
attending the solution of one gramme-molecule of Na,SQ, is
03°4—33°59=19°9, or nearly 20 cubic centim., or about 40 per
cent.; a quantity possibly sufficient to account for the thermal
change, exceeding, as it does, the volume-change on solution
of other anhydrous salts, which dissolve with either a slight
rise or fall of temperature. For a further extension of this
argument reference must be made to the papers of Miiller-

300 Notices respecting New Books.

Erzbach*, who has done much to extend our knowledge of
the connection between stability of compounds on the one
hand, and, on the other, the volume and thermal changes
attending their formation. |

XXXII. Notices respecting New Books.

Textbook of Practical Physics. By R.T. Guazmproox, M.A., LBRS.,
and W.N.Suaw, M.A. London: Longmans, Green, and Co.
(482 pp.). ?

ra‘HE development of practical work as a necessary part of the

teaching of science in its various branches is a very healthy

_ sign of the times, and promises well for the advantages we may

_expect from the increased study of science which has marked the
last twenty years. Formerly such things as laboratories for the
practical study of Biology, Physics, &c. were almost unknown in

England, except in a few of the more advanced educational insti-

tutions, and Chemistry was apparently the only science thought

worthy of being studied practically. Now there are numerous
well-equipped laboratories for the practical study of Physics ; and it
is for the use of teachers and students in such that this book is
intended, and it will be welcomed by all teachers of the subject as
supplying a much felt want. The Authors, in marking out the
limits of their work, had to decide between attempting to give
instructions equally applicable to all the different forms of instru-
ments to be met with in different laboratories, which must of
necessity have been very vague, and adhering to the descriptions of
the precise instruments in use in their own laboratory; and have
decided, we think very wisely, in adopting the latter course, trust-
ing that the necessity for adaptation to corresponding instruments
used elsewhere will not seriously impair the usefulness of the book.

Their general aim has been to place before the reader a description

of a course of experiments which shal] not only enable him to

obtain a practical acquaintance with methods of measurement, but
also, as far as possible, illustrate the more important principles of
the various subjects. There are two valuable introductory chap-
ters—one on the units employed and their dimensions, the other
on the arithmetic of numbers expressing approximate measure-
ments. We may quote, as an example, the application of the
principles of this chapter to find the effect on the value of a current,
as deduced from observations with the tangent-galvanometer, of an
error of a quarter of a degree in the reading.

The formula of reduction is
C=k tan @.
Suppose an error 6 has been made in the reading of 0, so that the
observed value is
C’=k tan (0+¢).

It is shown that when o is less than a certain small value, this is

* Poge. Ann., Berichte, Wied. Ann., Lieb. Ann., 1870-85. Sum-
marized in Meyer, Mod. Theor. d, Chem. p. 440.

-
2
oe
we bey
Bn
x
ty

Intelligence and Miscellaneous Articles. 801

equal to k (tan @+6 sec? 6); hence
O'—C=k6 sec? @,
and

a
mm eee ee

C k tan @ sinOcos@- sin 26@

The error 6 must be expressed in circular measure; if it be equi-
valent to a quarter of a degree, we have

via
Sip ie pea ASE -
5 a 00436 ;
_ C—O _ -00872
ny ORS sin 20

Hence we see, not only that when @ is known the effect of the
error can be calculated, but also that the effect of a given error is
least when the deflection is 45°. :

The different forms of balance, determinations of specific gravity,
and the mechanics of solids, liquids, and gases occupy 110 pages ;
various experiments connected with sound, 16 pages; the thermo-
meter, determinations of expansion, specific and latent heat, and
hygrometry, 58 pages; experiments with light, 100 pages; and
the rest of the book (130 pages) is taken up with magnetism, elec-
tricity, and electromagnetism. The theories of the balance and
thermometer are given with unusual completeness ; and the chapter
on hygrometry forms a complete elementary treatise upon the sub-
ject. But perhaps the most important part of the book is that
which relates to electricity, in which full and clear instructions are
given for the comparison and measurement of currents, resistances,
and electromotive forces, capacities of condensers, and so on. It
should be noticed that the book is well up to date. Thus, for
example, we have, in the description of Carey Foster’s method of
employing the B.A. form of Wheatstone bridge, an account of
Dr. Fleming’s form of bridge, described in the Proceedings of the

‘Physical Society, and in this Magazine for May 1884.

The sincere thanks of all teachers and students of Physics are due
to the Authors for their valuable book.

XXXII. Intelligence and Miscellaneous Articles.

ON THE ELECTRICAL RESISTANCE OF ALCOHOL.
BY M. G. FOUSSEREAU.

I EXAMINED the specific resistances of alcohol, and its mixture

with water and with salts, by comparing them with the known
resistance of a pencil-mark drawn ona plate of ebonite. I used
M. Lippmann’s method, and the experimental arrangement which I
employed in several previous researches*. Different specimens of
the absolute alcohol of commerce gave, at the temperature 15°, spe-

cific resistances comprised between 2°47 and 3°68 megohms. ‘These

differences may, @ priori, be ascribed either to the presence of a

* Comptes Rendus, May 26 and July 15, 1884.
Phil. Mag. 8. 5. Vol. 20. No. 124. Sept. 1885. Y

302 Intelligence and Miscellaneous Articles.

trace of water, or to the solution of substances from the sides of
the receiver.

In order to estimate the degree of influence of these various
causes, I first added to specimens of the same alcohol increasing
weights of distilled water, the resistance of which was nearly ten
times less than that of the alcohol used. I ascertained that the
resistance of the mixtures thus obtained goes on decreasing, and
attains a maximum which does not much differ from the resistance
of water, when there is not more than 3 per cent. of water, and
then increases up to the resistance of water. Buta notable alte-
ration of composition is always necessary to produce a distinct
change in the resistance of the mixture. ‘The difference between
the numbers cited above cannot therefore be ascribed to the small
quantity of water which commercial absolute alcohol might retain.

If, on the contrary, a trace of a standard solution of chloride of
sodium is added, an enormous alteration is produced m the resist-
ances. A fallin the ratio of 1: 0°527 is obtained by the addition

of a weight of salt representing 553000 that of alcohol. Weare
thus led to attribute the variations to the absorption by the alcohol
of some ten millionths of salts from the vessel which contains it. |

I have, in fact, observed that the conductivity of alcohol increases
in glass vessels more rapidly than that of water. It may double
in a few hours. Absolute alcohol prepared with great care and
kept for two years in a closed and full glass vessel in the chemical
laboratory of the Ecole Normale was ten times as good a conductor
as commercial absolute alcohol. Burnt upona platinum plate, this
alcohol coloured the flame yellow. It contained a trace of soda
salts from the glass.

I endeavoured to avoid these alterations by dispensing with the
use of glass vessels; and with this view I had absolute alcohol
collected in porcelain vessels in the manufactory of M. Billault. I
have observed the resistance to increase, and for the specimens in
question to 5°14 and 5°44 megohms at 15°. The second specimen,
contained in biscuit porcelain, retained a remarkable constancy for
several days.

M. Delachanal was kind enough to prepare for me, by a series of
distillations, alcohol at its minimum density like that which is used
for determining the point 100 of the new alcoholometers. Two
specimens of this were sent to me, one in a porcelain and the other
in a glass vessel. The former gave at 15° the resistance 7°031
megohms, the greatest I have observed; the other, which was
already altered by the glass, was 2-823 megohms. A fresh specimen,
prepared in the same way six days after the first and kept ina por-
celain vessel, gave 6°899 megohms, only differing from the other
by x6 of the value. Although traces of dissolved salts cannot ap-
preciably alter the density of alcohol, it might be desirable to distil
and preserve in porcelain receptacles the alcohol intended for
delicate chemical operations.

I have also examined the changes which alcohol experiences
when its temperature varies. I found that the resistance dimi-
nishes in the mean by 0°0145 of its value, when the temperature is

Rae

Intelligence and Miscellaneous Articles. 303

raised through one degree, near the ordinary temperatures. This
variation is not proportional to that of the coefficient of internal
friction, as is the case with distilled water; and for salts this latter
quantity varies, in fact, 0°0210 of a degree at the same temperature.
The mechanism of conductivity is therefore more complex for alcohol
than for salts and their aqueous solution.

This research was made in the Laboratory for Physical Research
at the Sorbonne.—Comptes Rendus, July 20, 1885.

THE VOLCANIC NATURE OF A PACIFIC ISLAND NOT AN ARGUMENT
FOR LITTLE OR NO SUBSIDENCE. BY J. D. DANA*.

In the remarks on this point in § 13 (Part I.) of my paper on
the Origin of Coral Reefs and Islands, I refer to the great depths
found in the ocean by soundings in the vicinity of Hawaii, and
speak of the facts as favouring the idea of more subsidence about
that south-eastern end of the group than along the north-western,
although the latter is the coral-island end. Another example of
similar character, but more striking, is afforded by the region of
the Ladrones. This north-and-south range of islands has its
largest volcanic islands in the southern part, and dwindles in the
opposite direction to islands which are little more than tufa conest ;
and 200 miles south of Guam, the largest island, the ‘ Challenger’
found a depth of 4475 fathoms (26,850 feet), one of the deepest
regions of the ocean. It hence may be that Guam, like Hawaii, is
a large island, not because of small subsidence, but because of con-

‘tinued eruptions that made it large in spite of the sinking that was

in progress. The question arises how far the depths in these par-
ticular cases are due to the undermining effects of volcanic eruption.
There are coral islands both north-east of the deep region, near
Guam, and also of large size, to the south-west and south-east, not
three degrees off; the former, those of an extension of the Pelew
range, and the latter, islands of the Caroline archipelago.

ON THE QUANTITY OF ELECTRICAL ELEMENTARY PARTICLES.
BY E. BUDDE.

There exists an attempt by Herwigt to estimate the magnitude
of electrical elementary particles; which, however, is open to the
objection of making some very arbitrary assumptions, and more-
over of confounding the ideas of “mass” and “quantity” of
particles. I believe that the following simple considerations lead
to a tolerably certain result for the quantity of electrical “atoms.”

In the sphere of ponderable masses we find that a definite body,
such as carbon, always enters into combination with a definite
relative weight, twelve. From this and from similar observations
with all other masses we conclude that carbon consists of atoms,
and that each of its atoms has the relative weight twelve. If we

* Communicated by the Author.

+ One of the two northern (Assumption Island) I give the outline of
on page 304 of my Expedition Geological Report.

}¥ Pogg. Ann. vol, cl. p. 381 (1878).

304 Intelligence and Miscellaneous Articles.

were in a position to deal with separate atoms, experiment would
not only refer to the relative weight of carbon atoms, but we
should see that carbon enters into all combinations with multiples
of a definite absolute weight; the above conclusions would then
shape themselves more definitely, and experiment would directly
teach us the absolute value of the atomic weight of carbon. |

We propose to ourselves the questions, (1) Are there in nature
discrete elementary particles of electricity ? (2) What is their mag-
nitude? According to the analogy of the conclusions just drawn
for carbon, this question may be answered as follows :—Ii there are
in nature discrete elementary particles of electricity, it is to be ex-
pected that an absolutely definite, very small quantity of electricity
occurs, and plays an important part in a large number of processes. -
If experiment shows us such an amount, then that amount of elec-
tricity is the probable quantity of the particles of electricity. The
region in which we must investigate, is that of those processes in
which electricity interacts with ponderable atoms, and defines the
action of those ponderable atoms—that is to say, the region of
electrolytical decompositions and combinations.

Here we meet with Faraday’s law, which, referred to individual
atoms, may be expressed as follows :—

Let KA be an electrolyte, which is separated by the voltaic cur-
rent into the parts K and A, of which each has the valency n; let
q be the quantity of positive electricity which goes with each sepa-
rate atom or radical K to the kathode; g/m is then for all bodies
and for ali currents the same absolutely definite magnitude.

On the basis of the above we may also say: g/1 is the absolute
quantity of an elementary particle of electricity with the same pro-
bability with which twelve is the relative atomic weight of carbon.

q/n can be easily calculated. Leth be the magnetic intensity of
that current which in unit time liberates a milligramme of hydrogen,
ch its intensity in mechanical measure, N the number of molecules
of hydrogen in a milligramme; the milligramme contains then 2N
atoms, and these bring the quantity ch /2 of positive electricity to
the kathode, by which n=1. Hence the quantity which is attached
to an atom is 4, |

In this we have approximately* in mm., mg., sec., ¢= 3°10", A=957 ;
and further, according to the theory of gases, N=14:10". This

aii E=0-00000051 mg.? mm.? sec.-!.

This value is thus the probable “Atomic quantity of Electricity.”
It may be a multiple, but with the same probability with which
C= 12, and not 6, or 3, is E the quantity of electrical elementary
particles. For even if electricity can be split into smaller parts
than E, it is not clear why such a smaller part is never met with
in experiment.— Wiedemann’s Annalen, No. 8, 1885.

_* Conf. Wiedemann, Galvanismus, iii. p. 450; and O, E. Meyer, Kine-
tesche Theorte der Gase (Breslau, 1877), p. 234.

THE
LONDON, EDINBURGH, ann DUBLIN

PHILOSOPHICAL MAGAZINE

AND

JOURNAL OF SCIENCE.

[FIFTH SERIES.]

OCTOBER 1885.

XXXIV. On the Electric Conductivity of Gases.
By }. STENGER™*.

[Plate IX.]

3 ers present research has two objects. In the first part,
the author endeavours to show, partly by use of the
unusually rich literature of the subject, and partly from his
own researches, that there exists no distinction universally
applicable between the arc-discharge and the glow-discharge.

Upon what factor the appearance of the one or other of the
two forms depends, and under what experimental conditions
the two forms of discharge can pass into each other, will form
the subject of the second part.

Part I.

I may be allowed, in the first place, briefly to place together
the essential characters of the arc-discharge as distinguished
from the glow-discharge ; and for this purpose I confine
myself entirely to the normal form of discharge.

(1) The gas-stratum in the arc-discharge possesses a much
smaller resistance than in the glow-discharge.

(2) In the arc-light the anode is more strongly heated
than the kathode, whereas in the glow-light the reverse is
the case.

(3) In the spectrum of the arc-light, the light due to the
substance of the electrodes overpowers that of the stratum of
gas between them; while, on the other hand, in the glow-dis-

* Translated from Wiedemann’s Annalen, No. 5 (1885).
Phil. Mag. . 5. Vol. 20. No. 125. Oct. 1885. Z

306 M. F. Stenger on the Electric

charge the spectrum shows only the lines of the stratum of |

gas, and the nature of the electrodes is a matter of indifference.

(4) In the arc both electrodes become disintegrated
although in different degree, whilst with the glow-discharge
disintegration takes place only at the kathode.

§ 1. On the Resistance of the Stratum of Gas.— With the
glow-discharge in exhausted gases it is, according to Hittorf *,
the first two layers of the kathode-light which oppose
resistance to the passage of the electric current, in comparison

with which the resistance of the positive brush-discharge is’

very small.

Since, as the pressure decreases, both layers continually
expand, the resistance of the kathode-light increases simul-
taneously. When Hittorf passed, on the other hand, to higher
pressures of gas, the thickness of the glow-light decreased
and the strength of current increased. With a vacuum-tube
filled with nitrogen of about 17 millim. pressure, there
appeared, as with high exhaustion, as soon as the circuit
was completed, a glow-light of 1 millim. thickness upon the

kathode of 17 millim. length, which speedily heated the

wire to bright redness, which increased to a white heat
sufficient to fuse the thick iridium electrode. If the density
of the gas was still further increased, the anode also became
white hot, and at the pressure of 53 millim. the anode was
even hotter than the kathode. Atthe same time the strength
of the current increased continuously up to 2 amperes, a
current-strength not greatly different from that obtained
with the arc-light. Certain other observations of Hittorf T
and Goldstein t, however, show very clearly that if the
experimental conditions are favourable, the resistance of the
gas in the glow-light is of the same order as in the Davy
are-light. |
Hittorf employed as kathode a platinum coil heated by
means of a strong current passed through it, and found that
the resistance remained unaltered as long as the platinum was
only red kot, but that the sudden diminution of resistance
occurred as soon as it became white hot, which continued as
the temperature of the platinum rose. If the experiment was
carried out in the same way with a Carré’s carbon-rod, it was
possible to obtain a discharge from 10 small elements with the
electrodes 4 centims. apart. At a distance of 15 centims.,
4() elements gave a constant current of 4) ampere. When
Hittorf employed as kathodes the carbons of the Davy are,
the action was much greater. If, on the other hand, the anode

* Wied. Ann. xxi. p. 97 (1884). + Ibid. xxi. p. 183 (1884).
{ Ibid. xxiv. p. 81 (1885).

: et te
an Zi
ii a
= — eee ee,

Conductivity of Gases. 307

was strongly heated, no change in the current-strength could
be recognized.

Goldstein has obtained exactly similar results. It appears
therefore from these investigations that it is only necessary to
employ suitable experimental conditions, in order to obtain
with the glow-discharge currents of the same order of magni-
tude as with the arc-light.

§ 2. Does the temperature of the electrodes in the arc-
discharge depend upon the pressure and nature of the
surrounding gas?

Gassiot was the first to make observations on the different
temperature of the electrodes. A few years later Grove™.
investigated the behaviour of the arc-light in different gases,
and in a vacuum such as could be obtained at that time. His
results only partially agree with my observations. According
to Grove, the temperature of both electrodes is the same in
hydrogen or in nitrogen, or in a tolerably perfect vacuum,
judging from the colour and rapidity of cooling. But, as he
himself remarks, in an atmosphere of hydrogen he only
succeeded with carbon electrodes in maintaining the arc
constant for any time, so that his result that then the electrodes
show no difference in temperature cannot be considered of
any great weight. I have also endeavoured in vain to obtain
a constant arc-light between metallic electrodes in an atmo-
sphere of hydrogen. As Liveing and Dewart+ have remarked,
the length of the are is much less in hydrogen than in air, and
hence a small increase in the distance of the electrodes apart
extinguishes the light. But that the temperature of the
electrodes may be very greatly different, when the current
is closed only for a short time, and when it is maintained -
burning, appears clearly from the statement of Moigno, that
after contact between the electrodes has been broken for pro-
duction of the arc, first white light flashes out from the point
of the negative electrode, and only then does the positive
begin to glow. With carbon electrodes, on the other hand, I
have established by a series of experiments that in hydrogen,
as in nitrogen, the anode always possesses a higher temperature
than the kathode, although the difference may not be so
marked as in air. There would appear then to be some
ground for Grove’s suggestion that secondary phenomena play
a part in gases which contain oxygen. ‘That in observations
of this kind the arc should be maintained for some minutes,
before making the comparison of temperatures, appears from
the fact that both carbons are often seen to cool simultaneously,

* Phil. Mag. [3] xvi. p. 478 (1840).
Tt Proc. Roy. = ee p-. 156 (1880).

308 M. I’. Stenger on the Electric

when the arc has been interrupted immediately after its

establishment. Most of these experiments have been made with
an extremely simple form of apparatus, which permits working
with pure gases in very various pressures, since the movement
of the electrodes can be effected without the use of stoppers,

which are never absolutely air-tight. The apparatus (fig. 1) |

consists of a bulb with two tubes blown on to it; each of the
tubes is further provided with a tube at the side, of which the
one communicated with a Bessel-Hagen air-pump and the other
with the apparatus for evolving the gas. The ends of the
tubes are provided with caps fitted by grinding, carrying the
electrodes, and the electrodes may be brought into contact
and separated again, so to form the arc, by simply turning
these caps.

This arrangement, however, could not be used for experi-
ments with an arc of greater length. In these cases I have
employed another arrangement, represented in fig. 2 in simple
form. A tube (0) 1 centim. wide and 80 centim. long, was
melted on to the centre bulb-shaped part a, and communicated
by means of an india-rubber tube with a second tube 06’ of
similar dimensions. ‘The short side-tube c¢ communicated
with the mercury-pump. At the top a short tube d, also
1 centim. wide but 10 centim. long, was melted on, upon which
a wide tube e was placed, from which lastly a tube /, 80 centims.
long, led downwards and was connected with another (g) by
means of an india-rubber tube. One of the two electrodes (the
apparatus was used only with carbon points) was inserted in
the tube d, so that its end reached to the middle of the bulb-
shaped portion, and was held in position by means of four
copper bars, attached to a copper ring which was clamped
close to the carbon. The wide tube e was then melted
together at the top. The current was conducted to the
upper carbon by means of mercury which filled the system
of tubes /, g and the space between c and d, and surrounded
the four copper wires. ‘The lower carbon floated on mercury
in the tube b, so that the distance between the carbons could
be altered by simply raising or lowering JU.

It appeared to me of importance by means of this apparatus
to control the statement of Grove, that in a vacuum the two car-
bons show no difference or very little difference in temperature.
For this purpose air, which had been repeatedly and care-
fully dried, was introduced into the apparatus and exhausted by
means of the air-pump to a pressure of less than 75 millim.
As soon as the current of the Gramme machine had established
the are-light, a considerable increase of pressure took place
in consequence of evolution of gas from the glowing carbons.

:
v5
Ox
, i

Conductivity of Gases. 309

The level of mercury in the communicating tubes b b/ con-
sequentiy sank, and the distance of the electrodes increased
considerably, often up to 2 centims., without extinguishing the
light. On the contrary, the light was remarkably steady.
The temperature of the carbons was then only a little different,

but the difference could be recognized with certainty, and

upon interruption of the current the kathode ceased to glow
somewhat sooner than the anode. When the are had been
several times restored, the apparatus being each time exhausted
as much as possible, the evolution of gas gradually decreased ;
and when at last no further increase in pressure could be
perceived, the difference in temperature between the carbons
disappeared at the same moment. But since the slow disin-
tegration of the electrodes continually exposes new portions
to ignition, I have not succeeded in obtaining pressures
under 1 to 2 millim., so that it remains undetermined whether
at still lower pressures the kathode would be more strongly
ignited than the anode.

In the usual forms of glow-discharge it is always observed
that the kathode becomes red or even white hot, while the
anode remains dark. That, however, forms of glow-dis-
charge exist in which the temperature of the anode is higher
than that of the kathode appears from the experiments of
Hittorf described in the first paragraph.

§ 3. The Spectrum of the Arc and of the Carbons.—If we
observe the spectrum of the arc-light as it is formed in
atmospheric air of ordinary pressure, it appears continuous,
except that we occasionally see a few metallic lines appear.
But at the moment that the current is interrupted, a very
large number of bright lines blaze up, which have their
origin in the mineral impurities of the carbon. Spectroscopic
observations of the carbon-light are made more conveniently
by producing the are in vacuo, and with comparatively feeble
currents ; we have then at once the advantage that the
distance between the carbons may be increased to 1 centim.
without the arc becoming broken, and that consequently the
spectrum of the kathode, anode, and are can be easily
observed separately. But the main advantage is that the
extraordinarily intense continuous spectrum of the particles of
carbon which exist in the arc at a brilliant white heat, does
not then obliterate the bright lines of the metallic vapours
present in the arc.

It has always been supposed that the gaseous atmosphere
in which the arc-light is produced is without influence upon
the spectrum of the are ; Liveing and Dewar * first observed

* Proc. Roy. Soc. vol. xxxv. p. 75 (1888).

310 M. F. Stenger on the Electric

the lines C and F in the arc produced by the alternating

current of a De Meritens machine in an atmosphere of
hydrogen ; but with the continuous current of a Siemens

machine, on the other hand, only the line C was observed at the —

moment of interrupting the current, and the F line not even
then distinctly. The reason of this behaviour is to be found
only in the overpowering brilliancy of the spectrum of the
white-hot particles of carbon ; for in my experiments in vacuo
the hydrogen lines as well as the metallic lines stood out extra-
ordinarily sharply from the bright background of the are
just the same, whether the apparatus had been filled with
hydrogen or with dry air. If, on the other hand, we observe
the light emitted by the electrodes themselves, the phenomena
are essentially different according to circumstances. I will
endeavour briefly to describe the process. If the carbons
had been already in use for some time, so that the small
quantities of gas set free from them by the high temperature
no longer produced any rapid change in the pressure, the
image of the are of light was thrown by an assistant upon
the slit of the collimator of the telescope by the aid of an
achromatic lens, and the light from the upper or under carbon
was cut off from the slit by means of a suitable screen. If
the current used was weak, the spectrum of the carbon was
feeble, and showed at first no bright lines ; after a short time,
however, the hydrogen lines as well as a number of metallic
lines became sharply visible, first on the one carbon and then
on the other. The cause of these sudden changes is to be
sought, no doubt, in the fact that the arc does not burn

uniformly round both carbons, so that the parts of the carbon —

points turned towards the collimator at times glow intensely,
at times are only feebly luminous. If the current is strong,
the spectrum of the carbons is a continuous one.

It results from the observations described that hydrogen
is set free at both carbons ; we thus obtain a confirmation of
the conclusion that the carbons employed for electric lighting
always contain hydrocarbons, which the high temperature of
the arc decomposes, partially at least. The investigations of
Dewar * have shown that it is impossible to free carbon from
hydrogen compounds, even by lengthened heating in a current
of chlorine. We are altogether ignorant what composition
these compounds may have, or what changes they may suffer,
at the high temperature of the arc, except that there is no
doubt of the formation of acetylene in the are in hydrogen, as
shown by Berthelot. But whether in this case we have really

* Proc. Roy. Soc. xxx. p. 87 (1880),

Conductivity of Gases. 311

to do with a synthesis of carbon and hydrogen to C,H, or
whether acetylene is not also formed in a vacuum, and com:
sequently we have to imagine a partial decomposition of the
hydrocarbons present in the carbons, appears still uncertain.
§ 4. The Disintegration of the Electrodes.— Under the usual
conditions of research the process of disintegration produced

_by the glow-discharge is confined to the kathode ; as far as

this extends the glass wall of the vacuum tube is covered with
an extraordinarily thin reflecting deposit, which forms very
quickly when thin platinum wires are used. The surface of
the electrode after some use is found to be eaten away into
fine points and hairs, as is seen most distinctly with the
difficultly volatile aluminium.

But when Hittorf sent the current of his battery without
interposed resistances through a vacuum-tube filled with
hydrogen or nitrogen at about 50 millim. pressure, both
iridium electrodes became white-hot, and even began to melt ;
at the same time the glow-light disappeared from the kathode,
and with it the metallic deposit. Hence it appears to me
that we must ascribe the metallic mirror formed in normal
cases, not to a superficial volatilization produced by the high
temperature of the glow-light, but to some peculiar action of
the glow-light. An observation of Dewar’s would seem to
show the same thing, that a metallic deposit formed with
magnesium electrodes disappeared again some time after the
interruption of the current ; possibly the gaseous molecules,
whilst conveying the glow-discharge, possess greater affinities,
and form compounds with the metal of which the kathode
consists, which under favourable conditions are again gradually
absorbed by the electrode. Warren De La Rue and H. Miiller
obtained a similar result with palladium electrodes in hydrogen-
tubes. Under the experimental conditions, however, where
both electrodes are heated to intense white heat and even to
the melting-point, both emit metallic vapours, as occurs in
a greater degree in the arc-light between metallic electrodes.

That with this latter mode of discharge the volatilization
of the electrodes takes place as well in air as in other cases
and in vacuo has long been known. The high temperature
of the electrodes is the chief cause of their wasting away,
and the loss is only secondarily due to combustion. In air,
in consequence of this higher temperature the anode is used up
more rapidly than the kathode ; and, further, the electrodes
become disintegrated so much the more rapidly the more
volatile they are.

The process is somewhat more complicated with the carbon
light. With very powerful currents (400 to 500 Bunsen

312 M. F. Stenger on the Electric

elements) according to Despretz *, the carbon of the electrodes
is changed into vapour, and on the surface of the carbon are
seen certain round globules molten together. With feebler
currents, such as are generally employed for electric lighting,
there can be no possibility of the volatilization of the electrodes.
As already remarked, all these carbons contain hydrocarbons ;
at the high temperature these become decomposed and leave
the electrode as a porous mass of carbon, which gradually
crumbles away. In my apparatus (fig. 2) accordingly I
found numerous carbon particles imbedded in the fine
deposit on the glass walls, which easily dissolved in nitric
acid, and must therefore have consisted of compounds of
carbon, and not of carbon itself.

Since, further, in a vacuum no difference of temperature
between the electrodes could be detected, the waste becomes
equal from both, disregarding, of course, differences in the
porosity and composition of the carbons.

Collecting, then, the results so far obtained, we see that
there exists no certain criterion by which to distinguish a gwen
discharge as glow-discharge or arc-discharge, but these very
different typical forms pass insensibly into each other.

Before passing from this first part to the intimately
connected considerations of the second, I wish to mention in
§ 5 an observation, which appears to me to be not without
interest.

§ 5. On Changes of Pressure in the Arc-light.—According to
Warren De La Rue and H. Miller t, an increase of pressure
occurs in the arc-light upon completing the circuit, which
disappears immediately upon interrupting the current.

According to their data the increase of pressure at an initial
pressure of 13 to 28 millim. amounted to from 25 to 50 per cent.
Nevertheless some doubt remains as to the existence of the
phenomenon, since the experiments { hardly leave it doubtful
that the authors did not work with a normal arc-light. By
accident I made an observation which seems to show in
fact that the discharge produced a momentary increase in
pressure. I had led hydrogen into the apparatus shown in
fig. 4, so that the pressure amounted to about 50 millim. : as
soon as, by raising the mercury, contact between the carbons
had been established and the arc had been formed, the level
of mercury sank, and with it the lower carbon, until the
length of the are amounted to 3 centim. Fig. 3 may
represent the appearance of the arc at this length. The

* Comptes Rendus, xxvii. p. 757 (1849); xxix. pp. 48 & 709 (1849).

8

tT See Goldstein’s criticism in Fortsch. der Physik, 1880, p. 858.
{ Phil. Trans. clxxi. p. 65 (1879) ; Proc. Roy. Soc. xxix. p. 286 (1879).

Conductivity of Gases. 313

centre is surrounded by a helmet-shaped strongly luminous
envelope, about which again is a less luminous layer. Then
suddenly the lower carbon rose, and came into contact with
the upper one, and the process began anew. ‘This occurred
some 50 times in a minute. Evidently an increase of

_ pressure occurred which increased the distance between the

carbons, so that the arc was extinguished ; but immediately
the pressure returned to its former magnitude, and the
carbons came into contact again.

This phenomenon would be very well explained by the
hypothesis of A. Schuster *, according to which the process
of electric discharge in gases is accomplished by the dis-
sociation of the molecules, but that as soon as the current is
interrupted the old condition of the gas is restored.

Part II,

The final result of the first part of our investigation was
that the arc-discharge and glow-discharge cannot be sharply
distinguished, and that in particular the usually enormous
difference in the resistance of the layer of gas does not always
exist. In this second part I shall endeavour to show upon
what conditions the different magnitude of resistance depends,
and that the same cause is at work in all cases of gaseous
discharge in which the resistance of the gas is small. It will
be the more convenient at once to state my view, and then to
show that it is correct in particular cases (§ 6-11).

When in gaseous discharge the resistance of the gas is small,
hot metallic vapours are present which conduct the current.

§ 6. I would first of all recall certain experiments which
show that incandescent metallic vapours conduct infinitely
better than nitrogen, hydrogen, or air.

De La Rue femployed for this purpose a globe-shaped
vessel with four tubulatures ; two opposite each other served
as electrodes of an induction-coil, and the two others for the
production of the arc ; the pressure of the nitrogen amounted
to from 2 to 3 millim.

A galvanometer included in the induction-cireuit served to
measure the current-strength, before the voltaic arc was
formed. Then the arc-light was produced, and the new
current-strength read off.

The result obtained was a great increase in the conducting
power, when the arc passed between silver and copper elec-
trodes ; the change was less with aluminium electrodes, and

* Proc. Roy. Soc. xxxvii. p. 317 (1884).
~f Phil. Mag. iv. pp. 29, 553 (1865) ; Comptes Rendus, |x. p. 1002 (1865),

te

314 M. F. Stenger on the Electric

feeblest with zinc, cadmium, or magnesium poles. The
increase in conductivity was very remarkable with carbon —
poles. On the other hand, iron and platinum electrodes gave
no perceptible difference ; whence it results that the increase
in conductivity in the other cases cannot be brought about
by the high temperature of the nitrogen. The experiment
is, however, so far unfavourable, and allows no conclusion to
be drawn as to the true conductivity of the different metallic
vapours, since, as Hittorf’s work sufficiently proves, the chief
resistance of the glow-discharge occurs in the neighbourhood
of the kathode ; and with the great distance between the
electrodes in De La Rue’s experiment, the formation of
metallic vapours takes place exclusively in the neighbourhood
of the positive brush-discharge.

Moreover, it is to be remembered that in an experimental
arrangement of this kind we have to consider, not only what
metallic vapours are present, but without doubt also their
quantity, that a less volatile metal may therefore afford better
conduction than a more volatile one, although the true con-
ductivity, in the first case, may be far worse than in the
second.

As a second proof of the relatively good conducting-power
of two metallic vapours, I may quote certain observations of
Hittorf’s * :—‘‘ At the temperatures of our flames, at which
the gases of which they consist possess so considerable an
electric resistance, other vapours have a much greater con-
ductivity. Of all gases, the vapour of potassium conducts
best. Next comes sodium. The other metals, so far as they
are volatile under these conditions, produce, in the state of
gas, little change in the deflection.”’

Hittorf shows in another place that mercury vapour, in the
non-luminous Bunsen flame, conducts much worse than
potassium vapour.

That most metals in the Bunsen burner produce no percep-
tible change in the resistance, is no doubt explained by the
fact that the temperature is too low for a sufficient evolution
of vapour. |

After these preliminary remarks, I will now pass on to con-
sider, in the particular cases where a stratum of gas possesses
a relatively small resistance, whether incandescent gaseous
vapours are present, and to show their influence.

§ 7. As far as the arc-light between metallic electrodes is
concerned, it is universally known that the colour of the light
varies essentially with the nature of the metals employed,

* Pogg. Ann. cxxxvi. p. 229 (1869),

Conductivity of Gases. 315

and that in the spectrum of the arc the corresponding
metallic lines appear sharply.

According to Casselmann™ the light-arc is of different
length with different metals, and is so much the longer the
more volatile these are.

He arranges the metals in the following order:— Potassium,
sodium, zinc, mercury, iron, tin, lead, antimony, bismuth,
copper, silver, gold, platinum. So that potassium yields the
longest arc, and platinum the shortest.

As already remarked, there are essentially two factors to be
taken into account—the true conductivity of the vapours, and

_ the quantity in which they are formed ; since further, the

conductivity depends in a high degree upon the temperature,
it is not to be wondered that Hittorf’s experiments in the
Bunsen flame, and Casselmann’s in the arc-light, have given
different results.

§ 8. It is less easy to understand the conditions which
obtain in the carbon light. That here also metallic vapours,
which are introduced into the arc, increase the conductivity,
is of course obvious from the preceding results, and it is not
even necessary to refer to the data of Casselmann. Hence
we obtain a longer arc between carbons which have been
saturated with metallic salts than between carbons as they
are found in commerce.

But that metallic compounds are abundantly present also
in the carbons which are now so largely used for electric
lighting may be easily shown. As is well known, the
spectrum of the electric arc shows a multitude of bright
lines, which are particularly distinct at the moment of inter-
rupting the current, when the continuous spectrum of the
white-hot particles of carbon has faded.

The sodium-line is especially brilliant ; besides which we
always observe the lines of calcium, iron, and magnesium.
The result is always the same, of whatever manufacture the
carbons may be; and it may perhaps be of interest to compare
different opinions as to the preparation of good carbons.
According to Carré, good carbons may be obtained by
compressing an intimate mixture of lampblack and powdered
coal, and strongly igniting after addition of iron, antimony,

or tin, or by boiling the carbons for a long time in a solution

of metallic salts. Archereau and Sandain recommend a
similar method. Jacquelain { proceeds in an entirely different
way, since he endeavours to remove all mineral constituents

* Pooe, Ann. Ixiii. p. 576 (1844). -

+ Comptes Rendus, |xxxiv. p. 346 (1877).

t Pogg. Ann. Jubelband, cecexl. (1874).

316 M. F. Stenger on the Electric

from the carbon by ignition in a current of chlorine, or
treatment with molten potash, or by soaking in hydrofluoric
acid. As, however, Liveing and Dewar* have shown, these
methods are not successful in completely removing iron,
magnesium, sodium, and calcium.

§9. Whilst electric discharges in air, hydrogen, nitrogen,
or carbon dioxide require a large electromotive force, because
they have to overcome a very high resistance, it is possible
in the presence of good conducting metallic vapours to obtain
the arc-discharge with comparatively few elements. Thus
Hittorf +, with cylinders of retort-carbon, the ends of which at
a distance of 3 to 4 millim. were brought into a Bunsen
flame, in which a bead of potassium-salt was heated, obtained
the are-light with 80 of his elements, without its being
necessary to first make contact between the carbons. For
this it was of no importance whether both points or only

the kathode was surrounded by potassium vapour, because,

as with the glow-discharge in exhausted gases, the neighbour-
hood of the kathode offered a much greater resistance also
in the gases of the flame.

Further, Gassiot, employing a battery of 400 Grove’s ele-
ments, observed that upon making contact, the discharge
between balls of carbon or of metal was discontinuous, but
very quickly passed into the continuous arc-discharge,
evidently because metallic vapours had been formed by the
sparks which leapt across, which by their small resistance
permitted the establishment of a constant current of great
intensity.

Lastly, Herschel’s artifice is well known, of producing the
arc, not by contact of the electrodes, but by means of a spark
allowed to pass between them.

§ 10. I have already had occasion to mention observations
of Hittorf{, in which he obtained in nitrogen of 53 millim.
pressure currents of considerable strength by means of his
battery of 1600 elements.

The considerable decrease in the resistance of the stratum
of gas, as I have already mentioned, was accompanied by
intense white heat of the electrodes, so that here also we may
very well regard the abundant formation of metallic vapours
as the cause of the phenomenon.

The influence of the white heat of the electrodes is seen sti]]
more distinctly in later researches of Hittorf § and Goldstein ll,
which were made with very small pressure of gas. The older

* Proc. Roy. Soc. xxx. p. 155 (1880) ; xxxiii. p. 406 (1882),

+ Comptes Rendus, Ixxxiv. p. 346 (1877).

{ Wied. Ann. xxiv. p. 81 (1885). § Lbid. xxi. p. 133 (1884).

| Ibid. xxi. p. 111 (1884).

Conductivity of Gases. 317

observations do not show the decrease in resistance so
strikingly, no doubt because, whilst the resistance at the
kathode is greatly diminished by the formation of metallic
vapour, the resistance of the positive light is increased by the
higher pressure of gas.

But if by suitable means we heat the kathode in exhausted
gas to the point of volatilization, we may obtain powerful
currents even with small electromotive forces. In fact,
Hittorf and Goldstein have observed these phenomena; I
will only mention of their results, that with a distance between
the electrodes of 6 centim., one small element produced a
constant current with high exhaustion and heating of the
kathode to white heat.

§ 11. Warren DeLaRue and Miiller*, in a comprehensive re-
search, have endeavoured to show that glow-discharge and arc-
discharge pass into each other in consequence only of change
in the density of the gas. J wish here to express again my
doubt (referring also to the criticism of Goldstein f) whether
the authors did in fact have arc-light; their assumption is
also in direct contradiction with the previously published
observation of Gassiot, on the passage of glow-discharge into
arc-discharge at unaltered pressure.

Moreover, in accordance with the preceding consideration,
the pressure is of small importance ; but everything depends
upon whether the space between the electrodes is filled with
incandescent metallic vapours or with ordinary gas. I have
further had opportunity in my experiments to observe the
passage of an arc-discharge into a glow-discharge, which is
easily explained on the same principle.

If with the apparatus figured in fig. 2, at a pressure of
10 millim. the are was lengthened until the resistance increased
to the point of extinction, there appeared immediately in front
of the anode a stratum of blue glow-light about 1 millim. thick
and 10 millim. long, which, after a few seconds, disappeared
with the light between the carbons.

If, then,the resistance became too great for the arc-discharge
to continue, the conductivity of the still glowing metallic
vapours was still sufficient for a short time to permit a much
feebler current in the form of glow-discharge.

§ 12. In this concluding section, I should like to make
some remarks on the luminosity of gases and vapours, which
connect themselves with the experiments, already several
times mentioned, of Hittorf. If, in electric glow-discharges,
the electrodes came to an intense white-heat, the blue glow-

* Phil. Trans. clxxi. p. 65 (1879).
+ Goldstein, Fortschr. der Physik, 1880.

|
|
j

318 Mr. R. H. M. Bosanquet on Electromagnets. —

light disappeared each time. It follows from this, as well as
from former observations of Hittorf* and W. Siemens f on
the luminosity of flames, that up to the temperature of
melting iridium, gases possess no perceptible emissive power,
if they are not concerned in chemical changes or in electric
discharges. The cessation of luminosity at the white heat of
the electrodes may perhaps be explained by the hot metallic
vapours taking over the conduction, so that only a small
fraction of the current goes through the hydrogen or
nitrogen, which is not sufficient to bring it to luminosity.
The conditions under which metallic vapours emit light may
be quite different from those which hold good for ordinary
gases. The investigations of Schuster { have shown, at least,
that mercury vapour in a vacuum tube, from which every trace
of other gas has been expelled, presents essentially different
and, in fact, simpler phenomena, so that the electric discharge
takes place without glow-light, dark space, or stratifications.
Whether Schuster’s explanation is correct, that the reason is
to be found in the simpler constitution of mercury vapour, or
whether metallic vapours generally do not show this simpler
mode of discharge, seems to be worthy of further investigation.
Since I have myself always seen an unstratified coherent
mass of light filling the space between the electrodes in an
are of 3 centim. in length, the second possibility appears to
me the more probable.
Physical Institute,
University of Strassburg.

XXXV. Electromagnets. —1V. Cast Lron, Charcoal Iron, and
Malleable Cast Iron. By R. H. M. Bosanquet, St. John’s
College, Oxford.

To the Editors of the Philosophical Magazine and Journal.
GENTLEMEN,
{ew following measures of permeabilities have been made
on a number of rings of cast iron, charcoal iron, and

malleable castiron. ‘The magnetizing forces used varied in all
cases from small forces to forces large enough to raise the
metal to a condition but little removed from saturation. Ido
not include at present any comparisons with formule, as the
facts are perhaps best kept separate in the first instance.
These measures complete all the experiments on rings which
I have contemplated so far.

It may be convenient to summarize shortly some chief
points of the results of all the experiments on rings.

* Wied. Ann. vii. p. 587 (1879). { Ibid. xviii. p. 311 (1883).

t Proc. Roy. Soc. xxxvile p. 317 (1884).

Tse
oh

Mr. R. H. M. Bosanquet on Electromagnets. 319

Permeabilities or Susceptibilities.

Small Maximum

Permeability for For
inductions.| permeability.

93 =10,000. | 33=14,000*.

ee 5

: 50 170 | 30
Cast UPON .eeeeeeeeeeeees 80 O50 30
Malleable cast iron ... an -” ns Zz
200 1800 1200 500
Wrought iron ......... 450 | 2500 92000 1000
450. | 2900 600
Charcoal iron ......... 470: | 3000 | 2000 260
ok Dest; eal |_| — ——— | —____—__|—_____
Cast tool steel (hard). { £P ee ee °
GO. =| 420 | 350
» » (soft) .. | 120 | 460 250 | -
Saturation-points. |
BEM rd PR ik. sk.aie vo keu cu enidh vaneenss 9,000—14,000.
Miatlcable Cast ION 2.20.2. 2.02. vce cesnno a's .-- 14,000—16,000.
Wrought iron
eon | AS SE Be aes ee 17,000—19,000.
Cast tool steel, both hard and soft ...... 20,000.

The malleable cast-iron rings were supplied to me by
Messrs. Crowley, of Sheffield. The materia! cuts exactly like
charcoal iron; and it seemed probable that its magnetic qua-
lities might be similar to those of charcoal iron. In this case
the possibility of casting the material would have been of the
highest importance with reference to dynamo machines.

The results for ordinary cast iron show that it is wholly
unfit for use in dynamo machines; the low saturation-point
being particularly objectionable. The malleable cast iron
would undoubtedly be a great improvement on ordinary cast
iron ; but still inferior to wrought iron.

It may be noted that the variations of temper in the differ-
ent rings of malleable cast iron did not appear to correspond
to any recognizable differences, either mechanical or magnetic.

About the time these experiments were drawing to a close
I became aware, by a notice in ‘ Nature,’ that Dr. Hopkinson
had communicated to the Royal Society a series of investi-
gations apparently similar to my own. I have not yet been
able to learn any thing about the contents of the paper in
question ; but considering the diversity of the properties of
different specimens, as well as of the methods in use, [I still
hope that my results may not be without usefulness.

* As before, 93 stands for the magnetic induction, or number of lines of
force across unit area; yw for permeability or susceptibility.

Mr. R. H. M. Bosanquet on Electromagnets.

320

£0 919%
ag 98101 1IZ Bugz
TL L6¥6 GIG LL¥G
€6 828 681 SGLT
IIT CLLL 6ST 9TST
OFT o8e9 Gal OFZT
SsT Shae PGT 9ZET
661 F68P IZI SOL
ai 680% 6 SIP
9G G09e GL 6G]
61 CE0E 19 ¥G
; ‘sqo wf OF ‘sqo a 4
By
= ‘QeTT= eebcceces S[IOo jo Ioquan Ny
M 66 ee eeescccceces SSO YOIY]-LVq
ne “UI1qUe0 LOT —— Stee e ones 19yoWVIp uve]
; “(panyy wnapepy) Ni ‘Pury wowr-2s09
Fa GLI ZC6P
rs OF I9TGI 881 ISP
<j OF 99101 161 Sach
H £9 S8901 10 CIP
GL 616 Z6L E898
8d 6988 G0 FILE
6 1298 FIG SI0g
FOI GIFL O1Z LOSS
981 1189 OOT ZO8
ial 6619 zal 9
6ST 089g GL SL
‘sqo a a =) ¢ “sqo vl ms) 4
‘980, "Byron Jo soquin yy
ce OGF-T — eeorcereccce SSOUYOIY}-AV__
"UIJUOD GZG.[Z= “'''''* LOJOULVIP UBOTAT

‘(puvpy wnrpayy) %) ‘bump Uuouy-7809

“WUIJU9O 1.0, =

STLOD jo 19Q UNL NT
eeerceee eee *SSOUYOIy-LVq
sewn teve.e IOJoULVIP Weep

‘(puny wnipayy) gq ‘buy uowy-3s09

&& 8606
bv T818
GE 669L
09 LLIL
GL 6979
&6 G69G
FOL LESG
ole 888P
Lot 1097
ovl IFIP
“sqo a “oF

“6C6L=

~~ 98h: =

"WII}UO) Z9.TS=

‘(pungy) uouy wapuylig “J, ‘bung wouy-7s09

| ot
on ESE
Fal

O9T

“sqo a

cc6&
8996
LOGS
EG6E
6E66
I88G
VGPG
L006
LIG
L9

“SE

ee eeceece SOO jo 19q TAN NT
eee ee SSOUYOIYI-Ivq
coccccccs I9J9UIBIp UvdT

LiL POL

OF COLFT SG, 0062 pray reotiagee
Ae at Ne ee LOT LLCS LLC T#88
sal 196 FEOLL £9) £009 SSP 98SE1 OFGG 918
a gga POSPT 88az STFS
oa ae Peay si oe gcg 18061 £296 6FIS
ae ie A ae SZEI GOZEL 66LZ £089 <
rh Sone soy ba aa 69CZT C682 esis n
ae mane oar een ePLl EZLIL ||. 9982 CLEP
OL8T ISZ11 916T Laat
999 TTO8 19), 6298 rie Bach tap a
619 628! CP 6&8 Poaseh 1 cena Oem Lae Rae" ye
90L SPLL 1G 78
T.Gr tA CEES a ee Ue ee | IES Caocansa errors: 25 *sqo nl OE *8qo nl a) $
"sqo 1 Of "sqo 1 OF
"LIG= eeareeeee STIOO jo JOQ UU AT
Saal a ere og 100 0 zoquan yy ce POET — eoeeeesecovs SSOUyOIY}-legq
“cea, = ote i pee caeniicditer "WI}USO G.QT= “""''' ceyeueIp uve
Uhteo 7.OL=. LIUIVIP UVET\L

e 66 ; % ;
‘(pioqyy) 7, ‘bua wouy-7sv) arqvayopy (1fog') Wy ‘bun wouy-pn0ounyD ysrpang

Mr. R. H. M. Bosanquet on Electromagnets.

Phil. Mag. 8. 5. Vol. 20. No. 125. Oct. 1885.

06 OPTLT 1626 1066 096 oLg¢
G1Z 99ecT EGGS I1¥8 8¢ E8661 08% 6098
Og COFFI 029% C88 a) 69LOT 616 F18G
909 GCOFT C196 $808 101 126 661 98ZZ
1@8 1881 G61Z 9T9L OT ChF8 OLI FOL
CLOT 08621 F466 1029 SaT OBE FOI OFST
CET OSFZT — L808 GFIS L61 F819 ESL cog
8061 SFIIT OGLZ L6ES 916 porG CTT 81g
E106 FIGOL OLLI CFE PIG Livy 98 9&6
6E1Z 61801 Gah 18 976 1688 LL OF
SBS SIO EE fe et OA Se CM ee CPA aes Boy eee || ee ue Be :
‘sqo. nl "OF ‘sqo OF ‘sqo as) “sqo of "OE
‘09G= "tt BIO Jo zoquanyT ‘9606 "1" S[109 Jo zoqunyy
6s JEG = oct ssouyoryy-aeg. ‘é GZEO-T = octet ssouyoryy-aeg

—_
—
*

. ¥ e@eesee 41949 wuetp UBOT, .

a

“THTFUCD Z.OT= ""'""* JajoureIp ueP

on ee

ae? = ; Mer sa oF

- * f la 2 , _* j in = a a“ 7 r
+ c ey oe 4 + e r . v ‘ { * — P A y pe!
= a Aaa 6 fF a % 1535 as fen fpnags fr &Aaaar an@rr aansr, |
a a eee y Ox 2 ees Speirs 2h al Ae O%& { DULL WMOMTARISTC A

"UIIJUOD FT]

Mr. R. H. M. Bosanquet on Electromagnets.

322

a

ov BEOCT + 67S O8F8
60T 89gET 8ag Go18
GT GTOET TLg 8882
G6T [e961 Ig 090L
OLG SLLIT 679 S669
986 6990T ¢99 6TLg
G8E O6G0T 169 PS9V
LOV G6SOT G19 GGOP
ShV LEOOT SLG PIGS
[sp V896 LYG [686
667 8966 L8¢ VIPl
GOS 1606 SEG G67
GGG 6168 ELT OL
"sqo “Op ‘sqo “OF
IGT alee ce ke BROS, JOG UEDA
pet = oC ssouyory-seg

"WIJUIO [0-OL=
‘fog') A ‘bury uou-ys0g ajqvayopy

La49Wvip UBoTL

Gg9 Shas
ST C6661 Lg 8866
8h Te9ET L6P GPG6T
GIG LESiT 669 GIGG
O&S OLTIT OS? TéGT
SIP 69G0T LLE CLIT
09g S6°6 cIé 668
DLG C3808 LEE LeL
618 [604 885 IhL
18Z 8664 1&6 006
LEL GLaP GG &&
‘sqo 1 “Of *sqo 1 a 5
Sl UC(Gy call Gene ere
see \ S[IOO Jo zoquan Ny
CCe.T = ctittttt ssouygoryg-aegy
‘MAI}UG0 Y.QT= “''*! Loqowuelp uve

"(puny uinipayy) -) ‘buy uosy-3sv) a)qva7)vpr

Ty

910¢T

. GOL b6GL
OOT C1661 68h 9989
O<cI 60FGT S64 9989
T6T LEviT LLL 9F99
166 SPOTT 164 PL8g
Och 90P0T F8L 196g
GOP VITOT v6L LGOG
LLG SEV6 9LL LOLF
6&9 LV¥68 PLL CEP
STL 6968 61L Els
6gL GOLL 81g SLOT
664 IpLL 186 SLY
GGL G8ed SOL GL

‘sqo ass ‘sqo 1 "OE
LUGS See so Seon
66 {et SS ae eee cee ssoUyoryy-cVgq

"ong H99 ¢/0.0T =

eeoovrveeres 1ojaue1p UvoTL

‘(punpy wnipapy) K ‘bury uosy-3svy apqvayvyg

889 1608
OV CEOST 194 PEt
L6 6GTET PLL 8FE9
aE G8Icl 684 2689
961 G6LIT 964 GPSG
ELE OcEIT 092 OZoP
G68 OZcOT 9LL SIEP
OV Gc66 96h SIGs
Gog 9666 &8G T1GG
609 C088 666 804
699 6988 66T v8
*sqo 1 "Of ‘sqo a "OF
4 og a aac Pe * S198 70. ae NT
73 Oke. T = oct ssouyoryy-regq
"MI}M09 GN.QT= "*"*"*** ~TOPOULVIP UBT

(puny) M “bury Wout 380) 2192. > FT

Influence of Heat on the Rate of Chemical Change. 323
Malleable Cast-Iron Ring, X (Soft).

Mean diameter ......... =10-115 centim.
Bar-thickness ............ ee ae
Number of coils......... = F220
| | |
: | 33. | # obs. | 33 | obs
| | |
| se | ie || 714A 735
| O76 146 7628 724
1293 491 7669 15D
| 3784 712 759 715
| 8976 701 9142 626
| 4895 "17 9526 618
| 4622 767 10373 495
| 5045 746 11080 417
| 5247 746 11907 276
| 54.49 458 12412 217
| 5853 754 12876 | 173
6196 759 13562 115
| 6781 a 162 15762 | 47

XXXVI. On the Injluence of Heat on the Rate of Chemical
Change. By Joun J. Hoon, D.Sc. (Lond.)*.

a. accelerating action of heat on the rate at which a
chemical change progresses has been made the subject
of investigation by several chemists. The amount of work ~
that has been done towards the elucidation of this side of the
great problem of chemical action, expressing the rate as a
function of the temperature, is as yet too insignificant to
allow of any deductions being drawn relative to the probable
- mechanical processes that take place in a medium containing
a chemical system undergoing change. + Considered in the
light of molecular interdiffusion, it would seem probable that
the relation between the rate of change of different chemical
systems comprising two or more active bodies (active in a
chemical sense) and the temperature should be of the same
type; and the experiments that have been made on this subject
all seem to show this in common, that the rate of the change
increases very rapidly for small increments of temperature.

Warder, in his experiments on the speed of saponification
of ethylic acetate, was led to infer that the increase of speed
in this reaction is approximately as the second power of the
temperature.

Mills and Mackey{, studying the evolution of hydrogen
from zinc and hydric sulphate at different temperatures, a
case of a chemical system of a non-homogeneous character,

* Communicated by the Author.

+ American Chemical Journal, vol. iii. { Phil. Mag. [5] xvi.

2A 2

:

324 Dr. J. J. Hood on the Influence of Heat

found a relation between what they call the “ zero strength”
and the temperature, represented by the equation
b+ ct
d+t
Urech *, from experiments on the speed of inversion of cane-
sugar by acids at different temperatures, expresses the rate as
a function of the temperature of the form at+a/?+a’é.
More recently Menschutkin f has studied this subject, em-
ploying the rates of formation of ethylic acetate, acetanilide,
and acetamide. His results show that if e and é’ represent
the amounts of these bodies that are formed during the interval
of one hour at temperatures 0° and 6°+7n”%, the value of e’ —e
passes through a maximum at a definite temperature.
Lemoine}, with reference to the simplest case of a chemical
change represented by the equation

s=at

such, for instance, as the action of hydric peroxide ona soluble
iodide in presence of sodic thiosulphate, as worked out by
Harcourt and Hsson, says:—‘‘ Un fait tres général et tres
digne de remarque est l’extréme accroissement du coefficient
A, c’est-a-dire de la vitesse de décomposition avec la tempé-
rature. Si @ est la température, A parait varier avec elle
suivant une formule exponentielle A=a’‘a’. Ce fait, résultat

d’une multitude d’expériences, parait corrélatif de la nature

méme de ce mouvement intérieur qui constitue la température
d’un corps.”

The “ multitude d’expériences,’” however, would seem to
imply the common observation of the great increase in rapidity
of a chemical reaction with small increase of temperature.

In a paper published in this Journal§, in which the oxida-
tion of ferrous sulphate by potassic chlorate was the subject
of investigation, it was shown that the course of the oxidation
was represented by the equation :

di

= =—pf(O)y. . +. +.
Integrating and writing it in the form y(a+-t)=8, it is clear

that for two experiments made at different temperatures 6”

and 6°+ 7°, the relation nae ist holds.
es ect
* Berichte d. d. Chem. Gesell. xvi. p. 762.
T Journ. Prakt. Ch. 1884.
} “ Etudes sur les Kquilibres chimiques,” p.178 (Fremy’s Ency. Chim.).
§ Phil. Mag. [5] vi. Gotoh. )

oo

on the Rate of Chemical Change. 325

It was stated that the few experiments that were made to
determine the form of this temperature function seemed to
indicate that /(@) =6@", or that the rate of change varies as the
square of the temperature.

_ This reaction has again been made the subject of investi-
gation, with regard to temperature, and the results obtained
are comprised in the following tables. |

The mode of performing the experiments has already been
fully given in several papers in this Journal, so that it will
be sufficient merely to give the strengths of the solutions
employed in the work. Hach experimental solution consisted
of ‘5637 gram iron (Fe”) as sulphate, °2057 gram KCI1O,,
and 3°099 grams free hydric sulphate, made up to a volume
of 260 cubic centims. contained in a glass flask and main-
tained at the required temperature in a large water-bath.
The course of the oxidation was observed by withdrawing 10
cubic centims. of the solution, and titrating with perman-
ganate at indefinite intervals of time, to determine the equa-
tion y(a+t)=6; y being cubic centim. permanganate and t
time in minutes. ‘Two, and in many cases four, experimental
solutions were employed at each temperature, and the values
for a and 6b that are given are the means of these. LHvery
precaution was taken to. keep the temperature of the mea-
suring pipettes, permanganate, &c. the same as that of the
experimental solution,—the thermometers employed being
very fine ones, by Negretti and Zambra, divided to tenths,
on which ‘02° could be read easily; and in every case where
an experiment fluctuated by so much as +°1° C., the results
were rejected.

Although the experiments tabulated below are only com-
prised between the narrow limits of 10° and 32° C., it was
found that outside these extremes the results were liable to
be vitiated by several sources of error. Below 10° C. the
progress of the oxidation was so slow that during the long
period that the experiments had to be continued there was
liability of atmospheric oxidation taking place; besides, a small
error in the permanganate titration made a comparatively
large one in the time, whereas above 32° C. the speed was
much too rapid to obtain accurate measurements. Between
these narrow limits of temperature the rate of oxidation
increases over sevenfold.

In the figure, the course of a few of the experiments are
plotted from the experimental numbers, and these curves
show ata glance the great differences in the rates of oxida-
tion at the various temperatures.

326 Dr. J. J. Hood on the Influence of Heat

Curves showing the course of the Oxidation of Ferrous Sulphate by
Potassic Chlorate at different Temperatures.

i
|
{
i
i
H
|
H
|

ene
Onoxidi sed tron

LNs]

Time : minutes
yoo 2100

| : bn
emp. Cy | a. b. Ratio fae

10 330°8 BOT R. Ul iAaee
11 | = 3016 3025 1-100
12 | 274-7 2752-9 1-098
13 250 2503 | 1-099
14 997-5 9282-7 1-096
15 206'6 20557 | 1-110
16 194°3 1920°8 1070 |
17 174-2 1733 1:109
18 159 1577°4 1-098
19 147-1 14526 1-086
20 134-4 1325-4 1:096
21 124 12168 1-089
22 1149 1123 1:083
23 102°6 1002°3 1-120
24 94:8 9245 1-084
25 89-9 869 1-064
28 68:°5 654°8 1:099
80 58:7 551°2 1-090
32 50°3 465°3 1-088

| Mean“ ...... | 1:093

on the Rate of Chemical Change. 327

From the foregoing table it will be seen that the ratio 2
| ry
has, as nearly as possible, a constant value, the mean from all

the experiments being equal to 1:093. It would seem there-
fore that 7 (@) may be written in an exponential form and
equation (1) becomes

as aa’y”.

dt
In this particular reaction, taking the mean value for e@ as

given above, and the rate of oxidation at 10° C. as unity, the
rate p at temperature 0° C. is represented by the equation

pHtEovs ys.
In the following Table the rates at the different tempera-

tures, as determined experimentally by the ratio at and as

eo:
calculated from the above equation, are given ; and it will be
seen that there is a close agreement between the two series of
numbers. |

TABLE II.
. Calculated Caleulated
Temp. C°. — rate of || Temp. C°.| . ou rate of
s 2. | oxidation. oxidation. | oxidation.
10 Ci, nn nett 20 2°51 2°43
ll 1:10 1:09 21 218 2°66
12 NRA 1:19 22 2°96 2°91
+3 1-33 1-31 De rE 3:18
14 1:46 1°43 24 3°59 3°47
15 1-62 1:56 25 aoe 3°80
16 tis 1°70 28 5:08 4-96
17 1-92 1°86 30 6:04 5-92
18 aol 2°04 32 7:15 7:07
19 2°29 2:23

In a previous paper, already referred to, it was suspected
that p « @ ; but as the experiments were too few in number
and between too narrow limits of temperature, the actual re-
sults were not given. It may be of interest, therefore, to see
how far the present series of experiments bear this supposition
out.

Representing the relation between rate and temperature by
the equation p=x(@+n)’, taking pjy=«(10+n)’?=1, the
following are the values obtained for n :—

328 Mr. 8. Bidwell on the Generation of

TABLE III.
Rate of Rate of Rate of
oxidation, Value for oxidation, pitas for oxidation, Value for |
0. p. 0.
1°10 10°5 1:92 8-1 3°32 5:8
1-34 10 211 77 3°59 56
35 96 2°29 75 3°83 56
1°46 92 5] TL 5:08 4:3
1°62 8:3 2°73 6'8 6°04 37
1:73 9 2°96 66 T1b ae

From this Table it will be seen that n, instead of having a
constant value, gradually diminishes as the temperature in-
creases, but that fora few consecutive experiments the dif-
ferences are such as might be referred to experimental errors,
and so lead to the erroneous supposition that p « @?.

So far as these experiments go, it would seem probable that,
for this reaction at least, the temperature function is of an
exponential form. Indeed, in many cases of chemical change
it is a common observation that the speed increases very
rapidly with the temperature, being, as Lemoine remarks,
characteristic of an exponential form. . In the course of some
experiments on retardation, in which ferrous chloride was the
material employed, instead of sulphate as in the present case,
it has been found that here also, in presence of agents that
retard the oxidation very considerably, the relation between
rate and temperature may be represented in the form p=aa’.
A full account of these experiments, however, will, it is hoped,
be given at a future time.

——_———

XXXVI. On the Generation of Electric Currents by Sulphur
Cells, By SaetForD Bipwet., MA., LL.B.*

i a communication recently made to the Physical SocietyT

I ventured the suggestion that the electric conduction of
selenium, when prepared in such a form as to be sensitive to
light, was, in the literal sense of the term, electrolytic. Sele-
nium itself indeed could hardly be supposed to be an electrolyte;
but it was pointed out that. when selenium was “ annealed ”’
in contact with metallic electrodes, metallic selenides would,
at least in most cases, be formed in sufficient quantity to
account for the electrolytic phenomena observed ; and that
even if this were not the case when the electrodes consisted
of such a metal as platinum, yet the necessary metallic element

* Communicated by the Physical Society: read June 27, 1885,
* Phil, Mag. August 1885.

Electric Currents by Sulphur Cells. 329

might possibly be found in the lead, iron, and arsenic which
are contained as impurities in ordinary commercial selenium.
Little direct evidence was offered in support of this view;
but it was shown that sulphur, when mixed with a certain
proportion of sulphide of silver and arranged in the form of a
“cell” with silver electrodes, exhibited many of the pro-
perties of crystalline selenium, especially that of having its
electrical resistance temporarily diminished under the influ-
ence of light. Analogy therefore tended to confirm the opinion
which I had been led to entertain.

While observing the secondary or polarization-currents
which are generated by sulphur cells (as by those made with
selenium), after being disconnected from a battery, certain
effects were noticed which seemed to indicate that when the
electrodes consisted of two different metals, a sulphur cell
might be capable of originating and maintaining an indepen-
dent or primary current. Hxperiments were therefore made
with the object of investigating this point; and the present
paper contains an account of the results obtained. I have
hardly attempted to connect them together by any complete
theory: of some of them, indeed, I can offer no explanation
whatever ; others appear to be in direct opposition to what
might have been expected. But, so far as I have been able
to ascertain, they are entirely novel, and of sufficient interest
to be worthy of record. |

(1) A slip of mica was wound with two parallel wires of
silver and copper 1 millim. apart, and melted sulphur con-
taining a small quantity of precipitated sulphides of silver and
copper was spread over one surface. It is not known what
proportion of the sulphides was contained in the mixture,
because the bulk of them sank to the bottom of the crucible
in which the sulphur was melted. When cold, the cell was
connected with a refiecting galvanometer, and was found to
generate a small but steady current, indicated by a deflection
of about 20 scale-divisions. When the connections of the
cell were reversed, the current was reversed. It was therefore
not due to any thermo-effect in the circuit. The direction of
the current was from silver to copper through sulphur: its
strength was diminished by exposure to light, and increased
by rise of temperature. Connecting the free ends of the
silver and copper wires and heating the junction, it was found
that the thermo-current thus produced passed (as usual) from
copper to silver through the junction. The increase of cur-
rent by heat was therefore not to be accounted for by thermo-
electric action, for that would produce the opposite effect.

Saat

SSS SES

330 Mr. 8. Bidwell on the Generation of

The action is almost certainly of the same nature as that -
which occurs in an ordinary voltaic cell. |

(2) It appeared desirable to construct a cell which, though
unsuitable for experiments with light, would have a much
smaller resistance than one of the form last described. A
plate of copper 3 centim. square was heated, and upon it was
spread a mixture consisting of 5 parts of sulphur and 1 part of
sulphide of copper. A plate of silver previously heated was
then laid on the melted mixture and the two plates squeezed
i together, thus forming a sandwich-like cell. The thickness

of the copper plate was 2°75 millim., of the silver plate -60
: i millim., and of the completed cell 3°65 millim.; the thickness

4 of the layer of sulphur was therefore *3 millim. When this
| cell (after cooling) was connected with the galvanometer, the
spot of light was violently deflected off the scale. Dr. Fleming
i was kind enough to make a very accurate measurement of its
electromotive force by comparison with one of his standard
Daniell cells. It was found to be ‘0712 volt, and its internal
resistance was 6537 ohms. As in the case of the former cell,
the direction of the current is from silver to copper; and there
can be no doubt that it is of a voltaic nature. At the time of
writing, the cell has been in existence nearly seven weeks, and
it is now, I believe, quite as powerful as at first.

(3) It was thought that the internal resistance might be
further reduced by adding a larger proportion of sulphide to
| the sulphur. Another cell was therefore constructed similar
| in all respects to that last described, except that the sulphur
{| and copper sulphide were mixed in equal proportions. Its
i

a Ss

Re SS a ae ee ee
i ee ee eee

SSS EE TO
[Ss a a a
——

———

eee
z 2 Ste

es ea ee ——
Fs ao = Se

ee ee

internal resistance was enormously lower, being only 13 ohms,
but its .M.F. was also lower, being *0071 volt.

(4) A layer of precipitated sulphide of copper was placed
between plates of copper and silver which were squeezed
together in a screw-press. ‘The resistance of this arrange-
ment was a small fraction of an ohm; but when connected
with the galvanometer, it gave no indication whatever of a
current. It seems, therefore, that a certain amount of free
sulphur is necessary for the generation of an electromotive
force by cells containing copper sulphide.

(5) Iwo parts of copper sulphide were mixed with one of
| sublimed sulphur, and the powder was compressed between
7 plates of copper and silver. This cell gave a very small
| current, indicated by a galvanometer-deflection of 2 or 3 scale-
4} divisions. The deflection was reversed as often as the con-
{| nections with the binding-screws of the cell were reversed ;
i and the existence of a real, though very small, electromotive
| force was undoubted. ‘The internal resistance was ‘088 ohm.

” _. = sc et LL LT TOL OT
lll il. OOS

Electric Currents by Sulphur Cells. 331

(6) The last-mentioned cell was taken to pieces and remade
after the addition of about an equal part of sublimed sulphur
to the mixture of sulphur and sulphide. Its resistance was
now found to be many megohms, yet it produced a larger
galvanometer-deflection than before.

(7) Once more the cell was taken to pieces and a little
more sulphide added. When remade, its resistance was at
first about 2700 ohms; but it varied considerably. It pro-
duced a galvanometer-deflection of about 100 divisions, which
in a few minutes increased in a somewhat irregular manner
to 250 divisions. It was then disconnected from the galva-
nometer, and, when again connected after an interval of six
hours, it deflected the spot of light offthe scale. Shunting the
galvanometer (the resistance of which was 3483 ohms) with
a coil of 300 ohms, the deflection amounted to 130 divisions;
and on the following day with the same shunt, the deflection

was at first about 250 divisions, rapidly diminishing, however,

when the circuit was closed.

(8) A cell was made by compressing precipitated silver
sulphide unmixed with any free sulphur between plates of
silver and copper. When connected with the galvanometer,
this cell produced a deflection which, with a shunt of 35 ohms,
exceeded 400 divisions. But in this case the silver was the
negative plate, the direction of the current being from copper
through sulphide to silver. The H.M.F’. was less than that of
the ceil described in (2). |

(9) Another cell was made in the same manner as that
described in (2); but the sulphur was mixed with sulphide
of silver instead of sulphide of copper. This gave a strong
current in the same direction as that produced when sulphide
of copper was used, and opposite to that generated by the cell
containing silver sulphide without free sulphur.

(10) Pure sulphur was melted on a clean plate of copper,
and, when just liquid, a warmed plate of silver was laid upon
it and pressed down with a weight until cold. This cell gave
a strong current from silver through sulphur to copper.
Sulphides were of course formed during the process of
construction.

(11) A melted mixture of sulphur and silver sulphide was
spread upon a copper plate, and a plate of silver pressed upon
it. When cold, the silver plate was split off witha knife, anda
piece of silver-leaf, sufficiently thin to appear blue by trans-
mitted light, was attached (by rubbing) to the exposed surface

of the mixture. As before, there was a comparatively strong

current from the silver to the copper. The silvered surface
was then exposed to the light of burning magnesium wire, and

332 Mr. 8S. Bidwell on the Generation of

the immediate movement of the spot of light through 50 scale-
divisions indicated a diminution of the current. When the
magnesium was extinguished, the current at once increased
to its original strength. When a nearly red-hot brass rod
was held at a distance of 3 centims. from the silvered surface,
the current slowly increased in strength; and when the hot
rod was removed, the current was again slowly diminished.
The effects both of light and of heat were verified by many
repetitions of the experiments. These results which, so far as
regards the effect of light, were unexpected, are of the same
character as those described in (1). I have elsewhere* given
strong reasons for believing that the combination of sulphur
with silver is assisted by the influence of light. If this is so,
it is certainly a remarkable fact that increased corrosion of the
silver electrode should be accompanied by diminution of the
current.

Two days afterwards the silver-leaf had become much dis-
coloured, and was in some parts quite black. When con-
nected with the galvanometer the cell gave a current of nearly
the same strength as before; but now it was found to be
slightly increased by light as well as by heat; and it is pro-
bable that the light as such exerted no influence whatever,
the observed effect being really due to the incidental rise of
temperature.

(12) The silver-leaf was scraped off, and the surface of the
sulphur mixture having been cleaned from all visible traces
of free silver with fine emery-cloth, a piece of thin gold-
leaf was pressed upon it. It was found very difficult to make
it adhere satisfactorily. The cell, when connected with the
galvanometer, gave no indication whatever of a current.
When the cell described in (2) was also inserted in the circuit,
the spot of light was deflected, showing that the first cell was
quite able to conduct electricity, and that its failure to origi-
nate a current was not owing to bad contact between the gold-
leaf and the sulphur mixture. 7

(13) A cell containing a mixture of sulphur and silver
sulphide between plates of silver and iron was found to have
an H.M.F’. of 023 volt, or about one third of that of the silver-
copper cell described in (2). The direction of the current
was the same.

(14) A cell was made by melting sublimed sulphur upon
a plate of copper and pressing a plate of iron upon the melted
sulphur. On connecting it (when cold) with the galvano-
meter, there was no indication of any current. Nevertheless
the cell was found to conduct electricity very well when a

* Phil. Mag. August 1885.

Electric Currents by Sulphur Cells. 333

battery was also placed in the circuit. Sulphide of copper
was of course formed by the action of the hot sulphur upon
the copper.

(15) A melted mixture of 5 parts of sulphur with 1 part ef
copper sulphide was pressed between plates of silver and iron,

and cooled. When first made, this cell generated a sufficiently

strong current to deflect the galvanometer-needle as far as
the stop would allow; but two hours later, when an attempt
was made to measure the E.M.F., it was found to have almost
completely disappeared, being less than a hundredth of that
of the cell described in (2). The H.M.F. was temporarily
restored by connecting the iron and silver plates with the
positive and negative terminals of a battery of ten Leclanché
cells for a few minutes. The result was of course merely a
polarization-current. |

(16) Two silver wires A, B were imbedded ina fused mass
consisting of equal parts of sulphur and copper sulphide.
When cold, the wire A was connected with the carbon pole
of a battery of ten Leclancheé cells, and the wire B with the
zine pole. After the current had passed for about a second,
the cell was detached from the battery and connected with
the galvanometer. A current was at once indicated in the
direction A B, 7. e.in the same direction as that of the battery-
current which had been caused to pass through the cell. This
experiment was repeated many times and on different days,
with the same result™. A period of some seconds necessarily
elapsed between the separation of the cell from the battery
and its attachment to the galvanometer. In order to render
this interval as short as possible, the apparatus was so arranged
that by depressing a key the transfer could be effected ina
small fraction of a second. It was then found that the first
effect of the transfer was a strong momentary current in the
direction BA, which was immediately followed by the more
permanent current previously observed in the direction A B.

' With the view of retarding the first effect, a battery of

two Leclanché cells was used, and the connection was made
for a longer period. After the battery had been connected
for one minute, the cell was transferred to the galvanometer
by means of the key, and the swing of the spot of light through
280 scale-divisions again indicated a current in the direc-
tion BA. Though this was not, as before, a current of only
momentary duration, it rapidly decreased in strength, becoming
zero in almost exactly 30 seconds. After zero was passed, a
current was at once set up in the opposite direction A B, which

* The experiments described in the remainder of this paragraph were
made after the paper was read.

334 Mr. 8. Bidwell on the Generation of

in 380 seconds produced a deflection of 70 divisions, increasing
to a maximum of 131 divisions in 10 minutes. The current
then slowly diminished ; and in 52 hours after the commence-
ment of the experiment the deflection had fallen to 17 scale-
divisions. At this point the observations were discontinued.
In the accompanying curve the abscissze represent the time
in hours, and the ordinates the current in scale-divisions. The
current of 30 seconds’ duration in the direction BA is not
represented.

@ Current.

Hours.

On another occasion, when the two Leclanche cells had
been connected for 3 minutes, the secondary current from B
to A decreased still more slowly than in the last-mentioned
experiment, vanishing in 34 minutes. The current which
followed in the direction A B attained its maximum, indicated
by 155 divisions, in 63 minutes, and then steadily decreased
for nearly 4 hours, when it again became zero. ‘The spot of
light did not, however, remain stationary, but moved steadily
on, indicating a second reversal of the current. In one hour
after zero had thus been crossed for the second time, the gal-
vanometer-deflection was 12 scale-divisions; and 44 hours
later, when the last observation was made, it had increased to
20 divisions. In both these experiments the galvanometer
(of which the resistance was nearly 3500 ohms) was shunted
with a coil of 100 ohms.

These alternating currents are probably of the same nature
as those which Faraday found to be generated when copper and
silver, or two pieces of copper, or two pieces of silver, were
immersed in a solution of potassium sulphide*. The wires
are alternately protected from the action of the free sulphur
by an investing coat of sulphide.

* Exp. Res. §§ 1911 and 2036.

Electric Currents by Sulphur Cells. 330

Summary.

Plates of silver and copper imbedded in a mixture of sulphur
with sulphides of copper or silver constitute a cell which at
the ordinary temperature is capable of generating and main-
taining a constant current, the silver being the positive
element. Such a cell, in which the mixture consisted of 5
parts of sulphur with one of copper sulphide between plates
8 centim. square and ‘3 millim. apart, had an H.M.F. of -0712
volt and an internal resistance of 6537 ohms.

If the proportion of copper sulphide to sulphur is increased,
the internal resistance of the cell is diminished; but its H.M.F.
is also diminished.

A cell containing copper sulphide unmixed with free sul-
phur fails to produce any appreciable current.

A cell containing silver sulphide only generates a current
in the opposite direction to that produced when the sulphide
is mixed with free sulphur.

Copper used in conjunction with iron or gold gives no
current whatever at the ordinary temperature.

The current generated by a silver-copper cell containing
free sulphur mixed with sulphide is diminished by the action
of light and increased by heat. It has not been ascertained
whether the effect is upon the E.M.F. or the internal resist-
ance, or both.

If a battery-current is caused to pass for a short time through
a cell consisting of two silver electrodes imbedded in a mixture
of sulphur and copper sulphide, the cell, after being discon-
nected from the battery, will generate a current of very short
duration in the direction opposite to that of the battery-current,
followed by a current which may be maintained for several
hours in the same direction as the battery-current. A second
reversal has in one case been found to occur after an interval
of four hours.

The experiments, of which an account is here given, must
be considered as being merely of a preliminary nature. A
complete investigation of the phenomena in question would
require more time than I am at present able to devote to the
subject.

Addition, August 3rd.

From (4), (5), and (6) it appears that a cell consisting of
copper and silver plates is incapable of generating a current
unless a certain proportion of free sulphur is mixed with the
sulphide. Thinking that the function of the free sulphur

336 Dr. J. Hopkinson on the Seat of the

might be merely to form silver sulphide by direct combination
with the silver, I constructed a cell as follows :—A layer of
copper sulphide was spread upon a plate of copper ; a polished
steel plate was laid upon the sulphide, and the whole was
strongly compressed in a vice. ‘The steel plate was then
removed, and a thin layer of silver sulphide was spread upon
the smooth surface of the copper sulphide. The cell was
completed by pressing a silver plate upon the silver sulphide.
This was found upon trial to give a current which, with an
external circuit of low resistance, was many times stronger
than that generated by any of the cells previously made. It
seems to be exactly analogous in its action to a Daniell cell
consisting of plates of zinc and copper in solutions of zine
sulphate and copper sulphate. The quantity of the copper
sulphide would be gradually diminished, copper being depo-
sited on the copper plate, while the quantity of silver sulphide
would continually increase with consumption of the silver.

In conclusion, it seems probable that, by selecting such
metals as experiment might prove to be better suited for the
purpose than silver and copper, a cell might be constructed
upon the principle of that described in the above paragraph
which would be of practical and commercial value.

XXXVIITI. Notes on the Seat of the Electromotive Forces in a
Voltaic Cell. By J. Horxinson, J.A., D.Se., £AS.*

HE following is an expansion of some short remarks I
made when Dr. Lodge’s paper was read at the Society
of Telegraph Hngineers.

I. Lhe controversy between those who hold that the difference
of potential between zine and copper in contact 1s what ts deduced
by electrostatic methods, and those who hold that it is measured
by the Peltier effect, is one of the relative simplicity of certain
hypotheses and definitions used to represent admitted facts.

Taking thermoelectric phenomena alone, we are not impera-
tively driven to the conclusion that the difference of potential
between zinc and copper is the small quantity: which the
Peltier effect would indicate ; but by assuming with Sir W.
Thomson that there is an electric property which may be
expressed as an electric convection of heat, or that electricity
has specific heat, we may make the potential difference as
great as we please without contradiction of any dynamical
principle or known physical fact. Let us start with the
physical facts, and introduce hypothesis as it is wanted.

* Communicated by the Author.

ee

Electromotive Forces in a Voltaic Cell. 337

These are, as far as we want them :—(1) If a circuit consist of
one metal only, the electromotive force around the circuit is
nil however the temperature may vary in different parts ; this
of course neglecting the thermoelectric effects of stress and
magnetism discovered by Sir W. Thomson. (2) If the
circuit consist of two metals with the junctions at different
temperatures t4, ¢,, then the electromotive force round the
circuit is the difference of a function of t, and of the same
function of t. According to Prof. Tait the function is

i path a 338 t.”
(o—4)( rig ), or, as we may write it, A+ Bi,— at

2
| A.+ Ba — ce i ; the series may perhaps extend further, but,

according to Tait’s experiments, the first three terms are all
that are needed.
Now, but for the second law of thermodynamics we should
2

naturally assume that pees was the difference of

2
potentials at the junction of temperature ¢, and A+BA+ St
at the junctions of temperature ¢, ; we should further assume
that what the unit of electricity did was to take energy
aed
junction with disappearance of that amount of heat, and to

out of the region immediately around the hot

2
take energy A+By—-— into the region immediately

surrounding the cold junction, with liberation of that amount
of heat. Now apply the second law of thermodynamics in

the form 32=0, and we have

1 1 bo Og
AC: i) Z ”
whence it follows that A=0, which may be, and that C=O,
which is contrary to experiment. The current then must do
something else than has been supposed, and the hypotheses
differ in expression at least as to what that something else is,
The fact to be expressed is simply this: when a current
passes in an unequally heated metal, there is a reversible
transference of heat from one part of the metal to another,
whereby heat is withdrawn from or given to an element of
the substance when a current passes through it between
points differing in temperature, and is given to or withdrawn
from that element if the current be reversed. Sir W.

Phil. Mag. 8. 5. Vol. 20. No. 125. Oct. 1885. 2B

338 Dr. J. Hopkinson on the Seat of the

Thomson proved that this follows from the fact of thermo- —
electric inversions and the second law of thermodynamics,
and verified the inference by experiment, his reasoning being
quite independent of any hypothesis.

Suppose wires of metals X and Y are joined at their extre-
mities, and the junctions are kept at temperatures t,t. The
observed electromotive force around the circuit is /(t,) —f (4)
or within limits according to Tait, B(@é,—t,)—3C(é.’—t’).
The work done or dissipated by the current when unit of
electricity has passed is f(t,)—f (¢;), and this is obtained by
abstraction of heat from certain parts of the circuit and
liberation of heat at other parts by a perfectly reversible
process. Let F(é) be the amount of heat which disappears
from the region surrounding the junction ¢, when unit of
electricity has passed from X to Y. let an element of the
wire X have its ends at temperatures ¢ and ¢+dt, and let the
quantity of heat abstracted from this element when unit of
electricity passes from ¢ to t+dt be represented by @(é)dé,
and let the same for Y be represented by y(t)dt. By the
first law of thermodynamics we have

F(t) +("4Oa—FG) +4 “eae=e) “Sa,

and by the second law, since the transference of heat from
part to part is reversible,

f, te
()/t—F(uyitr+ f b(t. dt—| “Y(t. dt=0.

Differentiating we have |

MO+EO-¥O =F,

P'()—-FO/2+b(0/t—p(O/t=0 ;
{ Fé) =¢/ (t) = Bt—C?,

P(t) —P) =H" (t) = Ct. |

This really contains the whole of thermoelectric theory
without any reference to local differences of .potential, but
only to electromotive force round a complete circuit. But
when we come to the question of difference of potential
within the substance at different parts of the circuit, we find
that according as we treat it in one or the other of the
following ways we may leave the difference of potential at the
junctions indeterminate and free to be settled in accordance
with hypotheses which may be found convenient in electro-
statics, or we find it determined for us, and must make our
electrostatic hypotheses accord therewith.

whence

Electromotive Forces in a Voltaic Cell. 339

The first way is that of Thomson, as I understand it.
Assume that there is no thermoelectric difference of potential
between parts of the same metal at different temperatures,
at all events till electrostatic experiments shall show that there
is. It follows that we must assume that the passage of
electricity between two points at different temperatures must
cause a conveyance of energy to or from the region between
those points by some other means than by passage from one
potential to another. Such conveyance of energy may be
very properly likened to the convection of heat by fluid in a
tube, for although convection is in general dissipative, it is
not necessarily so, e.g. a theoretically perfect regenerator.
Suppose, then, that in metal X unit of electricity carries with
it ((¢)d¢ of heat, and in metal Y, xp(¢) dé, this will account for

the proved transference of heat in the two metals. When a
unit of electricity passes across a junction at temperature ¢
from X to Y, it must liberate at that junction a quantity of

heat \p (t) dt —|\y(é)de ; but the actual effect at this junction

is that heat F'(¢) disappears ; hence the excess of potential at
the junction of Y over X must be

F()+\o(@dt—\y@dt or A+Be—-4C?,
A being a constant introduced in integration. If, then, we
assume a “‘specific heat of electricity,’ the actual difference
of potential at a junction may contain a constant term of
any value that electrostatic experiments indicate.

But the facts may be expressed without assuming that
electricity conveys energy in any other way than by passing
from a point of one potential to a point of different potential.
This method must be adopted by those who maintain that the
Peltier effect measures the difference of potential between two
metals in contact. Define that if unit-electricity in passing
from A to B points in a conductor homogeneous or hetero-
geneous does work, whether in heating the conductor, chemical
changes, or otherwise, the excess of potential of A over B
shall be measured by the work done by the electricity. This
is no more than defining what we mean by the potential
within a conductor, a thing we do not need to do in electro-
statics. ‘This definition accepted, all the rest follows. Be-
tween two points differing in temperature dé the rise of
potential is f(¢)dé in X, W(é)dt in Y; at the junction the
excess of potential of Y over X is F'(¢)= Bé—Ce’.

The second method of arranging one’s ideas on this subject
has the advantage that it dispenses with assuming a new pro-
perty of that hypothetical something, electricity ; but there
is nothing confusing in the first method.

7 2B2

340 Dr. J. Hopkinson on the Seat of the

II. The thermodynamics of the voltaic circuit may be dealt
with on either method of treatment ; in the equations already
used, instead of speaking only of the heat disappearing from
any region, we have to consider the heat disappearing when
the unit electricity passes plus the energy liberated by the
chemical changes which occur. Consider a thermoelectric
combination in which there is chemical action at the junctions
when a current passes.

If Gt be the function of the temperature which represents
the energy of the chemical reaction which occurs when unit
of electricity passes from X to Y across the junction, we have

H(t) + G (to) — F(t) — G4) +4 aut a $(dt=/(4) fl),

P'() +) +4) -¥ =F, i
F@)t—F@ P+ oO k—bO/t=0 5

whence

FO =7"' (4) —16'(0), \
POs) =H{ fe )-— GN}. 7

If now we proceed on the hypothesis of specific heat of
electricity, we are able to make the differences of potentials
at the junctions accord with the indications of electrostatic
experiments. We are, then, by no means bound in a voltaic
cell to suppose that there is a great difference of potential
between the electrolyte and the metal because there is a
reaction there, for we may suppose the energy then libe-
rated is taken up by the change that occurs in the specific heat
of electricity.

III. Adopting the second method of expressing the facts,
we may consider further the location of the difference of
potential in a voltaic cell. In the case of a Daniell’s cell
consisting of Cu | CuSO, | ZnSO, | Zn, at which junction is
the great difference of potential? Dr. Lodge places it at the
junctions of the metals and the electrolytes. For this there
is really some experimental reason, but without such reason
it is not apparent why there may not be a great difference of
potential between CuSO, and ZnSO,. In that case, in an
electrolytic cell with zinc or copper electrodes and ZnSO, or
CuSQ, as electrolyte there would exist a small difference of
potential between the metal and the electrolyte. Take the latter
case, an electrolytic cell of CuSQ,, and let us leave out of
account the irreversible phenomena of electrical resistance and
diffusion. Tirst, let us assume, as is not the fact, that the only
change in the state of the electrolytic cell when a current has

Electromotive Forces in a Voltaic Cell. 341

passed is addition of copper to one plate, loss of copper
from the other plate ; what could be inferred? Imagine a
region enclosing the anode, when a current has passed, what
changes have occurred within the region? An equivalent of
copper has disappeared from the anode, and that same quan-
tity of copper has departed and gone outside the region.
But by our supposition, nothing else has happened barring
increase of volume for liquid by diminished volume of metallic
copper ; there is no more and no less CuSQ, in the region,
the same quantity therefore of SQ, All the work done in
the region is to tear off a little copper from the surface of the
anode and to remove it elsewhere. If the fact were as as-
sumed it would follow that the passage of the current did
little work in the passage from copper to sulphate of copper,
and consequently that the difference of potential between the
twois small. But the fact is, other things happen in the cell
than increase of the kathode and diminution of the anode.
In contact with the anode there is an increase of CuSQ,, in
contact with the kathode CuSO, disappears: this is a familiar
observation to every one. Reconsider the region round the
anode. Assume as another extreme hypothesis that after
a current has passed we have in this region the same
quantity as before of copper, but more CuSO,; SO, has
entered the region and has combined with the copper. A

large amount of energy is therefore brought into the region,

which can only be accounted for by supposing that the elec-
tricity has passed from a lower potential in the copper to a
higher potential in the electrolyte. The legitimate conclusion
is, then, that there is between Cu and CuSQ, a difference of
potential corresponding to the energy of combination; and the
basis of the conclusion is the simple observation that the
copper is dissolved off one plate but remains in its neighbour-
hood, whilst it is precipitated on the other plate, impoverishing
the solution. In other words, it is the SO, that travels, not
the Cu. 3

Now consider the ordinary Daniell’s cell. Is there a sub-
stantial difference of potential at the junction of CuSO, and
ZnSO,? Is there, in fact, a difference apart from the
Peltier difference? Imagine a region enclosing the junction
in question ; it might have been that the effect of a current
passing was to increase the zinc and diminish the copper by an
equivalent of the electricity which passed, from which we should

- have inferred that the seat of the electromotive force in a

Daniell’s cell was at the junction of the two solutions. But
it is more nearly the fact that no change whatever occurs in
the region in question when a current passes, and that all that

342 Seat of the Electromotive Forces in a Voltaic Cell.

happens is that a certain quantity of SO, enters the region
and an equal quantity departs from it, from which follows
that there is no potential difference, other than a Peltier
difference, at this junction. |

Neither of the extreme suppositions we have made as to
concentration or impoverishment of the solution is in fact
true, but they serve to show that the position of the steps
in potential depends entirely on the travelling of the ions.
The fact is, that in general both ions travel in proportions
dependent on the condition of the electrolytes ; it is probable
that the travelling of the SO, depends on some acidity of the
solution. Given the proportion in which the ions travel and
the energy of the reversible chemical reaction which occurs,
and we can calculate the differences of potential at the
junctions.

In the preceding reasoning an assumption has been made,
but not stated. It has been assumed that the passage of a
current in an electrolyte is accompanied by a movement of
ions only, and not by a movement of molecules of the salt;
that is, when unit of electricity passes through a solution of
CuSO,, «Cu travels in one direction and (1—«)SQO, in the
opposite direction, but that CuSO, does not travel without
exchanges of Cu and of SOQ, between the molecules of CuSQ,.
In the supposed case when there is no concentration around
the anode, my assumption is that Cu is dissolved off the anode,
and that an equal quantity of Cu leaves the region around the
anode as Cu by exchanges between the molecules of CuSQ,.
But it is competent to some one else to assume that in this
case SQ, as SO, enters the region by exchanges between the
molecules of CuSQ,, and that at the same time a molecule of
CuSO, leaves the region without undergoing any change.
Such a one would truly say that there was no inconsistency
in his assumption; and that if it be admitted, it follows that the
difference of potential at the junction CuSO, | Cu is that repre-
sented by the energy of the reaction. I prefer the assumption
I have made, because it adds nothing to the ordinary chemical
theory of electrolysis; but it is easy to imagine that facts may be
discovered more easily expressed by supposing that an electric
current causes a migration of molecules of the salt, as well as
a migration of the components of the salt.

Feo 343.

XXXIX. A Comparison of the Standard Resistance-coils of the
British Association with Mercury Standards constructed by
M. J. R. Benoit of Paris and Herr Strecker of Wirzburg.
By R. T. GuazeBrook, IA., F.RS., Fellow and Assistant
Tutor of Trinity College, Cambridge*.

T° accordance with the resolution of the Electrical Congress

held in Paris in 1884, the Hlectrical-Standards Committee
of the British Association decided to have made a series of
resistance-coils in terms of the Legal Ohm; and the work of
testing and comparing these coils was entrusted to me. In

Paris, at the request of the Minister for the Post Office and

Telegraphs, M. J. R. Benoit undertook the same task. An

account of his experiments is published in the Journal de

Physique for January 1885. ?

M. Benoit started ab initio, and constructed a series of

glass tubes, the electrical resistance of which, when filled with

mercury, can be calculated from their dimensions. Four of
these were made, each having a resistance of about 1 ohm.

The plan adopted by the Committee was different. The
specific resistance of mercury in terms of the British-Associa-
tion unit has been carefully determined by Lord Rayleigh
and others.

For the purpose of constructing Legal Ohms, it was decided
to adopt a number for this quantity founded on their experi-
ments, and then work from the British-Association standards.
To adopt the other course would only have been to repeat at
some trouble the experiments of Lord Rayleigh, M. Mascart;
Herr Strecker, and others.

M. Benoit made a number of copies of his mercury-standards,
and after reading his paper, I wrote and asked him to send
me one or more of these, that | might compare them directly
with the coils which had by that time been constructed for
me by Messrs. Hlliott Bros. M. Benoit replied most cour-
teously, and sent me, about Haster, three of his copies. The
object of the present paper is to give an account of the com-
parison of these with the standards belonging to the British
Association.

M. Benoit’s copies were mercury-standards. Hach con-
sisted of a glass tube bent several times, AA, fig. 1. The
ends of this tube are ground flat and pass into two glass cups,
B, B, which for most of their length are considerably wider
than the tube, but taper down at their lower ends. The tubes
are connected to the cups by short bits of india-rubber tubing,

* Communicated by the Physical Society: read May 23, 1885,

344 Mr. R. T. Glazebrook’s Comparison of

CG, C, which fit tightly over the narrow ends of the cups, and
are secured by wrapping string firmly over the india-rubber.

R

ass Qe

MLM:

Q
oY”

\
x
SY

CEST SSS SSS SS EEE EEE : AS
~S
S

Fig. 1.
Gp 4 &
Z y
Q 4’
YU G
Y
G
b- a Y B 4-6
y
g
Z
y -
Z

SSOSSH Ag gg

SS]
S
eee SMS
4 ry

SSS wss5»uv

~~
VAS ILS
SSSA

aD NN

SSO "8 = "= Fé FZ” 7

QE
>
WIE

SSSVg

MOO

-
TQ Qu 0 0 lO BlFtEFt tl

To make the joint tight the india-rubber and glass were var-
nished over with shellac varnish. The cups open at the top,
and can be closed by glass stoppers.

The standards were accompanied by full directions for use.
I give a free translation of the greater part of the instructions.

1. On each of the cups are engraved two marks aa, bb.
The lower mark aa indicates the position which the extremity
of the tube introduced into the cup ought to occupy. The
upper mark 0 6 indicates the top of the mercury in the tube
when filled. These two conditions must be fulfilled approxi-
mately ; an error of ‘5 centim. in their sum changes the
resistance either way by about ‘00018 ohm.

2. Method of Filling.—The tube may be filled in air in the
following manner:—Place it nearly horizontally and pour
some mercury into one of the cups; then incline it gradually,
shaking it a little, until the mercury enters the tube. The
mercury should fill the tube slowly, without leaving a trace of
air. Add mercury always by the same cup until it reaches
the upper mark 66 in both cups. The method is not always
successful; it is better to work in a vacuum.

[ Various methods for doing this are possible ; it is perhaps
hardly necessary to give the details of the one employed by
M. Benoit. | |

Standard Resistance-coils with Mercury Standards. 345
3. Method of Using the Standard.—-The standard is intro-

duced into the circuit by means of contact-pieces dipping into
the mercury in the cups. These pieces should not reach more
than a few millimetres below the surface. Contact-pieces of
amalgamated copper cannot be used, because they render the
mercury impure and alter its resistance rapidly, diminishing
it appreciably in a few hours. Platinum, again, does not
make a constant contact with the mercury; and the uncer-
tainty produced by its variability is fatal to its use. To avoid
this double difficulty, M. Benoit employed contact-pieces of a
special form (fig. 2). <A glass tube A is Fig. 2.
drawn out to a point at its lower end;
through this point a platinum wire B
passes. The wire is held in position by
shellac C; and a small cup D of thin
glass is attached to its lower end also by
shellac. Thus the two separate glass
portions A and D are in electrical com-
munication by means of the platinum
wire B. The two portions A and D
are filled with mercury in a vacuum.
The mercury is thus brought into good
contact with the platinum, and the
mercury-platinum contact remains un-
changed during the observations. Into
the upper division A one end of a stout
copper rod E is plunged; the other end
of this rod forms the connection with
the rest of the circuit. The cup D dips
a few millimetres below the mercury in
the standard. The mercury which fills
it remains constantly in contact with the
platinum wire, and may be kept without
change almost indefinitely. These contact-pieces introduce
an appreciable resistance of 2 to 3 thousandths of an ohm into
the circuit ; but this resistance remains constant, and can be
eliminated by a method of substitution.

The standard is supported in a glass vessel, and can be
reduced in temperature to zero by filling the vessel with finely-
broken ice. The values of the standards are given at zero.
The value at any other temperature is given by the formula

R,= R,(1 + 0008649¢ + -0000011227),
t being the temperature in degrees Centigrade.
4. Preparation of the Mercury—The mercury may be

purified by the action of nitric acid, and dried by means of
sulphuric acid and caustic potash.

346 Mr. R. T. Glazebrook’s Comparison of

Hixperiment showed that mercury from different sources,
even when it had been rendered impure by the admixture of
copper, lead, or zinc, after being treated by this process, gave
the same results. The process, however, does not free the
mercury from silver or from the less oxidizable metals.

[In my experiments described below the mercury was
freshly distilled by the aid of the admirable piece of apparatus
designed by Weinhold, and introduced into England by Mr.
W. N. Shaw. |]

5. Method of Cleaning the Tubes—The tubes should be
cleaned by passing through them in succession :—

1. Distilled water.

2. Strong nitric acid.

8. Distilled water.

4, Ammonia.

5, Distilled water.

They must then be dried by a current of dry air.

To clean the tubes itis best to separate them from the cups.
To effect this, the pieces of india-rubber tubing which connect
them should be cut. ‘To make the glass slip into the india-
rubber tubing, it should be moistened with a drop of benzine.

In preparing the tubes for my measurements, I endeavoured
to carry out the above instructions as carefully as possible.
The resistance comparisons were made by Carey Foster’s
method, using the wire bridge of the British Association as
designed by Dr. Fleming.

In order to compare the resistance of one of the tubes with
a standard coil, it was necessary either to know accurately or
to eliminate the resistance of the contact-pieces. It was also
important to determine within what limits the resistance of
the contact-pieces remained constant, and whether they could
be repeatedly filled with mercury in such a way as to retain
the same resistance.

For this purpose they were compared with a short piece
of copper rod. The cups D (fig. 2), one belonging to each
contact-piece, were immersed in the same beaker of mercury,
being placed close together and covered to the same depth
as when actually in use; the other ends of the copper rods
Hj} connected in the usual way to the bridge. The contact-
tubes had been filled by being placed in a large test-tube
which was exhausted, and into which mercury was then
admitted by means of a tap attached to the test-tube. The
resistance of the contact-pieces was thus found in terms of
the divisions of the bridge-wire before and after a set of
comparisons of the mercury-tubes and standard coils.

Standard Resistance-coils with Mercury Standards. 347

It was found that the resistance of the contact pieces re-
mained very nearly constant during the time occupied in
making a series of comparisons. The variations in the values
found before and after rarely amounted to more than 1 bridge-
wire division, or about ‘00005 ohm; and the mean of the
two values obtained is pretty certainly correct to less than
this.

At the same time different fillings of the same contact-tubes
led to very different values for their resistances. This was
no doubt due to small irregularities in the contact between
the platinum and the mercury. Thus I found the following
values in terms of a bridge-wire division :— |

Date &e. Resistance of contacts.
DE LS 2 eg ene ee a 93°4 b.w.d.
ss Reaces, Tele so2 5c hes oe oe incense 69:4 ,,
Copper rods being left in the
man 2° mercury overnight ............... } oe
bs ,, Contacts refilled, one renewed ......... jig 3 ee
», 28 Copper rods being left in since 26th... 109-4 ,,

The differences between the 69°4 and the 68:2, and, again,
between the 114°5 and 109°4, are due, I believe, to the copper
of the contact-rods being dissolved by the mercury in the tubes.

On March 28th, the contact-tubes were cleaned and refilled,
and the resistance was found to be 57:2 b.w.d. Throughout
the experiments until April 9, when one of them was broken,
their values did not differ greatly from the above.

On April 11th, new contact-tubes were prepared and used
throughout the rest of the experiments; their value remained
fairly constant at about 58 b.w.d.

Thus the quantity actually determined by the observations
was the difference between the resistance of the mercury tube
+ the contact pieces and the standard coil. In order to
get the resistance of the mercury tube in terms of the standard
coil, the resistance contact-pieces found as above described
had to be subtracted.

Another way of eliminating the resistance of the contacts
from the result is to compare first the mercury tube and then
the standard with some other coil, using the same contact-
pieces to connect the mercury tube and the standard to the
bridge. This can be done by removing the contact-tubes
from the mercury tube, and dipping them into two mercury
cups, into which the electrodes of the standard also dip. Since,

348 Mr. R. T. Glazebrook’s Comparison of

however, my apparatus allowed me to measure the resistance
of the contact-tubes with all the accuracy I desired, the first
method was the more convenient, as it necessitated fewer obser-
vations. |

The comparisons were made both with the original standards
of the British Association—B.A. Units—and also with the
Legal-Ohm Standards of the same, recently constructed for
me by Messrs. Hlliott Bros., taking as a basis the value
for the resistance of mercury in terms of the B.A. unit
adopted by the British-Association Committee. According
to this,

1 Legal Ohm=1:0112 B.A.U.
The three mercury tubes Nos. 37, 38, 39, sent by M. Benoit,

were compared. Between two observations of resistance the
mercury was occasionally drawn through the tubes by con-
necting the cups alternately to the air-pump. I hoped in this
manner to displace any very small air-bubble or particle of dust
which might have got lodged in the tube. In the final set of
observations recorded below, it will be seen that no appreciable
change in the resistance was detected by this precaution. I
therefore suppose the tubes to have been properly filled.

In some earlier observations, which will be referred to again
shortly, variations of considerable amount were produced by
passing the mercury through. In the tables below the values
of the resistances which were found by comparison with the
B.A.U., and reduced to Legal Ohms by means of the known
relation between the two, are denoted by an asterisk.

TABLE I.

Tube No. 37. Value given by Benoit, 100045.

Date and Method of Treatment. Value in L. Ohms.

ET EL ee eS ae a -99991%

* 10.30 ee ee ee ‘99987

“ 11.55 pitas stasees gs ‘(99998

ter passing the :

ig eee ere through. } 99997

I ARs aie sabi ays the dine ellie RR EONS ‘99998
‘ After passing the

April lt sss, fae through. } 99996
EE EI) Be) Sess hissoapte nen cies aw aedee de ‘99979
RE AMEN Angelic oatunds tags Ri ins outs ste ‘99976

———- ——$—————— I | [eT RRs

Standard Resistance-coils with Mercury Standards. 349 ©
= | TABLE II.

Tube No. 38. Value given by Benoit, 1-00066.

Date and Method of Treatment. | Value in Legal Ohms.

|
UME ACM ewe c cles asecctyensdnyeucass 1-00006
SPS LOO AM. ooccccalevlscctoes i Re 100018
MR Ok eek att | 1:00023%
RM EY rece orca e see SN howe minis S03 1:00019
: {| Tube taken to pieces, re-
Bpet ty 1 cleaned and refilled. \
pen eI Pl oie acne 8 aa ins casey eos kao 10001 1x
LL LE eee SEE ee eee 100010
gi ee ee emer me a 100015
MEME NOE AEE ree Soca chs 55 d's 3 ees Gane eeames 100002
2200. LE A SSS 6 Sa eee ey een re 1-00003*
LSS yh 2 si ne eae ee ee es 1:00012
TaBLeE ITI.

Tube No. 39. Value given by Benoit, -99954.

Date and Method of Treatment. | Value in Legal Ohms.
Pepin be cn ace ee ee ae "99905
seapeetls Fh MRSA Aste se AK 0 int wae ated ‘99917
= After passing the :
| Apt 15, 11 a { ales et 99921
erapesh ts) 0055 WM... 2. GLI weed. & ‘29921
After passing the Hae!
oo AA AIA Reale ane ea 99921
TLD BAL es Oe Ss ee oC eee eee ‘99917
Micra Wale iiss tin 1 0 oy sao vs ‘99917
Thus we may collect the results as below.
TABLE LV.
No. of Tube. te es by ya ae by Difference, |
rs alesis 1:00045 ‘99990 ‘00055
5 eee 100066 100011 00055
2 Eppes 99954 ‘99917 ‘00037

Mean...... 100022 99972 00050

— 850 - Mr. R. T. Glazebrook’s Comparison of

Thus there is a difference of ‘0005 legal ohm between the
two. My results are based upon the value of the resistance
of mercury in terms of the B.A. unit adopted by the British-
Association Committee. If we denote by M the resistance at
0° C. of a column of mercury 1 metre long, 1 square millim. in
section, then, according to the number adopted by the Com-

MAIbICO, M. = S9oe Ba.
B.A.U.=1:04820 M.
The most recent values actually found for these relations
are given below:—

TABLE V.

M in B.A.U.| B.A.U. in M. Observer.
| Rayleigh * and
95412 1-04808 | Side vick
Mascartt,Nerville,
95374 1-04851 { Sane
‘95334 104894 Strecker f.
95411 104809 L. Lorentz §.

M. Benoit has not compared his tubes directly with the —
B. A. units. The mean of my comparisons gives for his

mercury tubes, M. = 95348 BAU,
B.A.U.=1:04878 M.

Thus there is a difference of -0005 ohm between standards
issued by M. Benoit as representing the legal ohm, and those
issued by myself on behalf of the B. A. Committee, M.
Benoit’s standards being less than mine by that amount.
Some other points remain to be noticed. The first time
that I filled Nos. 87 and 39, 1 obtained results for the resist-
ances differing appreciably from those given above. They
were filled each time with equal care, the mercury had been
treated in the same manner, and there seemed no reason,
taking each observation separately, why one should be better
than the other. For No. 87 I found on April 8th the value
100044 legal ohms ; I noticed on April 9 that this had gone
up to 1:00071 ; and on passing the mercury through as already
described, it rose further to 1:00080. These differences are
considerably greater than any possible errors of observation ;
* “On the Specific Resistance of Mercury,” Phil. Trans, pt. 1, 1883.
+ Journal de Physique, June 1884.

t rd gill eco Bd. xv.; also Wied. Ann. vol. xxv.
§ Wied. Ann. vol. xxv.

Standard Resistance-coils with Mercury Standards. 351

and it was clear that there was some small change progressing
in the tube. It was carefully examined, but no trace of an
air-bubble could be seen; and then it was emptied, cleaned
again, and refilled, with the results given in Table I. Much
the same was observed with coil No. 39. No. 88 was cleaned
and filledtwice, but gave perfectly consistent results all through.
This uncertainty seems to me to constitute one serious objec-
tion to the general employment of mercury tubes as standards.

I believe that for the observations recorded in the above
Tables the tubes were properly filled, and that there were no
bubbles of air in them. This, I think, is shown by the agree-
ment between the results of different fillings, and I suppose
that the resistance will not seriously alter so long as the same
mercury remains in the tubes; but it appears that, after fill-
ing, electrical experiments are required to make certain that
every bubble of air has been removed, and that the tube
really has its true resistance. I should imagine, too, that the
tubes would require somewhat frequent refilling to make sure
that the mercury may remain pure. My experiments showed
that it was almost impossible to keep the inside of the cups
above the mercury perfectly dry. It was necessary, when
making the comparisons, to remove the glass stopper and
insert the contact-pieces; and this had to be done when the
tube was immersed in the ice-water.

The cups were open to the air for a short time in making
this change, and that short time was quite sufficient to cause
a deposit of dew to be formed on the inside of the cup. So
long as this slight moisture remained in the cup and did not
reach the tube itself, of course it did not affect the results;
but it would be difficult to feel certain that after a time the
mercury in the tube was quitedry.

This difficulty would be avoided by working at the tempe-
rature of the room rather than at that of melting ice. The
large temperature-coefficient of mercury, from three to four
aa that of platinum-silver alloy, is, however, an objection
to this.

Another difficulty, caused by the necessity of working
at zero, was that the mercury in the cups was always
slightly warmer, 0°°2 or 0°38 C., than the ice. This was, no
doubt, caused by the conduction of heat down the copper con-
necting-rods, and from the upper portions of the glass of the
_cups which were exposed to the air. This, of course, would
necessitate a small correction to the values of the resistances
given in the Tables; but the correction must be exceedingly
small, for the temperature of the mercury in the tube itself,
where it is actually in contact with the ice, must be zero,

352 Mr. R. T. Glazebrook’s Comparison of

and it is only the small portions of the tubes which lie within
_the cups which will be at the higher temperature. 3

The necessity of using the somewhat complicated con-
necting pieces is also a great drawback to the practical use-
fulness of the mercury standards. I hope shortly to carry
out some experiments on the permanence of the contact be-
tween mercury and amalgamated platinum. Contact-pieces of
platinum with their end amalgamated with mercury would
be easier to work with, and should give consistent results.

I have also made some observations on the variation of the
resistance of mercury with temperature.

The formula quoted above from Benoit is given by Mascart,
Nerville, and Benoit as determined from experiments between
0° and 100°.

It is ,

R,= Ro(1 + 00086492 + -0000011227).
According to Strecker,
R, = Ro(1 + :000900¢ + -00000045:2?).

This is derived from observations at 0°, 10°, 15°, and 20°.
Lorenz finds that between 0° and 27°32 his experimental
results agree with the formula
and from 8°'32 to 35°31 with
R= R,(1 + 0009162).
While Siemens and Halske give
R,= Ro(1 + 0008523¢ + 0000013562").

I made observations on the two tubes, Nos. 37 and 389,
determining their resistances at temperatures of about 0°, 5°,
10°, and 15°; and the results of the observations show that the
average change for low temperatures is less than that given
by the above formule. Thus, if we call (R;—R))/Roé¢ the
average change between temperatures ¢ and 0, the values, as
found from my experiments and as calculated by the formule
of Siemens and Benoit, are as given below.

Average change Siemens and
between Halske. Benoit. R. T. G.
5 and 0 000858 000870 000834
19250 ‘000865 ‘000876 ‘000861

1D, yy DD ‘000872 ‘000881 000879

Standard Resistance-coils with Mercury-Standards. 353

The values given in the last column are the mean of those
derived from the two tubes; the greatest difference between
the mean and any one observed value was ‘000005. Strecker’s
values would be above Benoit’s. |

Lord Rayleigh found for the average change between 0°

and 12°, 000861.

Of course the value of this temperature-coefficient depends
somewhat on the glass of the tube; but the differences between
the coefficients of expansion of various kinds of glass are too
small to account for the whole of the differences shown
above.

The paper by Herr Strecker on the same subject has been
already referred to, and it will be noticed that the number he
arrives at as expressing M. in B.A. units differs by ‘00078
from that given by Lord Rayleigh ; or, if we take the num-
bers expressing the B.A.U. in terms of M., the difference is
°00086. Now Strecker had compared his mercury tubes
with a B.A.U. tested carefully by myself against the stan-
dards, and found to have the value of

"99937 B.A.U. at 17-7.
This coil is marked @, No. 54.

On the conclusion of his experiments, Herr Strecker re-
turned this coil tome, and kindiy sent with it a German-
silver copy of his tubes marked No. 138, and said to have,
according to his determinations, a resistance of

1:00189 M. at 10° C.,

with a temperature-coefficient of
‘000247 per degree.

It was thus in my power to repeat my comparison of & , No.
54, with the standards, and to compare the values of M. as found
by Herr Strecker with the B.A.U. The value found for €
54 agreed very closely with that given by Herr Strecker; the
results of my comparison of Herr Strecker’s coil with the
B.A.U. are given below. The comparisons were made as
usual, by Carey-Foster’s method; but since the difference be-
tween the two resistances was such as to require the use
of a long piece of the bridge-wire, a resistance of 20 B.A. units
was introduced in multiple arc with the standard 1 unit.
This resistance was composed of the two standard ten-unit
coils of the Association.

Phil. Mag. 8. 5. Vol. 20. No. 125. Oct. 1885. 2C

304 Lord Rayleigh on the Accuracy of Focus

TABLE VI.

Date Rese B.A.U. in M. |M. in BAU.
GT Ed Seta Flat. 1:04814 ‘95407
Pe a Tat es oe © 54. 104806 ‘95413
Biles A RE ae Flat. 104791 ‘95427
"tt aS Flat. 1:04820 95402
by Pinerelpae eect te P 54. 1:04811 ‘95410
oer Wes came re aes ® 54and20.| 1-04801 ‘95419
PCR ahs Flat and 20.| 104824 95399
I ae Ae Flat and 20.| 1:04797 95422
Moan \o.Gchercus. Beit eels | 95412

These numbers, it will be seen, agree with those found by
Lord Rayleigh and Lorenz, and differ greatly from Strecker’s
values. |

They represent the results of the comparison of a German-
silver copy of Strecker’s tubes with the B.A. units, including
the actual coil employed by Strecker, @ No. 54; while
Strecker’s numbers give the result of the comparison of his
tubes directly with No. 54. |

I have not been able to explain the discrepancy. Herr
Strecker suggests in a letter to me on the subject, that the
resistance of his coil No. 13 has changed somewhat in the
year which has elapsed since his comparisons were made. At
present I am in correspondence with Prof. F. Kohlrausch, of
Wiirzburg, in whose laboratory Strecker’s experiments were
conducted, and hope to obtain another copy of Strecker’s tubes,

or at any rate to have another comparison made between his
tubes and the B.A. standards.

XL. On the Accuracy of Focus necessary for Sensibly Perfect
Definition. By Lord Rayieien, /.L.S.*

N my “ Investigations in Optics”? t I have examined the
eftect upon definition of small disturbances of the wave-
surfaces from their proper forms. It follows, for instance,

* Communicated by the Author.
Tt Phil. Mag. 1879 & 1880.

necessary for Sensibly Perfect Definition. 309

that the aberration of a plano-convex lens focusing parallel
rays of homogeneous light is unimportant, so long as the
fourth power of the angular semi-aperture does not exceed
the ratio of the wave-length to the focal distance [a* < (A/f)],
a condition satisfied by a lens of 3 feet focus, provided that the
aperture be less than 2 inches. I propose at present to apply
similar principles to the question of focusing. |
The most convenient point of view is that explained* for
ealeulating the focal length of lenses. If the lens AB con-
verges parallel rays to a focus at F, the retardation of the

Db

G———

central ray EF, due to the substitution of a thickness ¢ of
glass for air, is (w—1)t; and this must be equal to the retar-
dation of the extreme rays passing the (sharp) edge of the
lens, 7. e. AF—CF. Thus, if AC=y, FC=f,
Ae
-DieV Pty) -f=3 2 + + @

approximately, which gives the focal length in terms of the
semi-aperture and the “‘ thickness ’’ of the lens.
If we suppose that w varies,

1 2
Su. t=— 5, ik cad te ee

giving the change of focus required to compensate the change
of uw. Let us, however, inquire what is the state of things
at the old focus. The secondary rays from the extreme boun-
dary of the lens arrive with the same phase as before the
change of index; but the central ray undergoes a relative
retardation amounting to du.t. This quantity tells us the

discrepancy of phase; and we know that if it is less than 3),

the agreement of phase is still good enough to give nearly
perfect definition. Hence from (2) we see that a displace-
ment of from the true focus will not impair definition, provided

fr
6 ant Ko i, a eS ee a
ak re (3)

* Loe. cit. p. 480.
202

356 Lord Rayleigh on the Accuracy of Focus

It appears that the linear accuracy required is the same
whatever the absolute aperture of the object-glass may be,
provided that the ratio of aperture to focal length be pre-
served. 7

In some trials that I have made the diameter of the object-
glass was 13 inch, and the focal length 12 inches. Taking
N=45.995 inch, we get from (3)

df <:0115 inch,

a result which corresponded very wel! with observation. The
instruments employed were the collimator and telescope of a
spectrometer, the object under examination being a slit backed |
with a soda-flame. A high-power eve-piece was used, and
the telescope was adjusted until the edge of the slit and the
wire in the eye-piece were seen well defined together. The
instrument was unprovided with an easy focusing motion,
so that it was not possible to try backwards and forwards
conveniently. In this way the setting corresponded more
closely to the suppositions of theory than if it were the result
of comparisons between appearances at equal distances within
and without the point chosen. It will be understood that
there is no theoretical limit to the accuracy with which a
focal point may be ultimately determined, if the lenses are
good, and observations are multiplied with suitable precautions
to avoid asymmetry.

In ten settings the extreme difference was only -02 inch, °
showing that a displacement of ‘O01 inch from the true focal
point was just recognizable.

By using various coloured flames, or by throwing a spec-
trum upon the slit of the apparatus, we may determine the
focal length for different kinds of light. With proper achro-
matic lenses the ditferences should be pretty small, the
minimum focal length corresponding to the yellow-green
rays. It so happens that my instrument is far from properly
compensated, and gives a fair primary spectrum, so that the
difference of focus for yellow and green is very easily recog-
nized. In the case of a single lens this method would give
the dispersive power of the glass with fair accuracy, By
comparison with the theory of the resolving power of prisms
we see that the dispersion is about as favourably determined
with a lens as with a prism of equal thickness. In either
case a change of index such that du.t=1A leaves the phase
agreement nearly unaltered at the original points ; but in
other respects the circumstances are probably rather more
favourable in the case of the prism.

It is generally considered that the most accurate way of

necessary for Sensibly Perfect Definition. 307

focusing a small telescope is to move the eye across the eye-

' piece, altering the adjustment until there seems to be no re-

lative motion of object and cross wires. I have tried this
plan in an improved form in order to see whether a higher
degree of accuracy of adjustment was really attainable, al-
though theory seemed tv show that no great advance was
to be looked for. A heavy pendulum, executing complete
vibrations in about two seconds, was fitted up in front of the
telescope, and carried with it a screen perforated by a slit.
The width of the slit was about a quarter of the entire aper-
ture, and the oscillations were at first of such amplitude as
just to bring the extreme edges of the lens into play. In the
earlier experiments the slit of the collimator was backed by
the clouds, a piece of green glass being interposed. This was
before I had discovered the remarkably unachromatic character
of the instruments, and I was puzzled to interpret the ap-
pearances presented. On one side of the focus the relative
motion of the image was (as it should be) in the same direc-
tion as that of the pendulum, and on the other side in the
opposite direction ; but the transition was not well defined,
and the image executed evolutions very visible to the ob-
server, who at the same time was not able to describe them
as swinging in one direction or the other. The effect upon
the eye was remarkably unpleasant and fatiguing to watch ;
it disappeared when recourse was had to sodium light, and
doubtless depended upon the variation of quality in the light.
It may be noticed that spherical aberration would show itself
by a swinging of the image in a period half that of the
endulum.

With the soda-flame the adjustment to focus by getting rid
of the swinging motion was pretty accurate ; but not much
advantage was gained in comparison with a setting by simple
inspection under full aperture. As before, the extreme
difference in a set of ten was about ‘02 inch.

The substitution of white for monochromatic light was in-
structive. In either extreme position of the oscillating slit
the light was seen to be spread into a spectrum of moderate
length, the blue and red being interchanged after each half
period. Under these circumstances the cross wires can be
made to maintain their position in that part of the spectrum
only for which the telescope is focused. If, for example, it
be the green of the spectrum, we may bring the cross wires
to this position when the pendulum is at rest, and then, in
spite of the oscillation, the position will be maintained. If,
without altering the focus, we move the cross wires to another
part of the spectrum, then, when the pendulum oscillates, the

308 Lord Rayleigh on an Improved Apparatus

wires will be seen on a different part of the spectrum after
each half period. In order to fix the new part of the spectrum
upon the cross wires, a change of focus is demanded. This
experiment would hardly succeed with properly compensated
object-glasses, but it could be imitated with the aid of single
lenses.

XLI. On an Improved Apparatus for Christiansen’s Expert-
ment. By Lord Rayueien, /.L.S.*

HE very beautiful experiment in question, described by |

C. Christiansen in Wiedemann’s Annalen for November
1884, consists in immersing glass-powder in a mixture of
benzole and bisulphide of carbon of such proportions that for
one part of the spectrum the indices of the solid and of the
fluid are the same. Being interested in this subject from
having employed the same principle for a direct-vision spec-
troscope (Phil. Mag. January 1880, p. 53), I have repeated
Christiansen’s experiment in a somewhat improved form,
which it may be worth while briefly to describe, as the matter
is one of great optical interest.

I must premise that the beauty of the effect depends upon .

the correspondence of index being limited to one part of the
spectrum. Rays lying within a very narrow range of refran-

gibility traverse the mixture freely, but the neighbouring

rays are scattered laterally much as in passing ground glass.
Two complementary colours are therefore exhibited, one by
direct, and the other by oblique, light. In order to see these
to advantage, there should not be much diffused illumination,
otherwise the directly transmitted monochromatic light is
liable to be greatly diluted. The prettiest colours are ob-
- tained when the undisturbed rays are from the green; but
the greatest general transparency corresponds to a lower
point in the spectrum.

The improvement referred to relates merely to the use of a
flat-sided bottle to contain the preparation. In order to get
a satisfactory result it is necessary that the sides of the con-
taining vessel be pretty good optically. This condition may
be satisfied with a built-up cell, but on account of the diffi-
culty of finding a suitable cement, it is rarely that such
cells remain in good order for any length of time. It
occurred to me that a bottle might be made to answer the
purpose, provided the precaution were taken of using the same
kind of glass for the bottle and for the powder. The outer sur-

* Communicated by the Author.

5 soo

for Christiansen’s Experiment. 309

faces of the glass sides of the bottle can be worked flat, while

the unavoidable irregularities of the inner surfaces are com-
pensated by the liquid, which, being adjusted to have the
same index as the powder, will have also the same index as
the glass of the bottle.

The bottles that I have used* are about 3 inches high,
14 inch wide, and about ? inch thick, outside measurement.
The outer surfaces are worked (like plate glass), and not
merely flattened upon a wheel, as is usual with ordinary per-
fume bottles. For my earlier trials I was provided with a
piece of flint glass from the same pot as the bottles; but
although the experiment succeeded well enough as regards
the elimination of the internal irregularities of the walls, the
glass-powder itself did not behave as well as I had seen plate-
glass powder do. It appeared ultimately that the flint was
not sufficiently homogeneous for the purpose, and another
specimen of flint was also a partial failure, from the same
cause ; but a sample of optical flint, kindly supplied to me by
Dr. Hopkinson, gave excellent results.

It is more important that the powder should be homo-
geneous in itself than that it should correspond very accu-
rately with the glass of the bottle. For ordinary purposes
plate-giass powder (all, of course, from one piece) may be
used in a bottle of soda-glass, or even of ordinary low flint.
In preparing the powder great care is required to exclude
dirt. With respect to the coarser grades there is no great
difficulty, but the finer powder is apt to be contaminated with
the substance of the mortar. I prefer to use one of iron, so
that a magnet will remove the foreign matter. The elimination
of fine dust is also facilitated by a blast of wind from bellows.

In order to get good definition it is necessary not only that
the powder be homogeneous, but that the temperature be
uniform ; for, as Christiansen has shown, the transmitted ray
rises rapidly in refrangibility with temperature. In order to
secure homogeneity it is sometimes necessary to shake up the
preparation, which (to prevent the formation of air-bubbles)
is best done with a rather gentle motion while the bottle is
held nearly horizontal, The proportion of liquids necessary
varies with the temperature and with the kind of glass. Flint
will require a higher proportion of bisulphide of carbon than
plate-glass. If the transmitted light is too high in the spec-
trum, add more bisulphide of carbon; if too low, more
benzole. .

The experimenter, working in a room lighted in the ordi-
nary manner, will be disposed at Srst to underrate the degree

* Supplied by Messrs. Harvey and Peak, Beak Street, Regent Street.

360 Comparison of Methods for observing Small Rotations.

of homogeneity of the transmitted light. If the shutters be
nearly closed, so that light finds entrance through a narrow
slit only, a better judgment can be formed, which may be
tested by prismatic analysis. |

In an otherwise dark room, lighted by a powerful soda-
flame, it is interesting to remark how very slight a change in
the critical colour manifests itself in the general appearance
of surrounding objects seen through the preparation. When
the ray of maximum transmission corresponds closely to that
of soda, the powder is almost invisible, and objects are seen
as through a clear medium. But so slight a change of tem-_
perature brings with it a hazy appearance, that it requires
some care to obtain the best effect. It is desirable also to
exclude by absorbing media the blue light which usually
attends a soda-flame in very sensible degree.

XLII. Optical Comparison of Methods for observing Small
Rotations. By Lorp Rayueten, /.R.S.*

N order to measure very small rotations, e. g. of the sus-
pended parts of a galvanometer or magnetometer, two—
methods are commonly employed. We may either observe
with a magnifier the motion of a material pointer; or, follow-
ing Gauss, cause the rotating parts to carry round a mirror
in which a scale is seen by reflection. In a modification of
Gauss’s method, well known from Sir W. Thomson’s galva-
nometers, the image of a dark or bright line is thrown ob-
jectively upon the scale. In deciding which arrangement to
adopt in any particular case, various circumstances would have
to be taken into account, but still a comparison of capabilities
from a purely optical point of view is not without interest.

In the mirror method the optical limit depends upon the
horizontal breadth of the mirror itself. The easiest road to
the desired conclusion, as well as the most instructive, is by a
direct application of the principles of the wave theory. To
take the simplest case, we will suppose the mirror rectangular.
Consider, then, a luminous point, and its image after reflection,
whether in the focal plane of a telescope, or formed directly
upon a scale. The optical work being perfect, the secondary
rays from every part of the mirror agree in phase at the focal
point. Now suppose that the mirror rotates through such
an angle that one vertical edge advances a quarter of a wave-
length (} 2%), while the other retreats to the same amount, and
consider the effect on the phase-relations at the point in ques-

* Communicated by the Author.

On the Thermodynamic Efficiency of the Thermopile. 361

tion. Itis evident that one extreme wave is accelerated and the
other retarded by § A, and that the phases are now distributed
uniformly over a complete cycle. The result is therefore
darkness; and the effect of the rotation has been to shift the
image through half the width of the central bright band

_ which, with accompanying fringes, is the representative in

the image of a mathematical line*. Such a motion would
be visible (with proper arrangements as to magnifying), but
the limits of revolving power are being approached. It is to
be noticed that the conclusion is independent of the focal
length of the mirror and of the employment of a telescope.
Provided of course that the full width of the mirror is really
used, a motion of its vertical edges through + > may be made
evident.

A comparison with the method by direct observation of a
pointer is now easy; for, as has been proved by the re-
searches of Abbe and Helmholtz,a motion of + X% may be
rendered evident in a very similar degree by direct applica-
tion of a perfect microscope to the moving object. If, there-
fore, we suppose the length of the pointer to be equal to the
half-width of the mirror, the two methods are optically upon
alevel. It is needless to say that it would be easy to give
the pointer a great advantage in this respect ; but the direct
use of the microscope would often be interfered with by
motions in the line of sight, making it impossible to preserve
the focus. And besides this, it is sometimes necessary for the
observer to remain at a distance.

XLII. On the Thermodynamic Efficiency of the Thermopile.

By Lord Rayieiew, F.R.S.T

URING the last few years the thoughts of many elec-
tricians have turned to the question of the possibility
of replacing the dynamo by some development of the thermo-
pile; and it is, I believe, pretty generally recognized that the
difficulty in the way is the too free passage of heat by ordi-
nary conduction from the hot to the cold junction. The
matter may perhaps be placed in a clearer light by an actual
calculation, accompanied by a rough numerical estimate
applicable to the case of German silver and iron.
If t, f, denote the temperatures of the hot and cold junc-
tions respectively, é the electromotive force per degree Centi-
grade, the whole electromotive force for n pairs in series will

* See “ Investigations in Optics,” Phil. Mag. 1879-1880.
+ Communicated by the Author.

362 On the Thermodynamic Efficiency of the Thermopile.
be represented approximately by
ne(t—t,).

The magnitude of the current (C) is found by dividing
this by the sum of the internal and external resistances
(R,+R); and the useful work done externally per second is
RC?. It reaches a maximum when the external resistance is
equal to the internal; and its amount is then

ne(t—t,)?
AR Osek,
The value of the internal resistance R, depends upon the
dimensions and specific resistances of the bars. Denoting the
latter quantities by 7, 72, and taking o,, a, to represent the
areas of section, the common length being /, we have
Ry=nl (2 +72),
ee
so that the external work per second is
ne(t—t,)?

41(" +7
ith tei
We will now compare this with the work dissipated by
ordinary conduction of heat along the bars.
If Q be the amount of heat conducted by the n pairs, ry,
ry the thermal resistances, then
ee ha Oo
=n|— +-— } (¢—1,).
Q n( 2 ag) eae
The fraction of this heat, supplied at temperature ¢, which
might be converted into work by a perfect engine workin
between the absolute temperatures ¢ and ¢,, is (¢—¢,)/t; so that
the work dissipated per second is

nJ(t—t,)?
Te

where J denotes the mechanical equivalent of heat.
The ratio of this to the useful work is |

4J Vy i) 07 on)

miei eed,
independent of (t—¢,), of n, and of 1. It is further evident
that the ratio in question does not depend upon the absolute
values of the sections, or of the electrical and thermal resist-

ances, but only upon the ratios of these quantities. Thus the
efficiency of the thermopile is independent of the absolute

Oo ox)
a ig PF
Ut i)

Electro-optic Action of a Charged Franklin’s Plate. 368

dimensions of the bars, and even of the difference of tempe-
ratures at the junctions. The power is increased by dimi-
nishing the length and increasing the sections to a limit not
indicated by these expressions (in which the terminal tempe-
ratures are regarded as given), and probably determined in
practice by the necessity of conveying the heat to the scene
of action.

The resistances being given, the ratio of sections a, /o, is
to be determined so as to make our ratioa minimum. This

happens when
ss eae.
ay ryry

AJ rh r\"
te 1} V9

To turn this into numbers for the case of German silver
and iron, we have the following approximate numbers in C.G.S.
measure” :—

and thus we get

mee x LO", ear EO
= 10x 10%, 1 — POX,
=" 107:

The value of J is 4:2 x10’, and for ¢ we will assume 500
(absolute measure). ‘The use of these gives, as the ratio of
the work that would be obtained by a perfect engine from the
conducted heat to that actually obtained from the thermo-
electric force, the number 300; from which we may conclude
that the steam-engine and dynamo are not likely to be super-
seded by a German-silver and iron thermopile, even though
considerable allowance be made for the admitted roughness of
the numerical estimate.

As regards other materials, it is interesting to note that the
ratio contains as divisor the square of the electromotive force
per degree.

XLIV. Electro-optic Action of a Charged Franklin’s Plate.
By Joun Kerr, LL.D., Free Church Training College,
Glasgow +.

HAVE examined this difficult subject in several sets of
experiments, the last ef which was finished about three
years ago. Having satisfied my own mind, I dropped the
subject, thinking that no publication of the results was

* Kverett’s ‘Units and Physical Constants’; Landolt’s Tabellen.
+ Communicated by the Author.

364 _ Dr. J. Kerr on the Electro-optic Action

required ; but on looking into M. Wiedemann’s treatise, now
in course of publication, I have been led to believe that a
short paper on the subject may be of some use.

In his‘opening remarks on Hlectro-Opties*, M. Wiedemann
draws attention to the experiments tried on Franklin’s plates
by MM. Quincke, Gordon, Mackenzie, and Grassman ; and
he seems to accept the absence of effect in these experiments
as decisive of the question of action or no action, in a uni-
formly charged plate of glass. Accordingly, when’ he comes
to the earliest known instance of electro-optic double refrac-
tion (the effect of intense electric stress at the centre of a
block of glass, which I discovered in 1875), he introduces my
experiment as an example of a non-uniform and optically
active electric field, in contrast with those other experiments f,
as examples of a field sensibly uniform and optically inactive.
Afterwards, at the end of his exhaustive account of the whole
subject, M. Wiedemann returns to the same point, and makes,
in substance, the three following statements :—

(1) When the dielectric is a solid (a block of glass), and
the distribution of electric force is not uniform, there are
certain strains known to be produced (simple expansions,
which are greatest at the place of shortest lines of force) ;
and these known strains may account for the observed
effects.

(2) There appears to be no electro-optic double refraction
in the case of a uniformly charged Franklin’s plate.

(3) But if one of the coatings be much reduced in size,
or if we use, instead of it, a little mercury, contained within a
small ring of glass, which is cemented to the upper surface
of the plate at its centre, then the charging of the plate gives
rise to a strain (a simple expansion, such as that produced by

ts ae ce
be) 4 ’ es <

local rise of temperature) in the part of the plate immediately —

between the coatings, and a ray which passes through the
plate and between the coatings is doubly refracted of course.
For some additional matter on the subject, not required

here, I must refer to M. Wiedemann’s treatise. The prece-

ding statements give a correct view of a theory of the pheno-
mena, which was advanced some years ago by M. Quincke ;
the jirst statement giving his explanation of my experiment
already referred to, and the third giving an explanatory
account of a subsequent experiment of his own. My principal
objection to this theory has always been, that it ignores

* Die Lehre von der Electricitit, vol. ii. p. 125.

+ All published subsequently to mine, and all, I believe, with explicit
reference to it; though Prof. Quincke’s experiment was performed some
ten years earlier.

a

of a Charged Franklin’s Plate. Bich a

those pure and intense effects which I obtained long ago in
liquid dielectrics, particularly in carbon disulphide, and in a
sensibly uniform field. The strains assumed in the first and
third statements are undoubtedly real ; and it must be left to
experiment to determine whether, and how far, they take

effect in any actual case. With regard to my own early

experiments on thick plates of glass, | am not aware of any
evidence to show that these very small strains contributed to
the observed effects in any degree whatever ; but I pass from
these points to our proper subject. 3

The second of the preceding statements is not true in any
other sense than as a record of several inadequate experi-
ments. I find that a charged Franklin’s plate, with uncoated
margin so wide only as to give a small, working difference
of potentials, acts upon transmitted light in the same way as
olive oil, amyl oxide, or any other negative dielectric. M
earlier experiments on the Franklin’s plate need not be de-
scribed, as they were chiefly of use in deciding between
methods, and in showing the difficulties that had to be avoided
or overcome. My last set of experiments on the subject were
finished in July 1882; and the following account of them is
drawn from full notes that were taken at the time.

One of my best plates was a rectangular piece of good
English plate-glass, exactly g of an inch thick, 23 inches
wide, 42 long. This piece was one of a set, which were cut
out of separate plates or fragments, and were then put together,
and worked in block as one thick plate. The smaller ends of
the thick plate were ground to sensibly parallel planes, and
were polished carefully, till the giass gave a perfectly clear
and undistorted view of neighbouring objects through the
length of the block. Before going further with the prepara-
tion of the plates, I tested them separately in the polariscope;
for I had found already that, unless the plates were tested
and carefully selected, the experiments in view were almost
hopeless.

Along with a couple of Nicol’s prisins, two other pieces
were employed—a rectangular slip of thin plate-glass used
as a hand-compensator, and a larger slip used as a fixed
compensator, the latter being placed in a simple screw-
press, and subjected to horizontal tensions and compres-
sions, which could be made to vary continuously through a
considerable range. The two Nicols were placed in hori-
zontal line with a good paraffin flame, and were exactly
crossed, their principal sections being at 45° to the vertical.
The plate was then fixed between the two Nicols, its faces

_ yertical, and its terminal polished surfaces perpendicular to

366 Dr. J. Kerr on the Electro-optic Action

the incident polarized pencil. The insertion of the plate
restored the light always from extinction ; and the fainter this
restoration, the more promising was the plate. ‘The fixed
compensator, which stood between the Nicols, and immediately
after the plate, was then strained by trial, with or without
small rotation of the second Nicol, till good extinction was
obtained permanently ; and the feebler the strain thus required,
and the purer and sharper the extinction obtained, the more
promising was the plate. ‘The hand-compensator was then
inserted between the fixed compensator and the second Nicol,
and supplied the last part of the test, which consisted in
observing the degree of strain (horizontal tension or com-
pression) which had to be applied, to give good restoration
from extinction. When the requisite strain was small or very
moderate, the plate was accepted as good, otherwise not.

The most of the plates that I have had in hand were pure
failures under this test, and were therefore thrown aside as
evidently quite unsuited for the electro-optic experiment. Of
the set of six plates last started with, four were rejected with-
out hesitation, the fifth was very imperfect, but worthy of
trial, the sivth was good, though not perfect. Of another set
of plates, which were only half as thick as the former, there
was none even moderately good; but one of them was re-
tained, as in a small degree promising. The serious failure
was generally in the last step of the test, that with the hand-
compensator. When a good or very fair permanent extinc-
tion had been obtained by the joint action of the plate under
trial and the fixed compensator, I found often that it required
the exertion of almost all my strength onthe hand-compensator,
to get anything like a good restoration in the polariscope.

The light should have been well restored by a barely sensible
effort ; but, in its passage through such a length of irregularly
strained glass, it had lost ever so much of that sensitiveness
under bi-refringent action which constitutes the whole value
of the plane-polarized ray in any really delicate electro-optic
work. No clear optical effect of moderate electric stress
could be reasonably looked for in a plate such as that now
mentioned, which allowed only a very small effect, or a mere
trace of effect, to be given by a comparatively very strong
bi-refringent action of the hand-compensator.

The plate chosen was coated with tinfoil in the usual way,
the margin left on each face all round being rather less than ‘
a quarter inch. Projecting from the centre of each coating
there was a small conducting tongue, made of a bit of tinfoil,
which was folded several times on itself, opened out at one
end, and pasted securely by that end to the coating. When

ee ne nee
*

Ta ,
na * ah

of a Charged Franklin’s Plate. 367

the paste had dried thoroughly, the plate was covered with a
coat of thick lac-varnish all over, with exception of the polished
ends and the conducting tongues. Another coat of the varnish
was applied afterwards ; and the operation was repeated more
than a dozen times, proper intervals of time being allowed
for the drying of the varnish. When all was done, the tin
coating was dimly visible in a good light, and the envelope of
lac appeared to be perfectly continuous.

The plate was placed edgeways on the top of a fixed pillar
of glass, and was attached to the pillar by a few drops of
melted lac. Two perforated balls of brass, supported on
separate pillars, were brought up to the plate at the centres of
the two coatings; and the conducting tongues, passing into
the perforations of the balls, were in good permanent contact
with the brass. Through these balls the two coatings could
be put into and out of connection with prime conductor and
earth, and without any pressure on the plate. Two small
screens of thin vulcanite were attached lightly to the plate on
opposite sides of it, so as to protect the eye from all useless
light.

” Final State of the Best Plate-—The insulating power was
not great. When one coating of the plate was connected with
prime conductor and the other with earth, the spark-length
between prime conductor and knuckle was at first about one
sixth of an inch ; but afterwards, when the lac had hardened
better on the plate, and in favourable weather, the spark-
leneth was fully a quarter inch. Spark and shock were of
course a good deal heavier than from the prime conductor
unconnected.

The optical action of the best plate (still uncharged) was
not sensibly different from what had been shown already in
the process of testing. The sensibility of the ray under the
action of the hand-compensator was indeed very satisfactory,
considering that the light had passed through more than four
inches of glass.

Method of Experiment.—In the simplest mode of working,
the pieces were placed on the table in horizontal line and in
this order :—paraffin flame, screen with narrow and _ short

vertical slit, first Nicol with principal section at 45° to the

vertical, the Franklin’s plate with coated faces vertical, and
terminal faces perpendicular to the incident pencil, fixed com-
pensator, a little space for the hand-compensator, second Nicol
at extinction. The axis of the small transmitted pencil of
light was equidistant from the upper and lower edges of the
coatings. In another arrangement, the collimator and lunette
of a spectroscope were used, the lunetite behind the second

i
oe

368 Dr. J. Kerr on the Electro-optic Action

Nicol, and the collimator (with very narrow vertical slit) before
the first Nicol, the light employed being either a sunbeam or
a beam from the bright sky. This arrangement served well
in the case of the very thin plate already mentioned, which
could not be managed well with the naked eye, because of the
confusion produced by internal reflections ; but for the larger
and better plates the first arrangement was preferred.

The electric force was applied most conveniently as follows.
A Leyden jar stood on the table near the Franklin’s plate ;
and the outer coating of the jar and second coating of the
plate were put permanently to earth. The jar was charged
by twenty to thirty turns of the plate-machine : a common
discharger, connected with the first coating of the Franklin’s
plate, was brought up to contact with the knob of the jar, but
gradually, so as to avoid spark-discharge over the margin of
the plate; and when the contact had been maintained for about
three seconds, the connection was broken, and the plate was at
the same time discharged. In this way the optical observation
lasted only for the few seconds before and after discharge of
the plate, so that small changes of intensity of the light in the
polariscope were the more easily detected.

First Hxperiment.—Particular attention was given to the
delicacy of action of the polariscope. The extinction was
made as pure as possible by proper strain in the fixed com-
pensator. The sharpness of the extinction was tested by the
hand-compensator, the requisite delicacy being obtained when
the light was restored clearly by feeble strains, horizontal
compression and tension successively. When these conditions
were fulfilled, the electric force, applied in the manner just
described, gave good restoration from extinction in the pola-
riscope, the light brightening gradually for a second or two,
and then, after discharge of the plate, falling back to extinetion
at much the same rate. The intensity of the restored light
was never great, being equal to that given by a very moderate
compression in the hand-compensator. When the insulation
was very defective, or the plate optically bad, the effect was
very poor indeed, and could not always be seized with cer-
tainty; but when these unfavourable conditions were avoided,
the rise and fall of the light, from extinction and back to it,
were quite distinct and regular.

Second Eaperiment.—The light was restored faintly from
extinction, as by horizontal compression of the fixed compen-
sator. I say as by horizontal compression, because the
action of the uncharged Franklin’s plate in the polariscope was
sometimes equivalent to horizontal compression, and had to be
neutralized by horizontal tension of the fixed compensator ;

of a Charged Franklin's Plate. 369

and then the desired effect was obtained by a small diminution
of that horizontal tension. In any case, the permanent initial
restoration required in this experiment, when tested by the
hand-compensator, was clearly strengthened by a feeble
horizontal compression, and clearly weakened by a feeble
horizontal tension. When these conditions were fulfilled,
the ‘light was regularly and very clearly strengthened by
charge of the plate, the intensity rising and falling gradually,
—hbefore and after discharge—as in the first experiment.
Under adverse conditions, the insulation bad, or the plate
optically imperfect, this experiment succeeded a good deal
better than the first, the effect being, as it ought to be,
distinctly stronger.

Third Experiment.—The light was restored faintly from
extinction, as by horizontal tension of the fixed compensator ;
and this initial effect was seen to be properly affected under
the action of the hand-compensator, clearly weakened by a
feeble horizontal compression, and clearly strengthened by a
feeble horizontal tension. ‘The electric force being applied as
formerly, the changes now observed in the polariscope were
contrary to those obtained in the second experiment, and
apparently of equal range. During charge the light was
weakened, its intensity falling regularly and very clearly for
a second or two, and then, after discharge, returning to the
initial value at much the same rate. In this experiment, as
in the other two, the intensity of the effects was very mode-
rate at the best ; but there was no uncertainty or indistinct-
ness about the results, except when the conditions were known
to be very unfavourable.

Ihave referred already to the earliest known instance of
electro-optic double refraction. In that instance, the electric
field lay between and around the ends of two straight col-
linear wires, at the centre of a block of glass, the thickness
of glass between the terminals being (as in the present experi-
ments) about 4 of an inch, and the thickness traversed by
the ray ? of an inch. It would be difficult to realize an elec-
tric field much more unlike that of the Franklin’s plate ; and
yet the optical effects obtained in that case were of precisely
the same kind as those obtained in the present experiments,—

negative double refractions with reference to line of electric

force as axis. It appears thus, that the fact of electro-optic
double refraction in glass is not dependent on form of the elec-
tric field, or mode of distribution of the electric forces.

There is one thing brought out by the preceding compari-
son which is at first sight not satisfactory. Inthe Franklin’s
plate as compared with the old dielectric, there was a con

Phil. Mag. 8. 5. Vol. 20. No. 125. Oct. 1885. 2D

370 Dr. J. Kerr on the Electro-optic Action

siderable increase of thickness of the dielectric in the direction
of the ray, and yet there was a large decrease of intensity of
effect. But if we assume, as we may fairly do, that the
quantity of optical effect varies as the square of the difference
of potential of the two coatings, we see that a moderate in-
crease of ray-length in the dielectric may be much more than
counterbalanced by a moderate decrease of difference of
potentials. Now in the old experiment and the new, the
ratio of ray-lengths in the dielectrics was about 1 to 6, while
the change of spark-length was from an average of 6 inches
to a maximum of + inch, implying a large fall of potential.
I may even observe that the increase of ray-length in the
present experiments appears to have told clearly on the in-
tensity of the effect, because, in the old experiment, and in
the most favourable circumstances, hardly any effect could be
obtained with as little as an inch of spark-length.

But, further, we must remember those permanent and irre-
gular strains which were sought to be eliminated in the pre-
liminary testing of the plates. The disturbing influence of
these strains may be expected to increase with every increase
of ray-length in the dielectric, and with every decrease of the
difference of potential. When the ray-length in glass
amounts to several inches, I believe that it is not possible to
get rid of this disturbance perfectly. In the best plates that
I have worked with, not only was the disturbance present,
but it always told seriously, weakening if not masking the
effect of any very small birefringent action to which the
transmitted ray could be subjected. Taking these things
into account, and viewing the new phenomena and the old
as cases of the same effect of electric force, | consider the
small effects now obtained with the Franklin’s plate to be as
good as could fairly be expected, and better than could be
expected confidently ; and I therefore hold it proved by ex-
periment, that electrostatically strained glass acts in the polari-
scope as if compressed along the lines of electric force, and thus
always, whether the electric field is uniform or not.

I conclude briefly with a simple and important argument,
which appears to be secured now against all reasonable ob-
jection. With reference always to resultant effect in the
polariscope, as characterized by (say) the hand compensator:—
Electrically charged glass acts as if compressed along the lines
of force ; electrically charged resin acts as if extended along the
lines of force ; electrically charged oil of colza, a negative di-
electric, acts precisely as glass compressed in a direction parallel
to the lines of force ; electrically charged carbon disulphide, a
positive dielectric, acts precisely as glass extended in a direc-

of a Charged Franklin’s Plate. 371

tion parallel to the lines of force ; and so forward uniformly
(with variations of intensity and sign, and with occasional
disturbance, mostly from known causes) through a list of
more than 120 different dielectrics, as I have proved by ex-
periments already published. ‘The similarity of all these
optical actions implies a similarity of condition in all the
acting media. We infer that glass, resin, oil of colza, carbon
disulphide, and all those other dielectrics, when electrically

charged, are also specially and similarly strained. And it

seems to be quite consistent with proper caution, in the ab-
sence of opposing evidence, to extend this conclusion to all
dielectrics whatsoever.

We are led thus by the facts of electro-optics, and inde-
pendently of all theory, to the conception and realization of
electrostatic strain,—a strain maintained by the action of elec-
tric force, or (at least) in exclusive and indissoluble connection
with that action, and sustained by the dielectric as an elastic
medium,—a strain of one type in all dielectrics,—a strain
essentially directional in liquids as in solids,—a strain swt
generis. And in all this, or rather in the facts themselves,
we have a remarkable confirmation of the theoretical views of

Faraday and Maxwell.
Glasgow, 21st August, 1885.

Postscript.—In the charged Franklin’s plate, the glass is
compressed by the electric attraction of the two coatings;
and it may be inquired how far this mechanical strain contri-
butes to the effect obtained in the preceding experiments. [
have had this question suggested to me by Sir William
Thomson, as too important and too obvious to be neglected,
and admitting also of a simple method of solution ; but I am
sorry that the best answer obtainable with my present means
is very vague and uncertain.

When the old Franklin’s plate is laid flat, compressed by a
weight uniformly distributed over it, and submitted to the old
test in the polariscope, I find that it requires a total weight of
14 pound to give an effect equal to the old effect of electric
charge. ‘This compressing force of 1} pound must be much
greater than the attraction of the electrically charged coatings,
whose dimensions are about 2x34 inches, their distance
4 inch, and their difference of potentials at the moment of
observation probably less than that corresponding to a spark

of 4 inch in air. It seems, therefore, that the part of the

optical effect of electric charge which is due to compression of
the glass must be very small, if not insensible. Ishould men-
tion, however, that by a simple calculation, on the assumption

2 D2

372 Prof. Oliver Lodge on the Seat of the

of an important determination made by Sir William Thomson*,
the preceding figures give a superior limit of the attraction
of the coatings, which is as large as one ninth of the optically
equivalent compressing weight. As to a definite solution of
the question, I hope to attempt that, or something equivalent,
very soon, in a set of experiments which I have had in view
for some time.
Glasgow, 26th August, 1885.

XLV. Sequel to Paper on the Seat of the Electromotive Forces
ina Voltaic Cell. Theories of Wiedemann and of Helmholtz.
By Professor OLIVER LopGEf.

| Sa a recent Beibldtter{, Prof. G. Wiedemann points out that

in my summary of views and work in connection with

the seat of E.M.F. in the pile I have ignored his theory. I
had in fact been mainly concerned with possible explanations
of the Volta-effect; while Prof. Wiedemann starts with this as
a datum, and on the strength of it explains the action of the
pile. Moreover, since most of the theoretical opinions held on
this subject are summarized at the end of the second volume of
his Hlektricitdt, while his own theory is given in the first volume,
along with the account of Volta-force experiments, I had over-
looked it. To make amends for this oversight, for which I
beg to express my regret, and because I find it difficult to
summarize his opinions briefly, I trust I may be permitted
to give a rough semi-literal translation of his most conspicuous
passages as a help to other students of the subject.

Theory of G. Wiedemann, 1870. (Wied. Elek. i. p. 251 et seq.)

“We have now to investigate in what way the electrical
forces in a closed circuit of two metals and a liquid (e.g. Zn, Cu,
dilute HCl) are excited. We know that the constituents of
the binary compound contained in the liquid separate by the
passage of the current in such a way that the one (H) is set free
at the copper and an equivalent quantity of the other (Cl) at
the zinc, and that the latter combines with an. equivalent
quantity of zinc to form ZnCl. This process shows that the
deportment of the binary body between the metals consists no
longer only in a simple overbalancing attraction of its whole
mass for one or the other electricity, as in metals, but also in

* Reprint of Papers on Electrostatics (§ 340).

+ Communicated by the Author, in continuation of paper in Phil. Mag.
for March, April, May, and June of the current year.

{ Beiblitter zu den Annalen der Physik und Chemie, 1885, No. 7,
Band ix. p. 538,

seca

Electromotive Forces in a Voltaic Cell. — 373

a special action of its constituents. Since the constituent Cl
is liberated where the + current enters the liquid, and H
where the negative enters, we suppose that each equivalent of
Cl in the compound HCl is charged with a definite quantity
of negative electricity, which conditions its attraction for the
opposite positive electricity. It is the electro-negative element
of the compound. So also must the equivalent H be charged
with + E and form the electro-positive element. These
charges could be produced by the combination of H and Cl
exactly in the same way as they are produced by the contact
of zinc and copper. Since the HCl formedis unelectrified,
we must suppose that atoms of the + and — constituents
contain equal quantities of + and —H.

“Tf now we immerse a Zn and a Cu plate into the dilute
HCl, we can conjecture that the zinc has a stronger attraction
for the electro-negative constituent (Cl) than for the H. In
consequence of this the molecule of HCl touching the zinc
will so lay itself that its electro-negative element is near the
zinc and its electro-positive element near the copper. At the
same time the molecules thus polarized act by electrical attrac-
tion on the constituents of the successive molecules of HCl,
arranging the whole row of molecules between the Zn and the
Cu plates as in the figure. :

Cl. H Cl. Cl. H Cl. TE

“‘ By induction the negative electricity of the zinc-touching
electro-negative constituent, chlorine, would act on the Hi in
the zinc in such a way that the near parts will charge them-
selves positively, the further parts negatively. In the same
way the copper adjacent to the electro-positive element (H)
of the HCl atom lying against it will receive negative elec-

tricity. The positive will remove itself to the remoter regions.

“Thereupon the positive electricity in the zinc would com-
bine with the negative of the next-lying Cl atom, and this
latter with zinc to form unelectrical ZnCU!l. The electro-positive
atom H which was formerly joined to that Cl atom would itself
unite with the neighbouring atom of Cl belonging to the second

374 Prof. Oliver Lodge on the Seat of the

atom of HCl, at the same time combining with the electricity
contained in this atom ; in the same way the hydrogen of the
second atom—Ithe welcome letters wu. s. f. are now not far
ahead, and we will let them serve for the next fifteen or six-
teen lines ].

“Tt is evident that in this process there occurs a continual
loss of vis viva, since the elements of the binary compound
rushing up to the metal move with a definite velocity to the
metal, and then are reduced to rest, either with formation
of ZnCl or by being set free(H). ‘This loss of vis viva is the
equivalent of the heat which is evolved by the obviously occur-
ring chemical process, which essentially is the solution of an
equivalent of zinc in the acid. ‘The same value must like-
wise be the work applied in effecting the redistribution of
electricity. ,

‘When, therefore, electricity begins to flow, work must
be done, during the solution of Zn and liberation of an eq. of
H from the liquid in the whole circuit, whether it be in the
form of heat or of external work, which is equivalent to the
heat-production attending that chemical process.

“The electrical shearing-force E is an accelerating force,
so its work A in unit time is equal to its product with the
quantity of electricity moved in unit.time, m, and the distance
s which the same travels, or at

A=E.ms. .

The product ms expresses the quantity of electricity which
travels per second through the cross section of the conductor.

“If we have different combinations, as Zn, HCl, Cu; or
Mg, H,SO, or NaCl,Cu; or &c., we suppose that the same
quantity of H travels per equivalent... .[a long appeal to the
facts underlying Faraday’s law is made to prove this |; hence
ms= const.,

*. 21 const. 1:

or the H.M.F’. is directly proportional to the heat generated
by the decomposition of an equivalent of the binary compound
and the solution of the zinc.

“For the production of steady currents, that electrical force
can alone be active which depends on the unequal attraction
and polarization of the atoms of the binary compound in the
exciting liquid of the cell; the electrical force at the contact-
place of the metals which no mechanical changes can prelude
must, on the other hand, be inactive. That this latter, if it
perhaps opposed the metal liquid H.M.F. (as by immersion
of zinc and lead in KCy solution), is not compensated by
a definite part of the tension force, is proved by the afore-

ae

rs 4 . /
es See
Ripe eet er Gy)

be

Electromotive Forces in a Voltaic Cell. 375

said complete proportionality of the total electrical force (and
H.M.F.) in the closed circuit with the said heat-equivalent
of the chemical processes. It must then be neutralized in
some other way. ‘This would most simply happen on the
hypothesis that, by contact of the exciting fluid with the
metals, the H.M.}’. is excited in a double way: first, by
an unequally strong attraction of the masses of fluid and metal
as a whole for the one or the other electricity ; and, secondly,
by the unequal attraction of the metals for the oppositely
electrified components of the liquid. (This double pro-
perty would be quite analogous to the following : a metal (e. g.
zinc) can not only attract to itself by adhesion the whole mass
of any given fluid (HCl), but also can exert a much stronger
attraction for one constituent of the same (chlorine); just
also as a magnet attracts iron, not only by reason of gravi-
tation, but much more because of the magnetic polarization
of its individual particles.) |

“In consequence of the first unequal (mass) attraction for
the electricities, the decomposable liquids would behave exactly
according to the metallic law of tension, and in a closed circuit
of metals and liquids a complete neutralization of electrical
forces (and H.M.F.) would obtain; the second (chemical)
action would, on the other hand, alone produce the electrical
shearing-force effective in current formation and its accom-
panying H.M.F.

“We will therefore in the sequel denote the total H.M.F. at
the various junctions with strong letters, and that portion of
the same of which no part engages in exciting the current
by italics in brackets, and by italics without brackets the
portion corresponding to the chemical work. In the closed
circuit—zinc, sulphuric acid, copper—the whole active E.M.F.

would then be

Cu/S+ S/Zn + Zn/Cu= Cu/S+ S/Zn+(Cu/S) + (S/Zn) + (Zn/Cu).

“The experiments of Hankel and others give the total H.M.F.
between the metal M and water, which we denote by M/Aq,
and which is compounded, according to the foregoing hypo-
thesis, of the E.M.F. excited by the mass action of the water
(M/Aq), and that excited by reason of its chemical polariza-
tion M//Aqg. If therefore we put into water a copper and a
zinc plate, the total potential difference, or H.M.F. between its
ends, is Cu/Aq+Aq/Zn. This is the quantity we should
directly obtain if we connected the Cu and Zn immersed in
water with the same named plate of a zinc-copper condenser,
and after breaking the connection determined the charge of
the condenser.

376 Prof. Oliver Lodge on the Seat of the
“On our hypothesis,
Cu/Aq + Aq/Zn = (Cu/Aq) + (Aq/Zn) + Cu/Aq + Aq/Zn.
If we close the circuit by connecting the zinc and copper plate
by a copper wire, the force Zn/Cu adds itself to these H.M.F'.s,

and this force removes from the tension series of the metals

the following forces, (Cu/Aq) +(Aq/Zn) ; wherefore we get
Cu/Aq + Aq/Zn + Zn/Cu= Cu/Agq+Aq/Zn.

But this last is the part of the total H.M.F’. applied to the

formation of the current. Hence to the directly found contact-

forces between different metals and water we must add the

electromotive excitation of one metal on the other in order to get
the current-forming E.M.F. in the circuit of a closed cell.”

So far as I am able to see to the bottom of the foregoing
theory its foundation appears to be as follows :—Contact ex-
periments compel us to accept the summation law for the
Volta-effects, or total differences of potential, |

H=Cu/Agq + Aq/Zn + Zn /Cu.
The chemical theory constrains us to admit that the H.M.F. of
a cell is the equivalent of the chemical action going on, and

thus suggests that it equals the sum of some chemical contact
forces ; or, in Wiedemann’s notation, H ought also to equal
Cu/Aq+ Aq/Zn.
How are these two requirements to be reconciled ?
Assume that at every junction there is a total force made
up of two portions—a chemical force, such as Cu/ Ag, and
a physical force, which may be denoted by (Cu/Agq), and

write the whole E.M.F. of a cell equal to the sum of all these
chemical and physical junction-forces,

H=Cu/Aq+ (Cu / Aq) + Aq / 2n+(Aq / Zn) +(Zn /Cu).

We have then only to make the physical forces obey Volta’s
series law; so that

(Cu /Aq)+ (Aq / Zn) + (Zn / Cu) =0;
and we get the required relation
H=Cu/Aq+Aq/ Zn,

which harmonizes chemical and contact views.

Prof. Wiedemann’s reconciliation of contact and chemical
theories would thus seem to be somewhat of the same order as

the later and less complete one of Fleeming Jenkin, quoted
in my paper, § 6; but I cannot help feeling that the theory

Electromotive Forces in a Voltaic Cell. 377

of Prof. Wiedemann is rather fuller of hypothesis than most
other contributions to the subject with which I am acquainted ;
for it postulates

1. An unexplained difference of potential between metals in
contact.

. A similar difference of potential between the elements of
a compound.

. An unequal attraction of bodies for + and — EK.

. An unequal attraction of metals for chemical elements.

. A chemical contact-force between metals and liquids,
depending on No. 4, not obeying Volta’s law.

. A physical contact-force, depending on No. 3, obeying
Volta’s law.

. Proportionality of the chemical contact-forces and the
heats of combination.

. A mechanical falling together of atoms accounting for
heat-production at electrodes; and

. That electrical actions go on between the molecules of a
liquid, and between those molecules and metals im-
mersed in it, according to the ordinary laws of electro-
statics.

co oe) ~l or) Our Oo bo

Not all of the above hypotheses, however, are to be regarded
as independent : the dependence of Nos. 5 and 6 is stated, and
it is possible that some of the others are intended to be likewise
dependent, though I do not clearly see how.

: In some of these hypotheses, for instance Nos. 4 and 5,
: and, with considerable differences, Nos. 6 and 7, Prof. Wiede-
| mann’s theory bears a resemblance to the views set forth by
| the present writer in the paper to which this communication is
| a sequel.
| One obvious objection may be taken to the theory on a
| question of fact. According to it, zinc half immersed in water
is negative to copper half immersed in water (see diagram
! above): but experiment has failed to exhibit anything of the
kind ; and it is well known that Sir W. Thomson, Prof. Clifton,
and others have shown that zinc and copper immersed in water
or dilute acid are electroscopically at the same, or nearly the
same, potential.

Theory of Von Helmholtz, 1847 and later.

There is another theory which I did not indeed ignore, but
which I passed over with very insufficient mention, in the
paper referred to (see footnote to § 9, Phil. Mag. April, p. 258);
viz. the theory of Professor von Helmholtz, which he first stated
in his great memoir of 1847, Die Hrhaltung der Kraft, and

378 Prof. Oliver Lodge on the Seat of the

which he has improved and elaborated since (see Faraday
Lecture of 1881).

This fascinating theory has a small hypothetical basis, viz.
Nos. 3 and 9 of the above list; and on this foundation 1s erected
an explanation of the Volta effect, a theory of electrolysis, and a
great part of a theory of chemical combination.

Once grant, (1) that every substance has a specific attraction
for positive or for negative electricity which can be exerted
only through molecular distances, and (2) that the laws of
electrostatics may be applied to the charges of atoms in a
liquid ; and the ease with which a multitude of phenomena are
explained is surprising. I do not feel certain that a little too
much is not explained ; but we can postpone the consideration
of possible objections until the theory itself is briefly stated.

Metals, and “ electropositive’’ elements generally, attract
positive electricity or positively electrified atoms whenever
these come within range, 7. e. within some molecular distance
of the surface. Potassium and zinc attract strongly, copper
and gold feebly. Oxygen, chlorine, and other “electronega- |
tive ’’ elements, attract negative electricity in a similar manner.

Bring now two metals, say zine and copper, into contact,
and the superior attracting-power of zinc for positive elec-
tricity will at once charge it at the expense of the copper, so
that the zinc touching-surface becomes strongly positive, and
the copper surface strongly negative. Separate the metals,
and they mostly discharge into each other; but a feeble
charge remains if the surfaces in contact were large, and if
they have been neatly separated ; and this is the Volta effect.

Moreover since each metal has a definite attracting-power
depending on itself only, Volta’s series law is an obvious
necessity. | ;

If one of the substances had been an insulator much less
loss need take place while separating them, and accordingly
high charges are then found upon them even after separation
(“ frictional electricity ’’). |

Now consider similarly two atoms in contact, forming, say,
a water molecule. ‘The hydrogen will be positively charged,
the oxygen negatively. Jf some atoms should get knocked
asunder they retain their charges and therefore tend to unite
again, but, while separate, their mutual attractions and repul-
sions will preserve a uniformity of distribution throughout the
liquid—e. g. the repulsion of similar charges will prevent a
number of oxygen molecules crowding together ; and if any
cause should remove oxygen atoms from any part of the liquid,
the vacancy will be instantly supplied by similar atoms repelled
there.

Electromotive Forces in a Voltaic Cell. 379

The remarkable thing about the charges of atoms is, how-
ever, that they are all equal,and even for different atomsare still]
the same, or differ only according to simple multiples of some
one absolute quantity; so that one atom may have three times
as much negative electricity as hydrogen has positive, but not
any fractional number. Such an atom may therefore combine
with three of hydrogen atoms, and thus be called a triad.
Another may be a dyad, &c. A givenconstant charge belongs
to every unit of affinity which the atom possesses; and thus is
the fact of chemical equivalence stated, though not accounted
for. Faraday’s law compels us to believe that atomic charges
are thus multiples of one definite electric quantity, but why
they should be so we are wholly unable to say. It looks as if
electricity were atomic as well as matter; a sufficiently
startling idea, as Helmholtz says. |

Chemical affinity is thus due to the electrical attraction of
oppositely charged atoms ; not so much due to the attraction
of the atoms themselves. This latter kind of attraction Helm-
holtz would not indeed deny, and he considers it may account
for “ molecular ’’ modes of combination ; but he regards it as
much weaker than the electric forces, and as not effecting
definite chemical combination.

We have ideas now concerning the size of atoms, and can
calculate roughly what an atomic charge is. The charge of
each atom is but small, but the aggregate charge of an appre-
ciable number of themisenormous. It can be easily reckoned
that if you take the opposite electricities out of a milligramme
of water and give them to two spheres a mile apart, those
two spheres will attract each other with a force of ten tons!
Or, again, the electrical attraction of two atoms at any dis-
tance exceeds their gravitative attraction 71 thousand billion
times. Upon such charges as these even very feeble external
electrical influence may exert considerable force, and quite
overpower any other kind of chemical affinity.

Plunge, therefore, two platinum plates into the liquid and
keep them at some slightly different potential, say a volt or
two: their surfaces are oppositely charged, and even this
feeble charge may be sufficient to tear asunder the atoms of
molecules which come close up to it*. Ifitis not sufficient, it

* It is easy to calculate the surface-density, and the difference of po-
tential, on the above hypothesis, needed to tear asunder oppositely charged

atoms attracting each other across molecular distance x, Letit be o, and
let the charge of an atom be g; then
2
2ro0q= t.

Moreover we know that 19,520 electromagnetic units can decompose 18
grammes of water, which allows 536x383 x 10" electrostatic units of each

|

380 Prof. Oliver Lodge on the Seat of the

will at least hold a layer of oppositely charged atoms facing
toward it across molecular distance ; so that near each plate
we have an excessively thin condenser of enormous capacity ;
but in the mass of the liquid is no such strain, the slightest
Hi.M.F. being sufficient to redistribute the atoms here and to
produce perfect equilibrium. If the atoms are torn asunder
by the electrode and liberated, fresh ones are continually sup-
plied by the rest of the liquid, the state of strain being con-
stantly re-formed and instantly broken down.

Thus are the phenomena of polarization and electrolysis
explained. :

Now insert into the liquid a pair of plates of zine and copper
in contact. By the contact the zinc has become positive and
the copper negative, and accordingly the zine attracts the
oxygen (or SQ, radical) and combines with it: the copper
attracts the hydrogen, and does the best it can with it.

This crudely is Helmholtz’s view of a voltaic cell.

There is, however, one important detail not yet mentioned.
I have spoken as if the work done by an electrode were to tear
the atom from its combinations. This is not exactly Helm-
holtz’s view. He adduces arguments to show that the work
required to effect decomposition has for its object, not the
separation of the atom from the liquid, but the separation of the
atom from tts electric charge. It is to this that it clings, and
this that has to be taken from it. If it be made to give this up

kind of electricity per gramme; so if n® be the number of molecules in a
oramme of water, the charge of each dyad atom is

1:6 x 10"
g=-—— a3 &1:6x10~" probably.
Hence 1:6 x 1013
Bite or “peaeee, 3x10" probably ;

and, the difference of potential between either plate and liquid being given
by 4irxo, the whole difference of potential between the electrodes is

V=8rr0=6'4X 10” °

na?
or 1:92 volt.

In reckoning the approximate numerical values above, I have assumed
that m=10° and that nx=1. The resulting decomposition force comes
out, quite accidentally as far as I am concerned, very near that needed to
decompose acid-water. But I need hardly point out that order of mag-
nitude is all that is really calculated in these figures; and that not only
decimals but even numerals, prefixed to the power of 10, have very little
meaning. The value of « may perhaps vary somewhat for different sub-
stances; and as to the value of x, though that is definite enough, itis very
probable that it ought to be something more like 2 x 10° than what I have
above assumed it.

It is at any rate satisfactory that the electrostatic hypothesis should
give, as a polarization E.M.F., a number so entirely comparable with
actual results.

-0064 electrostatic units probably,

|
|
|
|

Electromotive Forces in a Voltaic Cell. 381

by a sufficient H.M.F., it is of small or no moment whether
the ion be really set free, or be dissolved, or otherwise kept in
the liquid. But when the electrode is of such a metal that an
ion can combine with it, that ion retains its charge, and
accordingly very little E.M.F’. is sufficient to decompose the
liquid under such circumstances.

Without this statement it might be objected, that when a
piece of zinc is plunged into acidulated water, since it has a
strong attraction for positive electricity, it ought to attract
the positively charged hydrogen atoms up to itself, instead of
oxygen as commonly supposed. I suppose Helmholtz would
admit that it must do this to begin with, when isolated from
other metals, on his theory, but that it has no way of sepa-
rating and liberating atoms because it cannot get rid of their
charges. If it be kept at a sufficiently negative potential
artificially it can indeed dispose of these charges, and it then
does attract and liberate hydrogen. But contact with copper
raises it to a positive potential, and it then attracts negatively
charged oxygen, rather than hydrogen, and combines with it ;
while the negatively electrified copper seizes hydrogen atoms,
tears their charges away from them, and sets them free.

I must confess, however, that I feel a difficulty here. The
natural tendency of zinc is, by hypothesis, to attract positive
and repel negative atoms; but, while a very feeble positive
electrification applied to it is sufficient to reverse this tendency,
a comparatively strong negative electrification is needed to
enable it to exert that force which by the hypothesis is sup-
posed to be natural to it. |

In considering a Daniell cell, zinc /ZnSO, / CuSO, / copper,
Helmholtz obtains the energy producing the current, by point-
ing out that electricity is removed from copper, which only
attracts it feebly, and given to zinc which attracts it strongly.
Zine goes into solution and becomes positively charged, while
an equivalent of equally positive copper comes out ; and since
this results in a gain of energy, it follows that electricity must
do more work in going to zine, than in going to copper.

But if zine attracts electricity so much more strongly than
copper, why is it so easy to drive electricity across a copper/zine
junction? And why does no energy manifestation result at
the junction from such an operation ?

The answer probably is, because of the large charge already
existing at the junction ; the zinc has pulled as much positive

electricity out of the copper as it wants, and there exists at

the contact an electrical double layer, whose existence makes
it quite easy for extraneous electricity to flow either way
across the junction.

But, then, if zine pulls so much electricity out of a piece of

382 Prof. Oliver Lodge on the Seat of the

copper with which it is put into contact, where does it all
come from? And why does it not leave the copper strongly
negative instead of only feebly so, as evidenced by Sir W.
Thomson’s electrified-needle form of experiment? Why, again,
- does this method of observing the Volta-effect, with the plates
permanently in contact, agree so well with the old method of
observing it by separation of the plates, if a great but uncer-
tain part of the charge leak back as soon as separation beyond
molecular distance is attempted ?

These and similar difficulties occur to me in connection with
this most interesting theory ; and I state them, not with any
idea that they are final and unanswerable, but because it is
always serviceable to point out the apparent defects of a theory
as well as its merits.

It will be seen that, whereas Helmholtz’s theory starts with
a hypothetical differential attraction of matter for electricity,
and explains the chemical attraction of matter for matter on
the strength of it, my view starts with a differential chemical
attraction of matter for matter (4. e. zinc attracting oxygen
more than copper attracts it), and on this basis works out the
theory of the voltaic cell and of the Volta-effect.

In electrolysis proper both views may be at one, because
here we are concerned with electrical attractions only ; but
Prof. Helmholtz’s has the advantage, because it explains the
fact of atomic charges, which mine has to assume. My view
of electrolysis has hitherto been much less electrostatic than
Prof. Helmholtz’s, but also it has been much more vague. A
great deal is to be said for the assumption of ordinary electro-
static attraction over molecular distances, and by adopting
such more definite ideas I am in hopes of improving my theory
further. And though, after all, none of these views can be
really absolute, because they begin and end with action at a
distance, still electrostatic attraction is a fact, and we know
that it can be accounted for by a strain in a continuous
medium. Hence phenomena reduced to electrostatics may be
held to be provisionally ‘ explained.”’ |

But beyond electrostatic attraction some other primary
force is necessary ; either chemical affinity, the attraction of
atom for atom, or electrical affinity, so to speak, the attraction
of atoms for electricity.

Chemical affinity is a fact, however it be accounted for ;
but, to enable it to explain everything, the fact of atomic
charge must be permitted to likewise go without explanation.
These are the necessary data for my view.

‘“ Hlectrical afinity’’ is not known to be a fact; but if it be

ranted, everything is explained (barring a few outstanding
difficulties), even the fact of atomic charge itself, though not

Electromotive Forces in a Voltaic Cell. 383

its numerical exactitude. This is the datum required for
Helmholtz’s theory.

It may thus be said that while the writer’s view rests on
two admitted facts which it does not attempt to explain, but,
assuming them, explains other things from them ; Helmholtz’s
theory rests on one unverified hypothesis, which, being granted,
everything else follows ; except indeed the numerical equality
of atomic charges, which has to be regarded as experimentally

demonstrated in both theories.

According to my view, the zinc of a cell pulls up the nega-
tively charged oxygen atoms (or SQ, radicals, whatever the
real ions may be) to itself, not because they are electrified, but
because they are oxygen. The electrical current produced
is a secondary result caused by the chemical action.

According to Helmholtz, the zinc pulls the oxygen atoms
because it is itself electrified by contact with copper, and
because they are oppositely electrified by contact with hy-
drogen. The chemical action resulting is a secondary result
caused by the electric forces.

Helmholtz’s is thus a true “ contact theory,” and is in many
respects like Sir W. Thomson’s. Sir William postulates an
attraction of zinc for copper, and from this explains the Volta-
effect and the production of a current. [am unable to picture
to myself exactly how the attraction of zinc for copper results
in a difference of potential when they are put into contact ;
but undoubtedly such a force, if granted, would put a supply
of energy at disposal which could account for the Volta-effect.
And a difference of potential thus set up may result in an
electrical decomposition of water and maintenance of current,
just as in Helmholtz’s theory.

I must confess that I am unable to feel quite comfortable
about energy considerations with either of these theories of
the voltaic cell; because it seems as if the metallic junction
were, after all, driving the current, whatever be said about
chemical energy. But in this I must certainly be wrong ;
that is to say, there must be some thorough way of reconciling
these, by such men advocated, views with energy considerations,
though I am unable satisfactorily to perceive it.

Both these contact-theories, in explaining the Volta-effect,
ignore the existence of the oxidizing medium surrounding the
metals. My view explains the whole effect as a result of this
oxygen bath, and of the chemical strain by it set up. The
other theory which likewise took account of the atmosphere
is the old chemical one; but this leaves the aspect of the
matter exceedingly vague. The strictures applied to it by
Prof. Helmholtz (Faraday Lecture, 1881) are well deserved:—

“The so-called chemical theory of Volta’s fundamental

384 Intelligence and Miscellaneous Articles.

experiment was rather indefinite; it scarcely did more than
tell us—here is the possibility of a chemical process, here
electricity can be produced. But which kind, how much, to
what potential, remained indefinite. I have not found in all
the papers which have been written for the defence of the
chemical theory a clear explanation why zinc opposed to cop-
per in liquids, where zinc really is oxidized and dissolved,
becomes negative, and why in air and other gases it becomes
positive, if the same cause (viz. oxidation) is at work.”

To the chemical strain theory, held by me, none of this
remark is in the least applicable. This theory does not say
that zinc opposed to copper in a liquid is negative. It says
that, changing the surrounding medium from air to acid-water,
makes scarcely any difference ; and that accordingly connect-
ing zinc and copper by a drop of water leaves them with the
same respective potentials as they had before. ‘The strain of
oxygen-atoms towards a metal exists whether in air or water,
but it can produce no effect in either case until at some point
the oxygen be swept away by contact with another metal.
Then instantly the chemical forces are able to do work, and to
produce, if in air or other dielectric, the Volta-effect ; if in acid
or other electrolyte, the voltaic current. The metallic con-
tact equalizes the potential of the two metals in the one case;
in the other it perpetually fails to equalize their potential.
The equalization of an otherwise disturbed potential is its only
effect.

And no indefiniteness exists as to either the kind or amount
of the Volta-effect as so produced, for I have shown fully: that
a Volta-effect can be calculated in absolute measure for any
pair of clean metals immersed in any medium, from purely
thermo-chemical data ; but whether this effect is the real one
or not is at present a matter of opinion.

These so-calculated effects undoubtedly agree in some cases
with experimentally observed ones; but whether they all so
agree, and, if not, the extent to which they are erroneous, are
matters mainly for future experiment to decide; and upon
such agreement or disagreement between calculation and
experiment my theory definitely stands or falls.

XLVI. Lntelligence and Miscellaneous Articles.

ON A DIFFERENTIAL RESISTANCE THERMOMETER.
BY T. C. MENDENHALL.
fs basaton determination or registration of the temperature at a distant

or not easily accessible point is so extremely desirable that
many methods for accomplishing this end have been proposed, and

Intelligence and Miscellaneous Articles. 885

to some extent made use of during the past fifty years. Naturally
enough electricity has been utilized in some way or other in the
majority of these systems of telethermometry.

The requirements of the problem seem to be that the device or
instrument used at the point, the temperature of which is to be
ascertained, shall be of the greatest possible simplicity of construc-
tion, involving little or no motion in its parts, so that the liability
to “ get out of order” shall be reduced to the minimum ; and that
at the observing or registering station, the necessary appliances
‘shall possess a maximum of durability and simplicity—so that a
minimum of time and skill will be demanded in making the obser-
vations. The whole system must be certain in its indications
and correct within a reasonable limit.

The first of these conditions is apparently sufficiently well satisfied
by the thermo-electric-junction, which has probably been more ex-
tensively made use of than any other form of electric thermometer.
It renders necessary, however, the use of a comparatively delicate
galvanometer, and as the electromotive force of a single couple is
small (it is difficult to use more than one in general practice) the
results are subject to considerable errors arising from unknown or
neglected sources of electromotive force. This source of error
becomes more important as the range of temperature measured
becomes smaller, although it may be almost entirely avoided by
care and skill on the part of the operator. The well-known
resistance method of Siemens satisfies the same condition very per-
fectly, and is certainly capable of giving good results when skilfully
applied, at least throughout moderate ranges.

The desire to possess some form of electric thermometer which
might be utilized in the study of certain problems connected with
meteorology, especially the observation of soil and earth tempera-
ture, and the use of which would not demand greater skill than
that of the ordinary meteorological observer, led to the device and
construction of the instrument to be described, which may be called
a * differential resistance thermometer.” It consists essentially of
a mercurial thermometer, not unlike ordinary forms, except that
the bulb is greatly enlarged so that the stem may have a diameter
of something like a millimetre, and still leave the scale tolerably
“open.” In one of the instruments already made 1° C. corresponds
to about 5 millim. of the scale. Running down through the stem
is a fine platinum wire about ‘08 millim. in diameter. The lower
end may be secured in the bulb so that it is kept straight in the
bore of the stem, and at the lower end a heavier wire is sealed in
the glass, so that metallic contact can be made with this wire both
at the upper end and through the mercury at the lower. It is
evident that the resistance between these two points will depend
largely (but not entirely) on the length of the platinum wire which
is above ihe mercury in the tube, and this will depend on the
temperature to which it is exposed. When this temperature rises
the resistance is decreased by an amount equal to the difference
between that of the platinum wire which disappears and that of

Phil, Mag.S. 5. Vol. 20. No. 125. Oct. 1885. 2 i

386 Intelligence and Miscellaneous Articles.

the mercury which takes its place—less the increase in the resist-
ance of the wire and mercury due to increase of temperature.

Let /=length of platinum wire exposed at 0°.
s=resistance per unit length (=length of 1°) of wire.
g=resistance per unit length (=length of 1°) of mercury.
k=temperature coefficient of platinum.
h=temperature coefficient of mercury.

kh, =total resistance at 0°.
R,=resistance (including all) at 7°.

Then
R,=R,.— {s(1 —kl)—ght—(ks—gh)®. . . . CG)

This equation is not quite rigorous, but the approximation is very
close. It is of the form

R,=2,—-Bi—-Ce. . . 2

The simplest and best way of dealing with it is to determine the
constants of the equation (2) by a series of observations making
use of the method of least squares. From the result the resistance
for any degree of temperature may be calculated, or, better, a curve
ean be constructed from which the temperature corresponding to
any resistance can easily be read.

The advantage of this method over the use of a simple resistance
coil is that the change in resistance accompanying a given change
in temperature is much greater, and in fact it may be made as
great as one desires. As a result the telephone may be substituted
for the galvanometer im the resistance measurements, thus greatly
simplifying the apparatus as well as increasing the rapidity with
which observations may be made; or, if preferred, a much less
sensitive galvanometer may be used. It also possesses the very
great advantage of allowing an increase of delicacy as the range of
temperature decreases. For earth temperatures this is very de-
sirable, and it will easily be seen that thermometers of this kind
can be constructed for a few degrees of range with which, by com-
paratively rude processes, the measurements may be correct within
a very small fraction of a degree. For use in earth temperature
measurements the thermometer will be enclosed in a strong brass
tube for protection, and the connection with the point. of observa-
tion made by means of a cable of heavy copper wire. ‘The cable
will of course form a part of R, in equation (2); but as it is a con-
stant, the substitution of one cable for another, if necessary, will
affect the position and not the form of the calibration curve. Its
resistance must be small, relatively, and the influence of tempera-
ture upon it may be neglected.

With this device a temperature observation may be taken in less
than a minute, no time being consumed in the preparation of liquids
of known temperature at the observing station, as in the use of the
thermo-junction or the resistance coil.—Silliman’s American
Journal, August 1885.

—s —

a

Intelligence and Miscellaneous Articles. 387

ON THE DETERMINATION OF THE COEFFICIENTS OF INTERNAL
FRICTION. BY WALTER KONIG.
The author determined the coefficients of a number of liquids by
the method of vibrating disks, and also by that of flow from capillary
tubes, and obtained numbers which agree with those obtained by

- other inquirers.

He also examined whether electrification and exposure in a
magnetic field had any influence on this coefficient, and obtained a
negative result—Wiedemann’s Anna/en, No. 8, 1885.

NOTE ON THE TRANSMISSION OF LIGHT BY WIRE-GAUZE
SCREENS. BY 8S. P. LANGLEY.

In the beginning of the present year a friend sent me a series of
wire-gauze screens, which he used to diminish the apparent bright-
ness of stars in making meridian observations, with a request that
I would determine photometrically the amount of light transmitted
by them. As such screens are occasionally employed in astronomical
work, particularly in the use of the heliometer, I have thought the fol-
lowing account of our experience of sufficient interest to make public.

I used for the measurements a photometer-box originally con-
structed for another purpose, and an opaque wheel or disk having
radial slits of variable width, which, placed in the path of a ray of
light and rotated with sufficient velocity, can be made to reduce the
light to any desired fraction of its original intensity. (This I have
employed for some years for photometric measurements when it is
desirable to avoid the use of polarizing-apparatus.) In the centre
of the photometer-box was a sliding Bunsen disk, which could be
viewed from above by a suitable arrangement of mirrors. The
open ends of the box were directed to two opposite windows, and the
disk placed in such a position that its sides were equally illuminated.
The wire screen was then placed over one end of the box, the wheel-
photometer in front of the other end, and the apertures of the
latter altered until the equality of Ulumination of the Bunsen disk
was restored. The screen then cut off the same amount of light as
the wheel. From several series of measurements made in this way,
it was found that

1 screen transmitted .......... *395 +004 of the incident light.
2 screens superposed transmitted -144+-004 ws be
= iy as sles 052 $-008

These numbers, as was to be expected, are nearly in geometrical
progression. The screens were returned to the sender and the
results communicated to him; but he wrote that, upon trial, he
found the reduction of ight very much greater than the above
values, three superposed screens reducing the light of a star by 7-1
magnitudes, which corresponds to a transmission of only :0014.

I was at that time absent, and my assistant, Mr. J. E. Keeler,
undertook the investigation of the cause of the discrepancy, which
he attributed to loss of light by diffraction under the circumstances
in which the screens were used by their owner, 2. ¢. in front of the
object-glass of a telescope directed upon a star. With diffuse

388 Intelligence and Miscellaneous Articles.

light, such as was used in the measurements with the photometer-
box, no loss due to this cause was possible.

In the apparatus devised by Mr. Keeler for an experimental
determination of the loss by diffraction, the star was replaced by an
illuminated pinhole in the focus of a 3-inch collimating-telescope.
This was viewed by an observing-telescope of nearly equal size, in
the eye-tube of which was an unsilvered plane-glass mirror, which
reflected into the eye-piece a comparison star—the image of an
illuminated pinhole produced by a coliimating-telescope at right
angles to the other two. In the path of the rays from this tele-
scope could be interposed the wheel-photometer. The light before
entering the first pinhole suffered reflection from an unsilvered
glass surface, in order to reduce its intensity to that of the com-
parison star.

The two images in the field of view having been adjusted to_

equality, the wire-gauze screen was interposed between the object-
glasses of the collimating. and observing-telescopes, reducing the
light of the star and producing around it the well-known diffraction-
image of a network. ‘The wheel-photometer was then introduced,
and the intensity of the comparison star reduced until it was equal
to the central image of the other. By enlarging the pinholes until
the superposition of the colours produced white light, the intensity
of the diffraction-images could also be estimated.

It was thus found that the central image had only :175 of its
original brightness, which would therefore be the proportion trans-
mitted by the screen under these conditions, and that the brightness
of each of the four first spectra was ‘05 of that originally possessed
by the central image. Two thicknesses of the wire-gauze transmitted
barely ‘02, as measured by the intensity of the central image.

The screens with which these experiments were made were much
coarser than the original ones, and it was expected that the effect
of diffraction would be less pronounced. The transmission of one
thickness, measured by the photometer-box, was ‘47; of two
thicknesses. ‘21.

Finally, the apertures of the screen and the diameter of the wire
were measured by a micrometer-microscope and the apertures found
to occupy °465 of the total area of the screen.

It was concluded, therefore, as the result of the experiments :—

1. That the transmission, as measured by the photometer-box,
was equal to the ratio of the sum of the areas of the apertures of
the screen to its total area, and therefore could be considered to be
the true transmission of the screen; and

2. That the much smaller transmission of the screen, when used
in front of the object-glass of a telescope to diminish the apparent
brightness of a star, is satisfactorily accounted for by the loss of
light caused by diffraction under these circumstances.

"3. That screens used for this purpose should have their constants
determined by special experiments of the nature of those just de-
tailed, and that their photometric use should then be limited to the
reduction of the light of bodies possessing a small angular magnitude.
—Silliman’s American Journal, September 1885.

A ee a

THE

LONDON, EDINBURGH, ann DUBLIN

PHILOSOPHICAL MAGAZINE

AND

JOURNAL OF SCIENCE.

[FIFTH SERIES.]
NOVEMBER 1885.

XLVII. The Luminiferous Atther. By De Vouson Woon,
C.E., M.A., Professor of Mechanical Engineering in Stevens
Institute of Technology, Hoboken, N.J.*

eee properties of the luminiferous ether appear to be

known and measurable with a high degree of accuracy.
One is its ability to transmit light at the rate of 186,300 miles
per secondy, and the other its ability to transmit from the
sun to the earth a definite amount of heat-energy.

In regard to the latter, Herschel found, from a series of
experiments, that the direct heat of the sun, received ona
body at the earth capable of absorbing and retaining it, is
competent to melt an inch in thickness of ice every two hours
and thirteen minutes{. ‘This is equivalent to nearly 71 foot-
pounds of energy per second.

In 1838 M. Pouillet found that the heat-energy transmitted
from the sun to the earth would, if none were absorbed by

our atmosphere, raise 1:76 gramme of water 1° C. in one

minute on each square centimetre of the earth normally

exposed to the rays of the sun§.

This is equivalent to 83°5 foot-pounds of energy per second,

* Communicated by the Author.
+ Professor Michelson found the velocity of light to be 289,740 metres

per second in air, and 299,828 metres in a vacuum, giving an index of

ae of 1,000,265. Journ. of Arts and Science, 1879, vol. xviii.
. 890,
{ ‘Familiar Lectures on Scientific Subjects,’ by Sir John Herschel,

. 6d.
: § Comptes Rendus, 1838, tom. vii. pp. 24-26.
Phil. Mag. 8. 5. Vol. 20. No. 126. Nov. 1885. a

390 Prof. De Volson Wood on

and is the value used by Sir William Thomson in determining
the probable density of the zether*. Later determinations of
the value of the solar constant by MM. Soret, Crova, and
Violle have made it as high as 2°2 to 2°5 calories. But the
most recent, as well as the most reliable, determination is by
Professor 8. P. Langley, who brought to his service the most
refined apparatus yet used for this purpose, and secured his
data under favourable conditions; from which the value is
found to be 2°8+ caloriest with some uncertainty still re-
maining in regard to the first figure of the decimal. We will
consider it as exactly 2°8 in this analysis, according to which,
there being 7000 grains in a pound and 15°432 grains in a
gramme, we have for the equivalent energy

2°8 x 15-482 9 772 x 144

7000 5 0155 x 60

per second for each square foot of surface normally exposed

to the sun’s rays, which value we will use. Beyond these

facts, no progress can be made without an assumption. Com-

putations have been made of the density, and also of the

elasticity, of the ether founded on the most arbitrary, and in

some cases the most extravagant, hypotheses. ‘Thus, Herschel
estimated the stress (elasticity) to exceed

17 x 10°=(17,000,000,000) pounds per square inch};

and this high authority has doubtless caused it to be widely
accepted as approximately correct. But his analysis was
founded upon the assumption that the density of the ether
was the same as that of air at sea-level, which is not only
arbitrary, but so contrary to what we should expect from
its non-resisting qualities, as to leave his conclusion of no
value. That author also erred in assuming that the tensions
of gases were as the wave-velocities in each, instead of the
mean square of the velocity of the molecules of a self-agitated
gas; but this is unimportant, as it happens to be a matter of
quality rather than of quantity. Herschel adds, “ Consi-
dered according to any hypothesis, it is impossible to escape
the conclusion that the ether is under great stress.””> We
hope to show that this conclusion is not warranted; that a
great stress necessitates a great density; but that both may
be exceedingly small. A great density of the ether not only
presents great physical difficulties, but, as we hope to show,

=133 foot-pounds

* Trans. Roy. Soc. Edinburgh, vol. xxi. part 1.

+ Am. Journ. of Arts and Science, March 1888, p.195. Also Comptes
Rendus.

¢ ‘Familiar Lectures,’ p. 282.

the Luminiferous ther. 391

is inconsistent with the uniform elasticity and density of the
zether which it is believed to possess; and every consideration
would lead one to accept the lowest density consistent with
those qualities which would enable it to perform functions
producing known results.

Ina work on the ‘ Physics of Atther,’ by 8. Tolver Preston,
it is estimated that the probable inferior limit of the tension
of the zether is 500 tons per square inch, a very small value

- compared with that of Herschel’s. But the hypothesis

upon which this author founded his analysis was—The tension
of the zther exceeds the force necessary to separate the atoms
of oxygen and hydrogen in a molecule of water; as if the
atoms were forced together by the pressure of the ether, as

_ two Magdeburg hemispheres are forced together by the external

air when there is a vacuum between them. ‘This assumption
is also gratuitous, and is rejected for want of a rational
foundation.

Young remarks:—“ The luminiferous zther pervading all
space is not only highly elastic, but absolutely solid.”* We
are not certain in what sense this author considered it as solid ;
but if it be in the sense that the particles retain their relative
positions, and do not perform excursions as they do in liquids,
it is a mere hypothesis, which may or may not have a real
existence. If it be in the sense that the particles suffer less
resistance to a transverse than to a longitudinal movement,
there are some grounds for the statement, as shown in
circularly-polarized light. Bars of solids are more easily
twisted than elongated, and generally the shearing-resistance
is less than fora direct stress. It certainly cannot be claimed
that the compressibility of the zther (in case we could capture
a quantity of it) is less than that of solids.

Sir William Thomson, in making a probable estimate of the
density of the «ther, made a more plausible hypothesis, by
assuming that “the maximum displacement of the molecules
of the xther in the transmission of heat-energy was =, of a
wave-length of light, the average of which may be taken as
sam Of an inch.” Hence the displacement was assumed to

be 5500000 of an inch; by means of which he found the weight

of a cubic foot to be 3x 10-” of a poundf. We also notice
that one Belli estimated the density of the ether to be $ x 10-*
of a pound}; but M. Herwitz, assuming this value to be too
small and Thomson’s as too large, arbitrarily assumed it as

* Young’s Works, vol. i. p. 415.

+ Phil. Mag. 1855 | 4] ix. p. 39.

t Cf. Fortschritte der Physik, 1859.

22

392 Prof. De Volson Wood on

10-18 of a pound per cubic foot; but arbitrary values are of
small account unless checked by actual results.

We propose to treat the ether as if it conformed to the
Kinetic Theory of Gases, and determine its several properties —
on the conditions that it shall transmit a wave with the velo-
city of 186,300 miles per second, and also transmit 133 foot-
pounds of energy per second per square foot. Thisis equivalent
to considering it as gaseous in its nature, and at once compels
us to consider it as molecular; and, indeed, it is difficult to
conceive of a medium transmitting light and energy without
being molecular. The Hlectromagnetic Theory of Light sug-
gested by Maxwell, as well as the views of Newton, Thomson,
Herschel, Preston, and others, are all in keeping with the
molecular hypothesis. If the properties which we find by
this analysis are not those of the ether, we shall at least
have determined the properties of a substance which might
be substituted for the wether, and secure the two results
already named. It may be asked, Can the Kinetic theory,
which is applicable to gases in which waves are propagated
by a to-and-fro motion of the particles, be applicable to a
medium in which the particles have a transverse move-
ment, whether rectilinear, circular, elliptical, or irregular?
In favour of such an application, it may be stated that
the general formulz of analysis by which wave-motion in
general, and refraction, reflection, and polarization in parti-
cular, are discussed, are fundamentally the same; and in the
establishment of the equations the only hypothesis in regard
to the path of a particle is—It will move along the path of
least resistance. The expression V*oc e+6 is generally true
for all elastic media, regardless of the path of the individual
molecules. Indeed, granting the molecular constitution of
the ether, is it not probable that the Kinetic theory applies
more rigidly to the zther than to the most perfect of the
known gases ?* |

The 133 foot-pounds of energy per second is the solar-heat
energy in a prism whose base is 1 square foot and altitude
186,300 miles, the distance passed over by a ray in one
second; hence the energy in 1 cubic foot will be

| aes TNS ON
186,300 x 5280 3x 10!

Where results are given in tenth-units of high order, as
in the last expression, 1t seems an unnecessary refinement to

foot-pounds. . . (1)

* See also remarks by G. J. Stoney, Phil. Mag. 1868 [4] xxxvi.
pp. 182, 133.

the Luminiferous Uther. . 393.

retain more than two or three figures to the left hand of the
tens; and we will write such expressions as if they were the
exact results of the computations.

If V be the velocity of a wave in an elastic medium whose
coefficient of elasticity, or, in other words, its tension, is e
and density 5, both for the same unit, we have the well-known

relation ye
de

V= \/ cat

7 dé

e= 07,

where y=1-4; and the differential of the latter substituted in
the former gives

And for gases we have

V = ve, CNB cis Sas oh domi
Z (2)
The tension of a gas varies directly as the kinetic energy of
its molecules per unit of volume. Ifv* be the mean square
of the molecules of a self-agitated gas, we have

ecdu, orwv=a-, ..... (3)
where x is a factor to be determined. Hquations (2) and (3)
give

PS Wels anil Yactiaon nes: wee
: &

Assuming, with Clausius, that the heat-energy of a molecule
due to the action of its constituent atoms, whether of rotation
or otherwise, is a multiple of its energy of translation, we
have for the energy in a unit of volume producing heat,

1.82
aydv ’

where y is a factor to be determined. Ife be the specific
heat of a gas, w its weight per cubic foot at the place where
g=32°2, J Joule’s mechanical equivalent, r its absolute tem-
perature ; then the essential energy of a cubic foot of the
medium will be cwrJ; and observing that w=g6d, we have

eGe — Goa A pees a NSO) ER
which, reduced by (4), gives

ZcegyTJd :
ay = Beko N\- aits2,

the second member of which is constant for a given gas. To

394 Prof. De Volson Wood on

find its value, we have
Hydrogen. Air, Oxygen.
Specific heat*. . . . 84098 0°2375 0:2175
Velocity of sound, feet per
second at T= 4932 ; } £168 1090 on

and g=82'2, y=1'4,J=772. These, substituted in the second
member of (6), give

wy forhydrogen . . « . 6599

yo. UE One woe 2 eee
sg. ORY SERS aah SS eee
3)19°901

Mean a... « t 5) ie

This value, which is nearly constant for the more perfect
gases, we propose to call the modulus of the gas, and represent
it by «; and for the purposes of this paper we will use

pw=6'6,

This relation of the product xy being a constant, has, so far
as we are informed, been overlooked by physicists, and is
worthy of special notice, since it determines the value of one
of the factors when the other has been found. Krénig,
Clausiust, and Maxwell give for # the constant number 3,
but variable values for y f.

Weare confident that the value of # is not strictly constant;
or if it is, it exceeds 3, since the effect of the viscosity of a
gas would necessitate a larger velocity to produce a given
tension than if it were perfectly free from internal friction,
For our purpose it will be unnecessary to find the separate
values of « and y; but if we have occasion to use the former
in making general illustrations, we will call it 3, as others
have done heretofore. Ifthe correct value of a exceeds 3, it
will follow that the velocity of the molecules exceeds the values
heretofore computed§. According to Thomson, Stokes showed
that in the case of circularly polarized light the energy was

* Stewart on ‘ Heat,’ p. 229.

+ Phil, Mag. 1857 [4] xiv. p. 128.

t ‘Theory of Heat,’ pp. 314 and 317. Maxwell states that the value
for y is probably equal to 1°634 for air and several of the perfect gases.
This would make v=4 nearly.

§ Maxwell gives for the mean square of the velocities (or, in other
words, the velocity whose square is the mean of the squares of the actual
velocities) of the molecules, in feet per second at 493°-2 F. above absolute
zero, hydrogen 6232, oxygen 1572, carbonic oxide 1672, carbonic acid
1570. Phil. Mag. 1873, p. 68. Our equation (4) gives for air 15938,

the Luminiferous AEther. 3995

half potential and half kinetic*; in which case y=2, and
therefore c=3'3.
The energy in a cubic foot of the xther at the earth being
given by (1) and (5), we have, by the aid of (4),
8 4

Se ig ie ke

ByOv Tat 3x10°° (8)
ds 4x14x2 Be tp8
~ 3x10" x 6°6 x (186,300 x 5280)? 35 x 104

which is the mass of a cubic foot of the ether at the earth, and
‘ which would weigh at the place where g=32°2 about

peaupornd, iis ar 4. C10)

8 lb.,. (9)

2
e107
compared with which Thomson’s value is less than 4000 times
this value. Thomson remarked that the density could hardly
be 100,000 times as small—a limit so generous as to include
far within it the value given in (9). According to equation
(10), a quantity of the ether whose volume equals that of the
earth would weigh about 5, of apound. If a particle describes
the circumference of a circle in the same time that a ray
passes over a wave-length A, the radius of the circle will be,
using equation (4),

Seo Pig dca
"Or ev: =
or the displacement from its normal position will be about
+; of a wave-length, or about ais of an inch at the earth.
HKliminating V between (2) and (8) gives
8 4
o= Bu x 10° = i108 eee eee ce) re a fe (1 1)

A,

for the tension of the sther per square foot at the earth,
and is equivalent to about 1*1 pound on a square mile.
The tension of the atmosphere at sea-level is more than
30,000,000,000 times this value. It somewhat exceeds the
tension of the most perfect vacuum yet produced by arti-
ficial means, so far as we are informed. Crookes produced
a vacuum of ‘02 millionth of an atmospheref without reaching ~

* Phil. Mag. 1855 [4] ix. p. 37.

+ “On the Viscosity of Gases at High Exhaustions,’ by William
Crookes, F.R.S., Phil. Trans. Roy. Soc. part ii. (1881) p. 400: ‘“ Going up
to an exhaustion of ‘02 millionth of an atmosphere, the highest point to
which I have carried the measurements, although by no means the highest
exhaustion of which the pump is capable.”’

396 | Prof. De Volson Wood on
the limit of the capacity of the pumps ; and Professor Rood

1 e e
produced one of 55555, of an atmosphere* without passing

the limit of action of his apparatus. The latter gives a

147x144
390,000 GODs, 7. 20-080
This, in round numbers, is 140 times the value given in equa-
tion (11). Even at this great rarity of the atmosphere, the
quantity of matter in a cubic foot of the air would be some
200 million million times the quantity in a cubic foot of the
eether—such is the exceeding levity of the ether.

Admitting that the ether is subject to attraction according
to the Newtonian law, and of compression according to the
law of Mariotte, we propose to find the relation between the
density of the ether at the surface of an attracting sphere and
that at any other point in space, providing that the sphere be
cold and the only attracting body, and the gas considered the
only one involved.

Let 59, &, Wo be respectively the density, elasticity, and
weight of a unit of the medium, whether ether, air, or any
other gas, at the surface of the sphere; 6, e, w, the corresponding
quantities at a distance z from the surface of the sphere; 7 the
radius of the sphere, go the acceleration due to gravity at its
surface, and g that at distance r+z from the centre of the
sphere. Then

pressure per square foot of of a pound.

04. ee eae

do €9 g Jo
and

ye
ITN p zy?
2
= 20 Soy te)

Wo g Wo rv

But

de=—wdz=—godz; . . . . (18)
de _ Joe yr

e €&y (r+zy
Integrating between e and @, r+z and r, we have

dz.

90% 72.

e=eye & rte. lw,
9080, _7%

d= Ove €9 r+zZ, ° ° e e e (15)

Neglecting the attraction of the earth for the ether, and

* Journ. of Arts and Science, 1881, vol. xxii. p. 90.

the Luminiferous Ather. 397

considering the sun as the only attracting body, we have gp
at the sun 28°6 x 32°2, and at the earth z=2107, r=441,000
miles, the sun’s radius; 8=<x 10-4, equation (9), and
e= =, x10-®; and these, in (14) and (15), give

28°6 X 32°22 33105. 210
gee aa Ty ee X98

1

Se sega kl 2s ea nd Aaa OS wa)
and Lime has
ee i on ys cite ie et at CLO)

for the tension and density of the ether at the surface of the
sun under the conditions imposed. But the millionth root of
e is practically unity; hence the elasticity and density at the
sun is practically the same as at the earth.

Now, starting at the sun with this result, and finding the
density at a distance z from it, then making z infinite, we shall
get the 995,000th root of e, the value of which is also sensibly
equal to unity; hence the density at infinity would be sensibly
the same as at the surface of the sun, the difference in the
_ densities at the sun and at infinity being less than wa part
of that at the sun. In order to make the density vary sen-
sibly with the distance, the attraction of the central body must
be something like a million times as great as that of the sun,
or have a diameter a million times as large; but there being
_ no such known body, therefore the density and tension of the
_ ether may be considered uniform throughout space. Such has
been our conception of it; and it is an agreeable surprise to
find it so fully confirmed by analysis.

If the density were uniform, the weight of a given volume
of it would vary as the force of gravity. At the surface of the
sun a cubic foot would weigh [equation (10) multiplied by
28°6, or] 57 x 10-*4; hence, for a height A it would weigh

EVES Reo ye DT eh 17

im |. (rtz)? “~ 10% ° r+h? ° (17)
which for h=« becomes Be of a pound, which is the pressure
upon a square foot of the sun of a column of infinite height
under the conditions imposed. This would compress the first
foot of the column about TT of its length, and would cause
a corresponding increase in the density, the value of which,
after this compression, will be found by multiplying the value
given in equation (9) by nana which will leave the result
sensibly the same as before. Hence, from this standpoint, we

398 Prof. De Volson Wood on

again conclude that the density of the ether may be con-
sidered as sensibly uniform throughout space, providing its
temperature be essentially uniform.

If we assume that the law of the resistance by which the
zether opposes the motion of a body varies as the square of the
velocity of the body, we are still unable to assign the coefii-
cient which will give the numerical value; but it is safe to
assume that the entire mass of the ether occupying the path
of a body moving through it, will not have a velocity imparted
to it exceeding that of the body; but to be on the safe side,
we wili assume that it imparts a velocity equal to itself. The
energy thus imparted will be lost to the body. To simplify
the case, consider a planet moving in a circular orbit: r the
radius of the planet, d its distance from the sun, D its specific
gravity compared with water as unity, v, the velocity in its
orbit; then the mass of ether occupying the place of the
planet during one revolution about the sun will be, using
equation (9),

2
35 x 1024 wr xX 20 d,
which, multiplied by 3v?, will give the energy imparted to it.
The kinetic energy of a planet, neglecting its rotation, will

be

tare? x 624D x 2

g7r X0L5 UX 2g
Dividing the former, after multiplying it by }v?, by the latter
gives

1 d

7x 10%" rD *
for the fraction of the energy lost during one revolution about
the sun. Applying this to the earth, we have

d~—rD=93,000,000 +3912 x 54 = 43000,

and (18) becomes

(18)

6 |

[ow nearly, oie 90 ae
for the fraction of the energy lost in one year; and hence
at this rate would require more than 1,666,0@@. trillion

(1,666,000,000,000,000,000,000) years to bring it to rest.
Hquation (18) is not applicable to the resistance offered to
a comet, on account of the elongated orbit of the latter ; but
some idea of the effect of the resistance of the ether to the
movement of a comet may be found by considering what

2 ——

it a. nn

Se

the Luminiferous ther. 399

it would be if the orbit were circular, having for its radius
the perihelion distance. According to Professor Morrison,
the perihelion distance of the great comet (6), 1882*, was
716,200 miles, its aphelion distance will be 5,000,000,000
miles, the diameter of its nucleus shortly before disappearing

on the solar disc was 7600 miles, the velocity at perihelion

295 miles per second and at aphelion 75 feet per second.
But little is known in regard to the density of comets ; but
to be on the safe side we will assume it as ;, that of water.
This data will reduce (18) to 13x 10-¥ for the fraction of
énergy lost during one of its revolutions about the sun; and
as it would make a revolution in, say, 20 hours, it would lose
in one of our years about 57x 10-16 of its energy, at which
rate it would go on for 170 trillions of years. Similarly, at
its aphelion its rate of loss would be less than 3 x 10-" of its
energy in more than 2000 years—the time of one revolution
in its orbit.

The most careful observations and calculations have failed
to detect any effect due to the resistance of matter in space ;
and the above analysis shows that, within historic times, it has
in any case scarcely amounted to an infinitesimal, certainly
not sufficient to be measured. And when we consider that
our assumptions have been very largely on the unfavourable
side, and, further, that the energy imparted to the ether may,
partly at least, be restored to the body, we assume that its
resistance never can be measured. Laplace, when he found
that the force of gravitation, if propagated by an elastic
medium, must have a velocity exceeding 100 million times
that of light, concluded that astronomers might continue
to consider its action as instantaneous (Mécanique Celeste,
B. x. ch. vill. p. 22, 9035) ; so may we, with as much con-
fidence, continue to consider the resistance of the ether as nil.

Equation (6) gives

ore 0 0(186800 x 5280)?
“a *2X B22 x 14% 772

from which the specific heat of the ether may be found if its
temperature were known. M. Fourier, the first to assign a
value to the temperature of space, assumed it to be somewhat
inferior to the temperature at the poles of the earth, or about
50° C. to 60°C. below zerot. M. Pouillet, considering the

=92x10", . . (20)

* 5 Monthly Notices of the Royal Astronomical Society, vol. xliv. 2,
. 54.
+ Ann. der Chemie, tome xvii. p. 155.

400 Prof. De Volson Wood on

atmosphere as a diathermanous medium, capable of absorbing
in different degrees the radiant heat from the sun and the
dark heat from the earth, deduced for the heat of space—or,
as he and Fourier called it, the stellar heat—approximately
— 142° C.* (—287° F.), which is about 174° F’. above absolute
zero. It is well known that Pouillet’s data were imperfect,
several important elements being neglected, notably that of
the humidity of the air; stiil it is not only the first, but, so
far as we know, the only attempt to formulate this relation.
It served to show, what has since been indicated by more
direct experiments, that the temperature of space is very low.
The delicate experiments of Professor Langley, before referred
to, show a great difference in the degree of absorption by our
atmosphere of different wave-lengths. The mean of the
values for nine different wave-lengths, treated by M. Pouillet’s
formula, gives 139° I’. above absolute zero, and the smallest
value of absorption, which was for the infra-red, gives only
71° EF’. above absolute zero for the heat of space.

The heat of space may be considered as composed of three
parts:—(1) stellar heat, (2) the heat contained in the dark
matter of space, (3) the essential heat of the ether. :

1. By the stellar heat we mean the heat received directly
from the stars. It is a matter of easy calculation that, if the
~ §0,000,000 of stars supposed to be visible with the most
powerful telescopes were all at the distance of the nearest
fixed star (a Centauri), or 221,000 astronomical units from
the earth, and if each radiated the same amount of heat as our
sun, the intensity varying as the inverse squares of the distances,
the earth would receive from them all less than ;4; as much
heat as it now receives from the sun. And when we consider
that only a very few stars are within measurable distances,
and that the remote ones may be, when compared with these,
well-nigh infinitely distant, it 1s evident that the amount of
heat received from the stars is insignificant and may be
discarded at the earth.

2. It is certain that there is a large amount of dark matter in
space, since the meteoric dust and meteorites must come from

* Comptes Rendus, 1838, vol. vii. p. 61. Pouillet’s formula is
; 2 2—6 :
a’ = 1-235 >, — 0-489,
in which 6'=the absorptive power by the atmosphere of the sun’s heat,
b=the absorptive power of terrestrial heat,
t'=the temperature of the stellar heat,
If b=1, its maximum, b'=0:2, we find ¢/= —235° C, (— 391° F.), or
71°F, above absolute zero.

the Luminiferous dither. 401

beyond our atmosphere. The zodiacal light is supposed to be
an evidence of meteoric matter between the earth and sun.
The tails of comets are visible by some action of light upon
some kind of matter. Matter in space not exposed to the rays
of the sun will be at about the same temperature as the ether;
but if in the rays of the sun and destitute of an atmosphere at
_ the distance of the earth from the sun, its temperature would
be very low. If present laws can be extended so far, and the
earth were without an atmosphere, and the heat received
were not conducted away, it has been computed that the mean
temperature at the equator would be about —70° C. (—94° F.);
and at the poles —221° C. *, or 114° F’. above absolute zero,
The last result is obtained on the supposition that the poles
receive heat directly from the sun a part of the year ; it is
further shown that if the poles were never exposed to the rays
of the sun, the temperature would fall to that of the ezther of
space. But the data is not uniform, and there is too large an
extension of empirical formulas to satisfy one that the above
_ numerical results are reliable ; still they point more and more
_ strongly to a temperature not many degrees above absolute
«Zero.

3. By the essential heat of the ether we mean the tempe-
rature which would be indicated by a thermometer graduated
from absolute zero in a room located in space beyond our
_ atmosphere whose walls were impervious to the passage of
_ external heat. It is the heat due to the self-agitated ether,
' just as air has a temperature when not exposed to the rays of
the sun. If the ether be perfectly diathermanous to the sun’s
rays, it will receive no heat on account of the heat of the sun
flowing through it, though it may be heated from other
sources. As direct evidence of an extremely low tempera-
ture of space, we cite the facts in regard to the meteorite which
fell at Dharmsalla, India, July 14, 1860 7. ‘‘The most re-
markable thing about it was, while the mass had been inflamed
and melted at the surface, the fragments gathered immediately
after the fall and held for an instant were so cold that the jin-
gers were chilled. This extraordinary assertion, which is con-
tained in the report with no expression of doubt, indicates
that the mass of the meteorite retained in its interior the
intense cold of the interplanetary space, while the surface was
ignited in passing through the terrestrial atmosphere.’ Since
this body had been exposed to the rays of the sun, its tempe-
rature must have exceeded that of the space through which it

: * Professional Papers of the Signal Service, U.S, A.: Washington,
- D, C. 1884, No. xii. p. 54.
+ Comptes Rendus, 1861, tome liii. p. 1018.

402 Prof. De Volson Wood on

passed, as well as been warmed by the heat developed at its
surface, from which it may be inferred that it had been intensely
cold. Direct investigations, given above, indicate that this
temperature is less than 200° I’, above absolute zero ; and we
cannot assert that it is not less than 100° EF’. above, or even
much less.

But, however low be the temperature of the ether, it can-
not be absolutely cold, or, in other words, it must have a tem-
perature above absolute zero, for otherwise it would be desti-
tute of elasticity, and hence incapable of transmitting a wave.
This is shown by eliminating V between equations (2) and (6),
giving

ecr= |, |
2g6J3 ’

in which if r=0, e will be zero, all the other factors being
finite, and if e=0, then V=O in (2). Indeed, this principle
is so well recognized in physics, that a proof in this place
seems superfluous. Being unable, in the present state of our
knowledge, to do more than assign the probable superior limit
of the temperature, we will, for the purposes of this analysis,
assume T=20° F., absolute, being confident that the actual .
value is between =, of and 10 times this value. This value in
equation (20) gives

c=46 x 1044=4,600,000,000,000. . . (22)

for the specific heat of the ether, that of water being unity.
This number so vastly—we might say infinitely—exceeds that
for any known gas, as to justify one, at first thought, in looking
with suspicion upon the applicability of the above analysis to
this medium. Assumptions in regard to the absolute tempe-
rature will scarcely improve the appearance of this number.
If it be assumed that the absolute temperature be only one
degree, the number in equation (22) would be only twenty
times as large ; and if the absolute temperature be assumed at
1,000,000° F’., the resulting specific heat would still be more
than a million times as large as for hydrogen. A few consi-
derations of other properties of the wether may aid one in being
reconciled to this paradoxical result. Is the result any more
incredible than the fact, everywhere admitted, that every par-
ticle of the ether, in transmitting a wave of light, continually

* We note that this equation shows that the specific heat for different
gases under the same tension, e, and temperature, T, varies inversely as
the density ; and for the same temperature and density the specific heats
e will be directly as the tension e. The more perfect gases, as hydrogen,
oxygen, and air, conform nearly to this law.

the Luminiferous ther. 403
makes 590,000,000,000,000 (6 x 104 nearly) complete cycles

of movements every second, for a wave-length of 5.000 of an
inch? The number of such complete movements in air for
the fundamental ¢ is only 264 ; and hence the ratio of the
former to the latter of these numbers is nearly 2x 10!%. The
ratio of the specific heat given in (22) to that of hydrogen
is nearly 14x10", which is not so different from that just
given for the ratio of cyclical movements in a second of the
ether and air. The velocity of sound in air at 493° F. above
absolute zero is about 1090 feet per second ; but if the tem-

‘perature could be reduced to 20° F’., absolute, the law being

extended so far, the velocity would be only

oy

but the velocity of light is 982,000,000 feet per second, a
number about 44 million times the former, and near a
million of times that of the velocity in air under ordinary
conditions. The ratio of the mass of air ina cubic foot at sea-
level to that of a cubic foot of the ether as computed, far ex-
ceeds any of these ratios. The fact is, the known qualities of
the ether in transmitting light and heat so far transcend
those of any known terrestrial substance, that we might anti-
cipate the fact that, in regard to magnitude, all its properties
will be extremely exceptional when compared with such sub-
stances. We mustaccept substantially the number in equation
(22), or subject this medium to different laws than those of
ases.

: We may deduce this result by another process ; thus, since
the specific heats of different gases are as the squares of the
wave-velocities in the respective substances, the other elements
being the same, if the specific heat of air be 0°23, we should
have for the specific heat of the ether

. 2
= 0:23( PE OREO 46 x 104,

as before. The correct value of the specific heat of air, 0°2375,
would give over 47x10", and nearly 48x 10!!; but these
differences are quite immaterial in this connection, the object
being to check the former result, and find chiefly qualitative
values.

On the other hand, in order that common air might be able
to transmit a wave of the known velocity of light, its specific

heat being taken constantly at 0°23, its temperature would be,

404. Prof. De Volson Wood on

according to equation (20),

Le XA. ‘a
sheaaiaell << paisa: 10 degrees F.

(=400,000,000,000,000° F.).

If the sun were composed of a substance having such spe-
cific heat, it could radiate heat at its present rate for more
than a hundred millions of centuries without its temperature
being reduced 1° F., exclusive of any supply from external
sources, or from a contraction of its volume. We know only
such substances in the sun as we are able to experiment with
in the laboratory; and if there be an exceptional substance in
it, we have no means at present of determining its physical
properties. It is, moreover, a question whether the ether
constitutes an essential part of bodies. We conceive of it
only as the great agent for transmitting light and heat through-
out the universe.

On account of the enormous value of the specific heat, it
will require an inconceivably large amount of heat (mechani-
cally measured) to increase the temperature of one pound of
it perceptibly. Thus, if heat from the sun, by passing
through a pound of water at the earth, would raise the tem-
perature 100° F'. and maintain it at, say, 600° F., absolute,
it would, under similar conditions, raise the temperature of
one pound of the ether, if its power of absorption be the same
as that of water, zaawoon of a degree*.

The distance of the earth from the sun being 210 times
the radius of the latter, the amount of heat passing a square
foot of spherical surface at the sun will be about 45,000 times
the heat received on a square foot at the earth normally ex-
posed to its rays, so that, under the conditions imposed, the
temperature would not be a billionth of a degree I’. higher at
the sun than at the earth. This, then, is a condition favour-
able to a sensibly uniform temperature even if heated by the
sun’s rays. We are now inclined to admit that the ether is
not perfectly diathermanous to the sun’s rays, but that its
temperature, however small, may be due directly to the
absorption of the heat of central suns; for we begin to
realize the fact that the ether may possess many of the
qualities of gases, such as a molecular constitution, and
hence also mass, elasticity, specific heat, compressibility,
and expansibility, although the magnitude of these pro-
perties is anomalous. We have already considered its

* This illustration is rather crude, since it discards the relative volumes
occupied by a pound of the respective substances, %

‘

the Luminiferous ther. 405

compressibility at the surface of the sun, due to the weight of
an infinite column, and found it to be exceedingly small ; now
it may be possible that the expansion due to the excess of tem-
perature of a small fraction of one degree at the surface of the
sun over that at remote distances will diminish the density
as much, or about as much, as pressure increased it, thereby
making the density even more exactly uniform than it other-
wise would be. According to what we know of refraction, it
is impossible for a ray of light to be refracted in passing
through the ether only,—at least, not by a measurable
amount ; for not only are the density and elasticity practically
uniform, but their ratio is, if possible, even more constant as
shown by equations (16) and (16’). But the freedom of the
ether-molecules may be constrained, or their velocity impeded,
by their entanglement with gross matter, such as the gases
and transparent solids ; in which case refraction may be pro-
duced in a ray passing obliquely through strata of varying
densities. Neither is it believed that the ether does or can
reflect light ; for if it did, the entire sky would be more nearly
luminous. The rays in free space move in right lines.

The masses of the molecules in different gases being inversely
as their specific heats, and as the specific heat of hydrogen is
3°4, and the computed mass of one of its molecules } x 10-?9*
of a pound, we have for the computed mass of a molecule of
the luminiferous ether,

11 o'4 1
ge ee thing! lip
18x10" 46x10" 22x10

* Stoney concludes that “it is therefore probable that there are not
fewer than something like a unit eighteen (10'%) of molecules in a cubic
millimetre of a gas at ordinary temperature and pressure” (Phil. Mag. 1868
[4] xxxvi. p. 141). According to the Kinetic theory the number of mole-
cules in a given volume under the same pressure and temperature is the
same for all gases. The weight of a cubic foot of hydrogen at the tempe-
rature of melting ice and under constant pressure being 0°005592 of a
pound, and as a cubic foot equals 28,315,000 cubic millimetres, the probable
mass of a molecule of hydrogen will be

6:005592 23 Be,
oo ee SIs eox 10" 1X10" -.
-. . 46 3 Se
Maxwell gives 0% of a gramme = 108 lb., which is about 2 the value

given above (Phil. Mag. 1873 [4] xlvi. p. 468).

' The difference in these results arises chiefly from the calculated number
of molecules in a cubic footof gas under ordinary conditions. Thomson gives
as the approximate probable number 17 x 10”, which is about 2 the value
given by Stoney. Thomson’s value would make the mass of a molecule
of ether about 5, x 10~*° of a pound, which is not much different from
that found above.

Phil. Mag. 8. 5. Vol. 20. No. 126. Nov. 1885. 2.&

406 Prof. De Volson Wood on
The mass of a cubic foot of the ther, equation (9), divided

by the mass of a molecule, gives the number of molecules in a

cubic foot, which will be

oo ni 22 x 10*
- pee 1

which call 10%. This number, though large, is greatly ex-
ceeded by the estimated number of molecules in a cubic foot
of air under standard conditions, which, according to Thom-
son, does not exceed 17 x 10°, a number nearly 17,000,000,000
times as large as that in equation (24) ; and yet, at moderate
heights, the number of molecules in a given volume of air
will be less than that of the ether.

Assuming that air is compressed according to Boyle’s law,
and is subjected to the attraction of the earth, equation (15)
will give the law of decrease of the density. Taking the
density of air at sea-level at 7, of a pound per cubic foot,
€)= 14:7 lb. per square inch, r= 20,687,000 feet, equation (15)
becomes

== x10% . . (24)

nN

85, x LOT ae 20

If c=, =; x 10-3, which would be the limit of the
density, and it is a novel coincidence that this limit is nearly
identical with the value found for the density at the height
of one radius of the earth according to the ordinary expo-
nential law wherein gravity is considered uniform*.

If the number of molecules in a cubic foot follows the same
law, then at the height z there will be

z

17 x LOT rte? oe) te cee oe

molecules per cubic foot. Similarly, the value of the length
of the mean free path would bef

2x10™+2-8 inches. . . . (27)

By means of these values, the following table may be
formed :—

* The ordinary exponential law results from dropping — compared with
unity in equation (15), giving
2 ft. 2 miles

z
S= dye 26221=§,10 60387 =x 107 14s

in the last of which, if z=3956, the exponent becomes 345.
+ Phil. Mag. 1873 [4] x]vi. p. 468.

—— ee

the Luminiferous ther. 407

Height.
Density or ten-| Number of Length
sion, that at the molecules of the
Fractional Approxi- earth being | in a cubic foot. mean free
parts of mate in unity. path.
earth's radius. miles.
0 0 1 Vixcl0 5, | 2x 107° inch
—4°3 20°7
5 50 10 17x10 210717 |,
#5 100 TT ee 17x10°°° | 21024,
a 200 100° 17x10°° | 792,000 miles |
= 282 100% 17x10? | 31x10",
ru 395 107 7 17x10-® | 31x10 ,,
1 800 1? 17x10—* | 31x104 _,
1 3956 10, + 17x 10-147, 31 «10160 |
2 7912 17 17x 10—2%) 31 10718 ,,
00 00 iG> 17 x 10-320) 31 x 10333

The numbers in the third column multiplied by = will

give the density (or mass per cubic foot) at the respective
altitudes ; and the same numbers multiplied by 15 (or more
accurately 14°7) will give the tension per square inch. <Ac-
cording to this law, at an elevation of 300 miles the density
of the atmosphere will be somewhat less than the density of
the ether as given by equation (9).

To find the height at which the tension of the atmosphere,
according to the above law, will be the same as that of the
zether, we have, by means of equations (11) and (25), sub-

stituting in the latter 2116 for z,

pee ed
2116 x 10 rH =a,

which solved gives
$k

3-947 1206 miles,

o =
w

so that at the height of 127 miles the tension would be less
than that of the ether, the temperature being uniform.

The mean free path according to the above law, in which
gravity varies as the inverse squares, is less, and for great
heights much less, than would be found according to the
ordinary exponential law. Thus Crookes states that the mean
free path of a molecule at the height of 200 miles is about

2G 2

«’#
: y.

408 Prof. De Volson Wood on

10,000,000 miles*; but according to the above law it becomes
about 792,000 miles. |

If a cubic inch of air at sea-level were carried to the height
of ; the radius of the earth, and then allowed to expand freely,
so as to become of the computed density of the atmosphere at
that point, it would fill a space of 4 x 1078 cubic miles, or a
sphere whose radius is 2,398,000,000 miles, which is nearly
equal to the distance of the planet Neptune from the sun ;
and there would be less than one molecule to the mile. Such
are some of the results of extending a law to extreme cases
regardless of physical limitations, or of the imperfection of
the data on which it is founded. For instance, a uniform
temperature is assumed, and, impliedly, an unlimited divisi-
bility of the molecules. The latter is necessary in order to
maintain a law of continuity. But modern investigations
show that not only air, but all the gases, are composed of
molecules of definite magnitudes whose dimensions can be
approximately determined ; and hence, if there be only a few
molecules in a cubic foot, and much less if there be but
one molecule in a cubic mile, it cannot be claimed that the
gas will be governed by the same laws as at the surface of
the earth.

To find the Height of the Atmosphere.—The atmosphere will
terminate at that height where the vertical repulsive force
equals the weight of the particles in the topmost layer. Asa
first approximation, conceive that the molecules are arranged
in horizontal layers and vertical columns, in a prism whose
base is one square foot, and whose height extends to the
height of the atmosphere ; the base of each column of moie-
cules being one of the molecules in the base of the prism.
Considering the number of molecules in a cubic foot of air
at standard conditions as 17x10”, and the weight of the
same as ‘08 of a pound, we have for

8

the weight of one molecule of ait =T7 Jou! (28)
The number of molecules along one edge of the bottom

layer will be VW 17 x 10 nearly ; and the number in the

bottom layer the square of this number, or 1703 x 10%,

which, according to the hypothesis, will be the number in

* Phil. Trans. Roy. Soc. London, 1881, Part II. p. 389.

+ This may be used as a unit for measuring the mass of a cubic foot
of the ether. Thus, dividing the value in equation (10) by that in (28)
gives 4250; or the mass of ether in a cubic foot is 4250 times the mass
of one molecule of air.

the Luminiferous ther. 409

the top layer ; and this multiplied by the weight of one mole-
cule will give e, the weight in the top layer ; and equation
(14) will give (the temperature of the column being considered
uniform )

as aa 170% x 10!6 x 8
14°7 x 144 x10" r+ = ether 3
c= 55-55 = 169 miles. mt es (29)

But the temperature is far from being uniform. In regard
to a definite mass of a gas, we have the well-known relation

ae ee 1" 29/
ae a constant ee ae aS (29’)

where p=e=the pressure on the base of a prism, and v=the
volume.
The value of 6 from this equation substituted in (13) gives

a aa)

But with 7 an unknown variable this cannot be integrated.
If r=7) we at once have equation (14). The relation between
T and z is unknown, if indeed there be any algebraic relation
between them. It is, however, known that, as a general fact,
the temperature decreases with the elevation ; although local
causes and air-currents often cause this law to be reversed
for moderate heights. The best that can be done, in this
case, is to find an expression that will represent, approxi-
mately, the mean values of the temperature. It is usually
assumed that the average temperature at the earth is about
59° F. or 60° F., and that for latitudes of, say, 40° N. to
40° 8. the perpetual frost-line is from 14,000 to 16,000 feet
above sea-level ; and observations indicate that the rate of
decrease of temperature decreases with the height. The last
fact is suggestive of an exponential law ; hence assuming

THT Ea, rete € ate 8) Coe)

and making r=493° F., absolute, at the height z=15,840 feet
and T7=520° F., absolute, we find a=296,000 (or 56 if z be

in miles), and our equation becomes

2 miles

a ee 8

410 Prof. De Volson Wood on
This gives
Heiner, absolute, oe steel)
0 520° FE. a ean oie ime
1—5 518 57 oT
2—5 5 iy 54
a5 513 52 oF
4-5 512 51 LAdg
1 510 4Y | Al
2 501 40 32
2 493 52 18
4. A84 23 8
5 475 14 — 2
6 467 6
q 458 — 3 —11°8
50 212 — JAY
i2 136 ao
100 87 —pi4
120 65 — 396
150 536 —495
294 9 —452

The temperatures given in twenty-five or more reports of
balloon ascensions, not only give values the mean of which
is fairly represented by the celebrated seven-mile ascent of

Mr. Glaisher, but his figures, given in the fourth column of |

the table, represent a more uniform law than is common in
such reports. Our computed values exceed his observed
values at all points except at the surface of the earth, where
they agree. In this ascent he reached the point of freezing
at the height of two miles, which is lower than the average,
as determined by many observations; and therefore it appears
that equation (31) probably represents the general law better
than this single set of observations. The effect, however, of
the exponential law is scarcely perceptible within the limits
of observation ; for the exponent of € is so small for eleva-
tions under seven miles, that it makes the law of decrease of
temperature nearly uniform with equal increments of eleva-
tion. Thus, omitting fractions, the computed decrease for
the first mile is 10°, and the average for seven miles is nearly
9° ; but to assume a uniform decrease throughout the column
limits the height of the column independently of pressure or
other conditions, for it could not extend beyond the point of
absolute zero. There is no objection to applying such a law

* «Travels in the Air,’ by James Glaisher, F.R.S., p. 50.

Wited-28

the Luminiferous ther. 411

provided it can be shown to be true—a condition which, at
present, is not accepted.

Substituting + from (31) in (30), and integrating between
the limits of z and z=0, gives

agsy z
—°F°0( ay
e=e,6 e ©); fer et ig CA)

which ultimately will equal the weight of the molecules in
the top layer. Hence, substituting numbers, we have

V4
8% 1703 x 1018 296000x 08 56
lg is

aarti

1) :
which gives
FMR YL yuh ae) pix a LO)

It is evident that the very low temperature of the higher
portions of the column will shorten very much the hypo-
thetical column of uniform temperature ; but there are other
conditions which will modify the preceding analysis. The
assumptions in regard to layers and columns would not be
realized even under statical conditions, and much less for the
conditions in nature. Statically, the molecules would arrange
themselves more like shot in a pile, each being over the space
between the molecules in the layer below, instead of over a
molecule. This arrangement would give a less number in the
horizontal layers than assumed above. But the hypothesis of
constancy in the number of molecules in the layers is open to
greater objections. Tor the distance between the molecules
will increase with the elevation on account of the diminution
of the pressure of that part of the column above the point
considered, and the elastic force will be correspondingly di-
minished ; while, horizontally, in the plane of a layer of the
molecules, the elastic force would remain constant. In other
words, in the medium arranged as assumed the tension would
not be the same in all directions, and hence would be in
unstable equilibrium. As a refinement, we notice that in
every heavy fluid the downward pressure at every point
exceeds the upward by the weight of a molecule.

Considering, now, that the molecules in the hypothetical
layers are distributed uniformly throughout the spaces imme-
diately beneath them, the number in the new top layer will be
less than in the former case, and the column will rise toa
greater height, and hence will exceed 86 miles ; and, in turn,
the new column would need another correction, and so on.
Assuming that the number in the top layer is 10'°, and that
the vertical component of the elastic forces follows the law of

412 Prof. De Volson Wood on-

equation (33), we find
z=95 miles ;

and if the number in the top layer be 10+, we find z=104
miles, and for one molecule, z=110 miles. In a similar
manner it would be legitimate to assume that the column was
capped by a fraction of a molecule, for that would be equi-
valent to one molecule at the top of a column having a base
of several square feet. We are unable to determine where
this process would end in nature; and hence this analysis fails
to fix definitely the extreme height of the atmosphere, even
for statical conditions.

Assuming that the distance between the contiguous mole-
cules would be inversely as the third root of the densities of
the medium, as they would be with sufficient accuracy where
the number of molecules in a cubic foot is immense, we have,
after substituting e, equation (33), and 7, equation (31), in
(29), |

5 [fetid] a
ag € 0 ] = a?

where dy is the distance between contiguous molecules at sea-
level and d the corresponding distance at the height z. Hence

2
298 (e¢—1)—?
a

d= dole “0

18 agSo
SS = ae aS Sie I) y aAV
Aix 10" eo ; we have
for z= 86 miles, d=, of an inch,
if ee Oy Wo d= 4:5 inches,
gray 0 o—1 12

These values of d are greatly in excess of the distances be-
tween contiguous molecules in the horizontal layers, according
to assumed conditions. Thus, at the height of 104 miles, it
was assumed that there were 10? molecules on the side of a
square foot, in which case the distance between contiguous
molecules would be about 2 of an inch instead of 11 inches
as above. ‘These results ought not to agree exactly, for one
analysis assumes that the atmosphere terminates with each
assumed number of molecules, while the other assumes that
the law is continuous to any height. It is apparent that
the laws represented by equations (33) and (35) both become
practically discontinuous at a height at or less than 95 miles.
For the sake of giving definiteness to the following remarks
we will assume that the mean height for statical conditions is

1

. 2%

f dy

7)

the Luminiferous ther. 413

95 miles. But the conditions in nature are not statical. The
changes in temperature in the column will be continually in-
creasing or decreasing its height; the air-currents also operate
to change it, first by increasing or decreasing the temperature
from the mean at considerable heights, and, secondly, by
operating dynamically to push the top of the column upward;
the aerial tides may operate to raise the column still higher,
and the molecules themselves are supposed to be flying with
great rapidity in all directions. An increase of temperature
of one tenth the mean value, which at the earth’s surface

- would be about 49° F., would elongate the column about ten

miles, and a corresponding decrease would shorten it about
the same amount, making it 105 miles in the former case and
85 miles in the latter. The effect of air-currents and aerial
tides cannot be so definitely calculated; but it is safe to assume
that they may produce a much greater increase of height
above the mean than they will depression below the mean ;
just as ina highly agitated sea, the depressions below the
mean surface-level may be small compared with the height

_ above the same level to which the spray from the top of a
wave may be thrown. It seems possible, therefore, that when

the temperature, air-currents, and aerial tides conspire to de-
press the column, the extreme height of the atmosphere may
be reduced to less than 85 miles; and when they conspire to
elevate it, it may possibly rise to a height exceeding 120
miles. .

If it be certain, as is assumed, that the meteors are rendered
incandescent by atmospheric friction, and the extreme height
at which they are visible could be determined by direct obser-
vation, it would fix a height less than the extreme height of
the atmosphere, independent of other physical considerations;
but the movement of these bodies is so extremely rapid that
it is impossible to determine their height with astronomical
precision. Still computations by Professor Herschel give a
height of about 118 miles*, and Professor Newcomb estimates
it to be about 100 milesf. It is possible that a meteor would
sometimes become inflamed by penetrating the atmosphere

* Professor A. S. Herschel gives the height of twenty meteors varying
from 40 to 118 miles. ‘ Nature,’ vol. iv. p. 504.

+ Newcomb says:—‘‘ The lightning-like rapidity with which the meteors
darted through their course rendered it impossible to observe them with
astronomical precision; but the general result was that they were first
seen at an average height of 75 miles and disappeared at a height of 55
miles. There was no positive evidence that any meteor commenced at a
height greater than 100 miles. These phenomena seem to indicate that
our atmosphere really extends to a height of between 100 and 110 miles.”
‘Popular Astronomy,’ 1878, p. 389.

414 Prof. De Volson Wood on

only a few miles, for although the atmosphere in the upper
regions is extremely rare, yet the actual number of molecules
in a cubic foot is large. Thus, according to our analysis, for
statical conditions, the topmost cubic foot of the 104-mile
column would contain about 1,000,000 molecules; and at the
height of 95 miles it would contain about 1,000,000,000,000,000
molecules; so that if the relative velocities of the meteor and
air be 20 miles per second, the meteor would encounter an
enormous number in the twentieth or even the hundredth part
of a second, after first entering the atmosphere.

The height of the auroral arch—supposed to be within our
atmosphere—bas been computed to be from 33 to 1000 miles
(see article ‘“‘ Aurora,’ Encyc. Brit.). But it has been shown
by experiment, that a vacuum may be produced through
which an electrical discharge cannot be passed, and yet the
atmosphere at the height of 150 miles under the most favour-
able condition, that of uniform temperature, is vastly more
rare than the most perfect vacuum ever produced by the most
perfect Sprengel pump; and at the height of 200 miles under
the same conditions the vacuum would be some 10,000,000
times as great as the most perfect vacuum yet made ; 5 while,
according to the probable law of the decrease of temperature
with the elevation, and in accordance with the probable mass
of a molecule of air, the extreme height falls far short of 150
miles. Itis evident, therefore, that the assumed determination
of the height of the atmosphere by means of the auroral arch
is, to say the least, unreliable*.

We have pursued this digression in regard to the atmo-
sphere partly for its own sake and partly to show, by way of
contrast and accumulative evidence, that the ether is a sub-
stance entirely distinct from that of the atmosphere,—that the
former cannot be considered as the latter greatly rarefied, as
some have supposed. Admitting the validity of the prece-
ding discussion, some of the distinctive properties are :—

1. The different modes of the movements of the molecules
in the two substances in the propagation of a wave; in one
the motion being a to-and-fro movement and in the other a
transverse movement. ‘These are distinctions recognized by
the best writers upon the subject, and are especially noticed
by Maxwell in an article on Hther in the Hncyclopedia Bri-
tannica.

2. It is impossible for a wave to be transmitted in air with
the known velocity of light, unless its temperature be increased

* Some writers incline to the view that the aurora is due to a cosmic
rather than a terrestrial origin, ‘Science,’ 1885, p. 396,

the Luminiferous ther. 415

millions of millions of degrees Fahrenheit above the standard
temperature; but such a wave is transmitted in the ether
although its temperature is far less than has ever been pro-
duced by artificial means.

3. The ratio of the elasticity to the density in the ether is
exceedingly large compared with the same ratio in air. The
temperature of air being taken at 60° I’. and the ether at 20°

_ F., absolute, the ratio is, with sufficient accuracy,

980,000,000\? , 4,
Smee —— x 10 nearly.

4. The specific heat of the ether is, at least, many million
times that of air, or of any other known gas.

5). The atmosphere is of variable density, elasticity, and
temperature, while the ether is well-nigh isometric throughout
space in regard to each of these elements.

6. A molecule of ether is well-nigh infinitesimal compared
with one of air.

7. Air is attracted to a planet with such a relative force,
that its extreme height is only a few miles.

8. The ratio of the density to the elasticity of the ether is
constant; but in the atmosphere, on account of the decrease
of temperature with the elevation, the density decreases less
rapidly than the elasticity, as may be seen by comparing
the first part of equation (35) with equation (33): we have

b Oe

Sars at
On this account a wave would be propagated with less velocity
in the higher regions of the atmosphere than in the lower,
while a wave in the ether has a sensibly uniform velocity
throughout space.

The question may arise, May not the resistance of the ether
drag away the remote molecules of the atmosphere, and so
scatter them in space along the path of the earth’s orbit?
Assuming that the atmosphere is moving with the earth
through space at the rate of 20 miles per second (which

exceeds the actual velocity), and that the resistance of the

ether ig measured in the same manner as for fiuids, we have
a2
2

for the resistance R=kwa =, where v is the velocity of a

molecule of air, a is meridian section, w the weight of a unit
of volume of the ether, and & a coefficient depending upon
the form of the body. Making k=1, which is greater than its

416 On the Luminiferous Aither.

i ee feet, which, again, is in excess of
the true area, w the value in equation (10), we find that

2 1 1

R= Jom x a x (20 x 5280)?. - 6h 08
The attractive force of the earth for a molecule of air is given
in equation (28), and hence the attraction of the earth for a
molecule of air will exceed 500 ,000 times the resistance of the
ether; hence the molecules accompany the earth in its orbit
as certainly as does the moon, and are more rigidly bound to
it than is its satellite.

The Kinetic energy of a molecule of air at standard con-
ditions is about

actual value, and a=

of a pound nearly.

1 8 2
9 . 55-9 x17 x 1077 000” = 1023 foot-pound ;
and of the ether, according to our results, about
N t yg |
5° x 59x 100 (2"8: 000 x 5280)*= 7x 102 foot-pound ;

which results are nearly the same; but in a pound of the
ether there is some 100,000,000 000 times the Kinetic energy
of a pound of air.

Considering the terrestrial atmosphere as equivalent to one
of uniform density and 54 miles high, each of whose mole-
cules has a mean square velocity of 1600 feet per second, and
the zther of uniform density each of whose molecules has the
mean square velocity of 286,000 miles per second, a rough
approximation shows that the ‘Kinetic energy of the sether in
a sphere whose radius is 92,000,000 miles (nearly the distance
of the earth from the sun) will be only about 100,000 times
that in our atmosphere.

The mean free path of a molecule of gas as given by
Loschmidt is

4, combined volume of the molecules
~ volume of the gas x } the diameter of a molecule’

and by Maxwell,

(the last member of which we have added), in which p is the
density of the gas, w the coeflicient of internal friction, and v
the velocity whose square is the mean of the squares of the
actual velocities of the molecules. In regard to the ether,
these equations contain at least three unknown quantities, J,
uw, and the diameter of a molecule, and hence they cannot be

On the Determination of the Ohm. A17

completely solved. Comparative results, however, may be
found by assuming that the density of the molecules of zether
equals those of hydrogen, or is any multiple thereof; for then
the diameter of a molecule of the ether might be found (that
of hydrogen being 5°6 x 10—"° of a metre); and the combined

volume in a cubic foot will equal the number of molecules in

a cubic foot multiplied by the volume of one molecule, and
hence will be found the length of the mean free path and the
coefficient of internal friction.

We conclude, then, that a medium whose density is such
that a volume of it equal to about twenty volumes of the
earth would weigh one pound, and whose tension is such that
the pressure on a square mile would be about one pound, and
whose specific heat is such that it would require as much heat
to raise the temperature of one pound of it 1° F. as it would to
raise about 2,500,000,000 tons of water the same amount,
will satisfy the requirements of nature in being able to trans-
mit a wave of light or heat 186,300 miles per second, and
transmit 133 foot-pounds of heat-energy from the sun to the
earth each second per square foot of surface normally exposed,
and also be everywhere practically non-resisting and sensibly
uniform in temperature, density, and elasticity. ‘This medium
we call the Luminiferous Aither.

XLVI. A Determination of the Ohm.
By Prof. F. Himstept*,

UGH the means placed at my disposal by the Go-

vernment of the Grand Duchy of Baden, for the deter-
mination of absolute resistance, I have been enabled to carry
out a determination of the Ohm according to the method
recently published by met; and beg permission to lay an
abstract of my results in the following paper before the Royal
Berlin Academy.

According to the above-mentioned method, the constant
divergences of the magnet in the same galvanometer are
observed—produced in the one case by means of induction-
currents passing in the same direction through the galvano-
meter at the rate of n per second ; in the other case by means
of a constant current whose strength is a known fraction of
the inducing current. Let us call the observed angles of
divergence a, and a, the required resistance 7 will be found

* Translated from the Sttzungsberichte der k. Pr. Akad, d. Wissens.

Berlin, July 23, 1885.
+ Wied. Ann, Bd. xxii. 8, 281 (1884).

418 Prof. F. Himstedt on the
(as is elsewhere more fully stated) from the formula
** tan @,
wh SE: LO

in which V expresses the potential of the induction-coils here
used towards one another.

If we take as inducing coil a solenoid with only one iayer
of wire, in comparison with whose length both its own radius
and the dimensions of the inducing coil are insignificant,
we get

V=47’° RK .D. (14+ 2a);
whence follows :—

pA RKB n(1-49a) 282 oo

tan ay

R expresses here the radius of the solenoid, K the number of
convolutions in a unit of length, > the total number of convo-
lutions of which the induction-coil consists, and 2a a term of
correction which expresses the action of the ends of the
solenoid.

According to the experiments described Rat the value of
2a was always less than 0°03; so that for the determination
of r only the exact measurement of the quantities R, b, n, K,

tan 2 )
and ——* were taken into account.
tan 1

The advantages of this manner of experimenting are that the
number of the quantities to be ascertained is a comparatively
small one, and that all those quantities are omitted which
require great care in their accurate determination. I place
under this heading the constants and variations of terrestrial
and of rod magnetism, the coeflicients of induction of the sur-
faces of convolution of the wire-bobbin of many superposed
layers, the reduction-factor of the galvanometer, the moment
of inertia and logarithmic decrement of magnets in oscillation,
and, more especially, the exact determination of the resistance
of copper wires which are often not all in the same room,
and whose temperature can only be approximated to from the
temperature of the surrounding air*.

All length-measurements in the following treatise were
compared with a standard which Privy-Counsellor Forster
was good enough to correct at the Normal-Standards Com-
mission in Berlin by means of the normal metre standard.

The measurements of time were carried out by means of

* Compare Roiti, Nuovo Cim. ser. 3, vol. xv., “ Diterminazione della
resistenza elettrica di un filo in missura ansol ita,”

Fon. gus

— Determination of the Ohm. 419

a ship’s chronometer of Brécking’s, which was rated by
observation.

The solenoid is wound round upon a wooden block re-
peatedly coated with glue, such as are used in orchestrions.
It was made in the year 1868, and was again planed down
and polished on the lathe in May of last year. Its dia-
meter was determined in three different ways, which were so
chosen that it was possible at the same time to ascertain that
the section of the drum sufficiently approximated to a circle,
and that the whole block was a cylinder. They were :—

1. Every six diameters of the same section were compared,
by means of a micrometer-screw at thirteen equidistant points
along the length of the block, with a glass rod measuring
23°3264 centim. in length”.

The diameters found were:---

Lowest value. Highest value. Mean.
centim. | centim. centim.
23°3193 23°3286 23°3248

2. The circumferences were measured by means of strips
of paper at thirteen equidistant points, whence the calculated
diameters were:—

Lowest value. Highest value. Mean.
centim. centim. centim.
23°3186 23°3252 23°3229

3. The diameter was reckoned from the length of the wire
wound round it. During the winding the thickness of the
wire was at the same time measured at 332 points by means
of a microscope with ocular micrometer, and found to be
0:0472 centim. The diameter of the block was found to be

D=23°3262 centim.

The measurements in No. 2 were repeated after the winding,
D=23°3190 centim., and were carried out again after each
determination, both with strips of paper and asteel-band mea-
sure, D=23°3194 centim. and 23°3204 centim. As the mean
of all the measurements, we get for radius R the formula (2),

R= 11°6846 centim.

The isolation of the wire coils was tested by means of
Hughe’s induction-scale f.

The number of convolutions was 2864; they covered the
drum to a length of 185°125 centim.

* The length of this glass rod was determined at the Normal-Standards
Commission in Berlin.

+ Cf. Rayleigh, Phil. Trans. 1884, vol. clxxv. p. 419, “On the Electro-
chemical Equivalent of Silver.”

420 Prof. F’. Himstedt on the

We get then
| 2864
im 1851S

The induction-coil is also wound round wood, and consists
of 3848 convolutions in 15 sections, which can be combined
at pleasure. The width of the bobbin was 4:01 centim. The
mean radius of each section was reckoned from the circum-
ferences of the separate layers of wires, which had been
measured by means of strips of paper. To test the isolation
of the coils, and to possess a control over the paper-measure-
ments, the mean radius of each section of the wider one was
determined by Von Bosscha’s method*. As, however, the
dimensions of the bobbin are only expressed in the term
2a of formula (2), a mistake of one per cent. would only
falsify the value of r by at most 0°0003.

The resistance 7, which has to be determined in absolute
measurement, amounted either to 1 or 4, or 2 Siemens units,
and was composed of 2 Siemens units (by Siemen and
Halske), nos. 3618 and 3619, which were used either singly
or parallel, or behind one another. These were made fast in
metal cases, which were filled with oil, and stood in a large
bath of the same. Both before and after the experiments,
both standards were compared with a third unit, and with
each other.

The thermometers were divided into one-tenths of a degree,
and were three times compared carefully with the standard
thermometers of the Physical Institute.

The interruptor was driven by a toothed wheel ft. For
determining n, the number of interruptions, a counter was
made fast to the axis of rotation, by means of which eo of a
revolution could be directly read off. 7

At least 700 revolutions were counted during each experi-
ment, so that the necessary approximation of 0:01 per cent.
was easily attained. The apparatus worked very evenly, so
that the divergences of the galvanometer were constant
throughout. |

This appears to me an advantage that this interruptor
possesses over that used by Herr Roitit; and I believe that
the discrepancies between his separate observations are due,
if not entirely, at least in a great measure, to the unequal
action of the ‘* Schmidt ”’ water-motor used by him.

K

* Wiedemann, Ziec. iii. p. 2138.

+ F. Himstedt, “ Zwei Verschiedene Formen eines selbstthatigen Dis-
junctors,’ Wied. Ann. vol. xxii. p. 276 (1884),

{ Cap. I. page 11.

Determination of the Ohm. 421

The galvanometer was not artificially damped. The posi-
tion of rest of the needle was determined from its maximum
divergence. The divergences to be observed, «, and a, were
never less than 800 millim. on a scale 4 meires distant, and
were always within afew scale-divisions of each other; so that
for the reduction to arcs an approximate knowledge of the
distance of the scale was enough. All connecting-wires were
covered with caoutchouc and twisted lightly together; all the
current-reversers, &c. were made of paraffin and sealing-wax.

Altogether, sixty-seven experiments were made; only the

- solenoid is common to them all.

Here the only variation possible was that the induction-
bobbin could be placed 10 centim. further towards either end,
instead of in the middle. All other components of the
apparatus were altered in each experiment to a considerable
extent. ‘The sections of the induction-coil were taken either
separately or in combinations of 2—5, corresponding to which
the number of the intermittences was 5-13 per second. The

| strength of the inducing current was from 0:0005 to 0:01
' ampére. The sensitiveness of the galvanometer was con-
| siderably alterable. The time taken by the oscillations of
) the needle was from 15 to 34 seconds.

To ascertain if the induction-current passing through the
___ galvanometer produced any other cross-magnetization of the
| needle, besides constant currents of otherwise equal galvano-

: : . tana,
metric action, by which the relation ra =

5

would be rendered

unreliable, several experiments were made with magnets of
a diameter of 0°6 centim., those used.in the above-mentioned
experiments being hardly 0-1 in diameter. No difference
was observed after a certain number (43) of experiments had
been made; the interruptor was fitted up afresh, and all con-
necting-wires as well as current-reversers renewed.

As the following results tally with the former, it may be
assumed that the isolation was perfect.

The whole plan of experimenting is based on the constancy
of the source of electricity, be this periodic, as in the observa-
tions by means of induction-currents, or continuous, as in the
measurement by means of the constant current. Means had
therefore to be taken to meet the final objection that errors
might arise through possible polarization in the battery.
With this object, I used as sources of the current :—

1. 1 to 4 Daniell’s elements.

2. Two points of a wire-resistance connecting a number of

Daniell’s or Bunsen’s elements.

3. A thermo-element.

Phil. Mag. 8. 5. Vol. 20. No. 126. Nov. 1885. 2H

492 Mr. J. Larmor on the Molecular

All sources produced the same result.
The mean of all experiments performed was
1 Siemens unit=0°94356 ohm ;
or, one ohm is equivalent to the resistance of a mercury-
column having a section of 1 square millim. and a length of
10598 centim. at 0° Centigrade.
Of the values determined,

the lowest was 1 Siemens unit =0°94323 ohm,
the highest was 1 Siemens unit =0:94380 ohm.

XLIX. On the Molecular Theory of Galvanic Polarization.
By J. Larmor, Fellow of St. John’s College, Cambridge™.

1. JT was first pointed out by Varley and Sir W. Thomson

that the polarizing action of a galvanic cell may be
explained by considering the cell to act as an electrical con-
denser of very large capacity. The mechanism of this action
has since been examined in detail, especially by Helmholtz.

In the polarization of a water-voltameter with platinum
plates for electrodes, the action according to Clausius’s well-
known molecular theory consists in the transfer through the
fluid of the temporarily dissociated hydrogen and oxygen
constituents under the action of the electric force ; so that in
the course of time a layer of hydrogen particles with their
positive charges accumulates in the immediate neighbourhood
of the kathode plate, and the complementary layer of oxygen
particles with their negative charges at the anode.

Hach of these layers, will form a sheet, with positive or
negative charge, lying close to the metal plate. On the plate
will therefore appear an equal and opposite charge by induc-
tion. There is thus a double electric layer formed at each
electrode ; the charged particles forming one side of it being
prevented from coming up to and discharging themselves in
contact with the metal, in obedience to the electrical attrac-
tion, by chemical forces of repulsion.

A double layer of this kind forms an actual condenser,
whose capacity is inversely proportional to the distance
between its faces. ; And Gauss’s well-known theorem relating
to magnetic shells shows, when applied to this case, that the
effect of such a condenser is to cause a sudden rise or fall of
potential in passing through it without producing any change
in the distribution of the electric force in the neighbourhood.

* Communicated by the Author,
t See his Wissenschaftliche Abhandlungen, vol. i., section Galvanismus,

and his Faraday Lecture, in the Journal of the Chemical Society for
1882,

Theory of Galvanic Polarization. 423

The notion of a condenser, therefore, gives a complete account
of the principal feature of the galvanic polarization.

Direct measures of the charge by Kohlrausch showed that
on dividing this polarization-fall of potential equally between
the anode and kathode plates, the distance between the faces of
the condenser comes out to be about the fifteen-millionth part
ofa millimetre; while more careful observations by Helmholtz
on cells in which absorbed gases have been removed from the
fluid, give the greater value ofa ten-millionth of a millimetre.
And Helmholtz makes out the very important fact that for
- all electromotive forces which do not exceed a certain moderate
value, the capacity, and therefore the distance of the surface-
layers, is very sensibly constant”.

2. The most accurate and convenient method of observing
the polarization at the common surface of two liquids is pro-
bably the electro-capillary method invented and applied by
Lippmann. |

When a surface of separation can persist between the fluids,
the energy, reckoned as potential, of pairs of particles close
to the surface must exceed that of the same particles when
in the interior of their respective fluids. The difference may,
as Gauss pointed out, be reckoned as surface-energy, and
specified by its amount per unit area of surface. If T re-
present this amount, it follows, as is well known, that the
forces which arise from it may be represented by a surface-
tension equal to T across each unit of length, tending to con-
tract the surface in all directions.

Now, if the common surface is polarized with constant
charges + Q and —Q on its two faces, there will exist an ad-
ditional electrical energy, which is also reckoned by its amount
per unit surface, and whose total value is

H= QV,

or, what is the same, Q?

=209 3
where § is the area of the surface and C is its electrical capa-
city per unit area.
The effect of this surface-energy will therefore, the system
being conservative, be represented by a surface-tension T’,
where dE

* Faraday Lecture, p. 296; Wissen, Abk. i. p. 858.
202

A424 Mr. J. Larmor on the Molecular’

i. é. is equal to the electrical surface-energy per unit area of
the surface, with negative sign.
The effect of the galvanic polarization will therefore be to

1D)
diminish the capillary surface-tension by this amount <:

In the actual case in which the polarization is maintained
by a battery, the difference of potential V between the faces
of the condenser is what remains constant ; and the system 1s
no longer conservative, because the battery can be drawn
upon; we have then

H=iC8y?,
and
ali *) 2
7 2lY F
i e
= 3 5
therefore dK=—Td58 ;

t. €. this force T’ now acts so as to increase the total energy
of electrification E,and is measured by its rate of increase per
unit extension of area; for the work done by the contractile
force T’ in an extension of surface dS is equal to —T’dS,
which is now the increment of the energy, but under the
previous conditions was its decrement.

It follows that under these circumstances the battery is
drawn upon for an amount of energy equal to twice that
required to do the electrical work of extension, viz. the energy
required to do this work together with the equal amount used
up in increasing H, as has just been found. ‘This is a particular
case of a general theorem of Sir W. Thomson’s. We have gone
into the matter here to show the consistency of the propositions
which make the capillary surface-tension equal to the rate of
increase of the ordinary surface-energy per unit extension,
while the electric surface-tension is equal to the rate of decrease
of the electrical surface-energy per unit extension.

Once the surface-tension becomes negative, a free surface
becomes unstable, and therefore practically impossible. We
notice therefore that, as the polarization is made stronger
and stronger, this state of affairs would finally supervene were
not the polarization previously relieved by electrolytic separa-
tion of the charged layer.

3. We have proved that the surface-tension is diminished
by galvanic polarization by an amount equal to 4CV’, where
C is the electric capacity of the surface per unit area.

The polarization-charge is therefore zero when T is a maxi-
mum, and the surface is then most curved.

The method that we have employed to determine the capil-

Theory of Galvanic Polarization. . 495

lary effect of the polarization-charge is different from that
used by Lippmann and by Helmholtz. In their mode of pro-
cedure the variation of the energy of the system is expressed
in terms of the variations of the surface-area and the surface-
density, and it is claimed that this expression is an exact
differential, 7. e. that any series of operations whereby the area
or density, or both, are changed so as finally to come back to
the original values, will also bring back the energy of the
system to its original value. This assumption seems to require
justification when it is remembered how complex such a series
of changes really is, and what a number of other variations
besides those of volume and density may enter into it. Helm-
holtz appeals to Lippmann’s experiments on the influence of
extension of the surface-film on its electrification and vice versé,
and to his capillary engine, as pointing in a general way to
the truth of the assumption.

In the method adopted above, we have proved the general
theorem that the mechanical action of two layers of positive
and negative electricity of equal amounts, spread over the
two faces of a flexible sheet, may be represented by a nega-
tive surface-tension of amount numerically measured by the
energy of the electrification per unit area. It follows, then,
on this representation of the phenomenon, that no matter
what other changes are taking place, the effect of the existing
surface-charges is to diminish the surface-tension of the sheet
by the amount just mentioned.

The case contemplated in the present application of this
general proposition is that of a sheet of uniform thickness ;
but we can clearly extend the result to flexible condensers of
variable thickness of dielectric, provided always that the
thickness be small compared with any radius of curvature of
the surface at the place considered. In this case the mecha-
nical effect on the condenser of a charge to potential V is to
produce a negative surface-tension, numerically equal to
KV’/8t, K being the dielectric constant and ¢ the thickness
of the sheet ; this surface-tension varies from point to point
of the sheet, and is at any place inversely proportional to its
thickness.

This result may also be at once deduced from the expres-
sion for the stress transverse to the lines of force in the
dielectric on Maxwell’s well-known theory.

4, Lippmann’s original form of capillary electrometer con-
sists of two mercury electrodes in contact with acidulated
water. One of the electrodes is in an extremely fine capillary
glass tube, so that the surface of contact is very small ; and
the other is of considerable area. It follows that when a
battery is applied, all the polarization and consequent change

426 Mr. J. Larmor on the Molecular

in the surface-tension practically takes place at the fine
electrode, as the corresponding charge at the other electrode
is spread over so much greater area. The change in the
capillary constant is measured by the column of mercury
whose pressure is required to restore the meniscus to its
former position.

Lippmann has given a series of observations with this
instrument in his paper in the Annales de Chimie, vol. v.
p- 507, the electrolyte being water containing one sixth part
by volume of sulphuric acid. He finds that the maximum
surface-tension is attained when the applied electromotive
force is *905 of a Daniell’s cell. This value, therefore, cor-
responds to absence of polarization at the electrode meniscus.
The following Table, calculated from his results, gives de the
excess (positive or negative) of the electromotive force above
this value, dp the excess of the pressure required to neutra-
lize its effect over its value when e=*905 D, and (6e)*/dp,
which is proportional to the capacity of the electrode, and
therefore inversely to the distance between the two electrified
layers,—on the supposition that the condensing arrangement
remains analogous to an ordinary condenser, viz. consists of
two infinitely thin layers separated by a dielectric sheet.

(de)?
o€. Op. op

— 89D 3433 23°07
— ‘88 , ool 230
— °86 3183 23-07.
— ‘805 2693 23°4
— “76 24724 23°3
— 71 2103 24-0
i) 63 170 23:3
— ‘54 12323 23°6
— 405 70 23°26
— 32 442 23°01
— 07 2 24-05
+ °35 574 21:3
+ ‘48 794 23'°3
+ “54 1193 24°4
+ ‘81 2302 28°5
+ 93 2482 34:8
+ ‘98 25424 34:2
+110 2642 458

The pressure supported by the tension of the meniscus
when e='905 D was 11083 millim. of mercury, which is
therefore proportional to the maximum surface-tension of the
film. The surface-tension, as ordinarily measured, corresponds
to e=0, and is therefore proportional to the pressure, which
was then 750 millim.

The last column of this table is in good agreement with
Helmholtz’s result, that for electromotive forces from zero up

Theory of Galvanic Polarization. 427

to a limit of considerable magnitude the capacity of the con-
densing arrangement remains constant.

As 6p is measured from a minimum value, it follows that in
the immediate neighbourhood of that value (de)? must vary as
dp ; so that the discrepancies for small values of 5p in the third
column are merely to be attributed to the special difficulty of
the observations in that part of the series.

Taking the second line of the table to give the average
value of this constant, we may calculate the thickness of the
dielectric, supposed to have the properties of vacuum, and
- therefore to have unit specific inductive capacity, on the sup-
position that the arrangement acts as an ordinary condenser.
When e=0, Lippmann found by direct measurement that the
surface-tension was ‘304x981 C.G.S. units, which there-
fore corresponds to p=750 millim. When e=:024 D, we
have de= —°88, dp =337 ; therefore the change of the surface-
tension corresponding to ée is

a ss phe
ola x 304 « 981.

This, as we have seen, is equal to the energy of the polariza-
tion charge per unit area. Now, taking a Daniell to be 1°1
volts, z.e. 1:1 x 10° C.G.S8. electromagnetic units, which is the
same as 11 x 10°~+(2:98 x 10°) C.G.S. electrostatic units of
potential, we have,
(Co ae

a 87r. thickness ’

therefore thickness of dielectric

_ (6e)?
~ 89.81”
Pi oot oof
=(—so5 —) + (8% --755 x 304 981),
__ (00825)?
Kvdgeoneely
~=°313x 10-" metre.

This calculation has already been made by Lippmann
(Comptes Rendus, 1882, quoted in Thomson and _ Tait’s
‘Natural Philosophy,’ 2nd edit., Appendix, ‘On Size of
Atoms’’). It gives an estimate of a molecular distance, viz.
that at which the two electrified layers are held by molecular
chemical forces, which, notwithstanding the very rough sup-
positions on which it is founded, ought to be of the true order
of magnitude ; and Lippmann has pointed out that it agrees
sufficiently with the estimates assigned by Sir W. Thomson
and others from different considerations. ; '

It is a satisfactory verification of the general notions in-

428 | Mr. J. Larmor on the Molecular

volved in this discussion to find that, notwithstanding the
large factors occurring in the calculation, such as the ratio of
the electrostatic and electro-magnetic units, it yet agrees so
closely in order of magnitude with the result 1 x 10~-!° metre,
obtained by Helmholtz from actual measurement of the polari-
zation capacity of platinum plates.

5. But, on the principles we have been following, we may
carry the analysis of the phenomenon still further. The
polarization consists in the transfer of charged particles to-
wards the electrode under the action of the electromotive
force, and they are finally brought to equilibrium at a distance
from the electrode, whose order of magnitude has just been
determined. As these equally charged particles repel one

another, they will tend to settle down in equidistant positions —

along the electrode surface. Instead therefore of two elec-
trified sheets analogous to an ordinary condenser, we have

really two sheets, one consisting of equidistant electrified par-

ticles, and the other of the charges brought opposite to them
on the electrode by induction. Hach charged particle and its
corresponding induced charge will be brought by their mutual
attraction so close together that this attraction will just be
balanced by the chemical forces which hold them apart.

For polarizations of sufficiently small amount, the sidelong
action of the neighbouring particles will be so small as to
have no appreciable eftect on the distance of any one particle
from the electrode surface ; because, in the first place, the
distances of neighbouring particles must be at first large com-
pared with the distance of two opposed charges, and, in the
second place, the smaller forces exerted by these neighbouring
particles must be resolved along the normal to the surface, in
which direction they have no appreciable component. The
radii of curvature of the surfaces are of course extremely
great compared with the distance between opposed charges.

It follows that as the polarization is increased the number
of charged particles over unit area of electrode increases in
the same proportion, and these particles all come to rest at
the same distance from the electrode surface, whatever be the
amount of the polarization. | And we can clearly expect this
uniformity of distance to hold good until the neighbouring
particles come within a distance of one another which is of the
same order as the distance of a pair of the opposed charges.

The pair of opposed surfaces which is thus arrived at, not
uniformly charged, but each with a system of equal isolated
point-charges arranged uniformly all over it, does not, of
course, act as an ordinary condenser in the sense of pro-
ducing a constant fall of potential in crossing it at all points,
in positions whose distances from it are of the same order as

4
¥ aed
- . <

~~ ee =

Theory of Galvanic Polarization. 429

the distance between neighbouring particles. But when we
compare two points on opposite sides at distances from it
great compared with this latter distance, it is immaterial
whether the distribution is supposed to be in isolated points or
uniformly spread over the surfaces. Therefore, as regards
points not in the immediate molecular neighbourhood of the
electrode, the effect of this polarization is still to produce
simply a difference of potential on the two sides, which is just
the same as if the charges were uniformly spread over the
surfaces at the actual distance apart.

These considerations, then, give a reason for the fact which
is brought out by the table given above, deduced from Lipp-
mann’s experiments with the capillary electrometer, and also
independently by Helmholtz from direct measurement of the
capacity of platinum electrodes in fluid with no dissolved gas
(which would disturb the action); viz. that the polarization
capacity is constant for all values of the applied electromotive
force up to limits of considerable magnitude.

6. In order to form an estimate of the nearness of the
neighbouring molecules on a face of the double sheet when
they begin to exert an influence on one another comparable
with that exerted by the opposite charges, we must assign a
limit to the interval of potentials within which the capacity
remains constant. ‘The table in § 4 shows that we shall attain
the correct order of magnitude by taking it to be, say, 1 volt
in the case there considered.

_ We may now make the following calculation, bearing in
mind that the sign = is to be interpreted as meaning that
the quantities are of the same order of magnitude.

Let ¢ be the thickness of dielectric layer ;

d the distance between neighbouring atoms when their
effective mutual action becomes comparable to that
between opposed atoms (the important part of this
action being that between any atom and _ the
neighbours of its opposed charge) ;

t/ the mean molecular distance in the electrolyte ;

e the constant aggregate charge of a single atom or

: radical ;
so that 7
t’-°e=the electro-chemical equivalent of 1 cubic centim.
of water, 7
10° i.
=a coulombs, approximately,
10? 9 1 ,
=5 x 3.10? electrostatic C.G.S. units,

and d “e = surface density.

430 Mr. J. Larmor on the Molecular
We have, then, for the condensing sheet,

where V=1 volt,

= 108--(3 x 10°) electrostatic C. G. 8. units.

Therefore
1 s 1
5. 10M ed 3 10°. Aunt ”
therefore

e ] -16
13 2 SS == e 10 e
Part 12

If now we write for ¢ the value found above, :3x107°,
and put ¢/ and d equal to each other, both being molecular
distances of the same kind, we obtain for either the value

15 x 107° centimetres,

which is very exactly of the same order as the value for
molecular intervals obtained already from the other con-
siderations.

On looking through this calculation it will be seen that
quantities which we have designated as of the same order of
magnitude do not differ nearly so much as in the ratio ten
to one.

7. The two estimates of molecular distance which have thus
been found on independent considerations connected with
galvanic polarization therefore agree within very close limits;
and they come very close to the third value determined by Helm-
holtz on the same theory of galvanic polarization, viz. 1 x 10~
centim. ; and they are also just below the superior limit as-
signed by Sir W. Thomson to molecular intervals from various
considerations connected with different physical phenomena,

viz. 10~° centim., his inferior limit being BS x 107° centim.

Sir W. Thomson’s different arguments lead to the following
superior and inferior limits of the average distance of mole-
cules from one another in solid and liquid substances :-—

Theory of Galvanic Polarization. 431

centim. centim.
Miminct electricity . . . - . 1Lx10~° : x 10-*
mertace-tension . . .°. .°. 5x 10~°
Kinetic theory of gases : x 107°
sceril 7 3 1 ae
Mees and liquids... . 10* 10 5 x LO

to which we may now add

Meteiolin® , O2f6.6 28 Pode 10S

while Lippmann’s method places the mean at ox 10°

centim.; and the other method here given places it at

a x 107° centim., with as small limits of error as any of the

methods given above.

8. The chief value of this discussion seems, however, to be
not so much that it gives an estimate of molecular distance,
but that its very close agreement with the other independent
estimates derived from considerations connected with the same
phenomenon of galvanic polarization is strong evidence of the
substantial and ultimate truth of that representation of the
phenomenon which has formed the basis of the discussion.

This argument seems to derive very great weight from the
wide variety and very different magnitudes of the physical
constants employed in the three calculations, one depending
on the direct measurement of the polarizing charge, another
on the direct measurement of change in the capillary constant,
and the third involving, in addition, the knowledge of the
quantity of electricity required to decompose a gramme of
water; while they all involve in different ways a constant of
such large numerical magnitude as the ratio of the two elec-
trical units of quantity.

9. The critical value -905 D in § 4 appears to have an im-
portant bearing on the much discussed question of contact
electrification.

As was pointed out by Helmholtz, a discontinuous change
of potential in crossing a surface can only be produced by the
existence of an electrical double layer on that surface ; so long
as we look upon electrification or electric distribution as the
cause of electrical phenomena, this is the only explanation open
to us.

It has been seen that this electrification represents a distri-
bution of purely surface-energy ; and if its properties are to

432 Mr. J. Larmor on the Molecular

be investigated, it is to be expected that much light will be
thrown upon them by their relations to other purely surface-
distributions of energy, of which the best known is that lead-
ing to capillary phenomena.

e are not required to explain the manner in which this
double layer at the surface of contact of two dissimilar sub-
stances is brought about. We may illustrate it by the rather
crude hypothesis that each molecule of an electrolyte consists
of a positively charged cation radical and a negatively charged
anion radical held together by electrical forces, but partly alse
by their forces of chemical affinity, so as to be analogous to a
magnetic molecule with north and south poles; that along the
surface of the electrode these molecules are all turned into the
same direction (polarized) by reason of the greater chemical
affinity of one of their constituents for the matter of the elec-
trode ; and that they thus form a double sheet analogous to a
magnetic shell. This illustration will at any rate show that it
is possible to give an account of the matter which shall be in
unison with the commonly received ideas of electrical and
chemical action, without having to speculate on the deeper
question of the relation of the material atom to its electrical
charge.

The electro-capillary observations of Lippmann quoted
above, and the later ones of Koenig for various electrolytic
fluids, show that, for one definite amount of polarization, each
of these fluids in contact with mercury shows a maximum
surface-tension. As we have seen that the existence of an
electrical double layer on the surface must diminish the sur-
face-tension, it follows that the critical value -905 D for Lipp-
mann’s acidulated water is that difference of potential which,
applied from without, just neutralizes the naturally existing
double electrical layer on the surface. It would seem there-
fore that the natural contact-difference of potential between
Lippmann’s mercury-and acidulated water is ‘905 D, and that
an absolute measure of a contact electromotive force has thus
been obtained.

Appendia.

The result that the mechanical effect of the electrification
on a charged condenser with thin uniform dielectric, whether
flexible or not, is equivalent to a uniform negative surface-
tension, has been derived in § 2 from the Principle of Energy
without the use of any analysis.

The same result will of course follow from direct calculation
of the mutual forces exerted by the charged elements of the
surfaces on one another. As it forms a good example of the

Theory of Galvanic Polarization. 433

theory of surface-energy which Gauss has made the founda-
tion of the doctrine of capillary action, in a case in which all
the circumstances of the forces are known, and as it also
illustrates some other points, the direct calculation is here
appended.

Consider, first, an infinite plane electrified surface, and
imagine a straight line drawn across it. The mutual repulsion
of the electrified parts on the two sides of this line will result
in a tension tending to tear the parts of the surface asunder
along the line, and whose intensity, measured across unit length,
we can calculate as follows.

Imagine a unit of electricity situated ata point distant &
from the line of division; the repulsion exerted on it by the
other half of the electrification is easily expressed in polar
coordinates, 7,0; @ being measured from the shortest distance
to the line of separation.

It will, however, be more convenient to take this unit charge
at a distance h from the plane, and to measure 7, @ from its
projection on the plane as pole. The repulsion exerted on it,
resolved parallel to the plane, is

3 6 ? cos @
af arf d a (i? be ya }? ae yr? ?

where p is the surface-density of the electrification, and
cos 6=& /r.
Therefore the repulsion-
== 2° ee gine :
é (7? + h?)2
To integrate this, write
rP—L=(7+h*)2;

therefore Bee
re +h
rdr= ize edz;
and the integral
ae :
= Z
> 1-2
z=1
E log 2 —:|

This quantity becomes infinite at the upper limit; so that
for an infinite plane sheet the tearing-force due to the electri-
fication would be infinite; a result which would also follow
readily from simple consideration of the dimensions of the
variable involved in the integral.

Suppose, however, we take a finite sheet bounded on the

434 On the Molecular Theory of Galvanic Polarization.

further side by the circular arc r=r,; the repulsion now is.

R= 2p log {(r?, +h?)3 + (7) — &)*} —p log (1? + &)

a pear
os G + 2)

As we have seen, this expression increases indefinitely as 7,
increases. But if now, instead of a single electrical sheet, we
had a double electrical layer with an intervening vacuum
dielectric of thickness ¢, the repulsion exerted by it on the unit
charge in the plane of the positive face will be equal to
dk
dh

But on differentiating the expression for R, it is obvious that
the first and last terms give parts which become zero when 7,
is infinite ; so that the repulsion of the infinite double layer
on the unit charge is finite, and is equal to

2tph
i+ 2
The repulsion exerted on a strip of unit breadth of density
p and extending from =0 to = therefore

” Qtoh
Sot eee
P| W+e

= ges

t

which is independent of h.
The repulsion exerted on a strip of the same double sheet is
therefore
2arp*t,

7. €. it is the electrical energy of the distribution per unit
area. And this quantity that we have thus calculated is
clearly the surface-tension required.

It is clear also that the stress across any line drawn on the

sheet is wholly tangential, and has no component normal to —

the sheet; so that this surface-tension is its complete speci-
fication.

The calculation just made has been only for the ease of an
infinite plane double sheet. Fora single sheet the distant
parts exert a finite effect; and we have seen that the stress
increases indefinitely when the size of the sheet increases.
But for a double sheet the parts very distant relatively to the
thickness no longer contribute sensibly to the result, and the
integrals converge. ‘Thus, if the double sheet be of sensible
but finite curvature, we may calculate the integrals either

os
ee

Pressure in Electrical Conduction and Decomposition. 435

from the sheet itself or from the portion of the sheet which
coincides sensibly with the tangent plane at the place consi-
dered, or from an infinite plane sheet coinciding with that
tangent plane. This is on the assumption that the part of
the sheet which coincides sensibly with the tangent plane is
of large dimensions compared with the thickness of the
dielectric, 7. e. that the latter is small compared with any
radius of curvature of the sheet.

The result obtained therefore holds for curved double sheets
as well as plane ones.

Now, if a curved sheet be under a uniform surface-tension
T, it is well known that the stress experienced by any element
858 of its surface is along the normal, and equal to

1 BPR |
T( e+ iz) 88,
R,, R, being the principal radii of curvature where 6S is
situated. When we apply this to the electrical double layer,
we obtain the same result as comes from the direct expression,
on Green’s theory, of the force exerted by the electrical system
on the two charged faces which belong to the element 6S.

Forasingle curved electrified layer of finite dimensions,
open or closed, the surface-tension is different at different
points, and at the same point across different lines on the sur-
face; except in the case of an electrified spherical sheet, in
which it is easily seen to be constant and equal to —7p’a,
where a is the radius of the sphere.

September 11, 1885.

L. On the Influence of Pressure on certain cases of Electrical
Conduction and Decomposition. By J. W. CLARK, Assistant
Professor of Physics in University College, Liverpool*.

A” the subject of electrolysis is to be brought prominently
before the next Meeting of the British Association, the

following short outline of an investigation, commenced some

years ago, may not be without interest at the present time.

The subject properly divides itself into three branches, viz.:—
the influence of pressure on the electrical conduction and
decomposition of —

I. Dilute sulphuric acid (products of electrolysis occupying
much greater volume than when combined to form the elec-
trolyte).

Ij. A solution of (e. g.) CuSO, (products of electrolysis

* Communicated by the Author.

436 Assistant-Prof. J. W. Clark on the Influence of

occupying about the same volume as when combined to form
the electrolyte).

III. The conduction of mercury.

Of these three classes, it will be noticed that the liquids in
I. and II. are composed of complex molecules which, accord-
ing to Clausius, are in a state of continuous motion varied by
collisions with one another which cause dissociation*. Mer-
cury, on the other hand, is a metallic conductor, and a liquid
which is usually regarded as being composed of simple or
monatomic molecules, which are therefore incapable of under-
going dissociation ; but by the application of a sufficient
external pressure its molecules may be brought nearer together,
or, in other words, the lengths of the molecular mean free
paths may be reduced.

The investigation to which I now wish to refer relates
to the decomposition of dilute sulphuric acid under an ex-
ternal pressure which it is my object to make appreciable
in comparison with the molecular forces of the liquid itself, for
it is only when that condition is fulfilled that any direct results
are to be anticipated such as those which are here sufficiently
suggested by the questions:—Is Faraday’s law independent
of pressure ? Do the conduction and decomposition of an elec-
trolyte always vary together? Can electrolytic action be stopped
by pressure? It was to obtain a reply to such questions that
I turned from the region of speculation to the surer ground of
experiment, and in the first instance set myself to measure
the electrical resistance of acidulated water, and the amount of
gas liberated from it by a known current under pressure.

The method of experiment in outline is as follows :—The
dilute sulphuric acid to be electrolyzed is hermetically sealed
in a short glass tube (0°1 centim. in internal, and 0°7 centim.
in external diameter), through the ends of which pass thin
platinum-wire electrodes, of which the upper wire is usually
so encased in glass that only the portion in the electrolyte is
exposed : a precaution found to be necessary in order to avoid
continuous recombination. This roughly describes the “ elec-
trolytic tube,’ in which the pressure upon the dilute sulphuric
acid is generated by the gases liberated by the current as they
accumulate in the (determined) volume of the tube which is
unoccupied by the liquid. When the circuit includes a gal-

* If volatility and a large coefficient of expansion for heat may be
regarded as a probable indication of a long mean free molecular path, it
does not seem impossible, on the hypothesis of Clausius, that under the
influence of a sufficiently great external Ete such liquids as condensed
gases (CO,, NH, SO,, &c., fused HgCl,), might be rendered good elec-
trical conductors, or at least have their conductivity increased.

ne eee ee
oat 2 aS

hr

Pressure on Electrical Conduction and Decomposition. 437

vanometer as well as a silver voltameter, variations of current,
as well as total current, can be measured ; and, finally, the
quantity of gas liberated in the electrolytic tube under the
accumulating pressure exerted by it upon the liquid is deter-
mined by allowing it to burst in a stoppered eudiometer (of

special construction) over mercury, so that the whole of the

gas is collected. From the volume of gas thus obtained the
bursting or maximum pressure can be calculated.

The principal results may be very briefly stated as follows: —

1. When such an electrolytic tube containing dilute sul-
phuric acid is fixed in a vertical position, the lower electrode
connected with the positive and the upper electrode with the
negative pole of a battery, the evolution of gas appears to
become less and less, and the electrical resistance of the con-
tents of the tube increases, until, with an E.M.F. of 30 volts,
the needle of a galvanometer in circuit shows only a very small
deflection. This action appears to be due to the formation of
very dense sulphuric acid (H,SO,?) at the + electrode at the
bottom of the tube, whilst the water from which it has been
separated forms a layer above it of so high resistance as to
almost absolutely stop the passage of the current. This con-
dition must not be mistaken for the cessation of electrolytic
decomposition produced by pressure.

The strong acid is produced at the + electrode by the action
of the current, and owing to the small sectional area of the
tube, though it again mixes with the liquid above it by diffu-
sion, it does so less rapidly than it is separated by the current.
Moreover, under a sufficient pressure diffusive rate is lessened
as the length of the mean free molecular path is decreased.

The above singular action affords a means of concentrating
sulphuric acid without boiling.

2. When the poles are reversed, the dense sulphuric acid
is formed by the current at the upper electrode, and thus
becomes again mixed with the rest of the liquid in the
tube as it descends through the liquid, the mixing action
being facilitated by the bubbles of hydrogen rising from the
lower electrode. In this way the separating action of the
current is prevented, and my glass tubes (which burst at about
290 to 300 atmospheres) fail to withstand the pressure of the
gases liberated by the current. I sought to strengthen one
glass tube of special form by first coating it with pure silver
by Martin’s method, and then thickly electrotyping it with
copper; but it also burst. It is certain, therefore, that a
pressure of 300 atmospheres is insufficient to arrest the elec-
trolytic decomposition of dilute sulphuric acid. I am not yet
certain whether pressure exercises any direct influence on the

Phil, Mag. 8. 5. Vol. 20. No. 126. Nov. 1885. 21

438 Pressure in Electrical Conduction and Decomposition.

electrical resistance and on the quantity of gas liberated from
dilute sulphuric acid, but what there is is slight. There seems
to be a small decrease in the resistance, but I cannot yet say
to what this is due; for this part of the investigation is in-
complete, and it is exceedingly difficult to distinguish and
separate between possible causes.

3, Ozone is present in the gases liberated by the bursting
of the electrolytic tube. Ihave not estimated it quantitatively,
but I have no reason to think it present in unusual amount.
Under some conditions, e. g. during the formation of ozone
under the influence of the silent discharge, pressure might
facilitate the condensation of oxygen to ozone.

It may, finally, be needful to increase the pressures employed
in this investigation still more, and to employ vessels of steel
or of some other material. The whole investigation is one of
great difficulty, but I shall probably publish a fuller account
of the methods and results of which the above forms a short
and very imperfect summary”.

Addenda to the Author’s paper ‘‘ On certain Cases of Electro-
lytie Decomposition,” Phil. Mag, July 1885.

Page 38, line 5 from top, omit binary.

Page 38,add:—Dr. Gore has discovered that heated argentic
fluoride commences to conduct while still solid, and that when
fused its electrical resistance is very small. He appears to
regard the conduction in both cases as being unaccompanied
by decomposition, 7. e. non-electrolytic. 7

Page 48, line 15 from top, for (and Hg,I) read (and
Hol, ?).

Page 39, line 6 from bottom, add :—It is of interest to
remark, m connection with the electrolytic decomposition of
glass at 100° C., that I have recently learnt from Mr. Hicks
that it may also be annealed at this temperature.

Pages 40 and 45, add:—Amongst the properties of fused

* To the proof I have an opportunity of adding that, since the above
was written, I have found a short paper by M. Bouvel in the Comptes
Rendus, t. Ixxxvii. p. 1068, wherein he states that he has found experi-
mentally that :—

(1) La décomposition de l’eau par un courant est indépendent de la
pression.

(2) La quantité de l’électricité nécessaire pour décomposer un méme
poids d’eau est sensiblement la méme quelle que soit la pression & laquelle
s’opere la décomposition.

M. Bouvel verified these statements up to 200, and found that the
decomposition was not stopped by a pressure of 360 atmospheres.

rr

Influence of Gulf-Stream on Winters of Great Britain. 489

mercuric iodide and chloride which led me to infer that the
molecules of these substances may possess an unusually long
mean free path of liquids, [ omitted to mention their large
coefficients of expansion for heat.

August 8, 1885.

LI. On the Winters of Great Britain and Ireland, as influenced

by the Gulf-Stream. By Prof. Henry Hennessy, /.2.S.*

N the ninth volume of the Proceedings of the Royal
Society, p. 324, is printed a letter which I wrote to the
late General Sir Edward Sabine, on the influence of the Gulf-
stream on the winters of the British Islands. I pointed out
that if abnormally cold or warm winters are due to changes
in the condition of the comparatively tepid currents bathing
our shores, then during cold winters the differences of tempe-
rature between the Northern and remaining coasts should be
greater than during mild winters. A comparison of the
observations made in the months of December 1855, 1856,
and 1857 presented precisely this result. The winters of
some recent years having exhibited abnormal features as to
temperature, I have made a comparison of the results pub-
lished by the Meteorological Office, of which the following
may be taken as the leading facts.

In comparing the coast-temperatures of Great Britain, it
should be remarked that the island is shaped in an irregular
triangular outline, of which the summit is in North Scotland,
and the base the south coast of England. For temperature-
comparisons it is therefore best to compare the South coast
with the West coast and North-east coast.

In the winters of 1871 and 1881 the month of January was
cold in Great Britain, and the following results have been
grouped as indicated :—

1871. © (Cold.)

SoutH Coast. WEstT Coast. NortuH-East Coast.
Helston ...... 41-1 Barnstaple .... 37-0 Holkham .... 82-2
SS ae 38°9 Llandudno.... 36:1 Boston 3s. .... 31°6
Sidmouth .... 366 Liverpool ..,. 33°5 2) a o1'3
Eastbourne... , 35:2 Cockermouth .. 33°'7 Whitby 2... sit
eminor...... 36°8 Ballot 5... . 33°0 emetds.- 2... 33°0
Osborne ...... 35'0
Bournemouth... 35:1
Worthing .... 34:9

Mean .... 36°7 3 Mean .... 34°7 i Mean At sek eee

Excess of South over West coast.......... 2:0
Excess of South over North-East coast .... 4:8

Communicated by the Author.
212

440 Prof. H. Hennessy on the Winters of Great Britain

1881. (Cold.)

SoutH Coast. West Coast. NortTH-East Coast.
DOVE . ws ., 850 Pembroke .... 36-0 Yarmouth .. 32:5
Hastings... .. 35°2 Holyhead .... 36:4 Gelderston .. 31:7
Southampton... 33:4 Llandudno.... 34:9 Spurn Head .. 33:2
Hurst Castle .. 34°6 Douglas ...... 33°2 Shields...... 309
Plymouth .... 35°6 Liverpool .... 315 Leith ....26)eee
Falmouth .... 38°5 Siloth: ft 2% 27°5 Aberdeen.... 28°5
Prawle Point.. 37:1 Barrow |..%% .% » Svat Nairn?) 388 28:6

Laudale ...... 31:5 Alnwick .... 28°6
Ardrossan 31°6
Mean .... 85°6 Meany.4..4.4 32°8 Mean .... 80°5
Excess of South coast over West coast ...... 5:8

Excess of South coast over North-East coast.. 51

For Ireland, the following results have been obtained for

January 1881.

NortTH CoaAST.

(Cold.)

SoutH Coast.

Mullaghmore ...... 35:2 Roches Point....... 87-0
Londonderry ...... 30:9 Valentia»...: 44) ee 39:0
Donaghadee ...... 35°2
PRES AN hes ue aie 33°8 Mean: «i aureene 380
Dlferenen ay sit/statess ‘ £2

In 1883 the three weeks ending December 31 were warm,
and the following were obtained for Great Britain :—

SouTrH Coast.

RPGR LS Roa
Hastings......
Southampton..
Hurst Castle ..
Plymouth ....
Falmouth ....
Prawle Point. .

LLP PP PP
C9 C1 G9 AD RR Ro
G0 bD CO 0D © GO Od

Mean .... 43°0

West COAST.

Pembroke ....
Holyhead ....
Llandudno....
Liverpool ....
Barrow oo as
Silloth
Ardrossan ....
Lavgale «ss vs

oeee#ee

NortrH-East Coast.

Sumburgh Head
Stornoway ....
Wick: 3.17) aaa

Nairn

a oY eg ae
Bom tsts ahh.
SOOT OKEA

Scarborough .,.
Spurn Head....
Yarmouth. . 75

Excess of South coast over North-East coast .... J]

Defect of South coast compared to West coast ..

Excess of South over all the rest .

During the same month (December 1883) the following

results were obtained in Ireland :—

43-5
41-0

and Ireland, as influenced by the Gulf-Stream. 441

NortH Coast. SouTH COAST.
Londonderry........ 43-9 Roches Point...... 45°83
Mullaghmore ...... 44:6 Malema ee tise ys 46°7
Donaghadee ........ 43°]

Pera, res 43°9 Meant seh... 46-0
Excess of South over North coast...... rina

The mean temperature at Dublin was 43°°6 ; whence, if
this may be taken as about the temperature of the East coast,
the North exceeded the Hast by 0:3.

If we now compare the cold and warm periods, we observe
that both in Great Britain and Ireland the difference between
the Southern and Northern coasts was less in the warm period
than in the cold; and the influence of warm currents was also
exhibited in the higher temperature of northern over eastern
stations at lower latitudes. The conclusion arrived at in my
former comparisons has been thus supported, and the influence
of the thermal currents surrounding these islands in modifying
the winters appears to be further confirmed.

In a paper of General Sabine’s, to which I have referred in
the communication printed in the Proceedings of the Royal
Society, he mentioned the winter of 1845-46 as exceptionally
mild. Some confirmation of the law I indicated may be de-
rived from some observations made during that winter on the
south-west coast of Ireland by Mr. Daniel O’Connell. The
results were communicated by the son of the illustrious orator
to the late Dr. Humphry Lloyd; and having consulted the
record, I found that Mr. O’Connell observed the temperature
daily at Darrynane Abbey for 106 days during the winter of
1845-46, from November to February ; on applying the cor-
rections recommended by Dr. Lloyd in his paper on the
Meteorology of Ireland, I found that the mean for the winter
at Darrynane Abbey was 4/°. During the four winter months
referred to, the mean eeneteres were as follows at other
stations :-— |

oc et ihie) Daag
ra, Or eT aes
Swansea . 444
St. Heliers (Jersey). 46°4
Onin: 41°8

~The winter of 1845-46 seems therefore to furnish an illus-
tration of the law of temperature-distribution already men-
tioned, and therefore of the influence of the heat- -bearing

: Be rrenia which wash the shores of these islands.

| 442 J

LIL. On the Comparative Temperature of the Northern and
Southern Hemispheres of the Earth. By Prof. Henry
Henvessy, /.2.S.*

ORE than twenty years have elapsed since I published

a systematic proof of the superiority of water over
Jand as a substance for absorbing and gradually diffusing
throughout its mass the heat received from the sun. The
superficial portions of the earth as a whole, including the
atmosphere in contact with the water, would be thus influ-
enced, and in this way the physical properties of water were
shown to have a predominant influence on terrestrial climate.

The properties of water here referred to are :—(1) Its excep-
tionally high specific heat ; (2) its moderate permeability to
luminous rays of heat contrasted with its impermeability to
obscure heat ; and (3) its mobility. From Pfaundler’s obser-
vations and experiments it appears that the specific heat of
the soil, which forms the general surface of dry land exposed
to the atmosphere, to sunshine and planetary space radiation, —
varies from 0°19 up to 0°5. The majority of soils have spe-
cific heat of from 0°25 to 0°28. Thus the absorbing-power
of dry land for radiant heat is in all cases much less than that
of water. At night land more readily parts with the heat
acquired under sunshine, and from its immobility it cannot
directly transport by convection or by currents any of the
heat it acquires to other parts of the earth’s surface. The
conclusions at which I arrived were entirely different from
those which were generally received anterior to the time of
the publication of my views. The generally received notions
at that time are best summed up in the words of Sir John
Herschel :—‘‘ The effect of land under sunshine is to throw
heat into the general atmosphere, and to distribute it by the
carrying-power of the air over the whole earth. Water is
much less effective in this respect, the heat penetrating its
depth and being there absorbed, so that the surface never
acquires a very elevated temperature.” f :

These words occur in the latest edition of the work from
which they are quoted. They furnished Sir Charles Lyell
the groundwork of some of his ingenious reasonings regarding
Geological Climate ; but my conclusions were so different
that they might be ‘tly expressed in terms diametrically

* Translated, with additions, from the Comptes Rendus of the Paris
Academy of Sciences for September 1882.
} ‘Outlines of Astronomy.’

On the Temperature of the N. and S. Hemispheres. 443

opposite, and in such terms I have in fact expressed them
long since as follows:—The effect of land under sunshine
is to rapidly throw off the heat it receives into the upper
regions of the atmosphere and the interplanetary spaces both
by day and by night; and thus, although it causes a consider-

able increase of temperature in the strata of air over it by day,

it is not well adapted for storing up and retaining heat.
Water is much more effective in this respect; the heat pene-
trates to a greater depth within it than on dry land, and it
becomes more completely absorbed, owing to the far higher
capacity for heat of water and the difference between its
diathermanous action on the luminous heat-rays entering it
from the sun compared with its action on the obscure rays
quitting the particles in the interior of its mass.

Among the consequences following from my theory of

terrestrial climate was one which seemed difficult to recon-

cile with the facts accumulated at the time this theory was
brought forward. I refer to the supposed higher temperature
of the northern as compared with the southern hemisphere.
This was formerly attributed to the heating-power of the
great masses of land north of the equator. I could not
account for this inequality unless by supposing that oceanic
and aerial currents which had not been as yet fully examined
transported some of the heat absorbed south of the equator to
its northern side. This supposed superiority of temperature
in the northern hemisphere was employed as an argument
against my conclusions.

In 1875, at the meeting of the British Association at
Bristol, I brought forward my views for the purpose of estab-
lishing my priority in regard to some of the conclusions which
had been frequently reproduced at different inquiries in Great
Britain without any reference to their original source. In
the discussion which followed one of the speakers, Prof.
Hverett, is reported to have asked, “‘ How, assuming Professor
Hennessy’s theory without acknowledgment to be correct, it
could be reconciled with the generally accepted fact that the
temperature of the northern hemisphere was greater than that
of the southern. Professor Hennessy denied that this was a
fact, but supposing it were so, he would attribute it to oceanic
currents.”

It now appears that the notion of any superiority of mean
temperature in the northern as compared to the southern
hemisphere must be abandoned. When this opinion gained
currency the number of temperature-observations made south
of the equator was very small compared to the number
recorded in northern latitudes. Of late years a considerable

444 Dr. J. J. Hood on Retardation

number of observations on the temperature of the sea and air
have been collected and published by the Government of the
United States of America. From these observations the
result has been deduced that the difference of temperature
between the two hemispheres is insensible, and probably
slightly in favour of the higher mean temperature of the
hemisphere which possesses the largest water-covering*. In
this way my theoretical views have been fully confirmed by
the crucial test of leading to conclusions which did not seem
likely to be true at the time I originally placed them on
record.

Some further illustrations of the question under consider-
ation may he obtained from the results of observations on
the distribution of a well-known class of plants. The condi-
tions most favourable to the growth of the larger ferns were
recognized by Robert Brown and other eminent botanists to
be humidity, shade, and uniformity of moderately elevated
temperature. These conditions exist in their most perfect
form among the smaller islands of the great oceans. In
islands like New Zealand and others of inferior size the tribes
of plants alluded to are widely spread and highly developed.
The relative distribution of these plants in the northern and
southern hemispheres is highly instructive. The extratropical
regions of the northern hemisphere contain thirteen times as
much land as the corresponding portion of the southern hemi-
sphere, and in the latter arborescent ferns are known to grow
much further from the equator than in the former. |

Professor Lindley remarked long since that at the time of
the ade of the Lias formation, a geological epoch of
somewhat higher general temperature than the present, the
vegetation was similar to that of the southern hemisphere in
the Pines as well as in the Cycads. !

LIII. On Retardation of Chemical Change. By JouNJ. Hoop, |
D.Sc. (Lond.), Assoc. Royal School of Mines.

ae a short paper published in this Journalt some time ago,
on Retardation of Chemical Action, it was shown that

* Dr. J. Hann has published in the Proceedings of the Academy of
Vienna a good résumé of the facts as to the distribution of temperature
in the two hemispheres of the earth, and he concludes that their tempera-
tures are almost equal. See Ferrel, American Journal of Science, August
i page 89, “The Relative Temperature of the two Hemispheres of the

arth.”

+ Communicated by the Author.

{ Phil. Mag. [5] xii.

of Chemical Change. 445

the rate at which ferrous-sulphate solution is oxidized by
potassic chlorate is retarded in a remarkable manner by the
addition of various sulphates. ‘The principal results obtained
were these—that the amount of such retardation is propor-
tional to the quantity of the sulphate added, and that certain
groups of analogous sulphates produce, for equal weights, the
same retardation-effect. For instance, equal weights of the
sulphates of sodium, potassium, and ammonium were found
to produce the same amount of retardation ; so also did the
potash- and ammonia-alums ; whereas the sulphates of zinc
and magnesium, although classed together as analogous salts
in the same sense as those of the alkali metals, produced dif-
ferent retardations. The natural inference that was drawn
from these experiments was, that the study of retardation
might afford a means of classifying chemical substances on a
dynamical basis, and of determining for each salt or group of
salts a numeric, the coefficient of retardation, of a character
somewhat similar to what Mills* has termed the “bergmannic”’
of a salt. Owing to the small number of soluble sulphates
that can be employed in the above reaction for studying their
retardation-effects, it is very limited in its application ; conse-
quently search was made for other reactions that could be
employed for the same purpose, of such a nature that nitrates
or chlorides could be experimented with. Of the several
reactions that were tested for this purpose, it was found that
the oxidation of ferrous chloride by potassic chlorate was by
far the best, being analogous in all respects to that of the
oxidation of ferrous sulphate, and capable of being rendered
as quick or as slow, by alteration of conditions, as might be
necessary for the purpose in view ; besides, the list of soluble
chlorides whose retardation-effects could be studied is a com-
paratively large one. The results, however, that have been
obtained with these salts are such as have modified somewhat
the author’s original notions concerning retardation and the
molecular movements that are generally supposed to take
place in a system undergoing chemical change.

It has been usual, in establishing the formule that are em-
ployed in the study of the rate of chemical change, to ignore
the fact that the products of the reaction, however inactive
they may be in the chemical sense, may retard the rate of the
change considerably. In the experiments with ferrous sul-
phate referred to, it was shown that the introduction of so

small a quantity as one gram of the sulphates of the alkali

metals in 260 cubic centims., the volume of the experimental

rE nil, Bag. (5) 1.

446 Dr. J. J. Hood on Retardation

solution, caused the oxidation to progress at a rate 10 per cent.
less than when no such salt was added ; an effect which might
be produced even to a much greater extent by the products
gradually formed during the course of any chemical change,
especially where the quantities of material in unit volume are
comparatively large, as, for instance, in the many cases of
etherification that have been studied dynamically during the
last few years, and might consequently vitiate somewhat the
inferences drawn from such experiments. |

As it was found in the experiments alluded to that the
amount of retardation was proportional to the amount of the
sulphate added, it would naturally be supposed that the pro-
ducts formed during a reaction would produce a retarding
or accelerating effect (for the latter is possible, as these
experiments show) proportional to their quantity ; and it is
easy on this hypothesis to introduce this effect into the
equations.

Suppose in a chemical system undergoing change there are
n different bodies taking part in the reactions, and the mea-
surements that are made of the progress of the change relate
to one of these n bodies. Let A,, A,,...A, be the initial
quantities of the active substances, and at time ¢ from the
commencement of the change a, a,...a,, the quantities of
these that have become chemically inactive members of the
system, and let the measurements that are made relate to the
body A,. Attime ¢ the quantities of these substances that are
still capable of reacting one with the other will be A,;—«,
A,—@,,...A,—a, 3 and the inactive products resulting from
the changes that have taken place will be a,+a.+...@,. Then,
on the usual hypothesis that the amount of change that takes
place in unit of time with regard to any one of these bodies,
A,, is proportional to the product of the active substances, and
that the retardation- or acceleration-effect of the products of
the reaction is proportional to their amount, the equation
representing this would be written

Pthati: (A,—a,)(A,—a,)... (An—an) (1)
dt 8 BEG, FN ag c.g ee

where 0’, ”, &. are the coefficients of retardation or accele-
ration of a, a, &e., and + is taken according as these
effects are all of the first or of the second character. The
nature of the effect of each individual product a, ae, &e.,
whether +, could be determined experimentally by adding a
considerable quantity, and noting the difference in the rate of

| of Chemical Change. 447

such an experiment and of one where no such substance is
present excepting what is formed during the reaction.

If the initial quantities A,, A,2,...A, are in equivalent
proportions, they may be expressed in terms of A,, thus
A,=«A,...A,=e,A,. But suppose for the most general
case that they are multiples of these quantities, or Ay =v,e,A,...
An=VnénAn 3 %1, &2,...&n, however, are always in equivalent
proportions, or a, =€«,...é,=6€,0". Substituting these values,
(1) becomes

da, pu(vje; A, —€ye,) 2 (Ap —&) «- (Vnen An — En&x)

met B tat, (Ayey HAgeg. «+ + Anen)
Writing A,—«,=y, the amount of A, that remains active at
time ¢, and = =— oy. This equation may be written in the
form
dy _ _wyiu-lAtyt {m—-lAtyt ... {@n—DAty} (2)
dt b+y

The reaction under consideration (the oxidation of ferrous
chloride by potassic chlorate) is represented by the equation

6 FeCl, + KC10, + GHCl=3Fe,Cl, + KC1+3H,0 ;

_ and consists of a system of three active members, in which the
iron, or y, was made the subject of measurement. In the
preliminary experiments that were made, the ratios of the
_ chlorate and acid to the iron were varied, in order to find the
_ quantities, and consequently the rate, of the oxidation most
_ suitable for bringing out the retardation-effects of the chlorides.
_ Some experiments were made with all three materials in equi-

valent quantities which would be represented by the equation
| ay 2 eg

dt by

3

But such conditions were unsuilable; for the experiments
showed, as is evident from this equation, that, as the rate .
varies as the third power of y, at the start the reaction pro-
ceeds very rapidly, and quickly becomes very slow, a state of
matters most inconvenient for the purpose in view. It was
ultimately found best to employ different multiples of equiva-
lents of acid and chlorate for the different purposes of studying
retardation of sulphates, of chlorides, and influence of tempe-
rature, as given below.

The preliminary experiments with the chlorides showed

448 Dr. J. J. Hood on Retardation

that these substances, even in large quantity, influenced the
rate of the change comparatively little; and that consequently
the term in (2) relating to the retarding or accelerating effects
of the products of the reaction could be neglected ; as indeed
the calculations from the experimental numbers showed to be
true, as is illustrated in the experiments with MgSO, given in
detail below.

For part of the work relating to the retardation of sulphates
and influence of heat, the following equation was employed:—

Y= yi(n—1)A+y} { 4-1)

_ which, on integrating, gives

(vy,—1)A
Y

y+ (%—I1)A }
ft (v,—1) logy — (1) logit IS bo tt

whilst in the experiments with chlorides the iron and chlorate
were in equivalent quantities, and the acid several multiples,
represented by the equation

dh
ae = ay? {(%y—1)A+yt,

which, on integrating, becomes

f/{ (4-1)A- aie logy =y(gy +t). (4)

' Taking either of these equations (3) and (4), it is easy to
see that = af-1, or that the time required to oxidize the

iron from 7/ to y” is proportional to f; consequently in a
series of experiments in which everything is the same except-
ing the presence of inactive salts, by a comparison of the
values obtained for f with its value for an experiment in which

* no such extraneous salt is added, or a blank experiment, a
measure is obtained of the effect of such a salt on the rate of
the oxidation ; so also in the case of variations of temperature
a measure is obtained of the accelerating effect of heat on the
rate.

Sulphates.

The original design of this investigation was to determine
the retardation-effects of the various soluble chlorides. The

of Chemical Change. 449

results, however, obtained for these salts, to be found further
on, being of a rather anomalous character, the influence of
sulphates was made the subject of a long series of experiments,
the results obtained being contained in Table I.

Hach experimental solution had a volume of 110 cubic cen-
tims., and contained 3736 gram Fe” as chloride, and 1°321
gram free HCl. To this was added 25 cubic centims. KCIO;
solution, equal to °3406 gram, making the total volume =135
cubic centims., care being taken to have the solutions at the
same temperature before mixing. From such a solution 10
cubic centims. were withdrawn, run into a small flask contain-
ing a few cubic centimetres of a nearly saturated solution of
MgSO,"*, and titrated with permanganate.

The chemical conditions represented by these quantities of
materials are

6 FeCl, + (2°5)K C10, + 6(5°425 HCL.

The value for A was determined by diluting 10 cubic cen-
tims. of the stock solution of ferrous chloride (=°3736 gram
Fe”) to 135 cubic centims., and titrating 10 cubic centims. of
this with the permanganate, the number of cubic centimetres
required being taken as equal to 10°5, the value for A; but

_ although only approximately of this strength, the usual pro-

portional corrections were made on the values for y for each
experiment ; in which way all the experiments are rendered
comparable with each other, as if performed with the same
standard solution of permanganate.

Inserting these values of v4, =2°5, v,=5'425, and A=10°5
in equation (3), it becomes

f (4425 logio —1°5 logig

vee =gtt. (5)

+ 46°46
y : )

The following three experiments will illustrate this formula,
and how the retardation-effects are determined with magnesic
sulphate as the retarding agent. Employing the first and
third observations, the values obtained for f and g in equation
(5) are 115:1 and 77:2 respectively ; and side by side with
the observed values for ¢ are the calculated quantities.

* Besides tending to stop the reaction, the addition of MgSQ,, the
author has shown (Chemical News, vol. 1.), renders the estimation of iron
by permanganate in presence of free hydric chloride perfectly reliable.
Mr. W. H. Deering has drawn attention to the fact that Zimmermann
had previously proposed MnSO, for the same purpose.

450 | Dr. J. J. Hood on Retardation
Blank. 10° C.

Permanganate, Time, Time,

y. minutes. calculated.

10°31 0 ded
7:88 21°3 20°9
6°45 38°4 isd
4°72 69°3 68°6
3°08 iepiaeany 25 115-0
211 1621 1601
1°58 198°4 196°5
1:10 247°0 244:°0

In a similar manner the values for f and g when 1 gram
MgSO, was added, the total volume being the same as before,
are found to be respectively 131:9 and 88:3.

1 gram MgSO, Temp. 10° C.

Permanganate, Time, Time,
y: minutes. calculated.
10°33 0 ces
8°16 21:0 21-1
6°60 41:8 oa
5°30 65:2 65°5
3°54 114°3 1140
2°55 156°0 1573
1:98 191-0 192°6
1-40 241-0 —243°3
With 5 grams MgSO, the values found are f=201°3 and
== *6
| ghiates 5 orms MgSO, Temp. 10° C.
Permanganate, Time, Time,
Ys minutes. calculated.
10°34 0 ae
8:19 311 31°4
760 42°3 ag
6°51 65°7 66°1
4°91 113-0 113-4
3°81 158°1 159°9
| 3°16 194-4 196-4

2°50 245'2 244°3

of Chemical Change. 451

Now since / varies inversely as the rate of change, by com-
paring the above three values the ratios are found to be

115°1 ;: 131°9 : 201°3=100 : 100+14°6: 100+5x14°9 ;

showing that the amount of oxidation that takes 100 minutes
to be performed in the blank requires 114°6 minutes in the
presence of 1 gram MgSQ,, and 174°9 minutes with 5 grams
of the salt; or that the retardation is proportional to the
quantity of the salt added, and equal to about 14°7 per cent.
per gram. ‘These three experiments are shown graphically in

the figure.

Curves showing Retardation produced by 1 and 5 grams MgSQ,.
3 Temp. 10° C.

Unoxidized iron =y,

|

40 80 120 100 200 240
Time, minutes.

The following table contains the values of f and g for equa-
tion (5), the mean in each case of several experiments, when
2 grams of the various sulphates were employed, the tempe-
rature being in every case 10° C,

45De Dr. J. J. Hood on Retardation

TABLE I,
Salt, | e | g. Time to oxidize
2 grams. iron from y toy’.
ao ee fis8 772 100-0

Am,SO,..... 156-2 105-0 134'8
ORO! 61.42 155:9 104:4 134-6
Na,SO, ....-. 155-4 104°2 134-2
BO, °.,.:.: 155°3 103-4 134-1
MgSO, ...... 148-7 99-2 128-4
Ti oe | 141°3 94:1 122-0
SiS, cua ley Laas 88:2 114-0

The noteworthy points exhibited by these numbers are
that equal weights of the sulphates of ammonium, sodium,
potassium, and lithium produce equal retardation-effects of
approximately 17:2 per cent. per gram ; whereas the sulphates
of zinc, cadmium, and magnesium, metals usually grouped
together in the same sense as the alkali ones, all produce
different effects. If the percentage retardation produced by
one gram of a salt be defined as its retardation-coefficient, the
values for the above salts are, for the alkali group, 17:2 ; for
ZnSO,, Lt 3 CdSO,, ri 3 and for MesO,, 14°2.

The study of the influence of heat on the values of these
retardation-coefficients it was considered would be of interest,
in view of the general hypothesis regarding the intermolecular
motions that take place in a chemical system, as well as the
accelerating effect of heat upon these movements.

The numbers that are given in Table II. relate to MgSO,,
and are selected from a considerable number of experiments
in which various sulphates were employed; but the facts
brought to light were all of the same general character as
with MgSQ,. |

In these experiments, instead of 24 molecules of chlorate
only 15 was used, the equation representing the experiments
being consequently

oy: y+5°25 + 46°46
i 2 (4 425 logy y —"5 logyo ipo) =9 +.

of Chemical Change. 453

TaBLeE II.
Blank. 2 grams MgSO,,.
Temp. O. |

Cane eal wate gf.
6-13 366-7 Ly daa 478°6 199-3
10 256-4 106:4 335°8 139-7
Bas Wate? 80:7 256°6 106-9
17 134-3 56:3 178-7 74-7

The relations among these numbers are shown in the fol-
/
lowing table aa gives the retardation produced by 2 grams

MgSO, at the different temperatures, or the time required to
oxidize the iron from y to y’, being for the blank experiment
equal to 100. In column a the values for « for the blank
experiments are given, calculated on the assumption that the
relation between rate of change and temperature is p=pe’,
and in column 0 the values for 2 when the experiments are
retarded by magnesic sulphate.

TABLE III.
aeien. O: J @ b.

: é
6:13 130°5 |

10 130°9 1:097 1:096

13 1334 1:098 1:094

17 133:0 1:096 1:095

ET eee a, 1:097 1:095

From the above table it is seen that the retardation-
coefficient of MgSO, increases slightly with the temperature,
from 15:2 at 6°13 C. to 16°5 at 17° C., the rate of oxidation
between these limits of temperature increasing about three-
fold ; also that the rate increases in geometrical progression

with temperature both in the blank and in the retardation
experiments, the relation being in the former p=p(1-097)’,
and in the latter p=p’(1°095)*. It has been shown* that

* Phil. Mag. October 1885.
Phil. Mag. 8. 5. Vol. 20. No. 126. Nov. 1885. 2K

when ferrous sulphate is oxidized by chlorate, the same rela- |)

454 Dr. J. J é Hood on Retardation

tion p=4(1'093)* holds good. | =
Chlorides. am |

When the present experiments were undertaken, the prin- }
cipal object in view was to determine the retardation- |
coefficients of the different chlorides; and it was conjectured jj
that, with the evidence obtained from former experiments, |
these salts would be found to group themselves together as |
regards retardation-effects. Such a grouping, however, has |
not been found to exist. Indeed the anomalies that the expe- |
riments exhibit are of a very striking character. For instance, |
instead of a retardation, the presence of certain chlorides has
been found to produce a considerable acceleration on the rate
of oxidation ; whilst in another case, that of sodic chloride, |
neither retardation nor acceleration occurs. nee |
Instead of employing quantities of iron, chlorate, and acid |

in accordance with equation (5), it was found that, by using
only one molecule of chlorate and so rendering the rate much |
slower, the effects of the chlorides, being comparatively small, |
were better brought out. The acid being the same in amount
as in the experiments with sulphates, orv,=5°425 and A=10°9, |
inserting these values in (4), the equation becomes |

+46°46
f (46-46— ai login 1S )=¥(g +9)

In the following table are given the values for f and g, the |
means of several experiments, the temperature being 10°C. |

a

TABLE LY.
| i
| _ Salt, f Time to oxidize
| 5 grams. ; | g- iron from y toy’.
ae 49°36 136°3 100
aaa | 49°72 136-9 100°8
<8 Baie 60:30 166:3 ° 122-2
1 oS, ae | 51-70 142-2 104-7
ty ealieamaghi 44-44 1219 90-0
Mel... 40°65 111-9 82°3
CAO. sic. 4634 | 1275 93-9

From these numbers it will be seen that no two chlorides |
give the same retardation-effects. Sodic chloride practically —
produces no effect whatever, whilst the chlorides of magne-—
sium, cadmium, and zinc produce an acceleration of the rate

of Chemical Change. 455

of oxidation. In the case of the sulphates, the retardation-
effects of the three latter stand in order of magnitude thus :
MgSO, > ZnSO, > CdSQO,; and theaccelerations of the chlorides
are seen to be in the same order—MgC], > ZnCl, > CdCly.

In attempting to form a mental image of the state of mat-

ters in a chemical system undergoing change, it is customary

to think of the moving parts or molecules, their velocities of
translation, and the influence of various agencies, such as heat
&c., increasing the velocities of the molecules and multiplying
the chances of collision in a given time. Hmploying such
language, it is easy to account for the retardation occasioned

__ by the presence of chemically inactive substances in the system

by saying that, by the motions of the molecules of such bodies,
they interfere with the movements of the chemically active
molecules by coming into collision with them ; in fact, con-
tinually getting in the way of the molecules of the several

~ bodies undergoing chemical change, and so diminishing the

number of impacts in a given time between the latter.
To such an explanation, however, the experiments in the
first part of this paper are open to a grave objection. It

_ might be argued, that the introduction of a sulphate into a

solution of ferrous chloride in presence of much free HCl
would give rise to such double decompositions as would
account for the retardation observed in the rate of oxidation.
If this were so, however, it is difficult to see how the retarda-
tion is the same in amount for equal weights of different
sulphates, as well as proportional to the amount of the salt
present, being the same result as was obtained with sulphates
in a ferrous sulphate solution. In the experiments with
chlorides no such double decompositions could occur, but here
the results are anomalous. 95 grams of sodic chloride produce
no effect whatever, whilst the same weight of the sulphate in
ferrous chloride would give a retardation of 85 per cent., and in
ferrous sulphate a retardation of 50 per cent. If retardation-
effects are to be interpreted on the hypothesis of intermolecular

movements and interference, and consequent diminution in

the number of impacts, how comes it that so much sodic chlo-
ride produces no effect, whereas the same weight of potassic
chloride causes a retardation of 22 per cent.? And, considering
the subject in the same light, more remarkable still are the
effects of magnesic, zincic, and cadmic chlorides, which pro-
duce an acceleration of the rate of oxidation, or increasing the
number of collisions of the active bodies in a given time by
an amount varying from 6 to 20 per cent. It seems difficult
to account for such results on the hypothesis of interdiffusion
and simple contact between the active molecules. Instances

2K 2

456 Prof. C. Michie Smith on

of the difficulties that lie in the way of such a theory are
exhibited by the great differences observed in the rates of
somewhat analogous chemical changes. To take one example :
how could such a theory explain the fact that, under suitable
conditions of temperature and dilution, ferrous sulphate is
oxidized by potassic chlorate at such a rate that only a frac-
tion of the total work capable of being done is accomplished
in ten thousand minutes, whereas permanganate performs the
same work practically instantaneously? Interdiffusion and
simple contact among the active molecules alone would seem
incapable of offering any explanation of this and similar facts ;
at the same time it would be dangerous to speculate on the
tendencies that chemical compounds undoubtedly possess to
resist or undergo change under various conditions, in the face
of the small amount of work that has been done bearing on
chemical change.’ ~

In conclusion, I have to thank W. Crookes, Hsq., F.R.S.,
for affording me facilities for performing the above expe-
riments.

LIV. Atmospheric Electricity. By C. Micuiz Smita, B.Se.,
FIRS, PRAS., Prof. Phys. Sc. Madras Christian
College™.

N 1882 I presented to the British Association a short
report on observations made on atmospheric electricity in
Madras, which, though few in number, were of some interest,
as they seemed to show that negative electrification of the air
was not necessarily associated with broken or stormy weather.
At the same time it may be noted that the observations were
by no means conclusive, as they were in each case followed
by local showers which fell some hours afterwards. Since my
return to India, however, I have obtained a large number of
observations, many of which fully bear out the conclusion
that, even in fine weather, the air may at times be highly
charged with negative electricity.

Neglecting a few single observations, the first series obtained
was in September 1883, when, on twelve days between the
8rd and 25th, negative readings were recorded. At the time
that these readings were taken I thought they were in some
way connected with the green sun and the wonderful sun-
glows which appeared that month ; and an account of them
will be found in my paper on these phenomena in the Transac-
tions of the Royal Society of Hdinburgh, vol. xxxii. p. 389.
More recent observations have, however, shown that, under

* Communicated by the Author.

Atmospheric Electricity. . 457
certain meteorological conditions, similar readings may always
be expected.

_ The following table gives the electrometer-readings at 10 A.M.,

along with the observations of wind-direction, moisture, and
raintall made at the same hour at the Madras observatory.

TaBLe I,
Observations at 10 hours M.M.T. September 1883.

Electro- | Percentage Wroade Rain in
Date. | meter- of satura- ; 5 ast 24 Remarks.
readings. tion. eae ees j
inch.

Se 56 W. 0-01

a eee 57 W.S.W. 0-01

aE AT W.

4. —241 55 W.

5. = 93 4+ W.

6. —310 55 W.S.W.

i. 7.46 64 S.W. 0-03 Cloudy and cool.

8, — 6 46 oe Oe eee Rain at 7 P.M.

9. + 16 53 W.S.W. 0-03 Morning cool and cloudy.
| 10. — 2] 61 W.S.W.
a8 — 13 55 WENO TTS Se. Cloudy, but hot.
eel 2. Sa 49 a eae Thin haze; much dust.

Distant lightning in early morn-

| 13. eo 53 Seopa alee ing. Local showers in the
| afternoon.
a4, | — 11 56 W.S.W.
| 15. ? 49 W.N.W.
16. ? 55 W.N.W.
3 Ve , 73 S.S.W.

18. + 69 W.S.W.
| 19. ? 50 W.

20. — 34 44 N.W.
a. | — 19 49 W.
| 22. a0: 51 W.
f23. ? 52 W.N.W.
| 24, — 25 62 WEN ee | aa Bright.

25. — 15 57 NYE WE ds ERS Cloudy.

26. = 0 63 NIV Sc ee Dull and hazy.
mer. | + 5 59 S.E.

28. ae 56 S.W. 0-04

29. ? 58 W.S.W.

30. 2 72 W.N.W. 0°45

Note.—Column 2 gives the difference between the earth-
readings and air-readings in divisions of the scale, of which
about 24=110 volts.

|
;
Neglecting cases occurring in broken weather, the next
: negative readings were got in March 1884. During that
month negative readings were obtained on six days, and on
| each occasion the weather was bright and the wind south-
|| westerly; and no rain fell in Madras on these days, nor, in
|
|
| ;

458 Prof. C. Michie Smith on

fact, until nearly a fortnight after the last negative reading
was observed. ‘The readings in April and the first half of
May were all normal; and during the next two months
absence from Madras prevented me from continuing the
observations, which were not renewed till July 14. Between
the 14th and 3lst of July the readings were negative on
seven days. On only one of these days did any rain fall at
the place of observation, and then only 0°04 inch fell some
eight hours after. In each case the wind was westerly, and
in most cases raised great clouds of dust.

During August negative readings were obtained on nine
days, and on none of these could the negative potential be
traced to storms at or near the place of observation. On
August 1, for instance, the readings were negative from
11 a.m. to 2 P.M., and no rain fell till 4 p.m. on the 2nd; on
the 28th they were negative from noon to 5 P.M., ranging
from 8 to 467 divisions (say 37 to 2140 volts), and no rain
fell till the morning of the 31st. The following table (II.)
shows the observations at 10 A.M. |

TasueE II.
Observations at 10 hours M.M.T. August 1884.

Electro- | Percentage Wind Rain in
ind-
Date. | meter- of satura- | 4. section past 24 Remarks.
readings. tion. ‘ hours.
inch.
a +27 85 S.S.W. 0-11
2. — 34 44 W.
3. —95 45 W.N.W.
4. —71 47 W.N.W. | |
/ Thunderstorm and rain at night} —
Ba 56 W.N.w. | 0:07 { Mer ait
6. 62 W. 0°62
‘p + 2 56 W.N.W.
8. — 3 56 W.
9. —26 50 N.N.W.
10, —17 48 N.W.
Lis +32 55 N.
12 +27 51 N.N.W,
13 +24 78 S.W. 0:59
14 + 7 69 Wetienia. ut.’ dgereee Reading —4 at 7 a.m.
15 +10 84 W.S.W. 1-01
16 +66 79 W.S.W. 1°85
17 +30 68 W.S.W. 0-40
18 ? 77 W.S.W. 0°16
19 + 6 61 W.N.W. 0:09
20 ? 62 WwW, O11
21 - 4 60 W. |
92. 0 68 W.S.W. 0:09 { Reading negative at 1l A.M.
and noon.
23 —35 66 W

=, Atmospheric Electricity. 459

“Tt will be noticed that the effect of a shower is to make
the succeeding readings positive; and generally it is found
that negative readings are got only when the ground is dry
and also warm; for ona bright morning the readings become
negative earlier than when the morning is cloudy. |
In September there were negative readings on 13 days
almost entirely during bright fine weather. |

Taswe IIT,
Observations at 10 hours M.M.T. September 1884.

Electro- | Percentage Wind- Rain in

Date.| meter- | of satura- | 4. |. ast 24 Remarks.
| readings. tion. soe ine:
E inch.
I; +27 85 Se Wiki ecoes 0-16 in. rain fell on August 31st.
| 2. —16 44 W. |
| 3. —95 45 W.N.W.
| 4. —72 47 Wieser: aie Shower at 4 p.m.
t- Evening thunder-storm, with
a —12 56 W.N.W. 0:07 heavy rain at sea 12 miles off
| Madras.
) 6. ? 62 W. 0°62
7: + 2 56 i Ch 2 et gl Negative reading at 11 a.m.
8. — 3 56 W
9.. — 8 50 N.N.W
10, —17 48 N.W
11 +32 55 N
12 +27 51 N.N.W
13 +24 78 S.W 0-59
14 + 7 69 W.N.W |
Rain and thunder in early |
15 +47 84 W.S.W 1-01 morning. Thunder-storm at
night.
16. +40 79 W.S.W 1:85
17 +23 6s | Ws.w. | 0-40 |
18. peer. 77 W.S.W. 0-16
19. +29 61 W.N.W. 0-09
20. ? 62 Ww. O11
-21. — 4 60 W.
22. 0 68 WEEN Eo hota Negative reading at 11 a.m,
23. —33 66 WwW. /
ae aes 58 Ww. |

From September 24, 1884, to May 14, 1885, only one nega-
tive reading has been obtained in fine weather.

Looking over the tables of electrometer-readings, the first
feature that strikes one is that in only one case is the potential
negative before 9 A.M. ; and that in most cases, if the poten-
tial is positive at 10 A.M., it remains positive all day, though
on one or two occasions when a low positive reading was
obtained at 10 4.M., a negative one was obtained at 1! or 12

460 : On Atmospheric Electricity.

o'clock. Again, in all cases mentioned above, the negative
readings were got with a westerly wind, which in Madras is
usually very dry, having blown over a large extent of flat dry
land. Whenever the wind veered round so as to blow in the
least degree from the sea, the potential at once became positive.
This has been well marked in all cases, and when the sea-
breeze came on early in the day no negative readings were
obtained ; but when, as happened on several occasions, the
sea-breeze was delayed till late in the afternoon, even then
negative readings continued till the wind veered round. An
important question is, What is the influence of the clouds of _
dust that are usually flying on days when we get negative
readings? This isa question which I have not yet been able
to answer fully; but certain observations tend to show that
there is an intimate connection between the presence of dust
and the negative electrification of the air. Many observa-
tions were made extending over 10 to 20 minutes; and it
was almost invariably found that the negative electrification
was strongest during gusts of dust-laden air. At such times
the potential would often run up so rapidly that it was im-
possible to measure it accurately, while during lulls it would
often fall almost to zero. The mere friction of the dry dust
against the match and insulated umbrella would not account
for this; for the potential began to increase rapidly before
the dust-cloud actually reached the place of observation.
Again, it was found that higher negative readings were got
when the observations were made on the ground than when
they were made on the roof of a house; and, further, that in
parts of the town where there was little dust, negative read-
ings were rare, and were never very high. ‘Too much stress —
must not be laid on this last point, for simultaneous observa-
tions could not be carried out at two places, as I had only one
electrometer. A number of observations have been made in
dust with south-easterly and north-easterly winds; but the
air was never found to be negatively electrified. This
may be explained readily enough as due simply to the
presence of moisture; for one of the earliest facts im-
pressed on an experimenter on static electricity in this
part of India is that, however well an electric machine may
work, or a Leyden jar hold its charge, while a westerly wind is
blowing, the moment the wind veers round to the east the
Leyden jar loses its charge almost at once and the electric
machine ceases to act, be it a Holtz, a Voss, or a Wimshurst,
or any other machine warranted to work “jin all weathers.” |

When the weather becomes sufficiently dry I intend to make
some direct experiments on the action of dust, and on the
height to which the negative electrification extends.

Notices respecting New Books. 461

I have found it impossible, unaided, to carry out a suffi-
ciently complete series of observations Cy get the diurnal curve
auring these times of land wind; but the accompanying dia-
gram may be taken as a first iapproriisation to it between the

: Hat
AMREREEREREREE

hours of 6 A.M.and 11 p.m. It is prepared from the observa-
tions of September 1883 and August and September 1884.

Madras, July 16, 1885.

LV. Notices respecting New Books.

Handbook of Technical Gas-analysis, containing concise Instructions
for carrying out Gas-analytical Methods of Proved Utility. By
CLEMENS WINCKLER, Ph.D., Professor of Chemistry at the Fre-
burg Mining Academy. Translated, with a few additions, by
GrorcE Luner, Ph.D., Professor of Technical Chemistry at the
Federal Polytechnic School, Zurich. Van Voorst. London.

O those who have, during the last eight or ten years, been under
the necessity of analyzing the gaseous mixtures from the flues
of regenerative furnaces, and there are many such persons in this
country, the name of Dr. Winckler is favourably known as being
one of the first so to simplify the operations as to render it un-
necessary to employ the methods of Bunsen and Regnault. The
translator is, we think, scarcely aware of the extent to which tech-
nical gas-analysis is cultivated in England; for, in his preface, he
says that “itis now quite usual, at any rate in Germany, to perform
technical gas-analysis, not mer ely i in chemical works, but for testing
the efficiency of steam-boiler furnaces “ies such purposes. In
England some of these processes have also been introduced ; but
they are not as yet known and appreciated to the same extent
as abroad.” It would of course be difficult, if not impossible, to
compare the actual number of analyses made in the two countries ;
but we have no hesitation in saying that Dr. Bunte’s process, at all

462 Notices respecting New Books.

events, is well known, thoroughly appreciated, and frequently em-
ployed for analyzing the vite from regenerative furnaces in this
country.

In this little work of 125 pages the author has sueceeded in con-
-densing so large an amount-of information that any chemist, even
if previously entirely unaccustomed to the analysis of gaseous mix-
tures, may, with a moderate amount of practice, soon find himself
in a position to undertake the analysis of flue gases.

One of the great advantages of the new method is that water is
made to supersede mercury in the various manipulations; the
extent to which this substitution does away with the necessity for
personal skill in the operator can only be appreciated by those who
have had to analyze gases by the older processes.

The author very properly commences by describing the apparatus
employed by him in sampling the gases to be examined, and the
directions given are so complete as to leave nothing to be desired ;
on the other hand, a properly educated chemist would never have
any difficulty in devising a method of withdrawing a portion of gas
from a flue or chamber, and storing it In a convenient vessel from
which small quantities could be removed as required. It must be
remembered, however, that the work is intended by the author for
the instruction of students and others who have not previously
‘acquired skill in gas-analysis. Among the devices described is
Bonny’s automatic aspirator, which he has patented in Germany.
This instrument appears to us to have advantages which will make
it extremely useful under many circumstances. The chapter on
the measurement of gases will be found very useful, especially to
those who are unacquainted with the methods commonly employed
in scientific researches. The author (p. 30) states that ‘ Gas-
meters are never altogether reliable; but they give serviceable
approximate figures, especially if merely the number of revolutions
is noticed as shown by the dials, without looking for the absolute
volume of the gas passed.” To ‘this we may add that, although the
best gas-meters in this country are very well made, and quite suffi-
cient for every technical purpose unconnected with refined analysis,
it is very difficult to adjust even the most expensive ones to a less
error than one third of a per cent. on the total volume of the gas

assed.
: We think the author might with advantage have omitted the
description of his own gas-burette, as it is admitted to have been:
superseded by newer and better forms.

The descriptions of Honigmann’s, Bunte’s, and Orsat’s apparatus
are given clearly ; and the same remark applies to the ingenious gas-
pipettes of Hempel, founded, we believe, on that of Doyeére.

Altogether this is a very valuable little work, and is absolutely
indispensable to every chemist who undertakes thie technical analysis
of gases. ‘The author is to be congratulated on the excellence of
the typography and the admirable manner in which the engravings
have been reproduced. For the fidelity of the translation the name
of Dr. Lunge is a sufficient guarantee.

Owl Eee ee OEE eee rin

i Nile a. ee ee

Pe 46s 5
LVI. Intelligence and Miscellaneous Articles.

ON THE SEPARATION OF LIQUID ATMOSPHERIC AIR INTO TWO
DIFFERENT LIQUIDS. BY S. WROBLEWSKI.,

| 1 my note of April 13, 1885, I had occasion to remark that the

laws of the liquefaction of atmospheric air are not those of
the liquefaction of a simple gas, and that air behaves like a mix-
ture the components of which are subject to different laws of
liquefaction. If at first sight liquid air seems to behave in such
fashion that we might speak of the critical point of air, that arises
simply from the slight difference there is between the curves of the
tension of the vapour of oxygen and of nitrogen. Thus we may
designate the pressures between 37 and 41:3 atmospheres and the
temperatures between —140°8 and —143° as defining the critical
point of air. The curve of the tensions of air depends, among
other circumstances, on the manner in which the quantity of the
liquid is used for the experiment. Much more does this curve
cease to have any significance when we obtain temperatures of
—190° C., and especially at still lower temperatures. Thus with
air liquefied under high pressures, and then liberated and exposed
to the pressure of only one atmosphere, the boiling-point rises
gradually from —191°-4 to —187° C., and that owing to a change in
the composition of the liquid. As nitrogen evaporates more rapidly
than oxygen, the temperature of the boiling liquid tends towards
—181°5 C., which is the boiling-point of pure oxygen.
Still more striking are the special features shown by air evapo-
rated in oxygen, as seen from the following Table :—

Tempera- | Pressure | Tempera- | Pressure | Tempera- | Pressure

ture. in centim. | ture. in centim. | ture. in centim,
fo) fo) re)

— 195°02 ae —196°65 ony | —198°6 38
196°2 es 196-65 124 | 198°2 ‘2
197°1 sé! 196°65 1] | 197°9 3°6
197°5 ee 196-55 10 | 197°85 ov!
197°6 Re ieee 196°1 0 197-85 3°5
197°6 eh 196-0 a | 198-05 fae
197°6 ee 196-95 ie 198°5 32
197°5 1671 198 44 | 198°75 ae
197-5 a 198°6 oe | 199-28 ie
19738 152 198°65 yo 199-55 30
71S | ow 198°8 + ' 199°8 ee
19685 | 146 i9s99 | .. |) 200 | 28

While the tension of the vapours of liquids evaporated by the
pump diminishes steadily, the temperature, as shown in the Table
for equidistant intervals of time, passes through a series of maxima
and minima; under low pressures it ultimately attains values
which are scarcely higher than those of pure oxygen at the same
pressure. In these conditions the air only contains a very small
quantity of nitrogen.

But this is not all. Air can give two distinct liquids, different
in appearance and in composition, forming two separate layers

464 Intelligence and Miscellaneous Articles.

separated by a perfectly visible meniscus. I have attained this
result by the following method :—Having liquefied at — 142° C.
a certain quantity of air in the tube of the apparatus which I
employ for using permanent gases as cooling mixtures, I allow a
quantity of gaseous air to enter the tube, such that the pressure ot
the gas having become equal to 40 centim. and its optical density
equal to that of the liquid, the meniscus entirely disappears, after
which I slowly lessen the pressure; the moment the manometer
indicates a pressure of about 37-6 atmospheres, I see that a new
meniscus forms at a point in the tube much higher than the place
previously occupied by the vanished meniscus. A few minutes after-
wards the first meniscus reappears at the place at which we had
seen it disappear, and at the same moment two liquids are distinetly
seen, different in character, one on the top of the other. The
two liquids remain separated for several seconds. After this a
current of very small bubbles forms, which ascend, detaching them-
selves from the meniscus which separates the two liquids. In
consequence of this phenomenon the upper liquid becomes a little
opaque ; the meniscus, gradually destroyed by the current, ulti-
mately disappears altogether, and the last result is a single liquid
homogeneous in appearance. *

By means of a small metal tube introduced into the apparatus
L have been able, without destroying the meniscus, to take at will
either from the bottom or top layer a quantity of liquid sufficient
for analysis. While the lower liquid contains 21:28 to 21°5 of
oxygen, the upper one only contained 17°3 to 18°7.

This experiment is of great importance in the theory of the
critical state of gases. It shows in fact, contrary to the assertion
made some years ago by certain physicists, that the disappear-
ance of the meniscus of a liquid, when it is obtained by inereasing
the pressure exerted by a gas on a superposed liquid, does not effect
a solution of the liquid in the gas.—Comptes Rendus, Sept. 28, 1885.

ON THE SOURCE OF THE HYDROGEN OCCLUDED BY ZINC DUST,
BY GREVILLE WILLIAMS, F.R.S.

In my paper “On the Synthesis of Trimethylamine and Pyrrol
from Coal-Gas, and on the Occlusion of Hydrogen by Zine Dust,”
I showed that, under certain conditions, zine dust behaved towards
hydrogen like palladium; and I inferred, from the phenomena
observed, that the zinc dust occluded hydrogen at ordinary tempe-
ratures, and gave it off in an active condition when heated. In my
second paper, ‘‘ Note on the Occlusion of Hydrogen by Zine Dust
and the Meteoric Iron of Lenarto ” +, I gave the results of determi-

* In this experiment air, which is a completely colourless liquid, shows
moreover a singularly enigmatical optinal phenomenon, which immedi-
ately precedes the appearance of the upper meniscus. The part of this
tube where the meniscus should form, assumes a feebly orange coloration,
which disappears the moment the meniscus appears. A like phenomenon
never precedes the appearance of the lower meniscus, which separates the
two liquids,

t See ‘ Journal of Gas Lighting,’ vol. xly. p. 15. t Loe, ert. p. 485,

Intelligence and Miscellaneous Articles. 465

nations of the amount occluded in the dust by direct heating, and
also by combustion with cupric oxide. I likewise made the sugges-
tion that the hydrogen, in the cases both of the dust and the meteor-
ite, had probably been originally derived from water, and not neces-
sarily, in the latter case, from the meteorite having at one time

been exposed to an atmosphere of hydrogen at a high pressure.

The present paper is to be regarded as a study of the effect of
water, in the forms of liquid and vapour, upon the amount of
hydrogen occluded. As my later determinations were made on a
fresh specimen of the commercial dust, I considered it necessary,
in the first place, to determine the volume of the hydrogen con-
tained in it. In the course of the experiments it was found, on
leaving the apparatus to repose after the hydrogen had been ex-
pelled, that the volume of the mixture of gas and air in the
measuring-tube gradually diminished; and this is one cause of
the variations in the amounts obtained in the earlier experiments.
Variations in the volumes are also caused by differences in the tem-
peratures to which the retorts containing the zinc dust are exposed ;
and, at an early stage of the experiments detailed in this paper, I
was led to finally abandon the use of soft glass in favour of retorts
made from combustion-tubing.

Determination of the Hydrogen contained in a Fresh Sample of
the Zine Dust.

Experiment 1.—6°4790 grammes of a fresh sample of zine dust
were heated in a soft glass retort to as high a temperature as it
would bear. The arrangements were precisely the same as in the
experiments detailed in my second paper. The reason why I con-
tinued to use 6°4790 grammes (100 grains) was because this amount
had been previously found to occupy the volume of one cubic cen-
tim. All volumes are reduced toa temperature of 60° Fahr. (159-55
C.) and a pressure of 30 inches of mercury; and a correction is
made for the volume of that portion of the exit tube of the retort
which enters the measuring-apparatus. ‘The graduated tube before
the experiment contained 46°8 cubic centim. of air. After the
apparatus had cooled to the atmospheric temperature, the volume
was found to be 78 cubic centim. This gives 31:2 cubic centim.
for the hydrogen expelled from 1 cubic centim. of the zinc dust
at the temperature at which the experiment was made. ‘This is
the lowest result yet obtained, and was doubtless due to insufficient
heating. | o:

Experiment II.—In this experiment 6°4790 grammes were
heated in a horizontal retort made from hard combustion-tubing.
The air in the measuring-tube before heating amounted to 7°3 cubic

_centim. After heating and subsequent cooling to atmospheric

temperature the volume was 44°1 cubic centim.; after deducting
the air we have 36°8 cubic centim. from 1 cubic centim. of zine
dust. This result approximates pretty closely to the experiment
in my second paper, which gave 37°5 cubic centim.

Experiment IIT.—In this experiment the utmost heat of the lamp:
was continued for a long time—until, in fact, there was no indica-

466 Intelligence and Miscellaneous Articles.

tion of any trace of gas being expelled ; and the volume was read

off directly the retort had cooled to the temperature of the labora- .

tory, thus preventing reabsorption. The hydrogen after all correc-
tions amounted to 46°4 cubic centim.

Experiment 1V.—In a fourth experiment, made like the last,
47-4 cubie centim. were obtained.

We now have the following values for the volumes of hydrogen

expelled from the new sample of commercial zine dust :— |
cubic centim.

I. Soft glass retort, measured next day ...... 31:2
II. Hard glass retort, measured when cooled .. 36°8
et. .: i fs .. 46:4
IV. ,, 9 ” Z -. 474

The volume in the last experiment is the largest I have been able
to obtain, and shows that ordinary commercial zine dust may
contain nearly fifty times its volume of hydrogen.

Effect of Wetting on the Amount of Hydrogen Oceluded.,

Experiment I.—I took the usual amount of zie dust (6°4790
grammes) and thoroughly wetted it with boiling water. On drying
it in the water-oven until the weight was constant it was found to
have increased by 0°1924 gramme. The retort was then charged
with 6-4790 grammes of the dried dust, and heated in a retort
made from combustion-tubing. The air in the measuring-tube
before the experiment amounted to 3°9 cubic centim. The volume
of hydrogen evolved, after deducting the air and making all correc-
tions, amounted to 89°4 cubic centim., or 42 cubic centim. more
than in the highest result obtained from unwetted zinc dust. This
very large increase in the yield of hydrogen as compared with any
of the experiments made with the original zine dust points very
clearly to the source whence it was obtained, and strongly tends to
confirm the view thrown out in my second paper, that this source
was water. It was again found in this experiment that it required
the highest heat of the lamp to drive off the last traces of hydro-
gen, and it was observed that during the operation a certain amount
of water was produced. On keeping the apparatus in the state in
which it was left after the experiment, it was found that the volume
of the gas and air in the measuring-tube became less day by day,
until, at the expiration of a fortnight, it only measured 80-4 cubic
centim. ‘The absorption therefore amounted to 9 cubic centim.
As the air in the measuring-tube only amounted to 3-9 cubic cen-
tim., it became evident that the absorption was chiefly, if not
entirely, due to hydrogen. So that we have now conclusive proof
that zinc dust, in addition to taking up water and giving up its
hydrogen on heating, absorbs this gas at ordinary temperatures
when surrounded by it in a moist condition, thus confirming the
conclusion arrived at in my first paper from the phenomena ob-
served during the synthesis of trimethylamine. Whether it absorbs
the gas when it is in an anhydrous state I propose to determine
later on.

Lxperiment I1,—In a second experiment 6:4790 grammes of the

OT

Intelligence and Miscellaneous Articles. 467

dust, on being saturated with boiling water and then dried in the
water-oven until the weight was constant, gave an increase of
0-1561, or 0:0363 gramme less than in the first experiment. I
then subjected 6°4790 grammes to a dull red heat; and the result-
ing gas, in two hours after the operation, measured when all correc-
tions were made, gave 89 cubic centim., or 8:4 cubic centim. more
than in the last experiment. The two highest experiments with
unwetted zinc dust gave 46°4 cubic centim. and 47°4 cubic centim.,
the mean being 46°9. The two with wetted zine dust gave 89-4
and 89 cubic centim. The mean of these is 89-2, or only 4°6 cubic
centim. less than double the amount from unwetted zine dust.

It is proper to notice that the increase of weight of the zine
dust on wetting and subsequent drying in the first experiment was
0-1924 gramme, and in the second 0°1561 gramme. The mean of
these is 0°1742. Now, if 10 atoms of zine (65°3 x 10) had fixed 1
molecule of water, the increase would be 0:1785-—a number so near
to the mean of the experiments as probably not to be entirely with-
out meaning; but, on the other hand, I have not yet had time to
analyze the zinc dust, which, of course, cannot be taken as repre-
senting even moderately pure zinc.

Effect of Exposure of Zine Dust to a Moist Atmosphere.

Having shown that the wetted dust after drying gives off nearly
double as much hydrogen as was evolved from it in the condition
in which it was received, it became important to know how the
hydrogen had been occluded by the dust which had not been
wetted. It has long been known that shippers are unwilling to
carry large quantities of zinc dust in their vessels, owing to the
danger of its getting wetted, in which case it becomes heated to an
extent which may become dangerous. It being, therefore, extremely
improbable that the specimens with which I worked had ever been
wetted, or even rendered damp purposely, it struck me that it had
probably absorbed moisture from the atmosphere. This would per-
fectly account for the presence of hydrogen in the commercial
product; and, owing to its being kept closely packed, for the
amount being small as compared with that contained in the dust
which had been thoroughly wetted. ‘To determine the question I
placed 6°-4790 grammes in a watch-glass over a vessel containing
water, the whole being then covered by a bell-glass. The apparatus
was kept in a room having a nearly constant temperature of about
72° Fahr. It was weighed almost every day from the commence-
ment of the experiment (Aug. 6) until the 17th of September,_by
which time it had ceased to increase. it was found that for the
first fourteen days it gained about 3 centigrammesa day; the next
fourteen days the increase fell to about 14 centigrammes a day ;
and after this it gradually diminished, and at last entirely ceased.
It was then dried in the water-oven until the weight became
constant. The substance which had caked together was then
pulverized and taken for the following experiments.

Experiment 1.—One quarter of the usual amount, namely 1°6198
grammes, was heated in a similar manner to the previous experi-

468 Intelligence and Miscellaneous Articles.

ments. The corrected volume of hydrogen was 85°3 cubic centim., —
which multiplied by 4 gives no less than 341°2 cubic centim. of
hydrogen from 6:4790 grammes. This is about seven times the
amount afforded by the original dust, and 3°8 times more than the
mean of the results obtained with the dust which had been wetted.

Experiment 11.—In this experiment, in which the full heat of a
strong Bunsen flame was kept until the volume of hydrogen ceased
perceptibly to increase, 90°7 cubic centim. of hydrogen were ob-
tained from 1-:6198 grammes, or 362°8 cubic centim. from 6°4790
grammes= 100 grains.

The above results confirm in the strongest manner the views I
had entertained as to the source of the hydrogen in zine dust, and
conclusively show that exposure to a moist atmosphere at a mode-
rate temperature is eminently favourable to the condensation of the
hydrogen. It is not improbable that, by suitably modifying the
conditions, this amount may be exceeded. As zine dust thus
charged with hydrogen can hardly fail to become an important
chemical reagent, I shall endeavour to determine the maximum
amount that can be occluded, and the ratio to the amount of
metallic zinc present.—Journal of Gas Lighting, Oct. 13, 1885.

ON TWO NEW TYPES OF CONDENSING HYGROMETERS.
BY M. G. SIRE.

The accuracy of condensing hygrometers is known to depend on
the exactitude with which we cbserve the temperature of the
surface on which dew is deposited as well as on the more or less
distinct perception of this deposit.

I have succeeded in rendering these determinations very distinct
by observing the deposit of aqueous vapour on a cylindrical or a
plane brilliant surface, which gives rise to two new types of con-
densing hygrometer, where the fall of temperature is produced by
the evaporation of ether in which isa thermometer. .

The reservoir of the first type consists of a cylindrical tube of
thin polished metal, the ends of which are insulated in the interior
by two pieces of ebonite, so that the volatile liquid is-only in con-
tact with the metal side by a middle zone of about 1 centim. in
height. On this zone the dew is deposited; it shows itself by a
whitish ring which appears at half the height of the reservoir.

The reservoir of the second is entirely of ebonite; it is traversed
laterally by a circular aperture closed by a thin metal disk, polished
on the inside. On the part of this disk which closes the aperture
the dew is deposited, forming a dull white circle in the centre;
under a certain angle the disk appears of a pure black.

The bright surfaces are obtained by an electrochemical deposit
of palladium ; the black polish of this metal reveals the slightest
traces of condensation.

The essential character of these hygrometers is that the deposit
is made in the centre of a brilliant metal surface without a break.
On the other hand, the agitation of the liquid and the thinness of
the sides ensures perfect equality of the two parts in each instru-
ment.— Comptes Rendus, Sept. 28, 1885.

THE

LONDON, EDINBURGH, ann DUBLIN

PHILOSOPHICAL MAGAZINE

AND

JOURNAL OF SCIENCE.

(FIFTH SERIES.]

DECEMBER 1885.

LVII. On the Dilatancy of Media composed of Rigid Particles
in Contact. With Experimental Illustrations. By Professor

OsBoRNE Reyno tps, LL.D., F.R.S.*
| | Plate X.]
| ues rigid particles have been used in almost all attempts

to build fundamental dynamical hypotheses of matter :

__ these particles have generally been supposed smooth.

Actual media composed of approximately rigid particles
exist in the shape of sand, shingle, grain, and piles of shot; all
which media are influenced by friction between the particles.

of theoretical treatment by the aid of certain assumptions.

- The dynamical properties of media composed of ideal smooth -
particles in a high state of agitation have formed the subject
of very long and successful investigations, resulting in the
dynamical theory of fluids. Also the limiting conditions of
_ equilibrium of such media as sand have been made the subject

These investigations, however, by no means constitute a
to) y) y)

any attempts have been made to investigate the dynamical
properties of a medium consisting of smooth hard particles,
held in contact by forces transmitted through the medium.
It has sometimes been assumed that such a medium would
possess the properties of a liquid, although in the molecular

complete theory of granular masses ; nor does it appear that

to be in a high state of motion, holding each other apart by

| hypothesis of liquids now accepted the particles are assumed
|
|

* Communicated by the Author. This Paper was read before Section A
of the British Association at the Aberdeen Meeting, September 10, 1885,

, and again before Section B, at the request of the Section, September 15.
Phil. Mag. 8. 5. Vol. 20. No. 127. Dec. 1885. 21

470 Prof. Osborne Reynolds on the Dilatancy of Media

collisions ; such motion being rendered necessary to account
for the property of diffusion. |

Without attempting anything like a complete dynamical
theory, which will require a large development of mathema-
tics, I would point out the existence of a singular fundamental
property of such granular media which is not possessed by
known fluids or solids. On perceiving something which
resembles nothing within the limits of one’s knowledge, a
name is a matter of great difficulty. Ihave called this unique
property of granular masses ‘“ dilatancy,” because the property
consists in a definite change of bulk, consequent on a definite
change of shape or distortional strain, any disturbance what-
ever causing a change of volume and generally dilation.

In the case of fluids, volume and shape are perfectly inde- —
pendent ; and although in practice it is often difficult to alter
the shape of an elastic body without altering its volume, yet
the properties of dilation and distortion are essentially distinct,
and are so considered in the theory of elasticity. In fact
there are very few solid bodies which are to any extent dila-
table at all.

With granular media, the grains being sensibly hard, the
ease is, according to the results I have obtained, entirely dif-
ferent. So long as the grains are held in mutual equilibrium
by stresses transmitted through the mass, every change of
relative position of the grains is attended by a consequent
change of volume; and if in any way the volume be fixed,
then all change of shape is prevented. |

In speaking of a granular medium, it is assumed to be in
such a condition that the position of any internal particle
becomes fixed when the positions of the surrounding particles
are fixed.

This condition is very generally fulfilled, but not always
where there is friction ; without friction it would be always
fulfilled. |

From this assumption it at once follows that no grain in
the interior can change its position in the mass by passing
between the contiguous grains without disturbing these;
hence, whatever alterations the medium may undergo, the
same particle will always be in the same neighbourhood. _

If, then, the medium is subject to an internal strain, the
shapes of the internal groups of molecules will all be altered,
the shape of each elementary group being determined by the
shape of the surrounding particles. This will be rendered
most intelligible by considering instances; that of equal
spheres is the most general, and presents least difficulty. |

A group of such spheres being arranged in such a manner

composed of Rigid Particles in Contact. 471

that, if the external spheres are fixed, the internal ones cannot
move, any distortion of the boundaries will cause an alteration
of the mean density, depending on the distortion and the
arrangement of the spheres. For example :—

If arranged as a pile of shot (Plate X. fig. 2), which is an
oat tetrahedra and octahedra, the density of the

media is fot S taking the density of the sphere as unity.

a arranged in a cubical formation, as in fig. 1, the density

is = =, or ./2 times less than in the former case.

These arrangements are both controlled by the bounding
spheres ; and in either case the distortion necessitates a change
of volume.

Hither of these forms can be changed into the other by
changing the shape of the bounding surface.

In both these cases the structure of the group is crystalline,
but that is on account of the plane boundaries. —

Practically, when the boundaries are not plane, or when
the grains are of various sizes or shapes, such media consist
of more or less crystalline groups having their axes in different
directions, so that their mean condition is amorphous.

The dilation consequent on any distortion for a crystalline
group may be definitely expressed. When the mean condition
is amorphous, it becomes difficult to ascertain definitely what
the relations between distortion and dilation are. But if,
when at maximum density, the mean condition is not only
amorphous but isotropic, a natural assumption seems to be that
any small contraction from the condition of maximum density
in one direction means an equal extension in two others at
right angles.

As such a contraction in one direction continues, the con-
dition of the medium ceases to be isotropic, and the relation
changes until dilation ceases. Then a minimum density is
reached ; after this, further contraction in the same direction
causes a contraction of volume, which continues until a
maximum density is reached. Such a relation between the
contraction in one direction and the consequent dilation would

be expressed by
any
e—l= an / sin?

1

e being the coefficient of dilation, a that of contraction, and

é, the maximum dilation ; the + ve root only to be taken.
The amorphous condition of minimum volume is a very

stable condition ; but there would be a direct relation between

2L 2

472 Prof. Osborne Reynolds on the Dilatancy of Media

the strains and stresses in any other condition if the particles —
were frictionless and rigid. |

- If the particles were rigid the medium would be absolutely
without resilience, and hence the only energy of which it would
be susceptible would be kinetic energy ; so that, supposing
the motion slow, the work done upon any group in distorting
it would be zero. Thus, supposing a contraction in one direc-
tion and expansion at right angles, then if p, be the stress in
the direction of contraction, and p,, p, the stress at right
angles, a being the contraction, } and ¢ expansions,

Patt pyotpc=O 5
or, supposing b=c, py=Pz;
Prt + pylatc) =0. |
With friction the relation will be different ; the friction always
opposes strain, 7. e. tends to give stability.
It is a very difficult question to say exactly what part fric-

tion plays ; for although we may perhaps still assume without
error,

py 1—sing

— — ° 2
0, A Stamp
where ¢ is the angle of repose, we cannot assume that tan }
has any relation to the actual friction between the molecules.

The extreme value of ¢ is a matter of arrangement ; as in
the case of shot, which would pile equally well although
without friction.

Supposing the grains rigid, the relations between distortion
and dilation are independent of friction ; that is to say, the
same distortion of any bounding surfaces must mean the same
internal distortion whatever the friction may be.

The only possible effect of friction would be to render the
grains stable under circumstances under which they would
not otherwise be stable ; and hence we might with friction be
able to bring about an alteration of the boundaries other than
the alteration possible without friction ; and thus we might
possibly obtain a dilation due to friction. How far this is the
case can be best ascertained by experiment.

In the case of a granular medium, friction may always be
relaxed by relieving the mass of stress, and any stability due
to this cause would be shown by shaking the mass when in a
condition of no stress.

But before applying this test, it is necessary to make per-
fectly sure that during the shaking the boundary spheres do
not change position.

Another test of the effect of friction is by comparing the

:
:
:

composed of Rigid Particles in Contact. 473

relative dilation and distortion with different degrees of fric-
tion. If the dilation were in any sense a consequence of
friction, it would be greater when the coefficient of friction
between the spheres was greater. Where the granular mass
is bounded by solid surfaces, the friction of the grains against
these surfaces will considerably modify the results.

The problem presented by frictionless balls is much simpler
than that presented in the case of friction. In the former
case the theoretical problem may be attacked with some hope
of success. With friction the property is most easily studied
by experiment.

As a matter of fact, if we take means to measure the volume
of a mass of solid grains more or less approximately spheres,
the property of dilatancy is evident enough, and its effects
are very striking, affording an explanation of many well-
known phenomena. |

If we have in a canvas bag any hard grains or balls, so
long as the bag is not nearly full it will change its shape as it
is moved about ; but when the sack is approximately full a
small change of shape causes it to become perfectly hard,
There is perhaps nothing surprising in this, even apart from
familiarity ; because an inextensible sack has a rigid shape
when extended to the full, any deformation diminishing its
capacity, so that contents which did not fill the sack at its
greatest extension fill it when deformed. On careful conside-
ration, however, many curious questions present themselves.

If, instead of a canvas bag, we have an extremely flexible
bag of india-rubber, this envelope, when filled with heavy
spheres (No. 6 shot), imposes no sensible restraint on their
distortion ; standing on the table it takes nearly the form of a
heap of shot. This is apparently accounted for by the fact
that the capacity of the bag does not diminish as it is deformed.
In this condition it really shows us less of the qualities of its
granular contents than the canvas bag. But as it is imper-
vious to fluid, it will enable me to measure exactly the volume

‘of its contents.

Filling up the interstices between the shot with water so
that the bag is quite full of water and shot, no bubble of air
in it, and carefully closing the mouth, I now find that the bag
has become absolutely rigid in whatever form it happened to
be when closed. |

It is clear that the envelope now imposes no distortional
constraint on the shot within it, nor does the water, What,
then, converts the heap of loose shot into an absolutely rigid
body? Clearly the limit which is imposed on the volume by
the pressure of the atmosphere.

474. Prof. Osborne Reynolds on the Dilatancy of Media

So long as the arrangement of the shot is such thatthere is
enough water to fill the interstices the shot are free, but any
arrangement which requires more room is absolutely prevented
by the pressure of the atmosphere.

If there is an excess of water in the bag when the shot are
in their maximum density, the bag will change its shape quite
freely for a limited extent, but then becomes instantly rigid,
supporting 56 lb. without further change. By connecting
the bag witha graduated vessel of water so that the quantity
which flows in and out can be measured, the bag again becomes
susceptible of any amount of distortion.

Getting the bag into a spherical form and its contents at
maximum density, and then squeezing it. between two planes,
the moment the squeezing begins the water begins to flow in,
and flows in at a diminishing rate until it ceases to draw more
water.

- The material in the bag is in a condition of minimum den-
sity under the circumstances. This does not mean that all
the parts are in a condition of minimum density because the
distortion is not the same in all the parts; but some parts
have passed through the condition of maximum while others
have not reached it, so that on further distortion the dilations
of the latter balance the contractions of the former. If we
continue to squeeze, water begins to flow out until about half as
much has run out as came in; then again it begins to flow in. —
We cannot by squeezing get it back into a condition of uni-
form maximum density, because the strain is not homogeneous.
This is just what would occur if the shot were frictionless ; so
that it is not surprising to find that, using oil instead of water,
or, better (on account of the india-rubber), a strong solution
of soap and water, which greatly diminishes the friction, the
results are not altered.

_ On measuring the quantities of water, we find that the
greatest quantity drawn in is about 10 per cent. of the volume
of the bag; this is about one third of the difference between
the volumes of the shot at minimum and maximum density.

1
79 : 1, or 30 per cent. of the latter.

On easing the bag it might be supposed that the shot would
return to their initial condition. But that does not follow:
the elasticity of form of the bag is so slight compared with
its elasticity of volume, that restitution will only take place as
long as it is accompanied with contraction of volume.

So long as the point of maximum volume has not been

composed of Rigid Particles in Contact. 475

reached, approximate restitution follows quite as nearly as
could be expected, considering that friction opposes restitution.
But when the squeezing has been carried past the point of
maximum volume, then restitution requires expansion ; and —
this the elasticity of shape is not equal to accomplish, so that

the bag retains its flattened condition. This experiment has

been varied in a great variety of ways.

The very finest quartz sand, or glass balls ? inch in dia-
meter, all give the same results. ..Sand i is, on the whole, the
most convenient material, and its extreme fineness reduces any
effect of the squeezing of the india-rubber between the inter-
stices of the balls at the boundaries ; which effect is very.
apparent with the balloon bags, and shot as large as No. 6.

A well-marked phenomenon receives its explanation at once
from the existence of dilatancy in sand. When the falling
tide leaves the sand firm, as the foot falls on it the sand
whitens, or appears momentarily to dry round the foot.
When this happens the sand is full of water, the surface
of which is kept up to that of the sand by capillary attrac-
tion ; the pressure of the foot causing dilation of the sand,
more water is required, which has to be obtained either
by depressing the level of the surface against the capillary
attraction, or by drawing water through the interstices of
the surrounding sand. ‘This latter requires time to accom-
plish, so that for the moment the capillary forces are overcome ;
the surface of the water is lowered below that of the sand,

__ leaving the latter white or dryer until a sufficient supply has

|
|
|
|

been obtained from below, when the surface rises and wets
the sand again. On raising the foot it.is generally seen that.
the sand under the foot and around becomes momentarily
wet ; this is because, on the distorting forces being removed,
the sand again contracts, and the excess of water finds
momentary relief at the surface.

Leaving out of account the effect of friction between the
balls and the envelope, the results obtained with actual balls,
as regards the relation between distortion and dilation, appear
to be the same as would follow if the balls were smooth.

The friction at the boundaries is not important as long as.
the strain over the boundaries is homogeneous, and _ particu-
larly if the balls indent themselves into the boundaries, as
they do in the case of india-rubber. But witha plane surface
the balls at the boundaries are in another condition from the
balls within. The layer of balls at the surface can only vary
its density from 2/3 to1. ‘his means that the layer of

_ balls at a surface can slide between that surface and the adja-

cent layer, causing much less di lation than would be caused.

476 Prof. Osborne Reynolds on the Dilatancy of Media

by the sliding of an internal layer within the mass. Hence
where two parts of the mass are connected by such a surface,
certain conditions of strain of the boundaries may be accommo-
- dated by a continuous stream of balls adjacent to the surface.
This fact made itself evident in two very different experiments.

In order to examine the formation which the shot went
through, an ordinary glass funnel was filled with shot and oil,
and held vertical while more shot were forced up the spout
of the funnel. It was expected that the shot in the funnel
would rise as a body, expanding laterally so as to keep the
funnel full. This seems to have been the effect at the com-
mencement of the experiment ; but after a small quantity had
passed up it appeared, looking at the side of the funnel, that
the shot were rising much too fast, for which, on looking into
the top of the funnel, the reason became apparent. A sheet
of shot adjacent to the funnel were rising steadily all round,
leaving the interior shot at the same level with only a slight
disturbance. |

In another experiment one india-rubber bag was filled with
sand and water; at the centre of this ball was another much
smaller ball, communicating through the sides of the outer
envelope’ by means of a glass pipe with an hydraulic pump.
It was expected that, on expanding the interior ball by water,
the sard in the outer ball would dilate, expanding the outer
ball and drawing more water into the intervening sand. This
it did, but not to the extent expected. It was then observed
that the outer envelope, instead of expanding, generally bulged
in the immediate neighbourhood of the point where the glass
tube passed through it; showing that this tube acted as a
conductor for the sand from the immediate neighbourhood of
the interior ball to the outer envelope, just as the glass sides
of the funnel had acted for the shot. /

As regards any results which may be expected to follow
from the recognition of this property of dilatancy,— |

In a practical point of view, it will place the theory of earth-
pressures on a true foundation. But inasmuch as the present
theory is founded on the angle of repose, which is certainly
not altered by the recognition of dilatancy, its effect will be
mainly to show the real reason for the angle of repose.

The greatest results are likely to follow in philosophy, and
it was with a view to these results that the investigation was
undertaken.

The recognition of this property of dilatancy places a
hitherto unrecognized mechanical contrivance at the com-
mand of those who would explain the fundamental arrange-
ment of the universe, and one which, so far as I have been

composed of Rigid Particles in Contact. ATT

able to look into it, seems to promise great things, besides
possessing the inherent advantage of extreme simplicity.

Hitherto no medium has ever been suggested which would
cause a statical force of attraction between two bodies at a
distance. Such attraction would be caused by granular media
in virtue of this dilatancy and stress. More than this, when
two bodies in a granular medium under stress are near
together, the effect of dilatancy is to cause forces between the
bodies in very striking accordance with those necessary to
explain coherence of matter.

Suppose an outer envelope of sufficiently large extent, at
first not absolutely rigid, filled with granular media, at its
maximum density. Suppose one of the grains of the media
commences to grow into a larger sphere; as it grows, the
surrounding medium will be pushed outwards radially from
the centre of the expanding sphere. Considering spherical
envelopes following the grains of the medium, these will ex-
pand as the grains move outwards. This fixes the distortion
of the medium, which must be contraction along the radu,
and expansion along all tangents.

The consequent amount of dilation depends on the relation
of distortion and dilation, and on the arrangement of the
grains in the medium. At first the entire medium will un-
dergo dilation, which will diminish as the distance from the
centre increases. As the expansion goes on, the medium
immediately adjacent to the sphere will first arrive at a con-
dition of minimum density ; and for further expansion this
will be returning to a maximum density, while that a little
further away will have reached a minimum. The effect of
continued growth will therefore be to institute concentric
undulations of density from maximum to minimum density,
which will move outwards ; so that after considerable growth
the sphere will be surrounded with a’series of envelopes of
alternately maximum and minimum density, the medium at
a great distance being at maximum density. At a definite
distance from the centre of the sphere not more than

1.4R,

where R is the radius of the sphere, the density will be a
minimum, and between this and the sphere there may be a
number of alternations depending on the relative diameters of
the grains and the spheres.

The distance between these alternations will diminish rapidly
as the sphere is approached. The distance of the next maxi-
mum is 1.2K, the next minimum is given by 1.09 R, and
the next maximum 1.06R.

478 Prof. Osborne Reynolds on the Dilataney of Media

_ The general condition of the medium around a sphere which
has expanded in the medium is shown in Plate X. fig. 3,
which has been arrived at on the supposition that the sphere
is large compared with the grains. |

From a radius about 1.4 R outwards the density gradually
increases, reaching a maximum density at infinity ; and at all
distances greater than 1.8] the law is expressed by

de 1

dr”?

where n has some value greater than 3 depending on the
structure of the medium.

Within the distance 1.4 R the variation is periodic, with a
rapidly diminishing period. In this condition, supposing the
medium of unlimited extent and the sphere smooth, the sphere
may move without causing further expansion, merely changing
the position of the distortion in the medium ; for the grains,
slipping over the sphere, would come back to their original
positions. It thus appears that smooth bodies would move
without resistance if the relation between the size of the grains
and bodies is such that the energy due to the relative motion
of the grains in immediate proximity may be neglected.
The kinetic energy of the motion of the medium would be
proportional to the volume of the ball multiplied by the density
of the medium and the square of the velocity. |

But the momentum might be infinite supposing the medium
infinite in extent, in which case a single sphere would be
held rigidly fixed.

If we suppose two balls to expand instead of one, and sup-
pose the distortion of the medium for one ball to be the same
as if the other were not there, the result will be a compound
distortion, Since, however, the dilation does not bear a linear
relation to the distortion, the dilation resulting from the com-
pound distortion will not be the sum of the dilations for the
separate distortions unless we neglect the squares and products
of the distortions as small.

Supposing the bodies so far apart that one or other of the
separate distortions caused at any point is small, then, retaining
squares and products, it appears that the resultant dilation at
any point will be less than the sum of the separate dilations
by quantities which are proportional to the products of the
separate distortions.

The integrals of these terms through the space bounded by
spheres of radii R and L are expressed by finite terms, and
terms inversely proportional to L, which latter yanish if L is

composed of Rigid Particles in Contact. 479

infinite. Thus, while the total separate dilations are infinite,
the compound dilations differ from the sum of the separate
by finite terms, and these are functions of the product of the
volumes and the reciprocal of the distance.

Assuming stress in the medium, the difference in the value

of these finite terms for two relative positions of the bodies

multiplied by the stresses, represents an amount of work which
must be done by the bodies on the medium in moving from

one position to another.

To get rid of the difficulty of infinite extent of medium, if
for the moment we assume the envelope sufficiently large and
imposing a normal pressure upon the medium, then, since the
work done will be proportional to the dilation, the force be-
tween the bodies will be proportional to the rate at which
this dilation varies with the distance between them.

The force between the bodies would depend on the cha-
racter of the elasticity as well as on the dilation.

It is not necessary to assume the outer envelope elastic; this
may be absolutely rigid and one or both the balls elastic.

In such case the two balls are connected by a definite
kinematic relation. As they approach they must expand,
doing work which is spent in producing energy of motion ;
as they recede, the kinetic energy is spent in the work of com-
pressing the balls.

As already stated, the momentum of the infinite medium
for a single ball in finite motion may be infinite, and propor-
tional to the product of the volume of the ball by the velocity ;

but with two balls moving in opposite directions, with velo-

cities inversely as the masses, the momentum of the system
is zero. ‘Therefore such motion may be the only motion
possible in a medium of infinite extent.

When the distance between the balls is of the same order
as their dimensions, the law of attraction changes with the
law of the compound dilations and becomes periodic, corre-
sponding to the undulations of density surrounding the balls.
Thus, before actual contact were reached, the balls would
suffer alternate repulsion and attraction, with positions of
equilibrium more or less stable between, as shown in figs. 4
and 5 (Pl. X.).

We have thus a possible explanation of the cohesion and
chemical combination of molecules, which I think is far more
in accordance with actual experience than anything hitherto
suggested.

It was the observation of these envelopes of maximum and
minimum density which led me to look more fully into the
property of dilatancy.

i
|
fam ,

480 Prof. Osborne Reynolds on the Dilatancy of Media.

The assumed elasticity of the surrounding envelope, or of
the balls, has only been introduced to make the argument clear.

The medium itself may be supposed to possess kinetic
elasticity arising from internal distortional motion, such as
would arise from the transmission of waves in which the
motion of the medium is in the plane of their fronts.

The fitness of a dilatant medium to transmit such waves is
only less striking than its property of causing attraction,
because in the first respect it is not unique.

But as far as I can see such transmission is not possible in
a medium composed of uniform grains. If, however, we have
comparatively large grains uniformly interspersed, then such
transmission becomes possible. If, notwithstanding the large
grains, the medium is at maximum density, the large grains
will not be free to move without causing further dilation; and
it seems that the medium would transmit distortional vibra-
tions in which the distortions of the two sets of grains are
opposite. |

Such waves, although the motion would be essentially in
the plane of the wave, would cause dilation, just as waves in a
chain cause contraction in the reach of the chain. They
would in fact impart elasticity to the medium, exactly as,
in the case of a slack chain having its ends fixed but other-
wise not subject to forces, any lateral motion imparted to the
chain will cause tension proportional to the energy of distur-
bance divided by the slackness or free length of chain.

Distortional waves therefore, travelling through dilatant
material which does not quite occupy the space in which it is
confined when at maximum density, would render the medium
uniformly elastic to distortion, but not in the same degree to
compression orextension. The tension caused by such waves
would depend on the gross energy of motion of the waves
divided by the total dilation from maximum density conse-
quent on the wave-motion. All such waves, whatever might
be their length, would therefore move with the same velocity.

If, when rendered elastic by such waves, the medium were
thrown into a state of distortion by some external cause, this
would diminish the possible dilation caused by the waves.
Thus work would have to be done on the medium in producing
the external distortion which would be spent in increasing
the energy of the waves. For instance, the separation of two
bodies in such a medium, which, as already shown, would in-
crease the statical distortion, would increase the energy of
the waves and vice versa.

As far as the integrations have been carried for this con-
dition of elasticity, it appears, with a certain arrangement of

On the Refraction of Fluorine. i 481

large and small grains, that the forces between the bodies
would be proportional to the product of the volumes divided
by the square of the distance ; 2. e. that the state of stress of
the medium may be the same as Maxwell has shown must

exist in the ether to account for gravity. We have thus an

instance of a medium transmitting waves similar to heat-waves
and causing force between bodies similar to the forces of gra-
vitation and cohesion, in such a manner as to constitute a con-
servative system. More than this, by the separation of the two
sets of grains, there would result phenomena similar to those
resulting from the separation of the two electricities. The
observed conducting power of a continuous surface for the
grains of a medium closely resembles the conduction of elec-
tricity. And such a composite medium would be susceptible
of a state in which the arrangement of the two sets of grains
were thrown into opposite distortions, which state, so far as it
has yet been examined, appears to coincide with the state of
a medium necessary to explain electrodynamic and magnetic
phenomena according to Maxwell’s theory.

In this short sketch of the results which it appears to me
may follow from the recognition of the property of dilataney,
I have not attempted to follow the exact reasoning even so
far as | have carried it.

In the preliminary acceptance of a theory the mind must
be guided rather by a general view of its adaptability than
by its definite accordance with some out of many observed
facts. And as it seems, after a preliminary investigation, that
in space filled with discrete particles, endowed with rigidity,
smoothness, and inertia, the property of dilatancy would cause
amongst other bodies not only one property but all the fun-
damental properties of matter, I have, in pointing out the
existence of dilatancy, ventured to call attention to this
dilatant or kinematic theory of ether without waiting for the

completion of the definite integrations, which must take long,

although it is by these that the fitness of the hypotheses

ek.
must be eventually tested.

LVIII. On the Refraction of Fluorine.
By GEORGE GLADSTONE, /.C.S.*

ie his paper on the Refraction-Equivalents of the Elements,
_ published in the Phil. Trans. of 1869, Dr. Gladstone
estimated the equivalent of fluorine at 1°45, from the results

* Communicated by the Author, having been read at the Meeting of
the British Association at Aberdeen, September 1885.

482 . Mr. G. Gladstone on the

of an observation taken of a solution of fluoride of potassium.
He gave it, however, with all reserve, remarking at the same
time that “ fluor-spar and kryolite gave very small values for
fluorine, or rather indicate that this body has scarcely any
influence on the rays of light.”” In the same paper he gives
the specific refractive energy as 0°073, which corresponds to
the refraction-equivalent given above. As it stands, it is by
far the lowest on the list, none of the other elements having a
specific refraction of less than 0°1; but had it been calculated
from the results given by the minerals above mentioned, the
specific refraction would have been as low as 0-016.

As the discrepancy then noted was so very wide, and special
interest attaches to fluorine in this connection on account of
its singularly small refractive power, I have thought it worth
while to review all the evidence upon the subject, and to
obtain new observations of a solution of fluoride of potassium.

Brewster gives 1:344 and 1:349 as the index of refraction
of kryolite. The mean of these will give 24°63 as the refrac-
tion-equivalent for Na;Al,F,. Deducting 13:2 for the sodium
and 9°7 for the aluminium, which is the most recent value
assigned to this metal, there will only be 1°73 left for si
atoms of fluorine. Thus fluorine equals 0°29. |

Of fluor-spar there are several independent observers.
Brewsier gives 1°436 for the index of refraction for the bright
part of the spectrum; and Wollaston 1:433. Fizeau gives
1:435 for the line D; Stephan, in 1871, gives 1°4339 ; and
Kohlrausch, in 1878, gives 14324 and 1:4342 for the same.
The average of these is 1°4341, which, taking the specific
gravity at 3°183, according to Landolt’s tables, will give a
refraction-equivalent of 10°64 for CaF. Deducting 10:0 for
the calcium, fluorine will equal 0°32.

I am indebted to my brother for a redetermination of the
aqueous solution of fluoride of potassium, which gives 8°15 as
the refraction-equivalent for the line A.; deducting 7°85 for
the potassium, the fluorine equals 0°30. otis

These are all closely accordant. 3

- Messrs. Topsoe and Christiansen have taken the indices of
refraction and the specific gravities of a series of crystalline
fluorides of uniform composition. They are biaxial crystals ;
and the observations are given by them for the lines C, D,
and I of the spectrum, both of the ordinary and extraordinary
ray. In the following table I have taken the mean of the two
rays for F, and give also the mean value for A calculated
from the differences between the indices for the lines C and
fF’, The refraction-equivalent of the substance is given for

Refraction of Fluorine. 483

the line A, and the last column contains the value for fluorine
obtained by deducting the amount due to the other elements,
viz. 35°556 for the water, 7-4 for the silicon, 11°5 for copper,
9-9 for nickel, 9°8 for zinc, 6°7 fur magnesium, and 11°5 for
manganese, and then dividing the residue by six. The
refraction-equivalents for the above-named elements are taken
from the revised list published in the ‘ American Journal of
Science’ for January 1885.

Substance, | SBenife | Tndes, | Inder, | cotrralent equivae't
| CuB,, SiF,, 6H,O...) 2182 | 14131 | 14049} 58-16 0-62
NiF,, SiF,,6H,O ...| 2109 | 1-4027 | 13935) 57-60 0-79
InF,, SiF,, 6H,O ...| 2104 | 13926 | 13857 57°80 pes |
MeF,, Sif, 6H,0.../ 1-761 | 13553 | 13493) 5435 | 078
MnF,, SiF, 6H,0...) 1:858 | 1:3690 | 1:3621| 59-44 0-83

_ Four out of the five will be seen to give closely accordant
results, and agree well with the general average, which is 0°77.

The same observers give similar data for a silico-ammonio-
fluoride, 2NH,F,SiF,, which is a uniaxial crystal. The
index for the line F is 1°3723, and the calculated figure for
the line A is 1°3670. The specific gravity being 1:970, the
refraction-equivalent will be 33:16. The latest estimate for
NH, being 11:1, there will be 3°56 left to satisfy the six atoms
of fluorine, after deducting what may be due to the other
elements. This gives 0°59 for fluorine ; which is, as near as

ossible, the mean of all the observations here examined.

This still leaves a range of from 0°3 to 0°8 ; but in any case
the refraction-equivalent is of exceedingly small amount; and
the specific refraction, even at the highest limit, can scarcely
be the half of that of any other substance known. ‘Thus,
while the specific refraction of chlorine, bromine, and iodine
amounts to 0:279, 0-191, and 0-193 respectively, that of .
fluorine will lie between 0°015 and 0°044, as calculated from
the extreme results of the observations recorded above.

RA AgL 4

LIX. On Measurements of the Intensity of the Horizontal
Component of the Earth’s Magnetic Field made in the Phy-
. sical Laboratory of the University of Glasgow. By THomas
Gray, B.Sc., F.R.S.E.*
[Plate XI.]

N the course of some measurements of the horizontal com-
4 ponent of the intensity of the earth’s magnetic field, which
have recently been made in the physical laboratory of the
University of Glasgow, several improvements in the apparatus
and mode of conducting these experiments suggested them-
selves. The following paper is a description of these mea-
surements.

The method adopted was, in general principle, that of Gauss ;
that is, the determination by a deflection-experiment of the
ratio M/H, and by an oscillation-experiment of the product
MH, where M is the magnetic moment of a magnet, called in
this description the deflector, and H is the horizontal intensity
of the earth’s magnetic field at the place of experiment.
The earlier experiments, of which the results are given below,
were, with the exception of the mode of determining the
effective length of the deflector, conducted in the way that
has for several years been practised in Sir William Thomson’s
laboratory, and descriptions of which have already been
publishedf.

TuE DEFLECTION EXPERIMENT consists in finding the deflec-
tion produced by the deflector on a second needle, called the
magnetometer-needle, suspended at a known distance from it,
and thus furnishing an equation of the form

M

fa, *) =Htan 05°), .)
where /(2a,,7”) is a function of 2a, the effective length of the
deflector, and 7 the distance between the centre of the deflector
and the centre of the magnetometer-needle, and @ is the
angle of deflection. The function /(2a,,7) depends on the
relative direction of the line joining the centres of the deflector
and magnetometer-needles, the direction of the axis of the
deflector and the direction of the lines of force in the magnetic
field, and on the distribution of magnetism in the deflector and
magnetometer-needles. Strictly /(2a,, 7) should be f(2a,, 6,7),

* Paper read before the British Association at the Aberdeen Meeting,
September 1885. Communicated by Sir William Thomson, F.R.8.

+ Vide Phil. Mag. for November 1878; ‘ Electrician’ for July 8, 1882
‘Nature’ for November 9, 1882,

| +

Horizontal Component of the Earth’s Magnetic Field. 485

where 0 is the length of the magnetometer-needle ; but it is
one of the merits of the method here described that the mag-
netometer-needle is made so short that its magnetic length
may be neglected, and the resulting equation thus much
simplified.

The positions adopted in these determinations were, for the
line joining the centres of the deflector and the magnetometer-
needle, the magnetic meridian and a line at right angles to it ;
and for the direction of the magnetic axis of the deflector, a
line at right angles to the magnetic meridian. We then have
for the former of these positions,

2 72\2
J(2Qa,7)= ar We We Staines Mie
F
and for the latter position,
peer Aty. ot a ok ED)
From equation (1) we obtain, by (2) and (3),
2 _8\2 |
eg tan Oe rim ipied tay Pat ee
r
and
M Bin. ANE Deve
i (92 +: a?) tau 6,. Ps ° 2 e (5)

Now, besides M and H, the value of a, is also unknown ; but
when @) and 6, are determined at nearly the same time, we
ean calculate a, from the above equations. We have clearly

(7*—a?)? _ tan 0,

hn aes ; (6)
2r(7? + a?)? tan >

' Expanding the numerator and denominator on the left-hand

side of the above equation, and neglecting small terms, we
readily obtain, as a close approximation,

o_. Ir? — 26,7?

= ' 7
_ 26or a 36.71 ( )

There is in the method here described a departure from the

\ usual practice—namely, determining the effect of the length
of the deflector and deflected magnets, by double and triple

experiments with the deflector placed at different distances

along a line through the centre of the deflected magnet. In

the first place, the third experiment is rendered unneces-
Phil. Mag. 8. 5. Vol. 20. No. 127. Dec. 1885. 2M

486 Mr. T. Gray on Measurements of the Intensity of the

sary by making the deflected magnet so short that its magnetic
length may be neglected ; and, in the second place, the effect
of the length of the deflector is found by adopting two posi-
tions for which the length enters in the resulting equations (4)
and (5) above, with opposite sign, rendering-the observation
highly sensitive to that effect. By adopting this method a
good estimate of the effective length of different deflectors is
obtained ; and this is useful information outside of the parti-
cular object of the experiment.

The latest form and arrangement of the apparatus for the
deflection-experiment is shown in Pl. XI. fig. 1, where
T is the table on which the apparatus is placed, M the
magnetometer, A and B the deflector-stands, and C the scale
on which the deflections are read. The magnetometer M
consists of a light mirror about ‘8 centim. in diameter, on the
back of which two magnets, 1 centim. long and °08 centim. in
diameter, are fixed. A better form of needle would, as has
been pointed out by Sir William Thomson, be two thin disks
of hard steel mounted with their planes parallel and at a dis-
tance apart somewhat less than the diameter of the disk.

The mirror with attached magnets is suspended by a single
silk fibre (half a cocoon-fibre) in a recess cut in a block of
wood, W. ‘Two holes at right angles to each other, and pass-
ing through the position of the mirror and magnets, allow the
magnetic system to be accurately adjusted when setting up
the apparatus. The holes, with the exception of that in
front of the mirror, are plugged while the instrument is in
use, and the mirror and fibre are protected from currents
of air by means of a plate of plane glass. The sole plate,
P, is furnished with three brass feet, which rest on a “ hole,
slot, and plane”’ arrangement cut in the top of the glass
plate, p, fixed to the table. The deflector-stands, A and
B, consist of a base plate of mahogany furnished, as in the
case of the magnetometer, with three brass feet which rest on
“hole, slot, and plane ”’ arrangements cut in the glass plate p.
A centre pivot, c, is fixed in the sole plate and passes through
a closely-fitting hole in the glass plate, g, which rests on three
thin blocks of hard wood and is free to turn in azimuth. A
strip of wood having a V-groove cut along its upper side and —
furnished with an adjusting-screw, s, is cemented to the top of |
the glass plate g in such a way that the bottom of the V-groove
is vertically above the centre of the pivot c. The screw, s, gives
an adjustment for the centre of the magnet, which is necessary
if the magnets differ slightly in length. The centre of the
magnet is seldom exactly at the middle of the length of the
bar, but, with the arrangement here adopted, the adjustment

Horizontal Component of the Karth’s Magnetic Field. 487

ean be readily made by turning the screw s until equal. deflec-
tions are obtained on opposite sides of zero, when the deflector
is reversed by turning the plate g through 180° in azimuth.
The deflectors, d, are shown in positions east and west of
the magnetometers; and the distance between their centres
when in this position is 70 centim. Two plates similar to
are placed, one on the north and one on the south side of the
magnetometer, in such a position that, when the deflector-stand
is resting on them, the distance between the centres of the
deflectors is 60 centim., and the line joining them is in the
magnetic meridian. The feet on which the deflector-stands
rest are so adjusted that the deflectors are equidistant from the
magnetometers, and at the same distance apart when the posi-
tions of the stands are interchanged. Thescale Cis graduated
to millimetres on glass, and is placed with its centre directly
in front, and in the focus, of the magnetometer-mirror. The
distance between the scale and the mirror is 129 centim.

A paraffin-lamp, with a copper funnel which has a vertical
slit in front, with a fine wire in its centre, covered by a plane
glass plate, is placed behind the scale, and furnishes a beam of
light which is focused on the glass scale by the magnetometer-
mirror. The scale being transparent, the deflections can be
read either from the back or the front of it. When the deflec-
tions were read from behind the scale, it was found convenient,
although not absolutely necessary, to render the glass partially
obscure ; and this, acting on a suggestion of Mr. Bottomley,
was done by dusting lycopodium-powder over it. The deflec-
tions were taken in the following manner :—

The table T was first placed in such a position that the line
joining the centres of A and B (PI. XI. fig. 1) was exactly at
right angles to the magnetic meridian. This was done in one or
other of the following ways :—(1)A thin wire was passed under
the magnetometer and stretched along the line joining the
centres of A and B, and then taken back, either over the top of
the magnetometer, or beneath and at a greater distance from it,
in such a way as to form a vertical plane circuit. An electric
current was then sent through the circuit, and the table turned
until it produced no deflection on the magnetometer-needle.
(2) One of the deflectors was placed in its position north or
south of the magnetometer, and lifted out of its V by the sus-
pension-fibre. The table was then turned until the suspended
needle produced no deflection of the magnetometer-needle.
When this is the case, the direction of the magnetic axes of
the deflector and the magnetometer-needle are in the same line ;
and if the latter needle be in its proper position, this line also

' passes through the centre of the deflector when placed on the
: 2M2

488 Mr. T. Gray on Measurements of the Intensity of the

other side of the magnetometer. To ensure that this was the
case, the deflector was placed on the other side and the position
of the table for no deflection again observed. ‘The magneto-
meter-needle was then adjusted, by turning the levelling-
screws, until the two positions were coincident. A combination
of the methods (1) and (2) gives a ready means of testing
whether the plates g have been properly placed on the table.

Suppose A to be east and B west of the magnetometer.
The deflectors were turned by means of the plate g until their
lengths were accurately in the magnetic east and west line,
and their poles so placed as to produce a deflection to the
same side of zero. The deflection was then read. The plates
g were next turned through 180° and the deflection on the
opposite side of zero read. The plates g were then turned back
to their first position, and the deflection again read. The dif-
ference between the mean of the first and third reading and
the second gives twice the deflection. The same operation was
then repeated with A north and B south, with A west and B
east, and with A south and B west. The mean of the deflec-
tions for the east and west positions and the mean of the
deflections for the north and south positions were then found,
and from them the mean effective length of the two deflectors
calculated. This length was then substituted in equations (11)
and (12) below, and the value of H calculated.

Previous to the adoption of the above apparatus only one
deflector was used, and it was placed by hand in the positions
necessary for a cycle of operations similar to that above de-
scribed. The idea of using two deflectors placed on opposite
sides of the magnetometer, and the arrangement for them
shown in fig. 1, is due to Sir William Thomson. It has the
advantage of greater symmetry, it allows the deflector to be
placed at a greater distance from the magnetometer, and,
what is most important, the magnet need not be handled
during the experiment.

TH Oscruation HxPreRIMENT consists in finding the
period of oscillation of the deflector when suspended with its
axis horizontal, and free to perform horizontal oscillations
under the influence of the earth’s magnetic field. This expe-
riment gives the value of MH from the equation

An?
MH+e= ae .. a

The quantity « refers to the torsional rigidity of the suspen- _

sion, which was practically zero in these experiments; wis |

the moment of inertia of the oscillating system, and P the

_ la tt Een

a

Horizontal Component of the Earth’s Magnetic Field. 489

period of oscillation. The following is a description of the
arrangement and the mode of observation.

A length of single cocoon-fibre was taken, and a stirrup
like that shown at 8 (fig. 2) formed on one end of it by fold-
ing it twice, so as to make four fibres, and making a knot
about 3 centim. from the end. The other end was passed
through a small hole in the brass bow, 6 (fig. 1), and then
fixed to a small strip of sheet lead which simply rested on the
sole plate and held the deflector in the proper position. The
deflector was placed in the stirrup 8, and the loops adjusted
so as to suspend it horizontally. The magnet was thus sus-
pended in a stirrup almost devoid of inertia, and by a fibre of
negligible torsional rigidity. This leaves », the moment of
inertia of the deflector, and P, its period of oscillation, to be
determined. The moment of inertia can be very accurately
calculated from the mass, length, and thickness of the bar,
when, as is the case with these deflectors, the bar is a round
cylinder of small and perfectly uniform diameter. Greater

- accuracy in these calculations, and the advantage of small

diameter in the ‘‘side on”’ position when taking deflections, are
the main reasons, apart from simplicity, why a solid cylinder
is preferred to a thin tube for the deflector. The period of
oscillation P was in the earlier experiments obtained either by
observing, with the eye placed behind a narrow slit, the times
of successive transits of the end of the bar across a fixed mark
(usually a black thread stretched vertically near the end of
the magnet and in the same magnetic meridian) ; or by ob-
serving the times of the successive transits across the vertical
wire of a telescope. In the later experiments, one end of the
deflector was polished with the view of using it as a mirror to
reflect a beam of light to a scale placed at some distance.
This method was not successful, owing to defects in the mirror,
and it has been abandoned in favour of a light silvered-glass
mirror, m (fig. 1), about 03 centim. in diameter and 0:01
gramme in weight, attached to the stirrup with its plane
parallel to the length of the magnet. The same lamp and scale
are thus available both for oscillation and deflection experi-
ments. With this arrangement the amplitude of the oscilla-
tion need never exceed one degree, and hence no correction
for arc is necessary.

With regard to the effect of the inertia of the mirror on pu,
it is to be remarked that its total moment of inertia is about
001, while that of the deflector is about 40; so that, even if
the inertia of the mirror be neglected, the error is not more

lg 2
|| than z$o per cent.

When a second observer was available, time was usually

490 Mr. T. Gray on Measurements of the Intensity of the

taken from a watch with a centre seconds-hand moving over
a dial divided to quarter-seconds. ‘The watch was keeping
almost perfect time. In this case an observer counted the
oscillations of the magnet, and called now at the end of every
four or five periods, while the other observer noted the time.
Personal error does not enter into this method, as it is elimi-
nated in taking the differences. When the transits and time
were taken by the same observer, the time was taken from a
chronometer beating half-seconds. ‘The observer took time,
say, at the beginning of a minute, and then counted the beats
until he could observe a transit. By simply counting the ~
number of beats between two successive transits, the total
number of periods in one minute could then be estimated; and
the time of the first transit after each minute was then taken
as long as the amplitude was sufficiently large to allow the
time to be accurately estimated. [Fractions of half seconds
were estimated from the position of the magnet at the beat
next before and next after transit.

The results of the different observations of time were com-
bined in the following manner :—Tirst suppose an even
number 2n observations to have been taken. ‘The sum of the
intervals of time between the nth and the (n+ 1 )th, the(n—1)th
and the (n+2)th, and so on to the Ist and 2nth, divided by
the square of the number of intervals, and by the number of
periods between each pair of observations, gives the average
period as nearly as it can be obtained from that set of obser-
vations. Suppose next an odd number 2n +1 observations to
have been taken. The sum of the intervals between the Ist
and the (n+1)th, the 2nd and (n+ 2)th, and so on to the nth
and (2n+1)th, each divided by the corresponding number of
periods, divided by the number of intervals, gives the average
period. Two sets of observations were usually taken; and if
they agreed closely, the mean of the two was assumed to be
the true period.

We have then the following equations for the determination
of H when one deflector only is used :—

M 72 — a? 2
iT eo he MP
-: = (72 + a?) tan a, 1 cnn.

~ Agr?

d
an MH— zs

Horizontal Component of the Earth’s Magnetic Field. 491

where 2a is the length, d the diameter, and w the weight of
the deflector-bar. From (4) and (8) we obtain,

ma’ Hii Yr
ts bate Bei 159) OF (9)
& PA (zr? Se") Mamby oe ee

From (5) and (8),

é 3
Si
4 m(a + od hw

TES Ca st SES
= Boe rae eta ey ee ag)

. When two deflectors are used, as described above, equations

(9) and (10) take the form,

ais
, , mr( a? + ?)(Pew, + P2w,)

3 Pa PPIPytan ee si

and

4 a + z ?) (P22, + P22) M
ie A aes ies Sp iors SA ia nes bets ey
ee a (r?+a2)3P?P2tan 0,” Sie

where P,; and P, are the periods of oscillation of the two de-
flectors respectively, 2a, the mean of their effective lengths

_ (supposed nearly equal), and w, and w, their masses in grammes.

HKquations (9) to (12) give, of course, the same value for H

_ if a has been properly calculated. The value of = is always

calculated, as it gives an idea of the quality of steel used for
the deflector. ‘The results are given along with the values of
H in tabular form below.

The values of H derived from equations (9) to (12) require
correction, in the oscillation experiment, for are of vibration,
for virtual increase of inertia due to air moved, and for change
of inertia, due to change of dimensions of the deflector, if the
temperatures when the bar is measured and when it is vibrated
are different. Neither of these produced a sensible error in

the experiments, and hence they were neglected. There are

besides a correction for the effect of variation of temperature
during the experiment in altering the magnetic moment of
the deflector, and a correction for the change of magnetic
moment between the oscillation and deflection experiment, due
to the deflector having its length in the direction of the mag-
netic meridian in the former, and at right angles to it in the
latter experiment.

492 Mr. T. Gray on Measurements of the Intensity of the

The magnets were found to change by 5,1. of their value

per degree centigrade change of temperature ; but as the change
of temperature never exceeded two or three degrees, this cor-
rection was neglected.

The correction for induction, however, was found to be
considerable in some cases, and the results tabulated below
have all been corrected for that.

THE TEMPERATURE COEFFICIENT was obtained by placing the
deflector behind the magnetometer-needle in such a position
as to produce a deflection of 1000 scale-divisions, and the
change of deflection produced by raising the temperature
about 40° C. observed.

Tue INDUCTION COEFFICIENT was determined by placing the
deflector, surrounded by a magnetizing coil, near the magneto-
meter-needle, and observing the change of deflection produced
by passing a known current through the coil. The arrange-
ment for these measurements is illustrated in Pl. XI. fig. 3 ;
and the results of a number of experiments on the effect of the
length and hardness of the deflector on this correction are
given in Table II., and are shown diagrammatically in the
following curve.

Curve illustrating the effect of Ratio of Length to Diameter on the
Inductive Coefficient.

Percentage change of magnetic moment produced by unit field.

Ratio of length to diameter.

Horizontal Component of the Earth’s Magnetic Field. 493

Referring to the figure, m is the magnetometer-needle ; C,
and C, two coils, consisting of one layer of No. 30 B.W.G.
silk-covered copper wire, wound in glass tubes about 5 millim.
in external diameter ; C is a scale divided to millimetres, on
which the deflection of the magnetometer-needle is read, in
the manner described for the deflection-experiment above ;
R is a box of resistance-coils, and G a current-galvanometer.
The line D E represents a magnetic meridian, and A F a line
passing through the centre of m and at right angles to D E.

The coil Cy was placed with its axis parallel to A F, and its
centre on the line DE. The coil C, was placed with its axis
on the line A F, and at such a distance from m that a cur-
rent passed in the proper directions through both coils pro-
duced no deflection. In making this adjustment a current,
equal to about 30 times the greatest current subsequently used
in the experiment, was passed through the coils.

The magnet to be tested was then introduced into one of
the coils,adjusted to the proper position, and the deflection of
m read. A current of such strength as to produce a field of
about 7, C.G.S. units intensity was then passed through the
coils, and the change of deflection read. The current was
then reversed, and the deflection again read. This operation
was repeated with stronger and stronger currents until a field
of from one to two units intensity was reached. The magnet
was next introduced into the other coil, and a similar series
of measurements taken. The results plotted on a sheet of
section-paper show clearly that a field of considerably greater
intensity than that used is required to permanently alter the
magnetic moment of the magnets when they are hard-
tempered.

The changes of deflection obtained when the magnet is in
the coil Cy are always a smaller fraction of the total deflection
than they are when it is in the coil C,;. This is no doubt due
to a change of magnetic distribution. The equations to the
deflections in the two cases are :—

. M
H(7r? + a)?
for the coil Co, and
A 27M
ers H (7? —a?)?

for the coil Q,.

From these equations itis clear that either a change of M or
of a, will affect the deflection, but that the effect of the change
of a, is opposite in the two cases. The method of measurement

494 Mr. T. Gray on Measurements of the Intensity of the

therefore allows both the change of moment M and the change
of effective length 2a, to be determined. When the change of
moment only is required, this method has still an advantage
in the fact that the effect of want of exact compensation
between the two coils can be eliminated by means of the two
measurements.

Sufficient sensibility is, for most purposes, obtained by
using a deflection within the limits of the scale, but almost
any degree of sensibility may be obtained by using an inferred
zero. It is not advisable to place the magnets very near to
m, because, if the tangent method be adopted, the compen-
sation of the coils will not then be perfect unless m is always
brought to its original zero position ; and if this objection be
got over by using a sine method, in which the whole appa-
ratus, including the magnetometer, can be turned round a
vertical axis, there still remains the objection that the effect
of distribution becomes very pronounced. An inferred zero
method therefore involves, when small magnets are being
tested, a diminution, by artificial means, of the intensity of the
field at m. This, however, cannot be pushed far, on account
of the very inconvenient fluctuations of zero which then take
place, due to changes of declination in the earth’s field, com-
bined with the small variations of intensity of ‘‘ H”’ being
superposed on a field, the whole strength of which is not large
compared with these variations. |

The results of these experiments, given in Table II., show
that it is of great importance, so far as induction is concerned,
that the length of the magnets should be at least forty times
their diameter, and that they should be made as hard as
possible. It appears that for the steel employed in the experi-
ments, which was of the kind commonly called silver steel, a
much stronger magnet is obtained with a blue than with a
glass-hard temper*. ‘The magnets were originally magnetized
by placing them between the poles of a large Ruhmkorff
magnet excited by 24 tray-cells. They were again magne-
tized by placing them between the poles of the same magnet

* Note added October 26, 1885.—Since this paper was in type I have
received a copy of No. 14 of the ‘ Bulletins’ of the United States Geolo-
gical Survey, which is wholly devoted to a discussion of experiments on
the “ Physical Characteristics of the Iron-Carburets,” ‘by C. Barus and
V. Strouhal. In this work a large number of interesting experiments on
the effect of annealing specimens of this “silver steel,” which had been
previously tempered glass-hard, are quoted. Among other things the
large increase of magnetic moment which is obtained by softening the
steel, up to a certain point, is clearly brought out.—T. G.

Horizontal Component of the Earth’ s Magnetic Field. 495

excited by a dynamo having a low-resistance armature, and
giving a potential of 100 volts. The relative strengths of the
different magnets remained the saine, and the strength only
very slightly increased, thus showing that they were nearly
magnetized to saturation.

The greatest trouble in the determination of the intensity
of the earth’s magnetic field arises out of the variations of the
field itself—variations in declination causing changes of zero,
and variations of intensity between different determinations
causing apparent inaccuracy in the results.

This trouble was in the later measurements got over by
means of a permanent magnetic vibrator, the period of vibration
of which, at different times, gives a comparison of the intensity
of the magnetic field at these times. In order to render the
result of any one determination free from error due to diurnal
variation, the period of the vibrator is taken at the beginning
of the experiment, between the deflection and the oscillation
experiment, and again at the end of the experiment. If these
agree well and the results also agree well among themselves, the
value found is considered reliable ; if not, the whole experi-
ment is discarded. This vibrator has only lately been adopted,
and hence the results up to those taken on June 11th are not
corrected for diurnal variation. The results are tabulated in
the order in which they were taken on each day ; and it will
be observed that the earlier results are generally the smaller,
this being due to diurnal variation. The last three results
given in the Table are corrected to noon for diurnal variation.

It has been found impossible to get consistent results at all
on several occasions, owing to extraordinary variations, and
this was notably the case on the lst of September. It is
interesting to note how readily the vibrator shows extra-
ordinary variations which take place suddenly. ‘The observer
has only to watch for sudden changes of amplitude. These
changes of amplitude are sometimes extremely marked, the
vibrator being almost stopped on some occasions, and after-
wards gradually worked up to a large swing.

Table I. gives the results of the various determinations.
The meanings of the numbers in the different columns are
clearly indicated in the headings.

496 Horizontal Component of the Earth’s Magnetic Field.

(
‘poyjeut | x9GGT-
IOjoepep y

e[qno(d,

‘SHIVULOY

aANAAeOs
en) rR © FX 6 br ©

|
al

tH

=)
N
pen)
=

jO jos Yovo jo uve
‘syiun “S*p'O |
ut A}ISU9}UL [eJUOZIAO FL | :

*10}00]}
-op oy} jo summed

god Juewmlom oTjoUSeT
"10}O0Hop

oY} JO YASUE, SATIOBY
‘ IOLILUU-10}OUL
-OJoUseUL OY} WO]

6-821
6-821
6-81

6e

66

6

6e

n~
n

=
ee)
=)
am

g[vos oy} Jo cous

00-06
00-0€
00:08

66

GL-8G

useUl

“su UI
oy} wWOIZ AOJOPep

(‘suotjisod qynos

pue 7}10N)

‘e[poou-19j}0TM030
jO osjueo JO ooueySIC

a

ATAV I,

00-96
00:96
00-9¢

"sud Ul

(‘suorjzisod 4sost

puw 48egf)

‘UOHVIIVA [VUANIP LO} UOOU 0} PoyoIIOH y

seul

ey} Wor; AOJe_Op

jo 91300 Jo ooUuRAsICT
*SOoTUUIvIS

UI ‘LOJOPOp JO JYSIOM,

‘g[pooU-10}9U10JOU

iMcor do©
aes

mi oO <H

"SOLJOUITJUSD UL

‘oOpopep jo sojowveic
"s9149UII}U90

ul ‘aojoeHop jo yAsue'T
‘LOPOPep Jo 1oquin yy

‘G88T
‘aye

Prof. T. Carnelley on the Periodic Law. 497

TasLe [].—Showing the effect of Length and of Hardness
on the Induction-Coefficient of Magnets.

Unit field,

Length Apparent | Apparent | Mean of Magnetic

ee fy Aree x eee percentage | numbers in | moment = :
centi- § increase of| increase of | eplumns per emarks.

diameter.!moment for|moment for 3 and 4. | gramme.

a unit field: | unit field:
side-on end-on
position. | position.

3 10 0-80 0°90 0°85 27 Glass hard

4 16 0°67 0-73 0-70 32 i

4 16 0°67 0:70 0-69 35 Ps
. 6 20 0°51 0 67 0°59 36 i

7 31 0-51 0°58 0°54 39 9

8 32 0-51 0:58 0-54 54 A

8 3o2 0°51 0°58 0°54 52 ps
10 34 9°46 0:56 0°51 40 DB
10 44 0-40 0°56 0-48 43 9

7 47 0-46 0:51 0-49 57 ‘
10 50 0-44 0°58 0°51 67 z
10 50 0-48 0°54 0-51 60 BS
10 50 0-46 0°55 0°51 53 6
10 50 0:46 0°52 0-49 71 a
10 50 0-46 0°56 0-51 60 =
10 67 0-41 0°51 0:46 65 =

73 0-41 0°50 0°47 64 x

10 105 0:42 0°45 0-43 66 -
10 of. 0:47 0°53 0°50 41°5 |Glass hard.
10 34 0°63 0°67 0°65 44°5 |Yellow.
10 34 0°84 0°98 0-91 541 |Blue.
10 48 0°32 0-40 0°36 45 Glass hard.
10 48 0°43 0°55 0:49 46 Yellow.
10 48 ' - 0b3 0°67 0-60 el Blue.

LX. The Periodic Law, as Illustrated by certain Physical
Properties of Organic Compounds.—Part II. The Melting-
and Boiling-points of the Halogen and Alkyl Compounds of
the Hydrocarbon Radicals, By THomAs CARNE LEY, D.Sc.,
Professor of Chemistry in University College, Dundee*.

: a the present series of papers my object is to determine
whether the elements are in any way analogous to the
hydrocarbon radicals of Organic Chemistry. For this purpose

* Communicated by the Author.

—eSE - - — -

498 Prof. T. Carnelley on the Periodic Law.

I have made a careful comparison of the physical properties
(chiefly melting- and boiling-points) of a large number of
both organic and inorganic compounds, and believe that I
have been able to make out numerous and varied relationships
between these properties.

My first paper (Phil. Mag. [5] xviii. p. 1) dealt with the
melting- and boiling-points and heats of formation of the
normal halogen compounds of the elements, and I showed that
certain well-defined relationships existed as regards the above
physical properties of these compounds. My second paper
(ibid. [5] xx. p. 259) dealt in a similar manner with some of
the physical properties of the normal alkyl compounds of the
elements, which were shown to exhibit exactly the same rela-
tionships as those of the corresponding halogen compounds ;
proving, therefore, that the function of the alkyl radicals—
methyl CH;, ethyl C,H;, propyl C;H,, &c.—was exactly ana-
logous to that of the halogen elements—chlorine, bromine,
and iodine. ?

In the present paper I wish to show that the normal ha-
logen and alkyl compounds of the hydrocarbon radicals exhibit
relationships similar to those of the corresponding compounds
of the elements, and therefore that the elements and hydro-
carbon radicals are analogous and have the same function in
their several compounds. Ina fourth and final paper I intend
to take a general review of the whole question, and to draw
certain conclusions as to the nature of the chemical elements
and of the Periodic Law.

The following table contains the experimental data, with
the authorities, which have been used in drawing the conclu-
sions given below. Tor the purpose of avoiding minus signs
all temperatures are reckoned from the absolute zero (— 273).
The following abbreviations are used :—Me = methyl, CH; ;
Kit = ethyl, C,H; ; Pr* = normal propyl, CH;.CH,.CH,—;
Pr® = isopropyl (CH;),CH—.

Prof. T. Carnelley on the Periodic Law. 499
TaBLe 1.—Experimental Data.
| Melting-
Constitution. Baling: Authority.
point.
Chlorides :——
Methyl chloride ......... ibs ket 3 Sige seen b.p. 249 | Regnault.
Ethyl BER net si Bren Ob ssc te b.p. 285 %
2 ore CH,.CH,.CH,.Cl. ...| bp. 319 | Linnemann.
i, a a ee ae OH CHELCH. |~.43>?. b.p. 310 4
Butyl |. eeitgeges aR Set) (OH Ul: cts. b.p. 351 a
me ple de SE ae Evie. Cah OLY, 24s: b.p. 342 #
a i Sea ee ae ites Ob ete holes b.p. 324 | Perkin.
Amyl a ee ei (OH-), : Ob: 25:25:23: b.p. 380 | Lieben and Rossi.
s Lal ies oe EB, 1 OE: 22202 wistacses:. b.p. 377 | Wagner and Saytzeff.
Pe DS ees GiMePre: CL 12222000. .8 b.p. 377 ¥ iy
| ” =P Naat Re Ee eS CHMe,.CH, .CH,.Cl..| b.p. 374 | Carius and Fries.
5 ra be gl ee reverrs Gh J... |. b.p. 860 | Butlerow.
a ee Sa Mer HG > CL +: 2555053122 02. b.p. 359 | Perkin.
|. = eaemeerenee BE GOH,)2.00....0.2 Bile. | Lichen avid diesel:
t3 va RE 2 ae Me. (CH,), . CHC1. Me. b.p. 397 | Schorlemmer.
c: ee oe Shag | eels b.p. 387 | Friedel and Silva.
EE I a a a eee m.p. 271 | Kaschirsky.
Wetepiyh = UL... OH,. ’ (CH, 15. 5) Cee ee . p. 482 | Cross.
[ ‘s Byer ear ass is. CHMe,. (CH). Cl(?)..) b.p. 423
e rte os. 3 « CMe," GMe,.: CL-......... ae 409 | Butlerow.
| “ pp age a ee Prs . (OH), ‘CHMe. Cl b.p. 409 | Rohu.
Octyl eee «2 CH, - (CH), : Cl oes. 200. b.p. 456 | Perkin.
‘ ee te DO tg. C,H, .-CHMe.Cl....... b.p. 448 | Schorlemmer.
| eae As C.H.:. CHEE. Ol... .....: b.p. 448 4
| a eee STI bo BEE COL 5 credernsncgsotonde. b.p. 488 rt
| . TU pat at ee CEt, ‘Pre. sO] eeee wearer b.p. 428 | Butlerow.
| NGS 1 0.H,.Cl ddtdte cae fiat did owes bp 421 e
) Methylene dichloride ... ould SS Ee ee ener ee b.p. 315 | Thorpe.
| Ethylene me J 15 25.2 2 Bt 0) eee b.p. 357 :
| Ethylene __,, godt Wet SORIOD, ooeccac peeves do b.p. 333 $s
| | Propylene _e,, 124 Gl (GH), OCA 220 eie. b.p. 892 | Reboul.
| ss = spel, -CHCl-CH.Cl b.p. 870 | Linnemann.
a3 - or Orr. GAOL |... b.p. 359 | Reboul.
Z : pS GE OCL, ~ CH. oc... b.p. 343 | Perkin.
Fa a Ae goo 8 0 Bore emai eer b.p. 405 | Wurtz.
| %. bcs See eee: learn» Be Seow b.p. 895 | Kopp.
| oF 2 ue, CHC. /.. 43: b.p. 377 | Giconimides.
* : ...| CH, . CCL. CH, . CH... b.p. 369 | Bruylants.
| Amylene : ie: sg Meee eB bp. 430 | Buf.
% pepe CCL: . Me ao ccin ss 14: bp. 418 | Friedel.
| 5 ...| CHMe, . CH, . CHCl,...| b.p. 402 | Bruylants.
| Hexylene - Si CMe, C1. CMe, Chive tanh. m.p. 433 | Friedel and Silva.
| i - WARE SOL, A ee Rt b.p. 455 | Pelouze and Cahours.
: 2 i <2 ap eRe tee Fi b.p. 448 | Wurtz.
) + Co 0, EI pee? > Wir ee b.p. 437 | Henry.
|Heptylene Ate ee sOk at ik ST b.p. 463 | Schorlemmer.
| 3 | Pra. CGP red 8t)..'54 b.p. 454 | Tawildarow.
\j ce = Pe MNS i times Sd. oan hd b.p. 508 | Thorpe and Young.
: nee hee AS a tat. LY b.p. 481 | Nieson.
| 3 C Fi CCl, . Me .........| b.p. 470 | Annalen, evi. 271.

a
ig

J

500 Prot: Carnelley on the Periodic Law.
TABLE I. (continued).
Melting-
Constitution. B be e Authority.
point.
Trichlor-methane ......... EDO) Stand ave ction mone ce a8 ee reas
Trichlor-ethane............ CHAAVCHCL, .wish.ch b.p. 887 | Staedel.
ac JOaCHAL CH, . CCl, POH GRECO Ameen C b.p. 347 39
Trichlor-propane......... CH,Cl. CHCI. CH, Ol. .| bp. 431 | Perkin.

" Nat peat 2s CH,Cl. CH,. CHCl, b.p. 420 | Romburgh.

99 is usmeadaa eed CH,.CHCl. CHCl... em b.p. 413 | Friedel and Silva.

- Mire oo eae CLG 8 UR Gol Mee b.p. 398 i
Trichlor-pentane ......... ER: 1 ©, I eet b.p. 460 | Buff.

" BO the canceenke:s 6, he dein, Men cece meee b.p. 448 | Bauer.
Trichlor-hexane ......... C8 cat 6 Re ae yee b.p. 490 | Pelouze and Cahours.
Tetrachlor-methane...... oy Oe ee ia kh Ld Se Bere cle
Tetrachlor-ethane......... CHCl, . CHCl, Bihan hem. b.p. 420 | Staedel.

5 oi eee ae CH,Cl. OO ci ee cake re b.p. 403 a
Tetrachlor-propane ...... Se yea ite eaee tee { yt ee Berthelot.

‘ rth ak, Be Oy ER ARN era eset cont b.p. 453 | Romburgh.

. oy Gata ee CH,Cl. CCl, .CH,Cl ...| b.p. 437 | Henry.

%3 29 nc tie 1 BE Dane aes Fe 5 b.p. 426 | Borsche.

Fa gs Waa He S| wba) m.p. 451 | Schorlemmer.
Tetrachlor-butane ...... ORE Ue OR ae en ote = Liquid. | Thurnlackl.
oi 5) teteesecenenceees se dt.p, 046 | Bull, Chin) =a
Tetrachlor-pentane As tone C,H eCl, pits (ste ORES b.p. 508 | Bauer.
Trichlor-ethylene ......... CCl, : 560 @! Riemer 0 b.p. 360 | Paterno.
Trichlor-propylene, 6 . ‘| 0,ELCl, 52 Sa cae eee b.p. 415 | Pfeffer and Fittig.
Bi MON Tae MO nov n plea Ae eet b.p. 388 | Borsche.
Trichlor-amylene he oh SG hs -cee ceeemecteee cee b.p. 473 | Bauer.
Dichlor-acetylene ......... OOTe= CCl eee ~~ th | Watts Dict. i. 768.
Dichlor-allylene ......... G3 EE OES 6, 0, Mare ee je b.p. 851 | Pinner.
Ohlornicene ‘....3.5.4....,. FL i enceatesensehkeree b.p. 566 | St. Evre.
Bromides —

Methyl bromide ......... gdh. yiies vs ght oe exe b.p. 278 | Perkin.
Ethyl ot hate nn, le ACMA, DE cp resinecash b.p. 312 _
RN cc Paget Pies ve CH, .CH,.CH,. Br b.p. 344 -

a Binetankadk<.. CH. Obl ir. OH. yar: b.p. 334 | Linnemann.
Butyl Rem daa a » hg Py 6) 5 A Aa aaa Pee b.p. 373 ”

9 (a ee OH Me, . OF, « Br... s0.5 b.p. 365 99

oD ae ee ee CMe. d Et, x uncocattbacat b.p. 344 | Perkin.
Amyl ie Rade... OH. (CH.)y3 ir... oe b.p. 402 | Lieben and Rossi.

- itt Uibsedsth, oe. CHMe,. (OH,), 2 Bre ct b.p. 893 | Perkin.

9 A wiTlvidts ts t. - CHMe, .CHMe. Br...... b.p. 386 | Wurtz.

%9 i Mi eean. OMe, Et nhl Linnnsteergeecss b.p. 882 | Wischnegradsky.
Hexyl py OP PRIS. ce Ci. (CH)... Br b.p. 428 | Lieben.

” wei: ects eee Co sOH Ma. Br....uc., b.p. 417 | Schorlemmer.

” eC toa cs’. CHMePre. Ce ris. b.p. 416 | Monatschrift, iv. 34.
Rae piel oY five 6 imei a - OT tCH,), . Es b.p. 452 | Cross.

” wits SPR RR i CH, ‘ (CH), .CHMe. Br} b.p. 439 | Venables.

% sd MCN ts Ciie.: OMe, . Br %....20% m.p. 425 | Kaschirsky.

a

Prof. T. Carnelley on the Periodic Law.

TABLE I. (continued).

| Melting-
| Constitution. Boiling
point.
| Octyl bromide ............ SHUT), | Br ..c.-6, +.

Ss ny BEG. Oca Gs fey LSE Se Reece
Methylene dibromide ...| CH,Br, .................. { pes
Ethylene CH,Br . CH,Br { ewes

| z Hie, fa Ul: So! ee m.p. 282
Ethylidene .,, OH CP Br ao. vce t desv a. b.p. 382
Propylene é Bre (CEL )o 4 Breivcccss.ss. b.p. 438

ae " ...| CH,.CHBr. GH,Br b.p. 415
2 $5 were Or VO DE Sy b.p. 405
a + ee CH,. CBr,. os eeeeee b.p. 387
‘| Butylene a get CHMcBr . CHMeBr b.p. 489
| sl % 1 CHMthr . CH.Br-.......3. b.p. 431
” ' pC Me, be. CH. Br :........ b.p. 422
Amylene ne heise ee ee ee b.p. 459
= it CHEtBr.CHMeBr...... b.p. 451
- = CHMe, .CHBr.CH,Br | b.p. 448
| Hexylene ff Oo (CE Be. ce. kc35 b.p. 483
i yy et) Mees By -* essere. b.p. 458
a m A CMe,Br. CMe,Br......... ana 442
. if, ie OME BS 3 ee ee eee m.p. 413
5 og oe Pré. CHB. CHMe. Br |m.p. 412
Octylene és SPN SEE, GEMS cacacsscisin ts. .vs b.p. 473d
| Tribrom-methane CHBr,. { ee
eas. Bet ccesrcceierst anata! el
Tribrom-ethane ......... Cheri Br. ! 3.20200. b.p. 462
| 8 A ae Ia CHUB, fected. b.p. 461
Tribrom-propane CH,Br.CHBr.CH,Br pees
chee ae m.p. 289
| ‘5 TS oe Se CH,.CHBr. CHBr, b.p. 473
| a Rare. |! CH, . CBr, .CH,Br...... b.p. 463
Tribrom-butane ......... Me Cir. CHBr,. ...4..:. b.p. 489
a are Pel EE ie else gaan att b.p. 485
Tetrabrom-methane CBr { RPneee
JBLOLLWY csaecese 4 eee ea oaete des sesosasees m.p. 864
‘Tetrabrom-ethane......... per ee Or. .....8..: { eee ae
7 : b.p. 482
i ee. CHBr, .CHBr,......... { wet oo
Tetrabrom-propane ...... CH,Br . CBr, .CH,Br...|m.p. 468
C H Br m.p. 342
39 Se te eh. AS Oe 3 5 3 eoeceeet OS oessase b.p. 518d
os ae CHBr, .CHBr.CH,Br..| b.p. 524
RS et ee City) Cir, : CHBr, gas b.p. 503
etrabrom-butane......... C, H, Br, REE Beet ee ee ere m.p. 389
~ pe TOME ot CHBr, ; pene. CH, Me.. m.p. 387
af ett C,H, Br, Pas veress ss Fae5e: m.p. 372
etrabrom-pentane ...... CMe, Br. CBr, .CH,Br..| m.p. 388
i ee MAL Pra, ‘OBr, . CHBr ss b.p. 548
eeu. Pep OC Gr, . CN Gr, :.....-| bp. 548
etrabrom-hexane ...... ple FAs eee ee ee m.p. 415
ee age Se e Rorekoan. 58k) TPS OOM

| Perkin:

dO]

x,

Authority.

b.p. 476 | Perkin.
b.p. 461 | Lachowicz.

Reboul.

Wurtz.

Perkin.

Thorpe and Pore
Wagner.

Bruylants.
Annalen, exxiv. 293.

| Helbing.

Rizza.

Eltekoff.

99
| Nieson:

Thorpe.
Perkin.

‘Denzel.

\ Henry.
Reboul.

39
| Linnemann.

Caventou.

Bolas and Groves.

Kessel.
Burgoin.
Tawildarow.
Anschitz.
Hartenstein.’

Annalen, exxxyi. 64.

Reboul.
Oppenheim.
Caventou.
Prunier.

Annalen, clxxil. 291.
Hokeutin.
Bruylants.

Boughaedie:
Zeit. Chem. 1871.

2N

502 Prof. T. Carnelley on the Periodic Law.

TABLE I, (continued).

Melting-
Constitution. Soe Authority.
point.
Tetrabrom-hexane ...... Cp Tas hil,” dectohacass eee nes Be } Annalen, exxxix. 251.
¥ in elaieiakse (CH,).(CHBr.CH,Br)2 |m.p. 336 | Wagner.
Tribrom-ethylene...... .. CH Br oO Bry skeecentets- b.p. 435 | Anschiitz.
Tribrom-propylene ...... CHBr: CBr.CH.Br ...| non-vol. | Henry.
ss of Aasadiegt Bee LTE Ee padascesceonet b.p. 457 | Oppenheim.
Dibrom-diallyl ........... C, H, ‘Br. O FIG Br ncesceet b.p. 483 | Henry.
Brom-diallylene ......... C.H,Br Se err a eee a b.p. 423 iy
Paprylidene bromide’ ...| 01,70... .ustes.cabesteceu- b.p. 477 | Annalen, exlii. 300.
Lodides :—
Methyl iodide ............ Cig Bi RAE es teen Un b.p. 315 | Perkin.
2 6 ER Oa ae nee: ae Ci. POHL | civeet cee Se b.p. 345 | Linnemann.
ROY. voy) hsveas@etsd CH, .CH, . OHI ...:.2 . bps S7o 5
9 saessad ex ieaiaeheme 8 8 CH, sCHL WOM. : oSeeak b.p. 362 | Perkin.
Bitty AOGINE «saves snk ene CH, (CEL, an Varies ce b.p. 403 | Linnemann.

r eds ast Eig. 82 CH Me, OME). giieGs, : b.p. 393 | Perkin.

. SE Se eA? CH Melt .T .s:.9..nthess b.p. 392 | Lieben.

: Se ee Ons aan ies. Aes opis sene eee b.p. 873 | Puchot.

2202) eevee Aa neiannve re CH, SH). Mey Cease b.p. 429 | Lieben and Rossi.

e Pind pea ee cs CHMe, . (CH, "G | ree b.p. 421 | Perkin.

: eee tee CH, (CH). CHI. CH, b.p. 419 | Wurtz.

ts Dee Ne ie CHEt aR Od ae b.p. 418 | Wagner.

“ iis ie es eA ed CHMe, ‘Ciil CHa b.p. 411 | Wischnegradsky.

PL ee ee G4.) Se Oa ERA ere bate b.p. 401 $3
Ey SR a Sa Ci. CH). 0 si ee b.p. 454 | Lieben.

te ae EE ee ali (CE, yal Gece b.p. 446 | Pelouze.

‘3 Sige A <a UE; ),:« ‘CEL, CH,} b.p. 440 | Wanklyn.

A a ete a ae Cheer. Re ea ee b.p. 488 | Connick.

. 5a) aes Bee CMe, .CHI.CH,.........) b.p. 415 | Jahresb. 1873.

Aan tine Se «. OMe Pre . 0 inG.sage cick b.p. 415 | Annalen, execy. 254.

Finis, Girma REE, ED OMe,Pr8 To... {| °P* 973 | | Kaschiesky.
RAI ts once et EL, ACO Midus Coccictiess<e b.p. 474 | Cross.

a Pitt cies vusevsdeaen be « AE Has A seahts tics s. chasis b.p. 458 | Jahresb. 1869.

by eae Fs eae Prp .(CH, do CHI .CH.,| b.p. 443 | Rohu.

. Goi < Wwahddbana chee bs ¢ OMoMtP re. 1..iss.000642. b.p.419 | Kaschirsky.

a ey ee are CMe... CMe, . Licces is. ae b.p. 414 | Butlerow.

BER ds Nadia xneade lve OH Me. CH.) 60k 4.57. b.p. 494 | Zincke.

m is Hal Baphietivenha ts: Lae ei ‘OHI. Sic Mera b.p. 484 | Bouis.
Methylene diiodide .:. ..] OH,I,,....-sessecsseseee {| mb! o77 | | Hofmann,
Ethylene - OTT ot OGel tics. stoners: m.p. 855 | Aronstein.
Ethylidene _,, CH. CHI, bin FA oe wae b.p. 453 | Plimpton.
Propylene _,, CH, ‘a CH,. Gin EP rnes b.p. 500d} Freund.

C,H,I, Pe ees oe b.p. 421d) Semenoff,
Triiodo-methane ......... Ss k BE ESSE pr eae m.p. 392 Serullas.

F m.p. 420 |) Calculated.
Tetraiodo-methane CDT hal ee nataaybrtaeieus sh b.p. 620 \ Phil. Mag. [5] xviii.17.1]
Tetraiodo-hexane ......... ae ),(CHI.CH,I), ...|m.p.a.373} Annalen, c. 363.
Triiodo-propylene......... ena dd py RIIB ls Sadie tuadnd m.p. 337 | Liebermann.

s oe Sy. sadeaedne J BOE OG eis At Rae m.p. 314 | Henry.

Prof. T. Carnelley on the Periodic Law.

TaBLE I. (continued).

Melting-

Constitution. cae

point
Alkyl compounds (Hydrocarbons) :—

MUNSETRTEC ou. vias... ass 0-00 PRA Balen oak cad wach b.p. 245
NRT 25 = kes dak Skea er = > ig Re URETS ee Des ae: ae ae b.p. 274
ER eee Orebte treet iid. lop Dp. 206

OS eee LTT) ER ee he a ea b.p. 311
oO eee GiiMe... CH; . Me .:.3... b.p. 303
b.p. 282

Se CULE spanner ted rama { 5 253
MING ctig nino siajendiacine- PM MMII Pope Je gu 2. Sa ons b.p. 344
I sonia dcaet CHMe,. (CH,), . Me b.p. 335
eee 12 1), Sn ere b.p. 337
err CG We; .CHMe, ......:.. b.p. 331
Se eee ee SG 20S eek. kee. b.p. 318
7 IGEN ocrsarreisustawasnia- Le. b.p. 371
_ ee Noss et ds eich « 52 b.p. 370

2) LE See ae CH We BGP ee oc. anaes b.p. 364

1 ae ee ee C Peis” CCE. i. Hb: 70.2. b.p. 863

GE Pian US aclecsics ese dee'es CMe, Kt, ieee Ciatevotote cnaele oak ere b.p. 360
ere PR OMBAE cic. sonata asigcate as b.p. 398
Peelers tres ci. <- CHMe, .(CH,),.CHMe,) b.p. 381
CMe, . CMe ae

oh 2 - ceppe ees m.p. 369
b.p. 422

ANS, he. aes cawewceee... Normal 2.2... seeeeenees m.p. 222
EES ee CHMe, . (CH.,),.CHMe,| b.p. 405
re Pr&.CH,.CHMePr® ...| b.p. 403

7 b.p. 446

RROCAVIG yo2.0.-2-.c.seeeenese 1. | See eee aeest \ m.p. 241
RIECATIC oo cease. sc sesee- =. DRS nee { 4 Bye
le OMe;:CHMe ............. b.p. 309
MPEPOXYVICIC. 6-..02..0.cceerees Re OG fice nkavinceens b.p. 346
BW SFA) Bondew: oy eens CHMe: OMebt............ b.p. 343
ee ee, CH B0..).:.5....7 b.p. 839
REOPEVIONIG 00 cevesiicecsnss see CMe, : CMeHt ............ b.p. 366
gh Se ee Se CHMe, . CH : CMe,...... b.p. 355

oa ee ie ea CMe, .CMe: CH, b.p. 352
MEOW ICTIO. andgsiind sane .s..e- CE a hts. da dedeeine'de b.p. 395
| es pattie <shen's0 ClPre - CHPYB -... 2030. b.p. 391
| 2 ren: Beis en Cie. CH. CMe... ce b.p. 376
'} Methylethyl-acetylene ...) OMe: CEt......--..0....... b.p. 324
Se ee a 0D s Oe eae b.p. 318
ssopropyl-acetylene ......; CH : CPrp .........-..+¢. .p. 301
ee eee OH Cite. ©: OH..2.2. b.p. 823
eM lide Hees ackenss So. EARS 0) ce 1, San nee b.p. 355
a es Cite 7 CoCH Pr, 210.5 b.p. 376
Praciveaasarideinan ate OMe eC CMe, nrcsc0ns} BPs oto

By the normal halogen or alkyl compounds of the hydro-

carbon radieals we mean the compounds formed by the follow-

ing radicals with the halogens chlorine, bromine, and iodine,

or with methyl, ethyl, propyl, &. ; viz. :—
2N 2

a

o03

Authority.

Lefebvre.
Schorlemmer.
Butlerow.
Schorlemmer.
Pelouze.

} Lwow.

Schorlemmer.,

Wislicenus.
Schorlemmer.
Butlerow.
Thorpe.
Ladenburg.
Just.

Thorpe.
Friedel.
Thorpe.
Williams.

\ Lwow.

} Krafft.

Wurtz.
Silva.

\ Krafft.

9

Flawitzky.
Rizza.
Jawein,

Kaschirsky.
Markownikow.
Eltekoff.
Williams.
Fossek.
Butlerow.
Eltekoff.
Reboul.
Flawitzky.
Reboul.
Hecht.
Tilden.
Henry.

504 Prof. T. Carnelley on the Periodic Law.

| Series 1.| Series 2. | Series 3. Series 4. Series 5.
teri ie Sees (C, IDX; | (C, 5x, (Cae
apo eee eee (C,,)X, (C,H.)X, | (C,H,)x, | (Ce

“ae cae eae (O,H)X, | (C,H) AS Pr (C,H;)X, (| (Cee

» IV.... CX, (C,H, )X, (C,H, )X, (C,H,)X, (O;H,)X4
Vou. (CH)X; (C,H,)X, (C,H,)X, (C4H,,)X, (C5H,)X,
» VAI.... (CH,)X, (C,H,)X, (C,H,)X, (C,H,)X, (C5H,,)X,
», VII. ...| (CH;)X,) (C,H;)X, | (CsH,)X, | (C,H,)X | (C;H,,)X,

Series 6, 7, 8, &c. being |
similarly constructed.

In the above, X=Cl, Br, I, CH;, C.H;, or C3H,, &e.

The above series, of course, might be considerably extended ;
but I have not thought it necessary to carry my investigations
beyond the 8th series for the halogen compounds, nor beyond
the 5th series for the alkyl compounds*.

When there are several isomeric modifications of the same
empirical formula, only those which are strictly analogous
should be compared ; thus, normal compounds should be com-
pared with normal, and iso-compounds with iso-, &. In
some instances, however, this is not possible, owing to the
data being incomplete, or to the constitution being unknown,
or when the compounds belonging to different groups of the

same series are to be compared ; in this case the mean of the |

melting- or boiling-points of the several isomeric compounds
is employed. Asa general rule, however, the same results
are obtained, no matter whether we use the mean melting- or
boiling-point for all the isomeric compounds, or whether we
compare only those compounds which are strictly analogous.

Now the same general rule that we have already shown
(ibid. xvill. p. 2) to hold good as regards the corresponding
compounds of the elements also holds in the case of the above
hydrocarbon radicals; for in whatever way we may arrange
the melting- or boiling-points of their normal halogen or
alkyl compounds, provided only that we arrange them syste-
matically, we always find that certain definite and regular
relations may be traced between them. The following are
examples of such relations :-— |

RELATION 1, Lf the above hydrocarbon radicals be arranged
in the order of their atomic weights, then the melting-points and
boiling-points of both their halogen and alkyl compounds rise
and fall periodically.

This periodicity is such that the values increase from the

* The methide curve in the diagram which accompanies the paper has,
however, been completed to the end of the eighth series.

Prof. T, Carnelley on the Periodic Law. 505

first (monad) up to the fourth or middle (tetrad) member of
each series, and then diminish to the seventh (monad) or last
member (see Table II.*).

This periodicity is not exactly the same as that in the case
of the corresponding compounds of the elements ; I shall not,
however, discuss this point at present, but defer it to my next
communication.

TaBLeE I].—Illustrating Relations 1 and 2.

CHLORIDES. | BROMIDES. IopIDES. |

MP, | BP. || MP. 4 BP. || Mp. BP.

eee i. eee wee ih tee eeee oth eve fis. fe awtwead” |i «-aandee?. Ey eele eee

een a ne enenh | one. CS eee ee K "es o'sets ET . eleiwate © My 9 e.cesfaie ) OF eamnd ave

Ov... || 248 350 364 462 420 620
(CH)iit || 203 334. 281 424 SOM hye
(CH,)i || Liquid. | 315 || b. 261 370 277 453

) (CH,)i | Liquid. | 249 | Liquid. 278 || Liquid. | 315
DE SET eG) eee a ec | eee a
(C,H)™ || Liquid. | 360 || Liquid.| 435 |} ...... | uu...
(C,H,)¥|| Liquid. | 411 327 Zo NEE Gap aah Bags ae
(C,H.)#4|| Liquid. | 3867 || Liquid. ABs itt ic HR Soe

Bee eee yl) figad. |. 345. |) .-.2.... 393 355 453 d.
2 (C,H,)i || Liquid. | 285 || Liquid. | 312 || Liquid. | 345
TE ee ie each Gi ck Pn.
S |(C,H,)#| Liquid. Fa ee ae Rams Fe es ee ae eg Sete aie Ce eee
© | (C,H,)#|| Liquid. | 402 |) Liquid. | 457 SO Ph inact
& |(C,H,)|| 434 447 405 iE eh Se A Se eee
& | (C,H.)ii|| b. 263 415 289 A Veailee oo Oe ty co
3 |(C,H,)# || Liquid. | 366 || Liquid. | 411 | Liquid. | 460
| (C,H,)' || Liquid. | 315 || Liquid. | 339 || Liquid. | 368
Og is il ea eta 9 ee ee | Gee ee eee eee
(TEES Sg PSE SO | Ree VG ig reer amen ona Serene
UE ie CSS AS | OR a Geer Gece! Beeneeee
(O.H,)" || 346 0)... oo See eM! Mees Seen
SYS | eee eae ae OS EE ae io See
(C,H.)" |; Liquid. 386 Liquid. 1S Sa eg, Seems
(C,H,)i || Liquid. | 339 || Liquid. | 361 || Liquid. | 390
ei Bcceatd. yt 5G f)- |i: ssc ened ik seep ores Sox. Of oa
(C;H,)# || Liquid. ESS. Sikh cxondams b taate as SORTS Byer
meme Pid) 27 econ ness IR? vaneocs b «eras
(CH, )iv || Liquid. 508 388 Ss ig! Ras ta Saeeee Be
eee Paguad | 454, || Solid [oo essa | eas
GO tequid’ }* 417 ||: Tiquid. | 453 ff}... | wae.
(C,H,,)'|| Liquid. | 371 || Liquid. | 891 || Liquid. | 416

* The periodicity is rendered still more evident by the curves for the
methides, chlorides, and bromides shown in the diagram (p. 514),

506

Protet: Carnelley on the Periodic Law.

TaBLE II. (continued).—Illustrating Relations 1 and 2.

METHIDES. ErnipeEs. Ree PROPIDES.
M.P 134 M.P BP M.P B.P
y Solidifies

Ce cares } 282 Lae oo:
(CH) || Liquid. 256 || Liquid. | 3/0 ||: <a see
(CH,)# || Liquid. 245 || Liquid. | 311 || Liquid 371
Cole ee | oes Gas. |} Liquid. 245 || Liquid 274

Cy ce MW Aves | cones © PSS
(C,H) || Liquid B09 ll eeneee? | tees)
(O,H,)' |) Liquid Sy | MP
(O,H,)" || Liquid 2O3....||. woven | eeeees
(C,H,)# || Liquid 265 || Liquid. | 341 Liquid. | 398
(C,H;)' -|| Liquid 245 || Liquid. | 274 || Liquid. | 311

a 1(O,F Ul | dee oll Ge?) bins
= |(C,H,)°||Tiquid. | 814 |)... | ° 3. |
& | (C,H, )H | Liquid. B43 fl esses | | ine SH eee
©...1(C,H,)" || Liquid.| 860 |) occ2- 1 jae I ee
2 +1 (GH) || Lignid. | 380) |...) 2... 7
ep 6 | (C,H, )# || Liquid. | 299 || Liquid. | 366 222 422
a (C,H,)i |] Liquid. 265 = || Liquid. 307 Liquid. | 340
(C,H)! || Liquid.| (823° |) .....5 04) ..2ce 0
(O,H,)# |) Liquid. i a ee Me | ee
(C,H) )| Liquid... B58. jf cece |) eee
(C,H,)* || 369 BBO. i eaceee |, Get ees
(C,H,)iti || Liquid.| -864 || ss.c0 | «2s. ee
(C,H, )# || Liquid. 339 || Liquid. | 398 241 446
(C,H,)i || Liquid. | 299 |) Liquid. | 333 || Liquid. | 366
(C;H;)! wdveee ML Seevee |] seen 1° skeeee l®
(C,H, )# || Liquid. sa | TS |
(C,H,)@ || Liquid... S87-.-)| ssescs-- |] ..eeee “1 Gee
(O,H,)¥ |) Liquid. 404. || neces | sneeee <] eee
(C,H, )#|| Liquid. DOL: || wacsse iq | weeeeetadl ie en
(O;H,,)#)| Liquid. 365 222 422 247 468
(C,H,,)i|| Liquid. 380 || Liquid. 365 Liquid. 398

N.B.—In the above table the mean value is given when there are several
isomers having the same empirical composition. Sometimes, however,
in the case of the melting-points, there are several pairs or triplets of iso-
meric substances of which the melting-point of one only is known, the
others being liquid; a mean number is therefore impossible; in this case
the known melting-point is given.

1 This is the mean of 448°-475°, which are the temperatures given
in Watts’s Dictionary (i. p. 768) as those between which protochloride
of carbon melts and boils. Chlorine does not act upon this compound
even in sunshine. It is most probably a polymer (C,Cl,)n. Berthelot in
fact regards it as C,,Cl,,._ Its vapour-density, however, does not appear
to have been determined.

? This is the boiling-point of chlornicene, as given by St. Evre (Jahres-
bericht, i. p. 530). This substance, if a definite compound, is probably a
polymer (C,H,Cl),.

Prof. T. Carnelley on the Periodic Law. 507

RELATION 2. The melting-points and boiling-points of either
the halogen or alkyl compounds of any hydrocarbon radical
increase as we pass from the chloride to the bromide and thence
to the iodide; or from the methide to the ethide, and thence to

the propide, §c. (See Tables Il. and III.)

TasLe I]J.—Illustrating Relations 2 and 3.

| CHLORIDES. BROMIDES. 1oDIDEs.
com- ||
ae | MP. | BP. | MP. | BP. | MP. | BP.
(| CH, ...|| Liquid. | 249 || Liquid. | 278 || Liquid. | 315
: ace | Liquid. | 285 || Liquid. | 312 || Liquid. | 345
{Normal primary | | C,H, ...|) Liquid. | 319 || Liquid. | 344 || Liquid. | 375
halides of C,H, ...|| Liquid. | 351 || Liquid. | 373 || Liquid. | 403
Group VIL. 1|C;H,,.../ Liquid. | 380 | Liquid. | 402 ||Liquia. | 429
| OH,.(CH,),.-X. | CH,,...| Liquid. | 406 | Liquid. | 428 || Liquid. | 454
| || CLH,;...| Liquid. | 432 || Liquid. | 452 || Liquid. | 474
( \| CH, ...| Liquid. | 456 Liquid. | 476 |) Liquid. |a. 494
Alkyl
. | com- METHIDES. ETHIDEs. a-PROPIDES,
| pounds. |
| atin EEE EE | EEE
b sie.
| ead priviary { | C2 3 oe Cae Hh taht 245 || Liquid. | 274

a1 ae 245 || Liquid. | 274 | Liquid. | 311

lkides of Te rh iqui iqui
| a VIL. 7 ---|| Liquid. | 274 || Liquid. | 311 || Liquid. | 344

CH,
C,H
‘ C.H
| C,H, ...|| Liquid. | 311 || Liquid. | 344. || Liquid. | 371
CH, (CH )u-X- || oom -..| Liquid. | 344 | Liquid. | 371 || Liquia. | 398
(CH
(
(

eee

C
| Normal alkides C,H,)" || Liquid. | 274
| of Group VI. C,H,)" || Liquid. | 311
| X.(CHe),.X. | | (C4H,)" || Liquid. | 344
| (C;H,,)""|| Liquid. | 371

Liquid. | 344 || Liquid. | 398
Liquid. | 371 222 422
Liquid. | 398 241 446

|
So ae 245 || Liquid. | 311 || Liquid. | 371
| 922 | 422 || 247 | 468

_ RELATION 3. The melting- and boiling-points of either the
halogen or alkyl compounds of the hydrocarbon radicals of any
one group increase as the atomic weight of the hydrocarbon
radical increases. (See Table III.)

RELATION 4. The differences between the boiling-points (and
also between the melting-points) of the chlorides and bromides,
bromides and iodides, chlorides and iodides (or between those of
the methides and ethides, ethides and propides, propides and
butides, Sc.), increase algebraically from the first member of each
series of hydrocarbon radicals up to the fourth or middle (tetrad)

508 Prof. T. Carnelley on the Periodic Law.

member and then diminish to the seventh or last member. (See

Table IV.)

Taste [V.—Illustrating Relations 4 and 5,

Be) eee ie. Pr— Et. | Pea
Dit matte 266 47 91:5 | 14 | ‘ja | ee

EP | BP. | Be | BP. | BE. | Bee
Peo ¢ 112 158 270 ||... |)
(Ory... Tyas, poe Bee meyer
(poe... 3 BB 83 138 66 60 126
ROH A see. 28 39 66 eae og he ae
(Ot cl) See Pl 1 ec
(Cog) csiegell Useceee) [0 ches) [ saeees I) ccanae | eee
(CB: Bb le | ae a
(CoH) vos 66 Jou PE eS
(C\H, yi... 61 jae eee eee a
(CHa, 2s. pga 94 || "60 51 ill
ea ie 24 29 Ce (a 33 70

- The number of cases in which Relation 4 can be applied is
_ unfortunately very limited ; but, in so far as it is applicable,

the data obey exactly the same rule as those of the halogen
and alkyl compounds of the elements.

RELATION 5. The differences referred to in (4) increase as
the difference between the atomie weights of the halogens (or
of the alkyl radicals) inereases. In the case where the differ-
ence between the alkyl radicals is the same, the differences be-
tween the boiling-points (or melting-points) diminish as the
sum of the atomie weights of the several pairs of alkyl radicals

increases. (See Tables IV. and V.)

RELATION 6. The differences referred to in (4) diminish

algebraically as the atomic weight of the hydrocarbon radical
increases. (See Table V.)

Prof. T. Carnelley on the Periodic Law. 509

TaBLE V.—lIllustrating Relations 5 and 6.

Br—Cl. | I—Br. I—Cl. | t—Me. Pr—Et.| Pr—Me.

Diff. in e
eee \ 445 47 91-5 14 14 28
fe (ee er | ee) bee. Be.
(| OH, 29 37 Saal Saat doe fe
C,H; 27 33 60 99% | 37 66 *
Normal halides | | C,H, ...|| 25 ol 56 37 3 70
and alkides of || C,H; ...) 22 30 52 33 27 60
Group VII. 1) C,H, 22 27 49 27 27 54
CH, (CH), . X. | C,H. 22 26 48 27 24 5
CH. 20 22 42 24 24 48
\| OH, SUR ee en aa 24 29 46
| (| (CH,)# || 55 83 | 138 6a Psat Ps tes
Normal halides | | (C,H,)# 48 a.60 ja. 108 70 54 124
Wi and alkides of ! (C,H,)i|| 45 49 94 60 Pie ae
Group VI ‘SAS Ot 5 SPR Beck eee sarge eee 54 48 102
Meee OH.) se | | a 51 46 97
| EL) ae) ai aes 48 41 89

:
| * These four exceptions are all due to the boiling-point of propane, C,H,,
| being taken too high. The number used is based on the data given by

Lefebvre, who states that propane is condensed to a liquid at —20° to
| —30° C,

RELATION 7. (a) The differences between the boiling-points
(and also between the melting-points) of the halogen (or of the
alkyl) compounds of the hydrocarbon radicals of each group
diminish as we pass from the chlorides to the bromides, and
thence to the iodides (or from the metiudes to the ethides, and
thence to the propides, §c.). (See Table VI.)

(b) These differences also increase as the difference between
the atonuc weights of each pair of hydrocarbon radicals increases.
(See Table VI.)

In those cases where there is a common difference between the
atomic weights of several pairs of hydrocarbon radicals, then the
differences between the boiling-points increase as the sum of the
atomic weights of the several pairs of hydrocarbon radicals

diminishes. (See Table VI.)

510

Prof. T. Carnelley on the Periodic Law.

TaBLE VI.—Illustrating Relations 7a & b.

x=
Group VII.
CH, .(CH,)n.X | Diff. in
at. wts.
C,H, eC C_H,; \
7H; ec C,H, |
C,H; ie C,H,
C5H,, é O,H, 4 H,
OF s ee OM co ==9,
OH, 7 C,H,
C,H, mw 3
C,H, a C,H; |
oye 7 Oe |
Gala by tere 9 ICH.
C;Hy, nA O,H, paar
C,H, Ge C,H, |
OH, ae 3 )
CH, i? C,H,
C,H; as C,H,
C1., — C,H. 3CH,
C,H,, — C,H, = 42
C,H, aor? 3
oe ie ne
feels eee 4CH.
CHa; - C,H, ERG;
C,H,, — CH,
C,H, im) C,H, >
CH. 0-H. 5CH,
C,H,, — CH, =40
CO;H,,.— C,H, || 6CH,
C,H,, — CH, J =84
C,H,, — CH, (CE,
ds

Chlo-
rides.

B.P.

207

Bro- : Me- : Pro-
mides. lodides. thides. Kthides pides
pe. | Bre. | BP > Bee
24 | 20 o4 | Tage 1
24 | 20 7 | ge 4
“be | 29 a7 | 97 24
29 | 96 33 | 27 27
29 | 98 37 | 33 On
32 | 30 29% | 37 33
34 | 30 a9% | 37
4g | 40 | Bt AB
Bo. [eas 5a Uh * ee 48
ces 60.) ae 51
58° | 5A 70 | 60 BA
Paha 66% | 70 60
66 | 60 eo eee eee
mao pe 78 | "5 70
“9 | 71 87 | 78 Bs
sa |. 79 97 | 87 "8
oo 84 99 | 97 87
9574-88 a 97
103 | 91 itt | Mende ae
108 | 99 | ia4 ae 102
116 | 109 | 196 | od eee
joa |> 114 196 Wie
132 | 119 | 148 (eee
140 |. 129° | 15391) 44a
150 | 139 153 | 148
164 | 149. | 177. (go
174 | 159 177 | 172
198 | 179 201 193

* These exceptions are all due to the boiling-point of propane, C,H,, being
too high; see footnote to Table V.

RELATION 8. The differences between the boiling-points T of
the halogen or of the alkyl compounds of the hydrocarbon
radicals of the seventh group, and those of the hydrocarbon
vadicals of the other groups respectively, increase ulgebravcally
from the chlorides to the bromides, and thence to the todides (or
from the methides to the ethides, and thence to the propides, &c.).

(See Table VII.)

Tt Data are wanting as regards melting-points.

TaBLE VII.—LIllustrating Relations 8 and 9.

oa ee : :
ag E 8 | X= | Ga | GRRe | Todides.|. .Me- | mtniaes. [Propides.
< Aq . ° e
ang, + minus a ———_—| en .)]3 ee... es le
38 s= | pif.
one Pee RE.) BY | BP. BP. | BP. |) BP
tes ) oS wis.
See | 90! 141 | 106) .—41° |. —177. | —195 | — 97
CH, — ©,H,,||_ 9.) —U7 | — 82 | —21 | —153 | —11 | — 7%
au, — C,H, — 99 aot ye ae eee — [48 — 102 re iy
mes. C4, Oe ie eS ey oe ee |
See OH.|}—71) — 75 | — 47 |. ...... S194 | 78, | = 48
a. = 0.4, Se = Se Pee eat 11h | 2" 75°) S46
a - C,H, er We SO = OG Bh a 87
OH, — OH; || sl) — 49 | — 23 | ...... = OE | ae Baek OF
SH. + 0.4, et Va Eee ie Baad ea
Me fk i | ee fae. gai eth eS So
mel, —.- 0,4, — 36 | — 3] +50 | — 66*| — 33 0
Bae. ¢—’ C,H, BPO keg APL PR eo oy Shee he
ees |. 43] — 15 | + 10 | +46d. | — 60 | — 27 | Oo
aS | A ee ee omer le ee — 54) = 24 0
a S| eee ee (ame gy © i eM ate
mon, — C,H, re ie bee (| = 28%). 0 4 98
mi — CH, = 2S Sea On ee Be io 0 ay
meee Oe | 99) + 13 | + 36 | 471d. | — 33 | . 0 + 24
2 Sd | Ee eceres eae ae Gene WoT F 6 + 24
Sal | Ga | eee meneame een a a 4. 22
Ber — CH, 20 15 be] 4-108 0 4. SF (oe Ge
oa... 0,H. fee oes A eee 0 + 33.) 458
won. — C,H, |+$—15] + 42 | + 65 |+97¢d.] 0 +27 | +51
Se et | ee eer Meee 0 oF | 4 48
I ee | octewe | entees 0 + 24 | + 46
men, — CH, |) ae |) 4199 | 4186 + 66*| + 98
oe — CH, Te OB... 45 90%) 42-70) | 87
ee Ga |} 1 474 | 4 94 | +1950.) 4:97 | + 60 | + 78
SS Ti i a |e ee Gee a ea ae a
2. Se ay es ean Re Oe as aly Oe
nO.H, — OH, <: el een Beas + 99 | +124
CH, — OH, ||, 43] +108 | +126 | +155d.) + 66*| + 97 | +111
oa. = OF | be oeg ae Ree 9 fe ey pee ig eal.
1s. tL a a ae as la “60 | + 78:1. 9b
ec... — cH, yaa | 160. |. fae a sk. +126 | +148
Le | ee eee eee + 99 | +124 | +135
os BST 2 ly aie | eee Oe lem oe | ETE | ee
Se — oH, \ Oy ee eee ee eee +153 | +172
PE ULESSTS OM os an i MR +126 | +148 | +4156
loHn — OF; =." 3: | se Ws cone eee (Sana +177 | +193

N.B. The numbers followed by a (d} are not strictly comparable with the others,
as one of the substances is decomposed on boiling, so that its boiling-point can-
not be determined with any degree of exactness.

The numbers marked with an * are exceptions to Relation 9. They are all
due to the b.p. of propane, C,H,, being toohigh. Cf footnote to Tables V. and VI.

512 Prof. T. Carnelley on the Periodic Law.

RELATION 9. (a) The differences referred to in Relation 8
increase algebraically as the algebraic duference between the
atomic weights of the hydrocarbon radicals increases. (See
Table VII.)

(b) In the case where several pairs of hydrocarbon radicals
have a common difference in atomic weights, then the differences
between the boiling-points of their halogen conpounds increase as
the sum of the atomic weights of the several pairs of hydrocarbon
radicals increases. This rule also holds for the alkyl compounds
so long as the difference between the boiling-points is a minus
quantity (7. e. so long as the boiling-point of the alkyl com-
pound of the polyvalent hydrocarbon radical is lower than
that of the corresponding alkyl compound of the monovalent
hydrocarbon radical), but when this difference becomes a plus
quantity the rule is reversed. (See Table VII.)

For the purpose of illustrating the above relationships we
have chiefly made use of the compounds of the sixth and
seventh groups of hydrocarbon radicals, because the sets of
compounds belonging to these groups are much more complete
than in the case of the other groups ; the rules, however, hold
equally with the latter as with the former.

The above nine rules have been applied in 1997 cases, of
which 71 (or 3°6 per cent.) are exceptions. The exceptions
are distributed as follows :— |

HALOGEN CoMPOUNDS. ALKYL ComPounpDs.

Melting- Boiling- Melting- Boiling-
points. points. points. points.

No. of | N00, |No-of| Coue. |No. of | NOOfll No. of| oso!

Sere Teint Wr tions CAaBes. | tions, || “1” gions
Frelatvon: 1." .his.4.. 20 2 68 1 5 0 48 0
wh _ Dead ae 20 0 60 q 8 0 49 0
i 2 Ai ee a me oe 104 1 6 O 73 0
9 ‘hy Le 8 2 35 1 Pah, 26 0
be (a ls eee Be 1 0 66 0 30 1
5 . 5, ae oe oe 78 1 42 6
ac. Lee 100 0 73 7
" VIL. 3b; 103 ) 2 0 102 14
i | oe oe 108 0 285 0
* IX.a & b. 106 2 361 32
otal oboe): we en oe oa, 21 | 0 | 10389} 60
moet fe —()'8 0/, —5'5 oN

Prok © Carnelley on the Periodic Law. 513

Of the above 71 exceptions no less than
45 are due to the b.p. of propane, C;Hg, being too high.

4) Ma = isobutane, CH Mes, being too low.
2 ee n.p. of carbon tetrabr omide, CBr,, being
— too high.
52

So that 52 of the 71 exceptions would be accounted for by
errors in the boiling-points of two of the above compounds,
and in the melting-point of the last one. As regards propane
there is certainly a considerable error in its boiling-point, or
rather in the number which has been assumed as the boiling-
point in our investigations ; for its actual boiling-point has
not been determined, the value used being in fact the mean
temperature of condensation, viz. —25° to —30° (cf. footnote
to Table V.). With respect to isobutane, the boiling-point em-
ployed was that (— 17°) given by Butlerow(Annalen,cxliv.10).
Konovaloff states that it boils below —10° (Bull. Soc. Chim.
(2) xxxiv. 333). Finally, carbon tetrabromide is not only an
exception to some of the relations noted above, but it was also
shown, in my paper on the Halogen Compounds of the Kle-
ments, that, owing to its abnormally high melting-point, it was
the most marked exception to some of the rules there stated.
In conelusion it will be seen, by a comparison of this and
my previous papers, that the above nine relationships, together
with others of less importance, hold good for the following
four classes of compounds :—
(1) The Halogen compounds of the Elements (7. e. of ele-
ments with elements).
(2) The Alkyl compounds of the Elements (i. e. of elements
with hydrocarbon radicals).
(3) The Halogen compounds of the Hydrocarbon Radicals
(i, e. of elements with hydrocarbon radicals).
(4) The Alkyl compounds of the Hydrocarbon Radicals (1. e.
of hydrocarbon radicals with hydrocarbon radicals).
Though the periodicity of Relation 1 holds for all the four
classes of compounds, yet the nature of the periodicity is not
exactly the same throughout for the compounds of the hydro-
carbon radicals as for those of the elements. |
We are therefore justified in concluding that, with the ex-
ception just stated, the physical properties of the compounds
of the hydrocar bon radicals (so far as investigated) follow the
same rules and exhibit exactly the same relationships as the
corresponding compounds of the elements.
In my next paper I hope to indicate the theoretical signi-
ficance of this general fact.

>
~
S
MW
ay
or
a
Y
ag
bo)
=
OQ
a
aD}
—
—
aD)
=
Lee
cs
©
|
Cay
©
a

|

x
ie

a Sopry jay .

v

: |
| |

SE re

-|_+Sappzga7 LOg.Lp IL ple Hy £4 SJ

|

|
2 ee a
i
|

4A = a a

oT

a as
alestiemedaetihaicle

I

|
1
+ |
{

ie i |
| _ | |

Sicace

pop

7

SIPWMOVY)
a al

{
|

iu
|

NIS929

FIA

|

Buty?

}

|

vee ee ——

— Oe ai

‘s[VoIpey Uogrvo0ApAF] oY} JO Sopiwmdig, pure ‘sopLlopyyD ‘sepry.ey [eULION ey} JO szULog

ie

ao faa

ao eae

T

rTP rT

|

-SUTTIOG 94} JO WeRLocLI

S
7

Geer

LXI. Some Thermodynamical Relations.—Part I.
By Wittiam Ramsay, Ph.D., and SypNEY Youne, D.Sc.*

| Cage relations to be considered in the following pages are,

we believe, well founded ; but we hope to confirm them
by more exact experiments than have as yet been made.

It appears advisable here to state the share which each of
the authors has had in this work. For our purpose, the
equation

br dp ot

Ss, dt ; J

will be employed. The relations of the first term of this
equation were the subject of a communication to the Chemi-
cal Section of the Philosophical Society of Glasgow in the
year 1877 by Dr. Ramsay, which, however, he considered too
incomplete to be published ; while the application of the last
term of the equation to the vapour-pressures of substances,
which formed the subject of a joint research by both authors,
was independently discovered by Dr. Young. It is also right
here to mention that some of the relations discovered by Dr.
Ramsay have been pointed out subsequently and independently
by Trouton (Phil. Mag. 1884, vol. xviii. p. 54).

The term é z (where L represents the heat of vaporiza-

tion of a liquid or solid substance at its boiling- or volatilizing-
point, and where s, is the volume of the liquid or solid at
these temperatures, and s, that of the gas into which one or
other is converted), if stated in words, denotes the heat ex-
pressed in units required to produce unit increase of volume
of substance at the temperature of ebullition of the liquid or
volatilization of the solid.

Two laws have been discovered, representing certain rela-
tions between different liquids.

The first law may be stated thus :—The amount of heat re-
guired to produce unit increase of volume in the passage from
the liquid to the gaseous state at the boiling-point under normal

pressure is approximately constant for all bodies; or ao
— ik a <2

The data on which this law is based are imperfect. The
heats of vaporization of but few bodies have been determined

* Communicated by the Physical Society: read November 28.

516 | Drs, Ramsay and Young on es

with accuracy; and we have recently shown in a communica-
tion to the Royal Society that the results obtained by Favre
and Silbermann, by Regnault, and by Andrews for ethyl
alcohol are only approximations to truth. It is probable that .
the data given for other substances less easy to obtain in a
pure state are still less to be relied on. We have also shown
in that paper that the density of saturated vapour of alcohol
is normal, or nearly so, only at temperatures below 50°.
ie ees. in calculating by means of the above-mentioned
formula the constant for alcohol, it has been assumed that the
saturated vapour of alcohol possesses normal density at 784;
and from want of knowledge it has similarly been assumed
that at their boiling-points the saturated vapours of the other
liquids to be considered have also normal density. ‘It must
therefore be acknowledged that considerable doubt rests on

both the expressions in the term z , and this doubt pre-

Ss

1
vented the previous publication of these relations; but it will
be seen in the sequel that the much more trustworthy deter-
minations of vapour-pressures amply confirm the law which
has been stated.

The following table exhibits these relations at normal pres-
sure :—

L

Substance. S$, —S. L. | 5 saee
WWiditers uber k te ead: 1695 537 0-3166
Methyl alcohol ....:.... 8659 263°7 03046
Bthyi alcohol... ..; 2.97. 624-7 202-4 0°3240-
Hihiyl omile™ J o.a5isci- 322'3* 90°5 02808
Methyl formate......... 4171 1171 | 02808 |
Methyl acetate ......... 362°5 110-2 0°3040
Ethy! acetate............ 3230 92°7 0:2870
Hihyl oxalate ....:.... 256'2 72:7 0:2835
Amyl alcohol (?) ...... 375°5 121°4 0°3238
ONG) 115) 716) 6 Oh hr rr 318-0 101°9 0:3204
PPPOE 0 BVidp sc cals sien seo 370°6 92°3 0:2491
MPTITOAMIOR ch acc ce sc icebe « 169:2 456 0:2695
Phosphorus chloride..| 209°5 51:4 0°2453
Carbon disulphide ...) 343°3 86:7 02526
Methyl iodide ......... 181°5 46:1 0:2559
Ethyl iodide ............ 180°5 46°9 (0):2598
Kthyl chloride ......... 361°9 100°1 0:2767
Carbon tetrachloride . 186°3 46°9 0:2515
Giiloroform 0 ..3%.2) 229°1 61-2 0:2673
RURCOIT sp juninyscvsssiens 2580 775 03004
Witric peroxide ......... 309°6t 93:48 | 0:3019

* Determined by Horstmann, Annalen, 1868, Suppl. vi. p. 64.
+ Determined by Ramsay and Young.
t Density taken as 39 (T1=1) as an approximation.

some Thermodynamical Relations. 517

It will be seen that these numbers vary between 0°2453 for
phosphorus chloride and 0:3240 for ethyl alcohol. The alco-
hols, water, and acetic acid give all nearly 0°32 ; and benzene,
and bodies containing halogens and sulphur give lower num-
bers, averaging 0°2564.

The sequel will show that these numbers cannot be regarded
as experimental deviations from a constant; but that they
have significance can hardly be denied. It would also appear
that the nature of the element in the compound, and its
amount influence the absolute value; though the nature and
extent of this influence can hardly be formulated.

The second law is :—J/ the amounts of heat required to pro-
duce unit increase of volume in the passage from the liquid to
the gaseous state be compared at different pressures for any two
bodies, then the ratio of the amount at the boiling-point under a
pressure p, to the amount at another pressure po, is approxi-

_ mately constant for all liquids.

| Or,
$1 — So

- for all liquids, and probably for all solids.
In support of this statement the facts to be adduced are as
\ follows :—

|

at p;, bears a constant proportion to at Doy

81 — Sg

3000 | 1433 | 479 | 505-7 | 1:055
4000 | 1542 | 369 | 497-7 | 1°350
5000 | 162°8| 301 | 491-6 | 1633
| 6000 | 170°1 | 255 | 486-1} 1-906
: 7000 | 177-0 | 222 | 481°3| 2°168 || 148-0 | 31-4] 52°8 | 1684
|

|
| Waiter. Chloroform.
2. | S,—S L s,—s L

| mms. : ccs. | S;— Sy z C8. “ $,—- S$,
ern ici tlt th gee} fh
) 300 | 76°4| 4022 | 552°1 | 0°187 || 35:0 | 585 | 64:65} 0°121
| 760 | 1000 | 1695 | 586°5 | 0°317 || 61:0 | 229 | 61-2 | 0-267
1000 | 107-1 | 1813 | 531-5 | 0-405 || 69°3 | 178 | 60°4 | 0:339

2000 | 129°3| 695 | 515°6 | 0°742 || 92°5 | 95:3} 58-2 | 0-611

8000 | 182°9 | 197 | 477-0 2424 | 155°0 | 27-9} 52:2 | 1-871

The ratios, placing

at 300 millims. pressure= 1:0 for

Ma
each substance, are :—
| eee 300 760 1000 2000 3000 4000 5000 6000 7000 8000
Pe Water ......... 1:00 2°30 2°94 540 7°68 9°81 11:88 13°86 15°77 17°63

| Chloroform... 1:00 2:22 281 506 7:07 894 10°72 12:38 13:96 15°51
)) Phil. Mag. 8. 5. Vol. 20. No. 127. Dec. 1885. 20

|

| |

518 Drs. Ramsay and Young on

Another method of comparison is to regard the value of ,

wae for water as equal to unity at each pressure, when the ,
Lines
numbers for chloroform become

1:00 0-962 0-953 0-938 0:922 0-911 0-903 0:893 0-885 0-850.

Similar results may be obtained on comparing other bodies |
with water; it may be, however, remarked that the instance
chosen is one of those in which least concordance is to be
noticed. It is evident that here also the concordance is
merely approximate.

Another point worthy of notice is this. The heat of
vaporization is expended in at least two channels :—it pro-
duces expansion against pressure, thus doing external work ;
and it is partly expended in internal work on the molecules
of the body. Now it follows from what has preceded that
the internal and total work bear an approximately constant
proportion to each other for any one pressure, whatever be
the liquid. Thus at a pressure of approximately 760 millims.
the ratio of external to total work is shown as follows :—

g Total work. b Total work.

Eperane: External work. pease External work,
Feremine = cia tae 11:09 Ethyl acetate .....:...... 11°81
Phosphorus chloride ..._ 10:09 Ethyl! oxalate: —.-). ene 11:67
Carbon disulphide ...... 10°39 Amyl alcohol »aeeemee 13°32
Methyl alcohol ......... 12°53 Ethyl chloride:............ 11°39
Hthylvale@hol | 5. .....0283 13°33 Benzene .¢si..cssaaeeeee 10°25

*Ethy! oxide (ether) ...... 11°55 Chloroform. |, ...pesnenem 10°99
NY IETS ss shitls ote > tie caver HheoO2 Carbon tetrachloride ... 10°35
Methyl iodide ............ 10°45 Mercury <.:3. pee 12°36
Weyl LOdide, pies... -0ero- 10°69 *Nitric peroxide .....5... 12°42
Methyl formate ......... 1ibS Acetic acid’ <.. cea es 13°18
Methyl acetate............ 12°51 Water .... 5... eee . 13°02

* Vapour-density found at a few degrees above boiling-point.

Bhs
8;—8, dt J’
mately known for a number of substances. But before com- |
paring them it is necessary to reduce the pressures which |
have been given in the previous part of this paper in milli- |
metres of mercury, to grams per square centimetre. ‘The ||
comparison has been made for water, with the following |
results :—

In the equation both terms are approxi-

some Thermodynamical Relations. 519

Pressures, in

- L dp ¢
oo. is ae
ON Sects lyr? Glos 01395
BOO) icky, =) OPAOD 0-414
B00). 3.5. OTD 0-770
SOOO 25° Fe ODS 1-110
Mee te. we A BOS 1-743

The non-equality of these numbers may be ascribed to
various causes: want of accuracy in experiment, most pro-
bably in determination of the heats of vaporization, and the
assumption that the density of the saturated vapour of water
is normal at high pressures, are among the most likely.
With such results for water, it was thought unnecessary to
compare these terms for other liquids, where the constants
are probably less accurately determined. These numbers,
however, suffice to prove that the relations pointed out in the
first part of this paper, although derived from observations
and assumptions not always trustworthy, agree fairly well
with those deduced from the vapour-pressures, which have
been determined with a high degree of accuracy.

We now proceed to exhibit the relations between the
vapour-pressures of various liquids. The constants of the
bodies considered were determined by Regnault, Olszewski,
Naumann, Isambert, Moitessier and Engel, and by ourselves.

If a curve be constructed to represent the relation of tem-
perature to pressure for any substance, and if tangents be
drawn to touch the curve at various points corresponding to
certain temperatures, these tangents will give the rate of
increase of pressure per unit rise of temperature ; in other
words, the value “3 for those temperatures.

If we construct curves for a number of substances, and

determine the value of 2 for each of them at the same tem-

perature, it is clear that the values obtained will differ widely,
and will be greater for volatile substances than for those
which are less volatile. But if we determine the values of

7 for the same series of bodies, not at the same temperature,

but at the same pressure, the conditions under which the
comparison is made will be more similar, and the resulting
values may be expected to differ much less.

202

520 | Drs. Ramsay and Young on

In the calculation of the vapour-pressures of a number of
substances for each degree between certain limits of pressure,
it became evident that at any given pressure the rate of in-
crease was generally, though not always, greater for the
volatile substances than for the less volatile.

It was found that the product of the absolute temperature
into the rate of increase of pressure ) at any given
pressure was approximately the same for the bodies examined,
but the differences were evidently too great to be ascribed to
errors of experiment or of calculation. That this product
should be approximately the same for different substances
might perhaps be anticipated from the following considera-
tions :—If there are two bodies, the absolute temperatures of
which must be raised to 200° and 400° respectively, in order
to produce a certain effect, the same for both, it might be
expected that a further rise of temperature of 1° would pro-
duce a greater effect on the substance whose temperature was
200° than on that at 400°; for the rise of temperature in the
first case is greater in proportion to the temperature to which
the body has already attained than it is in the second, the
rise in the one case being from 200° to 201°, and in the other
from 400° to 401°. The rise of temperature would perhaps
rather be proportional if the temperature of the hotter body
were raised from 400° to 402°, or if the rise of temperature
in each case were made proportional to the absolute tempera-
ture. In order, then, to make the conditions as similar as

possible in the calculation of the value of = for different sub-

stances, the magnitude of the unit degree of temperature
should be made to vary in the same ratio as the absolute
temperature of those substances, corresponding to the par-
ticular pressure at which they are compared ; or, in other

words, keeping the unit degree constant, the value a should
be multiplied by the absolute temperature ¢.
d
The values of a . ¢ were determined for a number of liquids
at several different pressures ; and it was found that the pro-
ducts obtained for different stable substances at the same
pressure were always approximately the same, whatever that

pressure might be. Thus at a pressure of 400 millims. the
following values were obtained :—

some Thermodynamical Relations. O21

Carbon bisulphide. . 4436 1-000
WlGHIGR cos, aie cl var te. JOeS 1°324
Chlorobenzene. . . 4724 1:065
Bromobenzene. . . 4703 1-060
Pr yo ioe wi L24 1°155
Methyl salicylate. . 4959 1-112
Bromonaphthalene . 4930 ALLE
PESTCUP YS iyo ey) br,/ac 7 4012 1:085

This may be better seen by making one of these values
equal to unity, and reducing the others in the same ratio.
The second column is thus calculated.

Repeating this operation at another pressure, a second series
of values was obtained; and these were reduced in the same
way. It was at once noticed that the reduced numbers at the
new pressure were identical within the limits of error of expe-
riment and calculation with those at the first pressure.

The calculations were therefore continued so as to include
the widest attainable ranges of pressure, and at every pres-

sure the value of = .t for one substance was made equal to

unity; and the others reduced in the same ratio. The values

of for water were made =1, because the vapour-pres-

sures of this substance have been determined by Regnault
between wide limits of temperature with very great care; but
it was noticed that the values for both water and alcohol were
very much higher than for the other substances examined ;
and a similar comparison was therefore made by taking the
values for carbon bisulphide as equal to unity.

It was then found that for pressures ranging between 150
and 1500 or 2000 millim. the reduced values for each substance,
with the exception of mercury, were very nearly constant
for all pressures.

In the case of alcohol and water, employing the vapour-
pressures of water as determined by Regnault (A/émoires de
l’ Académie, vol. xxv.), and of alcohol as determined by Ramsay

and Young, the ratio of the values e .t was seen to be the

same at pressures between 150 and 20,000 millim.

Two methods of calculation were adopted. In the first and
more accurate method the vapour-pressures for each degree
were calculated by the method of differences; and from these

522 Drs. Ramsay and Young on

numbers the value of 4 at any pressure could easily be caleu-

lated with very small error. The time required by this pro-
cess is very great; and it was only adopted in the case of
water, for which the pressures for each degree between —32°
and 230° have already been calculated (Balfour Stewart,
‘ Treatise on Heat’), and of carbon bisulphide, alcohol, chloro-
benzene, bromobenzene, aniline, methy] salicylate, bromonaph-
thalene, and mercury between certain limits of pressure,
generally 150 or 200 to 700 millim. (Ramsay and Young,
Chem. Soc. Journ. 1885, p. 640).

In all other cases tangents were drawn to touch the vapour-
pressure curves at points corresponding to definite pressures,
several curves being required for each substance on different
scales to admit of this being done. With care, this method
yields fairly satisfactory results.

It has been mentioned that the identity of the reduced
numbers for any one substance holds at pressures ranging
between about 150 and 1500 to 2000 millim. At lower pres-

sures it is much more difficult to determine the values of z4

with accuracy, and the influence of experimental errors is
greater. Moreover, it has been necessary provisionally to
adopt the method of tangents for all substances except water
and carbon bisulphide at low pressures ; and it is doubtful,
therefore, whether much confidence is to be placed in the
values at pressures below 150 millim.

As regards high pressures, with the exception of alcohol all
the substances have been investigated by Regnault ; and it may
be worth while to mention one or two facts which appear to
throw some little doubt on the accuracy of some of his determi-
nations at pressures above 2000 or 38000 millim. In vol. xxvi.
of the Mémoires de Académie Regnault gives the vapour-
pressures of a large number of substances for each 5°, calcu-
lated from his empirical formule. Many of these series have
been examined by the method of differences ; and it has been
found that in several cases the numbers forming the third set
of differences increase slowly up to pressures of from 2000 to
3000 millim., but decrease at higher pressures. This is
notably the case with alcohol; and it has been shown by
Ramsay and Young, in a paper read before the Royal Society
in May 1885, that the higher vapour-pressures of this sub-
stance as determined by Regnault are too low. It has also
been pointed out by Vincent and Chappuis (Compt. Rend.
vol. ¢. p. 1216) that the vapour-pressures of methyl chloride

some Thermodynamical Relations. 523

above 7 atmospheres are much higher than those calculated
from Regnault’s formula, the difference at 140° amounting
to no less than 11°3 atmospheres.

Where this decrease is observable in the third set of differ-
ences, it seems therefore justifiable to doubt the accuracy of
the results at the higher pressures.

It may be stated in conclusion:—(1) That the values of

2 ¢ are approximately the same for all stable substances at

the same pressure, but that the differences are real, and are
not due to errors of experiment or calculation ; and (2) that

the rate of increase of this value oa ¢ with rise of pressure is

the same for all stable bodies, at any rate for pressures between
150 and 2000 millim., while for alcohol and water it is the
same for all pressures between 150 and 20,000.

In the tables which follow, the values of ae t (absolute

temperature), and - ¢ are given for a number of stable
substances at definite pressures, also the reduced values of

e . t, that for (1) water and (2) carbon bisulphide being made

=1°000 at each pressure, the values for other substances at
the same pressure being reduced in the same ratio.

The second series of tables contains similar data for several
substances which dissociate more or less completely on their
passage into the gaseous state. It will be seen that for such

bodies the values of - .¢ at any pressure are considerably

higher than for sciite substances at the same pressure.

The initial letter of the name of the observer of the vapour-
pressure of each substance is given in the table containing the
values of ~ at the foot of the vertical columns. R. stands for
Regnault, R.and Y. for Ramsay and Young, O for Olszewski,
N. for Naumann, I. for Isambert, M. & E. for Moitessier and
Hngel.

T-LOV G.00G eeroce eoeres eveses eeeoree
G.ScP G.GQP eeesee cease eescee eeeeee eeoreoe woeeee eeece cose eecsee @evces eereee eocees | eeeses . es eereece eevee 000‘°CL
L-G&P G9.c9OF eeeeces eeeree ee ee eescee eesees eeeces eeee eevee eoeeee eeeecse eeceee ge coe | weeees ee eoreve eeeeee 000‘0T
¢.80F CS.CEF 1-262 coos e ececece eeccece evcvece ecesee ecoree Z-FOP coerce eecsee P-ELe (NG 1-8FE .61E cecces 1-168 000¢
€-80F BEF | 868 Oe oe fo | ogg] oon | oot ferse| oi | ore | sage] Peon] eeze| o-ege| | e9| 0008
0-8LE G.ZOF 1-L9¢ coeveee ecccce eeecee ZS8L ele qiajets [Liem elm aie G.G9E eeecce eocees 0-8&2 @.GQE FE L-GFE eeucce &.668 000z
= | $:698| Gee] 6-896 | | see | Q.egy| “7 | e699] Lege] “77 | v7 | 8c | 6-6] 0-208 | B-2eE | 777" 6-1FS | 00ST
S | peceles-ose| Lee) 7 | 07 Prilomee sa ee Gey | 22° pop9| Lie) 7 | ot | O-9TS| 0-69 | 0-662 | 9-6TE| “7"" | 0-868 | 00OT
eo | 9-996} 8-2L8| L5re | 77" sees | vere | gen] ct | 8-669] Gees) 7 | TT | G-BTE| 0-998 | 1-066) €-9T8) 777" | ¢-F68 | 006
= leg.cce lap-F1G| L-1Pe| F-66r| 1-86F| 6299] 9-FGL| © | L889) BFeE) “7” 1er| ¢-608 | 0-19] 0-282 | 0-18] “7 | 8-068 | 008
5 | 1.6F6| LOLE| 8-288 G-FS | F-26h| -6F9 | TOTAL] “7°” 6-629 | 6-088 | O-ZOF| 6-2h| L-C0E| 8-9FE| S-E8z | 0-60E] “777 SLOTS | OOL
i lopcpe| c.g9e| 6-6ee| ¢-8PF| 0-98F| €-oF| 0-L0L| "777" | L419] G-9G8| 2-968] 00Gb] &-TOE | OTE) E-62G | 2408) aL-GTE | ONY
~s | L-lse| L198| ¢-626| 61rr| 2-91| L¥e2| 1-969] ““"""" | S809] 0-128) G06E] SETH) 196c| £988 | 6-712 | 8.66 | ECEE | GONE | Oe
S | ooge! ogce! oFee| ZFSF GO-OLF| ¢-F69| 1-689 | O-F9T | 0-269 | 0-G1E| 8-288 | F-S0F| 0-166 | 0-088 | 7-696 | 6-866 | 1-866 [22-008 | OOF
S | gecel ogre! 921s eneer| ¢-ecr| ¢21g| 0-899| 0-091 | 0-889} 2-20¢| L-8Le] 8-968] 0-786 | 8-1ZE] 6-29Z | 8-986 | L-LTE | G-E6C | OOF
p | orze] 9.628| 6-808| FIP] 2.9F| 9.06r| “| GFT] S98] 0-862 | 8-198] GEE) BFLZ | O-TIR) SFEC| Ouke | OaOE | Gade 002
2 |ecre| Leee| 0-808 | 0-F0F| F-98h| 698h] “| FIST] G199) 7" 6-ece | 6-FLE| 9-892 | 0-406 | L-8FZ | 0-122 | 0-008 | ¢-926 | OST
A | @.108) L468] 0-966] 6-368] S-FCF | 7" “+ JOLOPL | 868g | 777" G-EFe | $-F98) $-096 | 0-966 | “7 8-93 | 0-162 | 1-29 | OOT
S| 0-968 STIS] 0-283] 0-228 | 2-C0F | 1°" “++ | Geet | 6-909 | 777" seer Lgapel | QTee] 777" | O08] 77 G0-F92 | 0¢
; GZ-69G C.FRS 0-662 eccces eaccece eeccce | seece . 0-EZ1 0-0¢F seeepeo | ee eme oe | eeeeee noua ter ie win amd aes Jia seee a: ly a einle ee eocence OL
inal Si = fax ee ioe tien | eS egress |e ee ee PS ee | ay a a ee eee
at ce = 5 2 S By) a
OF] Ulepe Op lobaOsu 2) 2] Bl oF |OFaora & 2 OlOE Wel El as |_Zoal Bs
‘Be | pS BER me Ee SEee| Wf | O2|e7 | 2 ESE d) BE (Sess e Heel me Pee aI
Os Os —7>7 ia ee Pe Se lt 2 = Ho SS Qe oie = 5 6 oO = a ad) 28 w ES =a
Be aH) pe CRayeel Site | ae| ta leesit? gS ees ee a Ee

524

525

some Thermodynamical Relations.

‘(wut QOOT

aa0qe)
a
pue
("uLU QOOT
03 OL) “W
OF-80F
OF- LS
OF-8EG
92-8E1
60-26
19-99
GE-EG
84-8
GPUS
97-1
0Z-8Z

‘HO'H’O
poqooye [Aq

‘EU (AP UAR AAP A aA ‘O

FG-VE1
LG-88
6&-G9
6G-0G
89-96
GG-EE
69-06
1v-L6
11-66
99-06
9T-LT
OF-S1
VG-6
SVL

eereee

“TN H’O

QAR OLE one. 4& PROHREAES bid sala sees shame ay ta 8

“o* Ho

‘oye Aoryes

e@eenee

LAueT

"I it: ia 1 9
‘oueTeyyydeu

eneeere

-OWl0Ig

fol aie
my saelea
Oe ae
Oe i
eee
Set ae
Bo Ur ieee
ee
GI-L | L0-06
G9-G | 19-66
9-4 ss
iene
"129-6
Rae
eer
we. | BS

< Lr) x

ie o

‘a

a (APT AR ay

“IOHO
‘ULALOJOLOTYO

1O°H°0

eres | es | ES | | LY | |

-OL10] GD
. Ig°H?°9

-OWlO1g

‘a
00-9TT| 69:96
09-82 | 06-99
PG-8¢ | GL-8P
IL-P | 82-88
LG-E§ | 66-86
G¢-06 | 86-96
8G-L6 | 16-66
GP-GG | LP-1G
O&-66 | 09-61
16-81 | 06-91
66-91 | 88-E1
19-61 | 66-01
16-6 | 40:8
SI-4 | 08-9
10-G | G&-P
pe ays 1G

FIDO
“aplLopyoe.tyo}

‘ay ‘aT
08-611} 6-801
08-68 | O1-¢2
06-09 | 00-9
GG-8P | 9G-PP
GL-GE | oV-GE
LV-GE | O16
LI-66 | S&-LzG
16:96 | 66-16
06-66 | G9-1G
80-06 | 98-81
F6-91 | LV-ST
CFS | Ge-GL
F6-6 | 08-8
v9-L | GT-L

“" 198-6
as, lar
re) 5 esis a e9|
Sees

onye A

eeeeee

eoeeee

eeeeee

ereerte

“i G 9
‘eprpor [Aqyy

‘i
eeesee 000'0Z
eereee 000‘ST
eeeees 000‘OT
00-GOT| 000¢
L9-0L | 000€
OF-GE | 000G
ever | 00ST
08-08 | OOOT
09:86 | 006
L9-2G | 008
OL-E@ | 0OL
86-06 | 009 —
G9-LT | 00S
GL-F1 | OOP
OL-11 | 00&
86-8 | 006
g9-9 | OST
SL4-h | O00r
99-6 | 0¢
ole

S. ay
95-8) B
"Es| ed

| Ps

fo en SS

Drs. Ramsay and Young on

526

OLLOGBT O89Z8T eocce eccces eeerce eovces eee e ec-ee ye: “is se sis . a oe oo . — . oe . eevcee 000‘0Z
OFFI O89ZFT eoecee eceee eeoee . . eee oosee Se : eee a cS fe Bee cre ones on . . eecces 000‘cT
OLSE0OT OGZOOT ec are ey El talelctae A crete eistes oly \sieve kiese nleintelss sor wisiets ste ° paint ale - eng fo erent EI Ri js ote ajainione s 000‘0T
O9F9G SPEC OFFEE coves Beatie ty, \lteueiatswreca, “ifm aeials ayer |) meee pip ail) serene (sub OCkSF Oe ws OR evccce PISEr FEEIP ZOLIF 908 TF oe SZLIF 000g
oLege | gene | 999e8) °° | Bee Cum) ee oe Gisle = ee OELLZ| F9G9%| SZELZ| 9889c| “* | 9G09% | 000E
Ghee | LOR | PORE, ee Tey Tac eae ees poser) L8161| 8L81| IFI6T| SF88T 69FSL | 0002
SIL61 | L9G81 | TST8T} |" ag! via Selo eoroneiereny | GISFI| OLFFI| 9TLFI] 9CLFT| “| LOGFT OOST
LEG | GrEST | G8get| SE te Ua lee a 16GST 9P80T. E1GOT| OLLOT] O6ZOT| TLEOT ZOTOL | OOOT
G6GSI_ | GZSIL | TPPIT) ees OTZOT| "| G4TOT] BOL6 | | ecce | ose6 | Ocr6 | FEEE | “| 1826 |006
FEOIL | ZeGOT | SPFOT| ZLE6 | 4916) 6806 | 6EZ6 |" 6 | ZI88 | Zo9s| oecg | sors | c4es | Fees | “| gezs | 008
Gh86 -| 6186 | 6&6 | 8GE8|EsI8| GOIS | SITS |" S918 | GYSL |SL9L/S889L| SLOL | 9FPL | FZ9L | LEGL Ligh | O0L
1ocg | 9e1g | LI6L | TIGL|OLIL| POTL | 6FOL |" L90L | 0289 |0FL9|8199| $ZL9 | 6999 | 4699 | L6¢9 1989 | 009
L61L | 8069 | OLL9 | 1Ez9 | 909] Z909 | S909 |-""” 6g | 628g |ScLE|FaLE) ZI9G | F89¢ | Ose | FG9¢ | “| LIFE | 00g
eyag_ | gggg | ogce |FZ1¢| ear | Os6r | S98F |OZGF| GIF | GLLF |FGLP| SOL | SGLPr | O8Gh | FOL | LbSh | 98Gb] YFP | 00F
ipcr | tose | ocep [Flor lFise| 2sLe | PeLe |Leoe| 0998 | 9P9E |OL9E | LE9E | SEE | LeaE | Veee | ETgE | FOGE| OEFE | 00E
aeig | 910g | L6G | 8822] 9696] GO9S | "~” F1'Z| E8FS | GLGz |8eaz |ezae| Tees | OTGS | SZGS | BEFG | 9GFG| SLES | 00d
GOFS | GI&s | 146% | FL1G|6G0% | BOOS’ | 7" 6481 | G06T | Z8GI | 9FET| BZ6L | GIGI | OOBT | SE6T | F881) GEST | OST
apor. | LOOT | L198 ISPLPISrt.) 22° 00007 BIFL| Ocel | "7° |F8ST|G6EL| FORT | See | °° | P9ST |STET| O8ZT | OOT
EG8 G.0F8 098 OGL es) eoeeece | ceeeee FE9 £99 coeocse | 2 8ee%8 OFL vce: Col, eeeecee CLL eee 1.G19 0g
CQ 2.F81 I6I ate intsise || hevevernvreress|) =e ehe-s0 Bievevctens OLT 981 a Teigiohe'e) ogilGaieie.e se aie sora wilhn cree siete Le seein eiate | re'e tere PN i ACh eel (Ae eee ee | OT
Eth nha es)

EA eS S| 5 o| £| _F S a8 GO aed 5 ee t
Qe ers Ae oF 1O8 nD & ce be = ag a Sg OS, OF ef =. n
ma | aa BS HE ae ee ey SE | ms SE a8. 8. ae Q8.8 oes |} Hd oe [ae ofS Bg
Se | SE | oe [Bele Es] ME |Fel ee | SS SF Gel os es) oF | es |e) PEs | eg
HS Ps | ms | me meSes) SB) 418s) ek Fe) OF f Be) Hg) ae re} BP 3
: 5 E to io & : = ¢ . B : 8 E . B a: - ¢ =

——s

jo sonye A

‘soInssolg o}Uo(T 7B SeoUVysgnY o[qvIg SNOLTVA IOF (ommyzesrodurey oynfosqe ou} SI 7 QLOTLA) ? *

527

some Thermodynamical Relations.

FFO-1 6¢ ee © | seeree
T¥0-T 66 ww ww we

980-1 6 eee | serene
9&80-T 66 186: seenee
ea a ae
OFO-T 6é 166. eerees

eeeree

eeevee

G90-T C8 eta sll! Se eeietaie.” |] ale .s dime
PS0-T | “ 1/866. | 168. | TLS:
O90-T | ‘" | 286. | 168 | 998:
S90-T | “ |6L6. |. 668. | PLe:
chO-T | “ 1086. | G06. | LLS-
880-T | “ |286. | G06. | 92.
TP0-T'] “ | 926. | 916. | P28.
9P0-T | “© |4L6- | F26- | 68:
860-T | — {G46 | 686- | LL8.
GOT | “ | L001 | GG6- | T6-
FIO-T | “ | 20-1 | G68: | 16:
166. {000-1 #e0r [oo foe
S phe
G25 QE | = Qs
Be | S| Be | Pe [eae
=. ou 9 ool ga hae ae
Be | °8 | ms | Be (Pe
o : °
Rs 2

|

“OTYVL CULES oy} Ul poonpot earnssoid OUIGS OY} 4e soouvsqns

19Y}O LOJ sontva oy} puv ‘ornssoad yous 4v OOO.T= opeul Sug 1oywA\ TOF yey} 4°

“ag*H? if fe)
‘oueTeyydeu
-omloig

"HO
eueyAqya

eo | ee80068

bbl. | ot
Tog. |
cog. |
91g. |
geg, | ot
1zg. |
969, ies
S68: | GIs:
Gb8: | Sas:
VPs: | 668:
EPs: | GEés:
9E8- | GPs:
P98: | SPs:
Hee | gag,
oo 198:
al =
ra So)
cE ‘tS
a8, | oe
ote Qs
Oe oe be
P iS
®

e@eeree

eeesee

s v9
Ig HO
“guoZUEqouLolg

eeeree

se eee ew | ee eo oe

G6L- | GGL:
961, | GOL.
Zc8: | OS:
961. | GLE:
atm
0g: | 6a
18: | 662:
S19: | F6L-
9ze. | 6T-
Ze. | ze
Org. | 608:
Gaze. | 119:
Gag. | Zee
G8. | 1Z8.
IIs: | 0&8:
TT8: oo
a ot
in sO
HE Obs
joe ee
‘ a

eoVeeee | eeesee | seesee ee 000‘0%
eeeree | seeeee | ceoeeee | seenee 000‘ST
seeeee | ceoeeen | ceovonee | ceoevee 00001
GOL. | 8GL. |" GGL. | 000
664. | PLL. | OGL | 000E
G6L. | GBk | "°°" | LOL | 000Z
G61. | 6l- | °° | T8L | 00g
G6L- | TOs: | "7" | O82 | OOOT
L6L- | F6L |" | 8% | 006
96L- | 618 | "°°" | $8. | 008
eis: | pog| 7" "| O84 | 0OL
G8: | T18 | “°° | G8L- | 009
6620 | Siler pes P8L- | 00S
908: | 08. | OT8: | P8L | OOF
Tis: | 908 | 708: | 48k | 008
86g. | 808: | PIS: | 28L- | 002
1z8. | LE8- | PI8 | PEL | OST
Rate 6r8: | 618: | 964 | OOT
nie ee By
ae Le eee eae

a ee ae =
oF. os aie, ate) Be
J or) ae oe ‘Hd = RS rm
OF) Fe | RE EE) Ba

Beal calm al ete =

Ne °

}
Ip jo sonje A poonpory

528

Drs. Ramsay and Young on

‘HO*H'O

‘Joyooye [AIT

|

eooeeee

eeacece

"OT}BA

° ee ar 8
‘oyey Aoryes
Aq
. Ig’ Fo

‘queTeyyydeu

eoeeee

-OULOLT

eee

eoosue

esoree

TO
‘ouoT Aaya

‘STH

‘AINOIOT,

eeoteee

“OHO
‘MAOJOLOTYO

ee

s eee

“euezueq
-OWl01g,

eee

“190
“OplLo[YoVs4o}

Le Coren ace)
‘1O°H‘O
‘OpLLO[Yo

TAA

: Te ee 9

“Oprulodg
ua

vO0: [

| 660-1

1c0- I
¢T0-1
460-1
GI0-T
660-1
0&0:
LE0-1
Fr0- 1
GE0-1

OlIvs OY} UL poonper o1nsserd owes oY} 7B soouRySqns 1OY}O LOJ son[va

——_— —_—_—_—__——

THD
‘eprpor [Aqua

Fan
‘oprydnsiq

oy} pue ‘ornssoad youe ye QO(.[= epem Suteq eprydinsiq-uoqaey oF ye} “7° Ip jo sonjva poonpoey

mogrep

SULIT [TUL
UL OLNSSOL

some Thermodynamical Relations. 529

Values of 2 for various Dissociable Substances at

definite Pressures.

|

Cd . | Chloral methyl- | Chloral ethyl-
Nitrogen Ammonium peated eas aleoholnece

N.O, | NH,Cl. | CO,N,H,. OC1,.CH | o¢ COl,.CH fox

OCH,. OC,H,.
a0 as ne Sd oi" let a gee be ame i alee
50 ie 1°91 3°75 2:97 2°94
100 52 3°63 7:02 5°46 547
150 81 4°98 10:13 8:26 791
200 11°3 6°26 12°71 10°23 10°17
300 15-6 8:39 Os) a 14:33 14-08
400 20°47 10°74 23°48 | 19°30 18-06
500 24-27 13°25 Ce TN a eee tee 22°34
ae. | 15-70
and Y.| R. and Y. Noes R. and Y R. and Y.
and
| M and E. |
___ Absolute Temperatures (¢) of various Dissociable Substances
: corresponding to definite Vapour-pressures.

|

Nitrogen| Ammonium) Ammonium)

Chloral methyl- | Chloral ethyl-

Pressure : alcoholate. alcoholate.
|| in millim. a chloride. | carbamate. OH ‘OH
NO, | NH,Cl. | CO,N,Hg, | OCl,.CH | OCH, CCl.CH | O65.

10 nee ae PPR abr RE Pi Meese ie Thea

50 oS D247 | 2899 3312 3340
100 256-0 542°8 299°4 343°3 346°1
150 263°1 5546 305'1 350°8 353°7
200 268-1; |. 563°3 309°8 356'1 358'1
300 Boo + OFT 3161 364°2 367°4
400 281-4 | 587-9 321-0 370°3 3738
500 285°9 596°0 A Lie ae eet eee 378°7
600 | 6080

Products of ~ into Absolute Temperatures fe : t)

for various Dissociable Substances at definite Pressures.

Chloral methyl- | Chloral ethyl-

Nitrogen|Ammonium| Ammonium alcoholate. alcoholate.

Pressure :
peroxide.) chloride. | carbamate.

in millim. | : O1 OH ee

N,O, | NH,Cl. | CO,N.H,. | CCl,.CH ocr, |COlsCH OC,H,.
Loe ee ae 3g PERN taal a ATR SOL LNT
50 or. 1002 1087 984 983
100 1331 1970 2102 1866 1893
150 2131 2762 3091 2898 2798
200 | 3030 3526 | 3988 3635 3642
300 | 4304 4844 | 5870 5219 5173
400 5787 6314 | 7537 7147 6751
500 | 6939 "897° | PERG Sipe Shae ee 8460
600 |e. 9467 |

530 On some Thermodynamical Relations.

Reduced Values of (a t) ; that for Water being made =

1-000 at each pressure, and the values for other substances
at the same pressure reduced in the same ratio.

Chloral methyl- | Obhloral ethyl-

Nitrogen |Ammonium|Ammonium

ee peroxide.| chloride. | carbamate. leas apis
N,0, | NH,Cl. | CO,N,H,. | OCl,.CH 1-00 OGH,,| Cs-O8 { OC,H,
LSE? 4 Og ata, Beam yh 452 Ft
EP BE goat 1s 1:293 1:170 1170
100 0828 1:226 1:308 1-161 1178
150 0-921 1:1938 1°335 1:252 1:209
200 1:004 1-169 1:306 1°205 1:208
300 0987 1-110 1°346 1-197 1:186
400 1-018 1116 1°332 1:2638 1:193
500 1:004 ee Re ee OM Ried cadet inn 1:225
BOO EE rams 1:164
Reduced Values of fee PF ; that for Carbon-bisulphide being

made =1°000 at each sein and the values for other
substances reduced in the same ratio.

Nitrogen|Ammonium peers Chloral meine) as
& maillim Caran chloride. | carbamate. pila coor
N,O, | NH,Cl. | 00,N.H,.|CCl,.OH { OCH, COl.CH | OC:,

10 yee he wee Ob 2 he Rahs he

50 oie 1-483 1-609 1-456 | 1456
100 1-043 1540 1-648 1-458 1-480
150 1-159 1°502 1-681 1-576 1522
200 1-277 1:486 1-659 1532 1-535
300 1:255 1°412 17s 1-521 1:508
400 1-298 1423 1-699 1611 1-522"
500 1-281 1-458 bedeney nye? eee 1°562
600 He 1-488 |

A relation may be observed between the two substances
bromobenzene and chlorobenzene, and also between ethyl
chloride, ethyl bromide, and ethyl iodide. ‘The ratio of the
absolute temperatures of the bodies in either group corre-
sponding to any given vapour-pressure is a constant. Thus
the ratio of the absolute temperature of bromobenzene to that
of chlorobenzene when the vapour-pressure of both is 100
millim. = aed or 1:061; and at the other pressures up to
700 millim. the ratio remains absolutely constant at 1°059.
The ratio of the absolute temperatures of ethyl bromide and
ethyl chloride has been determined for pressures between 150

On the Velocity with which Air rushes into a Vacuum. 531

and 5000 millim., and the numbers found vary only between
1:089 and 1:091. The data for ethyl iodide are much less
complete; the comparison with ethyl chloride can only be
made at pressures between 150 and 500 millim., and with
ethyl bromide between 50 and 500 millim. In the first case
the ratio of the absolute temperatures varies between 1:296
and 1:209, and in the second case between 1°105 and 1:108.

[To be continued. |

LXII. On the Velocity with which Air rushes into a Vacuum,
and on some Phenomena attending the Discharge of Atmo-
spheres of Higher into Atmospheres of Lower Density. By
Henry WILDE, Lsq.*

ONSIDERING the present condition of our knowledge
7 respecting the mechanical properties of air and other
gases, some apology might appear to be needed in bringing
before this Society the results of an investigation touching
some fundamental principles in pneumatics, which for more
than a century have been considered to rest on foundations
as secure as the laws of gravitation of the heavenly bodies.
A survey of the history of the dynamics of elastic fluids will,
however, show that, great as are the advances which have
been made in this branch of science, the laws of the discharge
of elastic fluids under the varied conditions of elasticity and
volume are still left in much obscurity. The several circum-
stances which have combined to produce this anomalous state
of our knowledge of this subject are :—(1) The application of
the laws of discharge of inelastic fluids, without any -modifica-
tion, to those which are elastic; (2) the confusion of the
‘quantity of the discharge of elastic fluids after leaving the
vessel, with the velocity of discharge through the aperture in
the vessel ; and (3) the want of a sufficient number of expe-
riments, under varied conditions and through sufficient range of
pressure, to compare with the deductions derived from theory.
It has hitherto been assumed, as a leading proposition in
pneumatics, that air rushes into a vacuum with the velocity
which a heavy body would acquire by falling from the top of
a homogeneous atmosphere of the same density as that on the
earth’s surface ; and since air is about 840 times lighter than
water, if the whole pressure of the atmosphere be taken as
‘equal to support 33 feet of water, we have the height of the
homogeneous atmosphere equal to 27,720 feet, through which,
by the free action of gravity, is generated a velocity of 1332

* Communicated by the Author, having been read at a Meeting of the
Manchester Literary and Philosophical Society, October 20, 1885.

532 Mr. H. Wilde on the Velocity

feet per second. This, therefore, is the velocity with which
air is considered to rush into a vacuum, and is taken as a
standard number in pneumatics, as 16 and 32 are standard
numbers in the general science of mechanics, expressing the

action of gravity on the surface of the earth. 3
| Now, so far as | am aware, no experiments have hitherto
been made directly proving this important proposition. Itis
true that attempts have been made to determine the initial
velocity by discharging air at extremely low pressures into
the atmosphere ; but, apart from the conditions of the dis-
charge into the air and into a vacuum being different, the
history of physical science shows that it is unphilosophic to
predicate absolute uniformity of any law through the order
of a whole range of phenomena of the same kind ; as nature
is full of surprises when pushed to extremes, or when interro-
gated under new experimental conditions. :

It was long ago shown by Faraday* that, in the passage of
different gases through capillary tubes, an inversion of the
velocities of different gases takes place under different pres-
sures, those which traverse quickest when the pressure is
high moving more slowly as it is diminished. Thus, with
equal high pressures, equal volumes of hydrogen gas and
olefiant gas passed through the same tube in 57” and 135/"5
respectively ; but equal volumes of each passed through the
same tube at equally low pressures in 8’ 15” and 8 11”
respectively. 4 Again, while the velocities of discharge of |
inelastic fluids are as the’ square roots of the heads, some
mathematicians have justly considered that this law does not.
apply to those which are elastic, and have assumed with good
reason (though what appears unlikely at first sight) that the
velocity of air discharged into a vacuum is the same for all .
pressures. But whatever differences of opinion there may be
amongst natural philosophers on this point, all are agreed in
estimating the quantity of air discharged from a higher into
air of a lower density, from the difference between the two
densities, as in the similar case of the discharge of inelastic
fluids, by the difference or effective head producing the pres-
sure. ‘This mode of determining the amount of the discharge
from a higher to a lower density, like that of the velocity of
the atmosphere into a vacuum, has not, so far as I know, been
made the subject of experiment through any considerable
range of pressure. It therefore appeared to me that, as each
gas has its specific velocity of discharge, such a series of ex-
periments might be useful in confirming and extending our
knowledge of the dynamics of elastic fluids. In the course of

* Quarterly Journal of Science, 1818, vol. vii. p. 106.

!

with which Air rushes into a Vacuum. 533

these experiments I have met with some results which I
thought of sufficient importance to bring before the Society.
The apparatus employed in this investigation consisted of
two strong cylinders of cast iron, shown in the engraving.
The small cylinder, A, had an internal capacity of 573 cubic
inches, while the large cylinder, B, had a capacity of 8459

ag i a
we ye ey

me ato

Subic inches, or about fifteen times the capacity of the cy-
linder A. To the top of this cylinder was fitted a syringe for

Phil. Mag. 8. 5. Vol. 20. No. 127. Dec. 1885. 2P

534 Mr. H. Wilde on the Velocity

condensing the air up to nine atmospheres, and also a Bour-
don’s pressure-gauge of an improved construction, graduated
through every pound of the above pressure. The accuracy
of this gauge was tested in my presence by the constructors,
Messrs. Budenberg and Co., through the whole range of pres-
sure, by comparing its readings with a column of mercury
of equivalent height. For pressures of 15 pounds above, and
for pressures below the atmosphere, a mercurial gauge and a
Bourdon’s vacuum-gauge were employed, the readings of
which were compared with each other: 380 inches of mercury
were considered equal to one atmosphere, and 2 inches of mer-
-cury to one pound of pressure. ‘The upper part of the glass
tube of the mercurial gauge was fitted with a brass cap and
screw-stopper, so that it could readily be used as a pressure-
gauge, or as a vacuum-gauge when required. The discharging
arrangement on the cylinder A consisted of a stopcock and
union for securing a thin plate, through which the discharge
was made. ‘The orifice in the plate opened as required, either
directly into the atmosphere or into the end of a short iron
tube two and a half inches internal diameter, communicating
with the bottom of the cylinder. The thin plate was a small
disk of tinned iron, three quarters of an inch in diameter and
one hundredth of an inch in thickness. The centre of the
disk was pierced with a circular hole two hundredths of an
inch in diameter. ‘The size of the hole was accurately deter-
mined by means of a wire expressly drawn down to the above
diameter ; the wire being calibred by one of Hlliott’s micro-
meter-gauges, divided into thousandths of an inch. The hole
in the plate was enlarged so as to fit tightly the gauged wire,
and the burrs on each side of the hole were carefully removed,
as this small amount of projection, as Dr. Joule has shown*,
exercises a notable influence on the rate of discharge through
apertures in thin plates.

The general reasonings, and the inferences drawn from
the experiments to be described, are based on Boyle and
Mariotte’s law of the density of a gas being as the pressure
directly, and the volume as the pressure inversely for constant
temperatures.

I have said that the capacity of the cylinder A was 573
cubic inches, which represents the same number of cubic
inches of air in the vessel at atmospheric pressure of 15 lb.
on the square inch ; and, generally, n times 573 cubic inches

* Memoirs of the Manchester Literary and Philosophical Society,
vol. xxi, p. 104,

Mc

|

with which Air rushes into a Vacuum. HOO

of air forced into the cylinder would be the equivalent of n
atmospheres of absolute pressure.

In converse manner, 5 lb. of pressure, or one third of an
atmosphere, is the equivalent of one third of 573 cubicinches,
or the equivalent of 191 cubic inches of air at atmospheric
pressure ; and, generally, 5 lb. of pressure is the equivalent
of 191 cubic inches of air at atmospheric pressure and for all
the higher pressures. ‘The mode of experiment was as fol-
lows :—Air was forced into the cylinder to the required density,
and after the heat of compression had subsided, the time of
each 5 lb. reduction of pressure was taken by means of a half-
seconds pendulum, commencing its oscillations at the moment
of discharge ; and the stopcock was suddenly closed, and the
number of oscillations noted for every definite discharge and
reduction of 5 lb. of pressure. In my earlier experiments, it
was found that when the air was compressed to nine atmo-
spheres, and successive reductions of 5 lb. were made to the
lowest pressure, the cooling of the air produced a notable
effect in diminishing the rate of discharge. By commencing

the experiments with the lower pressures and increasing them

by 10 lb. successively after each discharge of 5 |b., the changes
of temperature atttending the changes of density of the air
were kept within the limits of 5 |b. of pressure till the highest
density was attained. ‘The smali changes of pressure attend-
ing each discharge by the addition and abstraction of heat to
and from the cylinder were after a little practice easily cor-
rected, so that each discharge may well be considered as
having been made under conditions of constant temperature.
The large cylinder B was first used as a vacuum-chamber to
receive the discharge from the small cylinder. The chamber
was fitted with an exhausting pump and suitable vacuum-
gauges, and the pressure within the chamber was reduced to
six tenths of an inch of mercury; and that degree of vacuum
was maintained during the experiments.

The following table shows the velocity of air flowing into a
vacuum, as deduced from the time and difference of pressure
for every 5 lb. from 135 lb. to 5 lb. absolute pressure. The
velocities of the first column are deduced from actual experi-
ment, and in the next cclumn the velocities are calculated

_ from the difference of the area of the discharging orifice and
| | the vena contracta by applying the hydraulic coefficient °62.
:

2P2

936 Mr. H. Wilde on the Velocity

TaBLE I.—Discharge into a Vacuum 0°6 inch Mercury.
Barometer 29°42. Thermometer 54° F.

Absolute pres-| Time of Velocity, in Velocity
sure, in pounds| discharge, fe coefficient
et per second. =
persquareinch.| in seconds. 62,
135 75 750 1210
130 775 753 1214
125 8-0 759 1225
120 8°5 743 1198
115 9:0 734 1184
110 9°5 726 Li
105 10-0 {24 1168
100 10°5 722 1165
95 11-0 725 1169
90 12:0 703 1134
85 13-0 688 1109
80 14:0 678 1094
75 15-0 675 1089
70 16°5 657 1060
65 18-0 650 1048
60 20:0 632 1020
5D 22:0 628 1011
50 24°5 620 1000
45 27:0 624 1007
40 31:0 613 985
35 36'0 602 971
30 43-0 589 950
25 53°0 573 924
20 69:0 550 887
15 97:0 522 842
10 170-0 446 720

From this table it will be seen that the time of discharge
of 5 lb. from 1385 Ib. absolute pressure is 7°5 seconds. Now,
as 5 lb. pressure is the 3 part of the total pressure, we have

AK = 21:22 cubic inches of air from 1385 lb. pressure dis-

charged into the vacuum chamber in 7°5 seconds: or, in
another form, since 5 lb. and 191 cubic inches of air at atmo-
spheric pressure are equivalents, so 191 cubic inches condensed

to 9 atmospheres Mt = 21-22 cubic inches of discharge, as in

the above calculation. Again, we have for a cubic inch ex-
tended into a cylinder 0:02 of an inch in diameter (the size of
the discharging orifice), 265°25 feet x 21:22=5628 feet.
Hence V= sees 750 feet per second for the discharge
7°5 seconds
of air from 135 lb. to 130 Ib. into a vacuum through a hole in
a thin plate. Or V= es
orifice is formed to the contracted vein. By the like method

=1210 feet per second when the

:

(

:
|
|

|
:
|

with which Air rushes into a Vacuum. 5Bin

of calculation the velocities for the discharge of each 5 lb. of
pressure from 135 lb. to 10 lb. have been found. 3

The velocity with which air rushes into the vacuum, as seen
from the table, is considerably less than that which has hitherto
been assigned to it by theory, and is not constant for all
pressures, as might have been expected from the known ratio
of elasticity and density : the difference in the velocities be-
tween each discharge for the higher pressures, as will be seen,
is so small as to be exceeded by experimental errors. The
amount of this difference will, however, appear more clearly
when we are considering the velocity of air discharged into
the atmosphere. Meanwhile I may remark that the velocities
increase with the pressures by small asymptotic quantities, so
that the theoretic velocity of 1832 feet per second would be
obtained at a pressure of 40 atmospheres if the law of Boyle
and Mariotte held good for so high a density.

While the rate of each discharge may be considered ap-
proximately uniform for the higher pressures, the initial and
terminal velocities of each discharge of 5 lb. for the lower
pressures would be much different. This is specially notice-
able for the velocity (842 feet per second) assigned to atmo-
spheric pressure of 15 lb. ; and as it was a matter of much
interest that this important constant of nature should be de-
termined with all the accuracy attainable, experiments were
made to ascertain the velocity of discharge for every pound of
pressure from 15 |b. to 2 lb. In these experiments the read-
ings were taken from the mercurial gauge, and the vacuum
in the chamber was reduced to 0:4 of an inch of mercury.

The results obtained are shown in the table.

TasLeE 1].—Discharge into a Vacuum 0°4 inch Mercury.
Barometer 29:96. Thermometer 60° F.

Absolute pres-| Time of ame Velocity-
sure, in pounds) discharge, aaa : nd coefficient

persquareineh.| in seconds. atl cients ‘62.

15 16°0 633 1021

14 17°5 621 1001

13 19:0 614 990

12 21-0 606 977

11 23°0 600 568

10 25°5 596 961

5 ae 28°5 593 956

8 32°5 584 942

7 375 577 931

6 45:0 | 563 908

4) 55°0 559 901

4 70°0 542 874

3 102:0 497 802

2 180°0 421 679

538 Mr. H. Wilde on the Velocity —

- By a calculation similar to that for the higher pressures,
we obtain for the initial velocity with which the atmosphere
rushes into a vacuum through a hole in a thin plate

V= hy x a ne per second,
ory = = = 1021 feet per second for the contracted vein.

That the differences between the theoretic and experimental
velocities was not caused by the friction of the stream of air
against the circumference of a smaller orifice being greater
in proportion to that of the circumference of a larger orifice,
was proved by discharging air of 15 lb. pressure through a
hole one hundredth of an inch in diameter in another similar
thin plate, when the times of discharge through the short
range of 1 lb. of pressure were found to be in the ratio of
4 to 1, or inversely as the areas of the orifices.

Taking into further account the difference between the
initial and terminal velocities due to the reduction of pressure
from 15 lb. to 14 lb., the results of these experiments show
that an absolute pressure of 30 inches of mercury, and at a
temperature of 60° Fahrenheit, the atmosphere rushes into a
vacuum with a velocity not greater than 1050 feet per second,
or less than the velocity of sound.

Some anomalous rates of discharge which I obtained when
air of different densities was discharged into the atmosphere,
induced me to repeat the experiments with the same apparatus

TasLE III.—Discharge into the Atmosphere.
Barometer 30°17. Thermometer 59° F.

Effective pres-| Time of Apparent Velocity-
sure, in pounds| discharge, velocity, coefficient
persquareinch.| in seconds. | per second. "62.

15 8:0 1266 2043
14 8:25 1318 2126
13 8:5 138738 2214
12 9°0 1413 2280
11 9°5 1454 2349
10 10-0 1519 2450

bo Co HB St > ST OO
~—
a selene
—_
oc
=~]
or
Su)
S
bo
or)

with which Air rushes into a Vacuum. 539

and under precisely the same conditions as those which had
been made into a vacuum as above described. The results
are shown in Tables III. and IV.

On comparing the times of discharge in Table ILI. and the
velocities calculated therefrom with the times and velocities
in Table II., a remarkable difference will be observed in them
for the same effective pressures. Thus, the velocity of dis-
charge from 15 lb. to 14 lb. appears to be double that
assigned to the same pressure when the discharge is made
into a vacuum; while in the discharge from 2 lb. to 1 lb.
(the lowest pressure in the table) the velocity appears to be
more than six times greater, or 4219 feet per second. No
less remarkable than this apparent increase in the rate of dis-
charge is the complete inversion of the order of the velocities
as compared with those when the discharge was made into a
vacuum for the same effective pressure. Now, we haye
knowledge of several causes competent to diminish the velo-
city of air of constant temperature flowing into the atmosphere,
but none to increase the velocity except the form of the
aperture, which in this case remained unchanged. Recog-
nizing the fact that when air of 15 lb. effective pressure was
discharged into the atmosphere the cylinder actually con-
tained two atmospheres of absolute pressure, we are led to the
conclusion that the phenomenal increase in the rate of dis-
charge observed is caused by the external atmosphere acting
as a vacuum, and offers no resistance to the discharge into it
of air of 15 lb. pressure, which thereby becomes 30 lb. effec-
tive pressure. The velocity of air of 15 lb. effective pressure
discharged into the atmosphere based on this conclusion is
1021 feet per second, the same as the velocity found for the
discharge into a vacuum. For effective pressures below
15 lb. the velocities are compounded of the rate of discharge
into a vacuum, and the resistance of the atmosphere without
any regular ratio, but approximating to the square roots of
the pressures.

That the atmosphere acts as a vacuum to the discharge of
air into it of 15 lb. effective pressure, is further evident from
the results obtained, and shown in Table LV.

D40 | Mr. H. Wilde on the Velocity — :

TasLe [V.—Discharge into the Atmosphere.
Barometer 29°64, Thermometer 58° F.

Effective pres-| Time of dis- | Apparent Velocity-
sure, in pounds| charge, in | velocity, per | coefficient

persquareinch.} seconds. second. "62.
120 75 843 1360
115 775 852 1374 |
110 8:0 862 1590
105 8:5 852 1374
100 9:0 843 1360

95 9°5 842 1360
90 10-0 843 1360
85 10°5 851 1372
80 11:0 863 1392 |
75 12-0 844 1362
70 13:0 836 1348
65 14:0 833 1344
60 £530 843 1360
55 16°5 : 837 1350
50 18-0 843 1360
45 20-0 843 1360
40. 22°0 863 1392
35 24°5 886 1429
30 27°0 935 1509
25 31-0 980 1581
20 36°0 1053 1699
15 43:0 1178 1900
10 58:0 1311 2114

In this table it will be observed that the times of each
discharge from 120 lb. to 15 lb. effective pressure into the
atmosphere are identical with the times of discharge from
135 lb. to 30 lb. absolute pressure into a vacuum. Hence we
are able to formulate and prove the general proposition that
the atmosphere acts asa vacuum, and offers no resistance to
the discharge of air of all pressures above two absolute
atmospheres.

Although the times of discharge for each reduction of 5 |b.
of pressure, as we have seen, are the same as those for pres-
sures one atmosphere higher, when the discharge was made
into a vacuum, yet it seemed to me that a table showing the
apparent velocities due to the effective pressure would be
useful as exhibiting some further points of interest, and re-
vealing the fallacy involved in estimating the velocities from
the effective pressures. On comparing the velocities of each
discharge from 120 Ib. to 40 Ib., it will be seen that the theo-
retic velocity of 1832 feet per second is as nearly attained as
the units of pressure and time adopted in these experiments

with which Air rushes into a Vacuum. Gail:

would permit. We have therefore in the table a measure of
the difference of the theoretic and experimental velocities
with which air rushes into a vacuum by the same method of
calculation. This difference, as will be seen, amounts to
exactly one atmosphere of pressure.

For each reduction of 5 |b. from 120 lb. to 40 lb. the times
of discharge are inversely as the pressures ; and as the density
of the issuing stream of air diminishes in the same proportion,
the velocity of discharge is the same for all the pressures from
120 lb. to 40 lb., as shown in the table. Hence it appeared
to me at the commencement of this investigation, that the
theoretic and experimental velocities with which air rushes
into a vacuum were rigorously exact. The anomalous and
apparent increase in the velocities from 40 lb. to 10 lb., how-
ever, led me to suspect that the atmosphere in some manner
affected the results, and induced me to make the discharge
into a vacuum with the results shown in Table I.

That the phenomenal rate of discharge which I have de-
scribed should not hitherto have manifested itself in some
form, or be associated with some facts explanatory of it,
would indeed be surprising considering the varied circum-
stances in which the discharge of elastic fluids comes into
play. Hence, it has long been known that a jet of air issuing
from an aperture in a vessel produces a rarefaction of the
atmosphere near to the discharging orifice. This phenomenon
was first observed on a large scale by Mr. Richard Roberts in
the year 1824, and is described in a paper read before this
Society in 1828*. Roberts noticed that when a valve was
placed over an aperture in a pipe used for regulating a strong
blast of air for blowing a furnace, the valve, instead of being
blown off by the force of the blast, remained a short distance
from the aperture, and required considerable force of the hand
to remove it to a further distance. Subsequent experiments
showed that the adhesion of the valve was caused by the
partial vacuum formed between the valve and its seating by
the expansion of the issuing air. ‘These experiments were
repeated and extended by Mr. Peter Ewart to similar effects
produced by the discharge of steam through various apertures.
Some of these experiments were described before this Society,
and afterwards published in the Philosophical Magazine in
1829. ‘The degree of rarefaction produced by the discharge
of air and high-pressure steam was carefully measured by

* Memoirs of the Literary and Philosophical Society, 2nd series, vol. v.
p. 208.

+ “Experiments and Observations on some of the Phenomena attend-
ing the Sudden Expansion of Compressed Elastic Fluids.”

542 Mr. H. Wilde on the Velocity

Ewart by means of gauges inserted in different parts of the
jet. He also noticed the sudden fall of temperature from 292°
to 189° F. in the rarefied part of the jet when steam of 58 |b.
pressure was discharged into the atmosphere.

Sir William Armstrong also, in his experiments on Hydro-
electricity in the year 1842”, described a singular effect of a
jet of steam by which a hollow globe made of thin brass, from
two to three inches in diameter, remained suspended in a
jet of high-pressure steam issuing from an orifice; and
when the ball was pulled on one side by means of a string,
a very palpable force was found requisite to draw it out of
the jet.

it is abundantly evident from these experiments, that when-
ever elastic fluids escape into the atmosphere a partial vacuum
is formed near to the discharging orifice, the degree of vacuum
depending on the density of the issuing stream. Hwart’s
ingenious explanation, that. the vacuous space formed near the
discharging orifice is caused by the joint action of elasticity
and momentum of the suddenly released particles repelling
each other beyond the distance necessary to produce equili-
brium with the external pressure, has a high degree of proba-
bility ; but that this vacuous space should have the effect of
increasing the rate of discharge could only be ascertained, as
we have seen, by a direct comparison, under like conditions,
with the amount of the discharge into a vacuum.

Having established the fact that the atmosphere acts as a
vacuum to the discharge of air of all pressures above two
atmospheres within the range of my experiments, it appeared
to me that this phenomenon might only be a particular case
of a general law of the discharge of elastic fluids, and that it
would be interesting to know through what range of relative
pressures in two vessels the one would act as a vacuum to the
other. With this object air was compressed into the large
receiving cylinder from two up to eight atmospheres absolute
pressure, while air was condensed into a small discharging
cylinder up to nine atmospheres of absolute pressure. The
air was discharged from the same orifice as in the former
experiments, and the time of discharge recorded for each
atmosphere was for a reduction of 5 lb. of pressure. The
results obtained are shown in the table.

* “On the Efficacy of Steam as a Means of producing Electricity, and
on a Curious Action of a Jet of Steam upon a Ball,” Phil. Mag. ser. 3.
vol, xxii. p. 1.

=

with which Air rushes into a Vacuum. 543

TABLE V.

Absolute

ee eg | bs deel Soak B lo 18

| spheres.
9 ma TH ee ORB re Pe Tb 9-0 [14-0 Isenonda.
3 a ees) Ses Balls eh. 4. 86. 10-0.| 1aS ‘
7 | 100 | 100 | 100 | 100 | 100 |110 /145 :
6 |120 /120 | 120 | 120 | 125 |16-0 ‘
5 | 150 | 150 | 150 155 | 205 f
4 |200 | 20:0 | 20:0 | 255 | ‘
3 | 27-0 | 27:0 | 31-0 :
92 | 43:0 | 43-0 | | ‘
1 97°0 |

In this table the first vertical column to the left shows the
number of atmospheres in the small cylinder from which each
discharge of 5 lb. was made into the receiver. The ordinal

| numbers at the head of the table indicate the atmospheres in
the receiver when the discharge was made, commencing with
vacuo ; and the time of each discharge, in seconds, is shown
| against the pressure in the discharging and receiving cylinders
__ respectively. The times in the second and third vertical
columns are obtained from those in Tables I. and IV., when
the discharge was made into a vacuum and into the atmo-
sphere. On examining these results, commencing with the
lower pressures, it will be seen that for two atmospheres of
absolute pressure, the time of discharge (48 seconds) was the
same for a vacuum as it was when made into the atmosphere,
as has already been demonstrated. It will also be seen that
a pressure of two atmospheres in the receiver acts as a vacuum
to four atmospheres in the discharging cylinder. This is evi-
dent from the equality of the time (20 seconds) when the
discharge was made into one atmosphere or into a vacuum.
The like ratio will also be observed up to three atmospheres
in the receiver, which act asa vacuum to the discharge of six
atmospheres of pressure from the small cylinder. As the
pressure in the receiver was increased, the diminution of
resistance of the recipient atmospheres becomes still more
marked, till for the highest pressures we have the remarkable
phenomenon of six atmospheres acting as a vacuum to the
discharge of nine atmospheres of pressure. That this peculiar
relation of the discharging and receiving atmospheres has not
reached its full limit will be obvious from a comparison of the
numbers in the table, from which it would appear that, for
pressures exceeding those used in these experiments, the

544 On the Velocity with which Air rushes into a Vacuum.

resistance of the recipient atmospheres would be still further
diminished correlatively with an increase in the amount of

discharge.

With the object of giving more completeness to this research,
experiments were made to ascertain through what range of
relative densities the air in two vessels would act as a vacuum
to the other for pressures below that of the atmosphere. The
results are shown in Table VI., which are arranged in the
same manner as those in Table V. The times in the second
vertical column are taken from those shown in Table II. when
the discharge was made into a vacuum for each pound of
pressure, and the other times in the Table are those obtained
for successive discharges into air of different densities below
the atmosphere, the large cylinder being again used as a
receiver. .

TABLE VI.
|
Pounds
pee 0 1 2 4 6 S) 0 ae
square
inch.

cm fs a a | | |

15 160 | 160 | 160 | 16:0 | 160 | 165) 18-0] 215) 35°5 \seconds.
14 LT Ss LT BUSHES a 19D. 1 Ly -bs), 164 20% ae
12 21:0 | 21-0 | 21-0 | 21:0) 21:0 | 22:5) 30:0

9

99
bed
99
Pp)
9

99

As equality in the times indicates equality in the quantities
and velocities of the discharge for constant pressures, a simple
inspection of the table shows that, for discharging pressures
as low as 6 lb., the recipient air still acts as a vacuum up to
half the density of the discharging stream, and the regularity
of this law is maintained within the limits of 6 lb. and 90 Ib.
absolute pressure, as shown in Table V. For discharging
pressures below 6 Ib. the relative times of discharge and the
resistance of the recipient air increase ; and as we have already

seen that the similar times and resistances for discharging —

pressures above six atmospheres diminish, the continuity of
regular law is broken at both ends of the series of pressures,
just as it is in the series of planetary distances and some other
quantitative phenomena of nature.

| 545. |
LXIII. Notices respecting New Books.

Elementary Mechanics, including Hydrostatics and Pnewmaties.
Revised Edition. By Ouiver J. Loner, D.Sc. London, Professor
of Experimental Physics in University College, Liverpool. London
and Edinburgh: W. and R. Chambers. 1885.

WHEN the first edition of this little book appeared, those who
knew the tendencies of the time in the teaching of Mechanics
were bold to predict for it a certain and speedy success. The call
for a new and revised edition is the evidence that these prophets
were right, and that the demand for sound, clear, elementary expo-
sition of the essential laws of matter and motion is a genuine and
growing one. Thomson, Tait, Maxwell, and Clifford took a fresh
look at the old and well-worn principles, and presented them to us
in a fashion that was implicitly full and complete as well as novel.
But their language is not that of the people—of the school-boy, of
the artizan-stud ‘nt, of the young engineer; and even the trained
mathematician finds that their terse statement does not reveal all
its meaning and all its bearing at the first glance. Maxwell’s
‘Matter and Motion’ is the best example: his booklet of text
would require a volume of commentary to develop its entire sig-
nificance ; and so while it gives practised readers all the delight of
severe healthful exercise, it is virtually sealed to the tyro. Some
work was wanted, then, which should lay open in ordinary language
what was hidden in the works of the masters. ‘To do this without
sacrificing a jot of precision, without blurring by over- or under-
statement the sharp definition of the originals, without intruding
misleading illustrations intended to be vivid, without practising the
easy device of simplifying difficulties by evading them, itself requires
a kind cf masterhood; and this we think Professor Lodge to a
great extent in the first edition, to a greater extent in this second,
has come very near to indeed.

He has a happy audacity which in work of this kind is almost
genius: at any rate it gives the capacity of ‘‘ taking a fresh look ”
which talent often lacks, especially the talent of talented textbook-
writers. They have told us for generations that ‘force is that
which tends to produce motion,” and that word “tends” covered
with its five Jetters a very chaos of doubts and difficulties to the
hapless beginner who thought. The word “ force’ was made more
mysterious by its definition. Lodge cuts the metaphysical knot,
and sends the student straight to the concrete by telling him that
“by the term force we are to understand muscular exertion, and
whatever else is capable of producing the same effects” (p. 13).
Again, after a brief and clear investigation of the normal accelera-
tion of a point moving in a circle, that intangible phantom to the
beginner, we are told—‘“‘Although the point is always gaining velo-
city normal to the curve or along its radius at this rate, it does not
follow that it ever possesses any such velocity. It is in fact impos-
sible for a point to possess any velocity except that along the curve

546 Notices respecting New Books.

or at right angles to the radius of curvature ; for as fast as velocity
along the radius is generated, so fast does the direction of the radius
change; in the same sort of way that a promise for to-morrow need
never be fulfilled, because ‘to-morrow never comes’ ” (p. 23). To
give a puzzled student a hint like this is to “make his face to
shine.” Once more, in the part on Hydrostatics the question of
the pressure on the base and sides of a vessel containing liquid is
discussed without algebra or calculation, yet with perfect precision,
and each of the student’s difficulties is hit off and dissipated by a
happy phrase or analogy: then, when by a kind of experimental
induction he is led to see how all the cases are converging to a

common principle, the iast step is taken almost by the student

himself—‘in symbols, P=sAh. There is nothing more to explain.
This simple formula contains it all” (p. 159). How much more
truly educatwe is this method than the other that begins by raising
a dust of symbols, and forces a reluctant consent for the bothered
and unconvinced reader.

There is a further excellence in Professor Lodge’s method which
teachers will value. He every now and then reminds the student -

that there are large and interesting parts of the subject which are
only touched on for the present; and not content with stating this,
he gives him a glimpse of the directions in which these parts lie.
He opens up vistas of future interest which encourage the student
to persevere, in the hope that if he is faithful he will some day
reach wider pastures. ‘Thus at p. 5V enough is hinted to let the
student guess ‘“‘ what Rigid Dynamics is about:” he is brought to
the point where he must feel that some larger calculus is necessary.
The mystery of gravitation is glanced at in pp. 15, 81, and glanced
at in the right way, not as hopeless, but as stimulating. The
Degradation of Energy is aptly illustrated at p. 84, and its Trans-
formation at p. 86; a bit of Graphical Statics is worked in at p. 110,
with a characteristic note that the bit is “an indication of quite a
large art... . which may well occupy the student’s attention at a
later stage ;” the measures of stability are touched on at p. 130,
and meet halfway many perplexities of the inquiring student.
Lastly, at p. 164, after finding the vertical fluid pressure on an im-
mersed cube, we are told—‘ If we did not care for simplicity the
same might be shown by the symbols for a solid of any irregular
shape whatsoever, and a most important mathematical theorem it
would be. You may make its acquaintance hereafter in a more
general form under the name of Green’s theorem.”

This, we repeat, is something entirely different from the method
of cramming a smattering of all things into a primer: no student
can by such allusions as these to advanced matters have his head
turned to the vain fancy that he knows all about them; one and all
they come in as baits and lures to more earnest effort now and
higher achievement hereafter.

We need say little here about the good and trustworthy appa-
ratus of exercises abundantly provided, or the clearness and patness
of the figures, or the terse Saxon of the writing. What we desire

ee ne eee ee

Intelligence and Miscellaneous Articles. 547

most is to call the attention of enlightened teachers to a truly
educative manual, which is no mere collection of dry bones, but
alive with suggestion, and interest, and insight.

Transit Tables for 1886. By Latimer Ciark, F.RAS., MICE.

London: Spon. (Pp. 71.)

These Tables constitute a multum in parvo for the use of those
who do not care to turn over the pages of the bulky Nautical
Almanac. Mr. Clark has added to the usefulness of his previous
issues by appending the Right Ascensions of what he calls
*« Additional Stars,” as well as the time of the Moon’s southing on
every alternate day of the year. He has also reverted to his
original plan of giving the times of transit to hundredths of a
second.

With the aid of this handy volume and one of the excellent
Transit Instruments, invented by the Author, the merest amateur
will be able to keep his clocks going in accordance with Greenwich
or his own local time.

LXIV. Intelligence and Miscellaneous Articles.

VISIBLE REPRESENTATION OF THE FOCUS OF ULTRA-RED RAYS
BY PHOSPHORESCENCE. BY E. LOMMEL.

WELL-KNOWN experiment for making the ultra-red rays
visible is that of Tyndall. He uses a lens obtained by
filling a spherical flask with a solution of iodine in bisulphide of
carbon. This solution is opaque to the ordinary rays, but trans-
mits the ultra-red rays and concentrates them in a focus; the in-
creased heating effect in this point is shown by the ignition of
tinder or gun-cotton, or the incandescence of platinum.

This focus can easily be made visible by means of certain phos-
phorescent substances. Balmain’s luminous paint is well adapted
for the purpose; or, still better, a greenish-blue phosphorescent
calcium sulphide, the properties of which I have already described*.

If this has been made slightly phosphorescent by ordinary daylight,

it is increased to a bright luminosity by the less refrangible, and
particularly by the ultra-red rays ; and when the radiation is mode-
rately strong this lasts for hours, and even after the radiation
ceases it gradually diminishes, but is visible for some time. After
this cessation a dark spot appears on the irradiation place, because
here, in consequence of increased emission, the luminous power is
diminished or entirely destroyed.

If a transparent screen be mde of this substance by spreading a

layer of the powder between two glass plates cemented at the edges,

the ultra-red region of the spectrnm may be shown in greenish-
blue phosphorescence near the red end of the spectrum, both on
the front and back of the glass.

The same screen can be used to render visible the dark focus.
* Wiedemann’s Annalen, vol. xx. p. 853,

948 Intellwgence and Miscellaneous Articles.

If it be received on the screen, it appears as a bright spot on a
feebly luminous ground, and changes after a lengthened action into
a black spot, which has the same appearance as if a hole had been
bored in a bright surface. _

Instead of a solution of iodine or bisulphide of carbon, both
which substances have their drawbacks, I use a solution of nigro-
sine in chloroform or alcohol. These solutions are quite black
and only transmit ultra-red rays. Alcohol, it is true, strongly
absorbs the ultra-red rays, so that the alcoholic solution only gives
a focus of small thermal action. This fact is, however, advantageous
for the above experiment; for the luminosity of the very sensitive
phosphoescent substance lasts longer, and the blackening does not
set in so rapidly, as if the focus were more intense. The far more
diathermanous solution in chloroform gives a sufficiently hot focus
for demonstrating the thermal action; nigrosine is insoluble in the
still more diathermanous bisulphide of carbon.

With the great sensitiveness of phosphorescent substances for
ultra-red rays, the flame of gas, of a lamp, or even a candle is
sufficient to show the phenomenon. By means of the spherical
flask, or, still better, by means of a lens with interposition of a
plane parallel trough filled with the black liquid, a sharp bright
(positive) image of the flame is received on the screen; after the
radiation has ceased this lasts with continually decreasing intensity,
and then turns into a dark (negative) image on a brighter ground.
This is a kind of photography by means of invisible ultra-red rays.—
Wiedemann’s Annalen, No. 9, 1885.

ON THE TONES PRODUCED IN A PLATE OR A COLUMN BY FREQUENT
DISCHARGES OF AN ELECTRICAL MACHINE. BY E. SEMMOLA.

The conductors of an induction machine are connected, by means
of two wires 5 metres in length, with two binding-screws fixed at
the opposite sides of a brass plate 1 millim. in thickness, which rests
on an ebonite funnel. If, now, the path of the current is broken
so that sparks strike across, the brass plate begins to sound. It
sounds also even if a Geissler’s tube or a lead wire is interposed in
the break; and also if instead, a wire is led to earth from the one
binding-screw, and in this a break is made. If the end of the
broken wire is so far removed from the binding-screw that sparks —
pass, the tone is stronger, and can be distinctly discriminated from
the sound of the sparks.

Even it the wires are connected with a metal plate which stands
opposite that on the hearing-apparatus, sounds are heard in the
last-mentioned experiments. If the plate on the hearing-tube is
connected with the earth, the note is stronger. The same is also
slightly the case when a glass plate is interposed. If the wire
leading to earth is not broken, no sound is heard.

If the wires are connected with the insulated wire of a sonometer
instead of with the brass plate, nothing is heard; if, however, an
ebonite ear-trumpet is placed on the box the sound is audible.—
Beiblatter der Physik, vol, ix. p. 671.

) Phil. Mag. 8. 5. Vol. 20. No. 127. Dec. 1885.

INDEX to VOL. XX.

/
|

ABNEY (Capt. W. de W.) on the

production of monochromatic light
on the screen, 172.

Adther, on the luminiferous, 389.

Air, on the velocity with which,
rushes into a vacuum, 531.

Alcohol, on the electrical resistance
of, 301.

Alkyl compounds of the elements, on
the periodic law as illustrated by
the, 259, 497.

Alums, on the specific refraction and

dispersion of light by the, 162.

2 a on the height of the,
408.

Atmospheric air, on the separation of
liquid, into two different liquids,
463.

electricity, on, 456.

B.A. unit, on the determination of
the, in terms of the mechanical
equivalent of heat, 1.

Bidwell (S.) on the sensitiveness of
selenium and sulphur to light,178;
on the generation of electric cur-
rents by sulphur-cells, 328.

Bonney (Prof. T. G.) on the so-

called diorite of Little Knot, 205.

| Books, new :—Hoppe’s Geschichte

der Electrizitat, 70; Lefrey’s Mag-
netic Survey of the North-western
Territories, 204; Glazebrook and
Shaw’s Textbook of Practical
Physics, 300; Winkler’s Technical
Gas-analysis, 461; Lodge’s Ele-
mentary mechanics, 545; Latimer
Clark's Transit Tables for 1886,
547.

Bosanquet (R. H. M.) on electro-
magnets, 318.

Bourbouze (M.) on a new form of
hygrometer, 220.

Bouty (M.) on the electrical conduc-
tivity of solid mercury and the pure
metals at low temperatures, 77.

Buchanan (J.) on the thermoelectric
position of carbon, 117.

Budde (E.) on the quantity of elec-
trical elementary particles, 303.
Cailletet (M.) on the electrical con-
ductivity of solid mercury and the
pure metals at low temperatures, 77.

Calculating-machines, observations
on, 15.

Carbon, on the thermoelectric posi-
tion of, 117.

Carnelley (Prof. T.) on the periodic
law, 259, 497.

Chemical change, on the influence of
heat on, 323; on the retardation

_ of, 444.

Clark (J. W.) on certain cases of
electrolytic decomposition, 37,
458; on the determination of the
heat-capacity of a thermometer,
48 ; on the influence of pressure on
electrical conduction and decom-
position, 455.

Colloids, on the disengagement of
heat in the swelling and solution
of, 220.

Coral-reefs and islands, on the origin
of, 144, 269, 303.

Dana (Dr. J. D.) on the origin of
coral-reefs and islands, 144, 269,

Davies (D. C.) on the North-Wales
and Shrewsbury coalfields, 73.
Density-numbers, on the law of, 19,

TOE.

Dilatancy of media composed of rigid
particles in contact, on the, 469,
Diorite, on the so-called, of Little

Knot, 205.
2 Q

550

Edmondson (J.) on calculating-
machines, 15. |

Electric conductivity of gases, on
the, 305.

currents, on the distribution of,
in networks of conductors, 221 ;
on the generation of, by sulphur-
cells, 328.

Electrical conduction and decompo-
sition, on the influence of pressure
on, 435.

conductivity of solid mercury

and of pure metals at low tempe-

ratures, 77.

elementary particles, on the

quantity of, 305.

machine, on the tones produced

in a plate by frequent discharges
of an, 548.

—— processes in thunder-clouds, on,

34

—— resistance of alcohol, on the,
301.

Electricity, on an unnoticed danger
in certain apparatus for the distri-
bution of, 292; on atmospheric,
456.

Electrolytic decomposition, on cer-
tain cases of, 37, 438.

Electromagnets, on, 318.

Electromotive force, on the use of
Daniell’s cell as a standard of, 126 ;
on the seat of the, in a voltaic cell,
336, 372. |

Electro-optic action of a charged
Franklin’s plate, on the, 363.

Elster (J.) on the electrical processes
in thunder-clouds, 54,

Fleischl (E. von) on the deformation
of the luminous wave-surface in
the magnetic field, 216.

Fleming (Dr. J. A.) on Daniell’s
cell as a standard of electromotive
force, 126; on molecular shadows
in incandescence lamps, 141; on
the distribution of electric currents
in networks of conductors, 221.

Fletcher (Dr. L. B.) on a determi-
nation of the B.A. unit in terms
of the mechanical equivalent of
heat, I.

Fluorine, on the refraction of, 481.

Focus, on the accuracy of, for per-
fect definition, 354.

Fol (H.) on the depth to which day-
light penetrates in sea-water, 74.

INDEX.

Foussereau. (G.) on the electrical
resistance of alcohol, 301.

Friction, on the determination of the
coefficients of internal, 387,

Functions, on the mechanical inte-
gration of the product of two, 175.

Galvanic polarization, on the mole-
cular theory of, 422.

Gases, on the electric conductivity of,
305.

(xeitel (H.) on the electrical pro-
cesses in thunder-clouds, 34.

Geological Society, proceedings of
the, 73, 205.

Gladstone (G.) on the refraction of
fluorine, 481.

—— (Dr. J. H.) on the specific
refraction and dispersion of light
by the alums, 162.

Glazebrook (R. T.) on a comparison
of the B.A. standard resistance-
coils with mercury standards, 5438.

Goldfields. of the Transvaal, on the,
206.

Gray (T.) on measurements of the
intensity of the horizontal compo-
nent of the earth’s magnetic field,
484,

Gresley (W.8S.) on pebbles of haema-
tite from the Permian breccias of
Leicestershire, 209.

Groshans (J. A.) on a new law ana-
logous to those of Avogadro and
Dulong and Petit, 19,191.

Gulf-stream, on the influence of the,
on the winters of Great Britain
and Ireland, 439,

Heat, on the determination of the
B.A. unit in terms of the mechani-
cal equivalent of, 1; on the bleach-
ing of iodide of starch by means
of, 168; on the mechanical equi- ©
valent of, 217; on the influence of,
on chemical change, 323. |

Hennessey (Prof. H.) on the winters
of Great Britain and Ireland as
influenced by the Gulf-stream,
439; on the comparative tempe-_
ratures of the Northern and
Southern Hemispheres, 442.

Henry (Prof. L.) on the polymeri-
zation of the metallic oxides, 81.

Himstedt (Prof. F.) on a determina-
tion of the ohm, 417. .

Hood (Dr. J. J.) on the influence of |
heat on chemical change, 323. ¥

INDEX.

Hopkinson (J.) on an _ unnoticed
danger in certain apparatus for
distribution of electricity, 292; on
the seat of the electromotive forces
in a voltaic cell, 336.

Hutton (Capt. F. W.) on the geo-
logy of New Zealand, 212.

Hydrocarbon radicals, on the melt-
ing- and boiling-points of the
halogen and alkyl compounds of
the, 497.

Hydrogen, on the occlusion of, by
zinc-dust, 464.

Hygrometer, on new forms of, 220,
468.

Hygrometry, observations on, 218.

Incandescence lamps, on molecular
shadows in, 141.

Iodide of starch, on the bleaching of,
by heat, 168.

_ Jamin (M.) on hygrometry, 213.

| Judd (Prof. J. W.) on the deep boring
at Richmond, 210.

| Kerr (Dr. J.) on the electro-optic

| action of a charged Franklin’s

plate, 363.

_ Konig (W.) on the coefficients of in-
ternal friction, 387.

| Langley (Prof. S. P.) on the trans-

mission of light by wire-gauze

/ screens, 387.

' Larmor (J.) on the molecular theory
of galvanic polarization, 422.

| Lépinay (M. de) on an optical me-

_ thod for the measurement of small
lengths, 79.

| Light, on the depth of penetration

| of, in sea-water, 74; on the spe-

cific refraction and dispersion of,

by the alums, 162; on the produc-

tion of monochromatic, on the

' sereen, 172; on the sensitiveness

| of selenium to, 178; on the trans-

mission of, by wire-gauze screens,

| 387.

}| Lippmann (G.) on a mercury-galva-

- nometer, 220.

Lodge (Prof. O.) on the stream-lines
of moving vortex-rings, 67 ; on the
seat of the electromotive forces in
a voltaic cell, 372.

4' Lommel (E.) on the visible represen-

tation of the focus of ultra-red rays

by phosphorescence, 547.

| Ludeking (C.) on the disengagement

} of heat in the swelling and solution

of colloids, 220.

'
tT
;
q
| e
: hy

a — ——

dol

Luminiferous ether, on the, 389.

Magnetic field, on the deformation of
the luminous wave-surface in the,
216; on measurements of the in-
tensity of the horizontal component
of the earth’s, 484.

Marr (J. E.) onthe lower Paleozoic
rocks of Haverfordwest, 208.

Mendenhall (T. C.) on a differential
resistance-t hermometer, 384.

Mercury, on the electrical conducti-
vity of solid, 77.

Mercury-galvanometer, on a, 220.

Metallic oxides, on the polymeriza-
tion of the, 81.

Metals, on the electrical conductivity
of, at low temperatures, 77.

Molecular shadows in incandescence
lamps, on, 141.

Newall (H. F.) on colliding water-
jets, 351.

Nicol (Dr. W. W. J.) on supersatu-
ration of salt-solutions, 295.

Ohm, on the determination of the, 417.

Optical method for the absolute mea-
surement of small lengths, 79. .

Penning (W. H.) on the goldfields
of the Transvaal, 206.

Periodic law, on the, 259, 497.

Phosphorescence, on the visible re-
presentation of the ultra-red rays
by, 547.

Polarization, on the molecular theory
of galvanic, 422.

Polymerization of the metallic oxides,
on the, 81.

Quincke’s (Prof.) method of caleu-
lating surface-tensions, observa-
tions on, 51.

Ramsay (Dr. W.) on some thermo-
dynamical relations, 515.

Rayleigh (Lord) on the accuracy of
focus for perfect definition, 354;
on an improved apparatus for
Christiansen’s experiment, 358 ; on
the optical comparison of methods
for observing small rotations, 360 ;
on the thermodynamic efficiency
of the thermopile, 361.

Read (T. M.) on the action of land-
ice at Great Crosby, 73.

Resistance-coils, B.A. standard,
comparison of, with mercury
standards, 548.

Reynolds (Prof. O.) on the dilatancy
of media composed of rigid par-
ticles in contact, 469.

552

Ricketts (Dr. C.) on some erratics in
the Boulder-clay of Cheshire, 207.

Roberts (T.) on the lower Paleozoic
rocks of Haverfordwest, 208.

Rotations, optical comparisons of.

methods for observing small, 360.

Salt-solutions, on super saturation of,
295.

Sarasin (E.) on the depth to which
daylicht penetrates in sea-water,
74. '

Selenium, on the sensitiveness of, to
light, 178.

Semmola (E.) on the tones produced
in a plate by frequent discharges of
an electrical machine, 548.

Sire (G.) on two new types of con-
densing hygrometers, 468.

Smith (C. M.) on atmospheric elec-
tricity, 4656.

Stream-lines of moving vortex-rings,
on the, 67,

Strenger ‘( F’,) on the electric conduc-
tivity of gases, 305.

Sulphur, on the sensitiveness of, to
light, 178; on the gener ation of
electric currents by, 828.

Surface-tensions, on Quincke’s me-
thod of calculating, 51.

Sutherland (W.) on the mechanical
integration of the product of two
functions, 175.

Thermodynamical relations, on some,
5165.

Thermometer, on the determination
of the heat-capacity of a, 48; ona
differential resistance-, 384.

INDEX.

Thermopile, on the thermodynamic
efficiency of the, 361.

Thunder-clouds, on the electrical
processes in, 34,

Tomlinson (C.) on the bleaching of
iodide of starch by heat, 168.

Vacuum, on the velocity with which
air rushes into a, 531.

Voltaic cell, on the seat of the elec-
tromotive forces in a, 336, 372.
Vortex-rings, on the stream-lines of

moving, 67.

Water-jets, on colliding, 31.

Watts (W. W.) on the igneous rocks
of the Breidden Hills, 211.

Webster (A. G.) on the mechanical
equivalent of heat, 217.

Wiedemann (Prof. E.) on the disen-
gagement of heat in the swelling
and solution of colloids, 220.

Wilde (H.) on the velocity with
which air rushesinto avacuum, 531.

Williams (G.) on the source of the
hydrogen occluded by zine dust,
464.

Wood (De Volson) on the luminife-
rous ether, 589.

Worthington (A. M.) on the error
inv olved i in Quincke’s method of
calculating surface-tensions, 51.

Wroblewski (S8.) on the separation of
liquid atmospheric air into two
different liquids, 463.

Young (Dr. 8.) on some thermody-
namical relations, 515.

Zinc dust, on the occlusion of, by hy-
drogen, 464,

END OF THE TWENTIETH VOLUME.

Printed by Taytor ein FRANCIS, Red Lion Court, Fleet Street.

5°

iG a:
Br esis O
—_ . x s. a * o»
1000 ioe +
Eien Syn NS ai Eh a
Pakaafo “ = as Be =
AOR 6) <A aay, Se 7
Be V4 ~ 4
| = Be
TT | F
| | Fa
pas Dangergi, 985
Swains wy ‘ ge Eos
AAS 2746 LOUe Tama |
2404 170° hee

Mentsehikoff
— a
eX ae
ae
, a y
ge
Dy
SR ct
Nemu 3% = %
SS
A N
al .
Odia**:4 he a
Ne
SS
N
/ Ss N 2555
| . |
| ~
{ N
“N
ms <<
a ~ 29265
5°
= } ~
a?
|
:
i
|
3110 .
|
}
0
ee 3020
| =~ ~
ie
~
~~

Nassau *

rc

a

5°

th

10]

=

ee
ny

‘
Tah

/
rarivri

o.
eS
2

6 Mataki
A terwe

PMoiann

iS
a

Lraputeours pera
YS 2 Byron

YF Onoatoa

CENTRAL PACIFIC

BATHYMETRIC MAP

1000 fath 2000 fath........ 2000 fath

oTawell
MS Speiden
Hudson

Nokutaifat

oStrathmore Shoal

Sophia:

r ED, a i har
é i ud rea
i i } AY
f% 5 j
* } Z
t r eK
4 me j
¥ j /
" . .
f :
‘
~ |

~

« ] . ce ‘ i «

, & ‘A ; mae) ny ’
‘i 3 \ i j , ' ? ;
ee :s
we ee
Fig’ uP i 5
4 LY
: Ms i , e We
. 4 \ ve
aly ‘
x ae
Cal RL Ey
rs ) ' as
f
iS a) " ‘
* ei? oe
% 9 : a)
oy
"
pa
we ie def.
“al rf
4 {in
!
°
Ric
\
* hat
. ‘ Fi
uy “
‘
I X s ke
{
\
.
‘A
t
9, Uae!
te
aes
,
Be
<é;
5
ad \
’
5 .
i
:
.
‘
}
.
.
“fc
‘
iV pie
. ’
We i 4
,
/

Ladys the gh! (cel

i
+
ee a

\ : ‘ ' 1

i

BP oa te ondstnen wie 8 Mie ety ky oF harem han» WAY URNA Aire AMI iene i
i” mrrerey ag ,

4
a
|
.

YUM somg Usa yay

Ill Id 02 T9A' $s SEW TH

te

Phil. Mag. S.5 .Vol.20.P1. WV.

z

ul [ [I 1 4

5 ; Vue '
: x
; i
4 i Me « ey) Er
= ' a sd i
Re: “1 }
y 7 $4 4 x Ge oF .
P
on: ; rs
) ik
ui :
“id mn
a
| 7
J ; od
f 4 He 0 E , "
‘ :
3 ‘ phd
} , .
tx } ; + As
i *
in

a y
i »

i 4 § i
: ;

2 '

i

Ad
“i fe
‘ ! ' é

A u
; a 1 ;
rT
; e
a9 5 ’
.

4 1
eit P « we y
1
’ { at
F e
f ’ «
a” r ¥]

= a

i
i.
i +
* ? *
rh. i
Gis ita
'
Ne Ff “~
Wy
ings

?
feet
7 th
4%

‘

‘

A
fgar4s,,

(

iy Vcore gaa

cre

Ri

Zme Sulphate-

Phil. Mag.S.5.

Vol. 20 PRY.

Sulphate of Copper
_—— a

Electrotyped
Copper oa :

a

i?

1 4 Il E
Se ees
ee
, M ais
: : =
B, TT JE iaSoe INSEE SSSoo00oo0R0munI AL
Me ceaeneaen ore Ea RDO SSS SSS Sa aOSe eee A’ B
Ss:
b 2D |b M
Cc R

hohe

& a a a M
R2 R# Re
Rt R3 R5 R1

Mantern Bros.hth.

Phil . Mag. 5.5. Vol. 20. Pl.V.

to

Pare Zine Rod-|-

r=} (Electrotyped
U tte (essa iat,

Zine Sulphate-|---

Phil. Mag. S.5.Vol.20. P1.VL

See ea

‘
,
;
|
|

Bugs.

B
a

TF. Te
H

Mintern Bros. lth.

“5
an

te

¢

a

CO

Phil. Mag. S.5. Vol. 20. Pl. VI

galvanometer
hi G
"od B
= «
Coil (.*
A
i Z Coil
{
|
| Mt Q 1
' \
\ Secondaries
) \
\ M , RE
\ = Pig. 16, e
|
| Coil Coil 6
= &
N: Z
me Primaries. ae?
R =
iz:
a
Pp
lb
B Ze
R
x

Phil. Mag. S.5 Vol. 20. Pl. Vil

galvanometer

r G

Secondaries

Fig. 18,
mas

Primaries.

Fig. 13.

Mintern Bros. lith.

eas Ati hha
+ ff ie “a

: ii

if

fe i et tal Magia
4 “f bes Sy)
Oy Ai i A ets a cont ' Anes Mm ;

) ‘ es as yt
Ps Pte te
Oris ; bl, Ae ¢
, aa) 4 r
* ¥- to ey Pe |

ofo

25° 30° 35° 40)
100 1

‘3U° 35° 40

Fea: Mag 05, Hol (Zee ee

70° 75° 30° 85° 90° 95° 100°105° 110° 115° 120° 125°

er broken curve shows the effect. of
corrections , and was the carve

caelatirg the results of the

SSP Deen Cee |
75° 80° 85° 90° 95° 100°105" ILO’ 115° 120° 125° 130°135°

30° 35° 40° 45° 50°

Garve A.

[ee Ee ee

20° 36° 40° 45°. 50° 55° GO° Gb’ 70° 75° 80° 85" 90° 95° 100° 105° NO" 5°

55° GO° G5" 70° 75° 80° S5° 90° 95° 100° 105° 110° 115° 120°

Curve B.

65 |—

eS eae

= Eee

|

ald

Phil. Mag .S © Vol.20. Pl VIll

30°85" 40° 45° 50° 55° 6

0° 65° 70° 75° 80° 85°

90° 95° 100°105° 110° 115° 120° 125°

Mintern Bros. Lith

Sal aerate es el ee el eG

The upper broke curve shows the effect of

thermometer corrections, and was the curve
used in caleulating the results of the

observations .

[ese ial eee

|

| ie

oR 30° 35° 40° 45° 50° 55° GO? G5° 70°. 75° 80° 85" 90° 95° 1007105" ILO" Is* 120° 125° 130° 185"

val

Lice, Mes Paar cd io it i Athy

_ »

a
‘
'
“
se
5
”
"
~*
i
’ ‘
.
’
e {
‘
we Pa
‘
.

——

Stason
Ae.

=

bth

Sy

t.Mag.S.5.Vol 20 Bae

1

Pre

VMentem Bros

pe ee

ol. 20. Fae

Phil. Mas. ee

“OzIs [My Jpey apeog

/

%

Phul. Mas ‘ is 2 § V | 20 Pl XJ
5 VOU. G i AL.

Scale half full size .

PH

oe 29 ems. ---------

SS 7 G¥e) SSS Sea UN OY og epee tf eee

Mintern Bros. lth

sites Pai AALS a a ag me

THB; | Ht

_ LONDON, EDINBURGH, ayo DUBLIN ae

rest MAGAZINE, |

A
AND (‘a
_ JOURNAL OF SCIENCE. = f-
E Bex a Continuation of Tilloch’s ‘Philosophical Magazine, : 4
oy Nicholson’ 8 ‘Journal,’ and Thomson’s ‘Annals of Philosophy. | Sa
CONDUCTED BY | it 3
SIR ROBERT KANE, LLD. F.RS.MRLA FCS. |
- SIR WILLIAM THOMSON, Kyr. LL.D. F-R.S. &e. see
| WILLIAM FRANCIS, Pu.D. F.LS. F.R.A.S. FOS, a ie
. Ze
ian
FIFTH SERIES. =

<i

N® 122.—JULY 1885.

‘WITH THREE PLATES. F Tee
@ _Mhustrative of Prof. Ourver Lopex’s Paper on the Spica Vinee 1 ng

of ee Vortex-Rings.
%
LONDON: ;
PRINTED BY TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET, _

s fe 1a by Longmans, Green, and Co.; Kent and Co.; Simpkin, Marshall, and Co.;
and Whittaker and Co. ;—and by A. and ©. Black, and T. and T. Clark, Edin-

Fi Ee hargh Smith and Son, Glasgow :—Hodges, Foster, and Co., Dublin :—Putnam, $
New York: :—Veuve J. Boyveau, Paris :—and Asher and Oo., Berlin.

A limited number of Copies of Prof. OLIVER LODGES Memoir OMe Q

‘THE SEAT OF THE ELECTROMOTIVE FORCES IN THE a“
VOLTAIC CELL, ;

Complete, with Appendix and Plates, can be obtained on eee to
TayLor and Francis, Red Lion Court, Fleet Street, E.C.

UNIVERSITY OF LONDON.

Price 4s. (post-free, 4s. 6d.).
THE CALENDAR for the YEAR 1885-6,

Uontaining the Regulations for each Examination, the Examination Papers set
| during the past year, Lists of Graduates, and other information.

Taytor and Francis, Publishers to the University,
Red Lion Court, Fleet Street, E.C,

Demy 8vo, cloth, price £1, Volume I., with numerous Woodcuts, four Plates,
and Portrait.

THE SCIENTIFIC PAPERS

OF

J. P. JOULE, DCL, PRS.
Published by the Physical Society of London.
TayLor and Francis, Red Lion Court, Fleet Street, E.C.

Royal 4to, cloth boards, price £1. —
FACTOR TABLE FOR THE SIXTH MILLION,

CONTAINING THE

LEAST FACTOR OF EVERY NUMBER NOT DIVISIBLE BY 2, 3, or 5,

BETWEEN
5,000,000 and 6,000,000.
By JAMES GLAISHER, F.R.S.

Uniform with the above,
FACTOR TABLES FOR THE FOURTH AND FIFTH aR ape &
: Price £1 each. |

TaYLor and Francis, Red Lion Court, Fleet Street, E.C.

Demy 8yo, cloth, price 15s., to Members of the Physical Society 10s.
THE SCIENTIFIC PAPERS OF

THE LATE |
SIR CHARLES WHEATSTONE, D.C.L., F RS, &e.
Published by the Physical Society of London.

TAYLOR and Francis, Red Lion Court, Fleet Street, E.C.

[ADVERTISEMENTS continued on 3rd page of Con .

- &
4%

Arpane —

.

ee ee i
OORT | Pow!

. : , e = “J * a ee —

THE
LONDON, EDINBURGH, AND DUBLIN

—
we

eS

ea
rrr

: a oa |
as Vs i . *
~~ - = di « r P
- _ — mene PE | :
* - a ee = «8 ,
=. < = 2 ins 4 P
7 oe a >A etnger > <he-- ° .
418 dap SA, 4 Sock te? BS yma eG = ;
eta oi Bct — 7h, q . Po
. a ; - ~ om

=

cn

%%
3 = cea
“5s —

e <_e ee

.;
‘rh:

- > ” = a - oie ~ S aae
Pg ; = aE = a - om, Pd as
* Fees at ES EA eon? Wie Si eS Ae
= = 5 a - ‘ —
a s 2 ¥ Pus ay wa ‘on _ s
et ak =4 — f = x
y ss ee ‘

CONDUCTED BY

SIR ROBERT KANE, LL.D. F.R.S. M.R.LA. F.C... fea

SIR WILLIAM THOMSON, Kyr. LL.D. F.RB.S. &e. ad a
AND a We

ae Me ree Cie oe eee my Ak dan ol faa ne, aa
Re Vers S 7. 4 © ek - " ft Ata ge a he ‘. a ae ee
’ - =

Ee

WILLIAM FRANCIS, PuD. F.LS. FRAS. ros. |
e eee
eS FIFTH SERIES. *
. N° 123.—AUGUST 1885, are

“WITH THREE PLATES.

| ‘Illustrative of Dr. J. D. Dana’s Paper on the Origin of Coral Reefs he he
o Islands, Dr. J. A. Fuemine’s on the Use of Daniell’s Cell asa jf
Standard of Electromotive Force, and Mr. J. BUCHANAN’S on th we

en _ Thermoelectric Position of Carbon. oe

LONDON:

PRINTED BY TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET,

ba Sold by Longmans, Green, and Co.; Kent and Co.; Simpkin, Marshall, and» ao A .
ee ‘: -and Whittaker and Co. ;—and by A. and C. Black, and T. and T. Clark, idin- |f |

burgh; Smith and Son, Glasgow :—Hodges, Foster, and Co., Dublin :— she,
B rc York :—Veuve J. Boyveau, Paris :—and Asher and Oo., Berlin. — meee |

Rea

Lp PD ‘Double Number. Price Five Shillings.

: 1 S: ae ae
1 Nie At + ie ee a fe ed yk Fey ‘ : 4 Sasa? oe

and whose Papers are accepted, will be furnished before the Meeting with ©

Owixe to the length of several of the Cotnrutintonieeaa with which v

have been favoured, we have considered it desirable, in order to avoid
delay in their publication, to issue a Dovsite NumBer this month instead —
of giving the usual Supplementary Number at the end of the Volume. 1 ;

BRITISH ASSOCIATION FOR THE ADVANCEMENT OF SCIENCE, —
22 ALBEMARLE STREET, LONDON, W. |

The next ANNUAL GENERAL MEETING will be held at ABERDEEN, j
commencing on WEDNESDAY, September 9. |

President-Elect,

The Rigut Hon. SIR LYON PLAYFAITR, K.C.B., M.P., Ph.D., LL. D.,
F.RS.L. & E., F.CS.

Notice to Contributors of Memou's.—Authors are me tS that, under an —
arrangement dating from 1871, the acceptance of Memoirs, and the days on which ©
they are to be read, are now, as far as possible, determined by Organizing Commit- —
tees for the several Sections before the beginning of the Meeting. It has therefore —
become necessary, in order to give an opportunity to the Committees of doing —
justice to the several Communications, that each Author should prepare —
an Abstract of his Memoir, of a length suitable for insertion in the pub-
lished Transactions of the Association, and the Council request that he will send
it, together with the original Memoir, by book-post, on or before August 12,
addressed thus: —“ General Secretaries, British Association, 22 Albemarle Street,
London, W. For Section.......... ”” Authors who comply with this request, 4

printed copies of their Reports and Abstracts. If it should be inconvenient to |
the Author that his Paper should be read on any particular days, he is requested
to send information thereof to the SECRETARIES in a separate note.

Reports on the Progress of Science, and of Researches intrusted to Individuals —
or Committees, must be forwarded to the Secretaries, for presentation to the
Organizing Committees, accompanied by a statement whether the Author will ©
be present at the Annual Meeting.

No Report, Paper, or Abstract can be inserted in the Report of the Associa- 4
tion unless itis in the hands of the Secretary before the conclusion of the Meeting. —

T. G. BONNEY, Secretar

Se EON iy nae ph age. sais o ee ee

Now ready, medium 4to, price ls.

PROBLEMS ON THE MOTIONS OF ATOMS.

By JOHN KINNERSLEY SMYTHIES,
Barrister-at-Law of the Inner Temple.

TayLor and Francis, Red Lion Court, Fleet Street, E.C.

Seventh Edition. Price 2s. 6d. !
HYGROMETRICAL TABLES
Adapted to the use of the Dry and Wet Bulb Thermometer.
By JAMES GLAISHER, F.R.S.

TAYLOR and Francis, Red Lion Court, Fleet Street, E.C.

ee OS ey BA AT EE Be
PR, ees

[ADVERTISEMENTS continued on 3rd page of Cover.

ae ioe KA Re Ae

lise oe First Day of every Month.—Price 26. 6d. ;

THE
LONDON, EDINBURGH, anp DUBLIN

AND
JOURNAL OF SCIENCE.

_ Being a Ciniabion of Tilloch’s ‘Philosophical Mowe | ‘ 2
_ Micholson’ s ‘Journal,’ and Thomson’s ‘Annals of Philompiy? : ae

CONDUCTED BY oN
SIR ROBERT KANE, LL.D. F.RS. MRLA. F. ee

AND

WILLIAM FRANCIS, Pu.D. F.L.S, F.R.AS. FOS,
) ae

FIFTH SERIES, ~.
‘Ne 124,-SEPTEMBER 1885.

WITH TWO PLATES. = \22e,, ae
B Mlustrative of Dr. J. A. Fremiye’s Paper on the Distributinn ° n of

aye as Electric Currents in Networks of Conductors.
oh i
‘ \ - * 3
ie

a a F
a LONDON:
hha

oo ‘PRINTED BY TAYLOR AND FRANCIS, RED LION tity FLEET shire.

BSS. '

we
AN

Nae ready, medium Ato, prite le ye a:

PROBLEMS ON THE MOTIONS OF ATOMS. ae ee
‘By JOHN KINNERSLEY SMYTHIES, ===

Barrister-at-Law of the Inner Temple. ef

Taytor and Francis, Red Lion Court, Fleet Street, Eg C.

Seventh Edition. Price 2s. 6d.
HYGROMETRICAL TABLES
Adapted to the use of the Dry and Wet Bulb Thermometer.
By JAMES GLAISHER, F-.R:S.

TAYLOR and Francis, Red Lion Court, Fleet Street, E.C.

Royal 4to, cloth boards, price £1.
FACTOR TABLE FOR THE SIXTH MILLION,
CONTAINING THE

LEAST FACTOR OF EVERY NUMBER N OT DIVISIBLE BY a 3, or 6,1
BETWEEN |

5,000,000 and 6,000, 000.
By JAMES GLAISHER, F.R.S.

Uniform with the above,

FACTOR TABLES FOR THE FOURTH AND FIFTH MILLIONS. »
Price £1 each.

Tayzor and Francis, Red Lion Court, Fleet Street, E.C.

Demy 8vo, 98 pages and 2 Plates, price 2s. 6d. each, post free,
A limited number of Copies of Prof. OLIVER LODGE’S Memoir on

THE SEAT OF THE ELECTROMOTIVE FORCES IN THE
VOLTAIC CELL,

Complete, with Appendix and Plates, can be obtained on application to
Tayior and Francis, Red Lion Court, Fleet Street, E.C.

: es Ra } . o . =
> Ae ae eae ene is sania.’ ne : a
- os eet an into eceere sapere Seaphugc hee ile ei Ke ee - Pe +My

® aang * Ree Rees a *
Oe Ss rg as Ee ee ag

Demy 8vo, cloth, price 15s., to Members of the Physical Society 10s.
THE SCIENTIFIC PAPERS OF |

THE LATE

SIR CHARLES WHEATSTONE, D.C.L., F.BS., &c.

. — ars
“gis
fee bas 3 ee
en rk
—

A

yO

Published by the Physical Society of London. wi |
TayLor and Francis, Red Lion Court, Fleet Street, E.C. ‘¢

Demy 8vo, cloth, price £1, Volume I., with numerous Woodcuts, four Plates,
and Portrait. z

THE SCIENTIFIC PAPERS

OF

J. P. JOULE, D.C.L., FBS.
Published by the Physical Society of London.

TayLor and Francis, Red Lion Court, Fleet Street, A.C. = =§ @
[ADVERTISEMENTS continued on 3rd page of Covers)

ows

-

THE
LONDON, EDINBURGH, ann DUBLIN

reais Ey ae, a os
. ~. * >. 2.
A P ¢
. ?

Mie 7:
Bey hd

‘ Bey D.
ae
“3 ‘

P HILOSOPHICAL MAGAZINE B

ea
AND

JOURNAL OF SCIENCE.

Being a Continuation of Tilloch’s ‘Philosophical Magazine) Tae a
WNicholson’s ‘Journal, and Thomson's ‘Annals of Philosophy.’ |

Pte
a
ae o>

Pe 4
Game

eee
see de

=
Leet
os De

ah
[as
ce

7 ie :
Hy Si Wears

CONDUCTED BY

SIR ROBERT KANH, LL.D. F.R.S. M.R.LA. F.C.8.

SIR WILLIAM THOMSON, Kyr. LL.D. F.RS. ée. |
AND ¥ ee |

WILLIAM FRANCIS, Pu.D. F.L.S. F.R.A.S. F.0.S..

FIFTH SERIES. | ee a:

N° 125.—OCTOBER 1885,

WITH ONE PLATE.

| i " Mhstrativ of F. Srencer’s Paper on the Electric Conductivity
of Gases.

LONDON:

"PRINTED BY TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET,

d by Longmans, Green, and Co.; Kent and Co. ; Simpkin, Marshall, and Co. ; it LS
_ and Whittaker and Co. ;—and by A. and C. Black, and T. and T. Clark, Edin- ©
"i burgh; Smith and Son, Glasgow :—Hodges, Foster, and Co., Dublin | ,
New York :— Veuve J. Boyveau, Paris :—and Asher and Co., Berlin.

Now ae penta ie price an ik SI
asi, _ PROBLEMS ON THE MOTIONS OF ‘ATOMS. ee
1, By JOHN KINNERSLEY SMYTHIES, |

| Barrister-at-Law of the Inner Temple.

TAYLOR and FRANcts, Red Lion Court, Fleet Street, E.C.

Seventh Hdition. Price 2s. 6d.
HYGROMETRICAL TABLES
Adapted to the use of the Dry and Wet Bulb Thermometer.
es, , By JAMES GLAISHER, F.RS. |
Rei Taytor and Francis, Red Lion Court, Fleet Street, E.C.

- Royal 4to, cloth boards, price £1.
FACTOR TABLE FOR THE SIXTH MILLION,

CONTAINING THE

BETWEEN
pepe at , 5,000,000 and 6,000,000.
ata By JAMES GLAISHER, F.R.S.
Uniform with the above,

FACTOR TABLES FOR THE FOURTH AND FIFTH MILLIONS.
Price £1 each.

Taytor and Francis, Red Lion Court, Fleet Street, EC.

ae Demy 8vo, 98 pages and 2 Plates, price 2s. 6d. each, post free,
A limited number of Copies of Prof. OLIVER LODGE’S Memoir on

THE SEAT OF THE ELECTROMOTIVE FORCES IN THE
lint VOLTAIC CELL,

3 ae Eaniletc. with Appendix and Plates, can be obtained on application to
Taytor and Francis, Red Lion Court, Fleet Street, H.C.

: Bunny 8yo, cloth, price £1, Volume I., with numerous Wo four Plates,
and. Oy

THE SCIENTIFIC PAPERS

OF
J. P. JOULE, D.C.L., F.RS.
Published by the Physical Society of London.
Taytor and Francis, Red Lion Court, Fleet Street, E.C.

Demy 8vo, cloth, price 15s., to Members of the Physical Society 10s.
THE SCIENTIFIC PAPERS OF

THE LATE

SIR CHARLES WHEATSTONE, D.C.L,, F.R.S,, &e.
Published by the Physical Society of London.

3 Tayior and FRANCIS, Red Lion Court, Fleet Street, E.C.

4
4
4
4

a - LEAST FACTOR OF EVERY NUMBER NOT DIVISIBLE BY 2, 8, or 5,

SERUM HY 2 Wa ea ne eee ak ie Ae ae
Gees eg ery ny, “ty Sete ok eae OA

Ww Sds Sy 5 em ae ar id
ve 7 fA se

ay fA . wo SAL ;

re oe ee 2. Ta ae . >’ 7

=
~~
a

mee ROWE

Published the First Day of every Month.—Price 2s. 6d.

THE
LONDON, EDINBURGH, anp DUBLIN

| PHILOSOPHICAL MAGAZINE,

AND

JOURNAL OF SCIENCE.

Being a Continuation of Tilloch’s ‘Philosophical Magazine,
Nicholson’s ‘Journal,’ and Thomson’s ‘Annals of Philosophy.’

CONDUCTED BY

SIR ROBERT KANE, LL.D. F.R.S. M.R.LA. F.C.S8.

SIR WILLIAM THOMSON, Knr. LL.D. F.R.S. &e.
AND
WILLIAM FRANCIS, Pu.D. F.L.S. F.R.A.S. F.C.S.

FIFTH SERIES. @ 9% —

N° 126.—NOVEMBER 1880.

LONDON:
PRINTED BY TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET, |
|] and Whittaker and Co.;—and by A. and ©. Biack, and T. and T. Clark, Edin-

burgh; Smith and Son, Glasgow :—Hodges, Foster, and Co., Dublin :—Putnam,
| New York :—Veuve J. Boyveau, Paris :—and Asher and Co., Berlin.

| Sold by Longmans, Green, and Co.; Kent and Co.; Simpkin, Marshal], and Co.;

‘eS “>
Lf

Jr
v

MBER 1885. ——No. 196,

i?

i "HANDBOOK oF TECHNICAL GAS- NALYS
“Containing Concise Instructions for carrying out Gas-Analytia Methoc
of proved utility. i Math

"By CLEMENS WINKLER, Px.D., &e.
BSS ite Translated, with additions, by
ale GEORGE LUNGE, Pz.D.,
ys Author of ‘ Coal-Tar Distillation,’ &e.

ite JoHun Van Voorst, 1 Paternoster Row.

mane | Price 33s.

; Cae TEXTBOOK OF ELEMENTARY MECHANICS.
eee y: By Prorsssor OLIVER LODGE, D.Sc.

at eo Revised Edition, with Additional Exercises.

i ‘; __ This book aims at a sound knowledge of principles. Although the treatment

is throughout very elementary, the author has endeavoured to explain in clear

uage the fundamental principles thi underlie even the highest develop-

-‘ments of the subject. |]
. W. & R. Cuampers, London and Edinburgh. f

: bs | Medium 4to, price ls. | |
es PROBLEMS ON THE MOTIONS OF ATOMS.
i a By JOHN KINNERSLEY SMYTHIES,

is Per % Barrister-at-Law of the Inner Temple.

Taytor and Francis, Red Lion Court, Fleet Street, E.C.

<
at rs 5 Oe WI ve a =
RASS Se eS oe

PR Me : Royal 4to, cloth boards, price £1.
Mien. FACTOR TABLE FOR THE SIXTH MILLION,

is a! CONTAINING THE

| “LEAST FACTOR OF EVERY NUMBER NOT DIVISIBLE BY 2, 8, or Gm
idee BETWEEN F
way 5,000,000 and 6,000,000. 8
ys ans By JAMES GLAISHER, F.R.S. ene 8 |

Pm om
ON er mE Eee

* +.

Uniform with the above,

FACTOR TABLES FOR THE FOURTH AND FIFTH MILLION 5. ee
Price £1 each. ae

TayLor and Francis, Red Lion Court, Fleet see EC.

: | Demy 8vo, price £1, with numerous Woodcuts, four Plates, and Portrait. ’

THE SCIENTIFIC PAPERS

OF | ot
ps P, JOULE, D.C.L., F.R.S. 3 4 |
Volume I. published by the Physical Society of London. ae |

a

Members who have paid their inthgp te Soa: for the current year are entitled -
to this publication.

; TaxLor and FRANCIS, Red Lion Court, Fleet Street, E.C. oa. 2
[ADVERTISEMENTS continued on 8rd page a Cover. oT

7 , by ae \!

DECEMBER 1885. No. 127.

Published the First Day of every Month.—Price 2s. 6d.

THE
LONDON, EDINBURGH, anp DUBLIN

PHILOSOPHICAL MAGAZINE,

AND

JOURNAL OF SCIENCE.

Being a Continuation of Tilloch’s ‘Philosophical Magazine,
- Wicholson’s ‘Journal,’ and Thomson’s ‘Annals of Philosophy.’

CONDUCTED BY

SIR ROBERT KANE, LL.D. F.R.S. M.R.LA. F.C.S.

SIR WILLIAM THOMSON, Kur. LL.D. F.R.S. &.
AND |
WILLIAM FRANCIS, Px.D. F.L.S. F.R.A.S. F.C.S.

FIFTH SERIES. <p “72 |

Ne 127.—DECEMBER 1885.. ‘ sy
WITH TWO PLATES. daas

Illustrative of Prof. Ossornz Ruynotps’s Paper on the Dilatancy of
Media composed of Rigid Particles in Contact, and Mr. T. Gray’s on
3; Measurements of the Intensity of the Horizontal Component of the

_ Earth’s Magnetic Field. 2

LONDON:

a : PRINTED BY TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET,
vl Sold by Longmans, Green, and Co.; Kent and Co.; Simpkin, Marshall, and Oo. ;
and Whittaker and Co.;—and by A. and C. Black, and T. and T. Clark, Edin-
‘| burgh; Smith and Son, Glasgow :—Hodges,. Foster, and Co., Dublin :—Putnam,
); New York:—Veuve J. Boyveau, Paris :—and Asher and Co., Berlin. 3

i a is
| “” P38 ~~ i
oc) ee
| Rae i) oa ~

Demy 8vo, price £1, with numerous Woodcuts, four Plates, and Portrait,

PE. gi'

Demy 8yo, price 7s., with 20 Plates. | ‘ ee

MEMOIRS OF THE =
MANCHESTER LITERARY AND PHILOSOPHICAL SOCIETY, |
THIRD SERIES, VOL. 8.

Also, price 10s.
_ A CENTENARY OF SCIENCE IN MANCHESTER.
zi By R. ANGUS SMITH, F.R.S., &e.

- Tayror and Francis, Red Lion Court, Fleet Street, E.C.

_ THE SCIENTIFIC PAPERS —

OF
J. P. JOULE, D.C.L., F.B.S.
Volume I., published by the Physical Society of London.

Members who have paid their subscriptions for the current year are entitled
to this publication.

TAYLOR and Francis, Red Lion Court, Fleet Street, E.C. na

Medium 4to, price Is. |
PROBLEMS ON THE MOTIONS OF ATOMS.
By JOHN KINNERSLEY SMYTHIES,

: Barrister-at-Law of the Inner Temple.

Tayuor and Francis, Red Lion Court, Fleet Street, H.C.

* Demy 4to, slit price £1 10s. each.

ILLUSTRATIONS OF THE GEOLOGY OF YORKSHIRE.
7 By JOHN PHILLIPS, F.R.S. &e.

- Part I.—The Yorkshire Coast. 28 Plates and 1 Map.
», 1.—The Mountain Limestone District. 25 Plates.

TAYLOR and FRancts, Red Lion Court, Fleet Street, E.C.

Now ready, Price 6d., post-free 7d.,
TAYLOR’S CALENDAR OF MEETINGS

OF

THE SCIENTIFIC BODIES OF LONDON
FOR 1885-86.
TAYLOR and Francis, Red Lion Court, Fleet Street, E.C.

[ADVER TISEMETS continued on 8rd page of Cover.
2 ri 1b | “ie fi

: oh r
wee

be)

=

6 a ee

egace pe ae

SMIT HSONIAN INSTIT

WH | l F | il ] |

43450